Way to Success

The new strategy of personal knowledge storage

2011-12-23T22:34:00.000-08:00

Six years ago back into 2005, I started to write articles to accumulate and collect important information or papers in my blog.

I wrote emotions, study discussion, reports, computer skills, and so on, anything worthy memory.

As time goes, I came to realize that I should have tag or catalog to manage my stuff, in order to facilitate my memory.

Then I reset it up again, now it looked much better.

And this blog did help me to construct my knowledge framework.

However, I could not access my blog later because of Chinese government's policy on internet censorship.

I tried several other services, including myspace, but none of them met my expectation just like blogger did.

The past first ten years of the twenty-first century is the era of internet information. I benefit just a piece of it, however that has already so profound that it has changed my entire life.

I got the the chance to access my blog again using some VPN tool. It was exciting really. just like meeting an old friend.

But things were changing with time, always.

I still love blogger, but Google could do more than just blogger. I started to use Google Document, Google Buzz, Google Reader to record my life.

Especially when some article is not appropriate to upload to blogger, like pdf, I prefer to store it in Google Document.

Google Reader is also a powerful tool to read papers and other articles.

The problem is when you want to share something, so where you would like to post it, in Blogger, Reader, or Buzz, or just Document?

This has ever haunted me for some time. I don't want to place things without order, just like a mess.

This issue became even more critical when Google finally wanted to rock into SNS and release the Google+.

Google even disabled some feature in Google Reader, just wanting to compel the user into G+. Of course other Google products have been updated as well.

Then I have reevaluate this structure of information again.

Previously I thought I could just share stuff in G+ just like I used to in Reader, those article not written by me, but really great to learn from;

Some are just news, some are discussions, it all depends upon the focus.

Now I get the new strategy of personal knowledge storage as following:

Post the original articles or discussions in Blog (Actually I have more than one blog);
Post the daily life stuff into Google Plus, just like Facebook.
Post the paper I read about study or research into Google +1, without Share.
Post the web I read irrelevant with study or research into Google +, with Share.
Post private information, such as mind storm, into Google Document.

So where is Google Reader?

It would just stay where it is supposed to be. I still have many valuable RSS, I could insist on read this and maybe share in G+, +1 or not depends upon the content.

The criteria for a article should be share as the same time +1 depends whether this article should be collected for a second read in the coming future.

Just like I would not care what is the definition of sociology, hence that is supposed to be shared, but could be +1 by someone interested.

Change is inevitable, and it should be better.

Characterization of LED-Based Sun Photometers for Use as GLOBE Instruments

2011-12-21T17:12:00.000-08:00

http://instesre.org/papers/Snowmass/snowmass.htm

-- Shared with Google Share Button

Carbon Dioxide - Human-Related Sources and Sinks of Carbon Dioxide | Climate Change - Greenhouse Gas Emissions | U.S. EPA

2011-12-21T07:56:00.000-08:00

http://www.epa.gov/climatechange/emissions/co2_human.html

-- Shared with Google Share Button

Configuration error for Oracle 11g SQL Developer with Jave

2011-12-18T00:45:00.000-08:00

I came into the following error when I was try to deploy a website using Oracle.

However I got this error that SQL Developer can not launch because of one missing DLL!

And we all know it is not about the dll, actually it is the problem with the JRE and JDK issues.

So I Googled the error, and I got this information as following:

Supposedly "Java SE 7 is strongly compatible with previous versions of the Java platform". See the official discussion of compatibility here:

http://www.oracle.com/technetwork/java/javase/compatibility-417013.html

That said, if one configures SQL Developer 3.0.04 to run under jdk1.7.0, a warning dialog comes up with the following disclaimer asking if you wish to proceed:

You are attempting to run with Java 1.7.0

Running this product is supported with a minimum Java

version of 1.6.0_04 and a maximum version less than 1.7,

This product will not be supported, and may not run correctly

if you proceed. Continue anyway?

So no guarantees from Oracle here. I haven't heard which SQL Developer version will officially support jdk1.7.0.

To try this out, you need to adjust your configuration as follows. For example on Windows do something like:

1) The sqldeveloper.conf file must SetJavaHome, e.g., SetJavaHome C:/Program Files/Java/jdk1.7.0

2) copy "C:\Program Files\Java\jdk1.7.0\bin\msvcr100.dll" to the same directory as sqldeveloper.conf.

I did as it suggested, and it works！

Activities and Awards

2011-11-23T05:59:00.001-08:00

Note:
This article serves as the extended material for application.

Outline
This article mainly includes some snapshots of my activity awards or certifications during the undergraduate or graduate period.

“English Public Speech Contest for the Graduates of Wuhan University”

“Wuhan University Social activists”

“Wuhan University Sports Meeting” for eight times, awarded many times

Scholarship of Wuhan University for three times

“Wuhan University Straight-A Student”

Students’ Scientific Research Program

If you have any question, please contact : https://plus.google.com/u/0/117209043337528009210

How to improve my English (2)

2011-11-23T01:39:00.000-08:00

In my last post "How to improve my English?", I have already discussed a great amount of tips and suggestions of how to improve our English.

However, that is still far from satisfactory, believe it not.

Generally speaking, the common referred English ability as we discussed is merely the level of how we deliver and use English. In another word, that is simply the way use it, but not the content at all.

One simple example to illustrate that concept is that when someone would take a test like TOEFL, the judge will only measure your ability in deliver. Whether you are arrogant or not would never matter your final score!

So, the point is shall we pay attention to the content other than the way? I know most speech contest will also pay more attention upon your pronunciation and delivery. But definitely the content in your speech weights as much if you really intend to inspire your audience.

Then that is exactly what I am talking about now.Pay attention to your knowledge!

Most of the time, the ability of delivery and the knowledge are combined together. Empty words seldom make any sense, as we both know that. And a topic would not be meaningful unless we put ideas into it.

Then what type of articles you read, musics you listen, websites you visit, or persons you speak to could generally determine your knowledge scope, as a matter of fact. And that is the reason I mentioned they are combined previously.

As for me, the main interests and energy have been paid upon major study, programming, NBA, internet technology (Google), so I really spent a lot of time in doing relevant stuff.

And I suppose some of the topics may not interest girls, like programming and IT.

But, you should pick up your style, your favorite taste of content. The more you get to know it, the more pro you could be in that realm.

The gap between brag and confidence simply lies in that whether you are really an expert in that or not. And I would encourage you to share your thoughts with friends or peers at the condition that you seriously have done a excellent homework.

Once you started this journey of enrich your brain, you will soon find the real happiness. And if you record this moment or share your achievement, that could be exaggerated again. Eventually, all the information would be interlocked as one network, and the interdisciplinary knowledge would be connected and enrich yourself in a way you never expected before.

Keep searching , don't settle!

Feedback to G+: https://plus.google.com/u/0/117209043337528009210/posts

How to improve my English?

2011-11-22T04:53:00.000-08:00

During the past several years, I was asked a frequent raised question that "How to improve my English?" by lots of people who use English as a second language. I have ever tried many ways to encourage them to try to use English as much as possible. Since I know, at least for me, the only secret is practice.

Personally, I received English courses when I was little in the primary school. Not until into the college, I came to realize what English should be like except exams.

Change has come to me when I wanted to speak it out. I still remember six years ago, I would got up one hour earlier to have a morning reading in front of the library.

Besides, I kept attending English corner every week."Don't be shy, just try" was the first maxim I heard from the corner. By the way, it also took me an hour to get to the English corner, but I never quit.

I could certainly feel that I am making progress. I have even participated the English speech contest, which was a really challenge for me. Friends are inviting me into their training classes as a teacher, and I am so pleasure to help.

Lots of people have ever helped me, I know that. And I am trying to do the same thing.

However, God helps those who help themselves! If you want to make the change, keep in mind, no pain no gain.

Some students prefer to attend training centers, they do spend lot of money on it, especially for those who would take TOEFL, GRE, GMAT and IELTS,.

I have also finished the GRE and TOEFL, of course I didn't take any courses since they are just unnecessary for me.

So how shall we put it into action, simply put, what direction or point we shall pay attention to?

Since "practice" itself could be both abstract and concrete. I will list some tips or guides to show you what types of practice may work on you.

Morning reading. It's never too late to read some real great work independently. Campus are extremely cool place to take a morning reading. Take a deep breath, and you can enjoy one or two hours' wonderful moment in the wood with fresh air.
English corner. As I said, it is main form that I carry out my speech or speaking practice. Here you can meet lots of great friends, enjoy their stories, taste their accents, and so on. It is very important to open up to let your voice fly in the air. It's only by that way you can adjust your own pronunciation, your speech pace, you skills of delivery. and so on.
Change all of your electrical devices or services into English language as soon as possible, including computer operation system, smartphone, email interface, search engine, blog, even contact information. You have to touch English and expose yourself to English as mush as possible. As long as you are doing this, you will be able to understand and feel about the language. And this practice could shape your thinking as well, not like that the English corner could serve.
Listen English musics and movie, especially those classical ones. By doing this, you could hear and imitate some touching dialog and obtain some culture elements.
Speak to yourself in English as much as possible. I know that we are always short of chance to speak English. English corner doesn't open everyday, and you may not be so lucky to have a partner who accompany you always. Hence, you will be the best audience of yourself. It sounds silly, but it is even sillier if you dare not take a shot.
Read English articles. I would recommend Google News. It collects all the best information at one single page and I could browser them efficiently. Pick up one or two of your interests like science or technology and keep doing this everyday. Click and check the details if you are entranced in one of them. Good article can always provide you very valuable reading experience and Google News apparently would be the best tool to mine articles.
Be ready to meet friends who love English always. That requires you to be active and passionate towards English.
Keep writing journals, like Blog. It is completely different to write something compared with say something. And record your progress of life is a good way to review and move forward. It could be even more important if you want to publish some paper officially. Read those term of use or help manual, the standard way of deliver formal information.
If you are lucky, make some friends who can talk in English frequently, even in cellphone.
Some girls have a gift in language? Never say that, you could be as good as them, even better, trust me!

I do sincerely hope that these tips and guides could be a great help to some of you, and I would be very pleasure to contribute more suggestion if necessary.

If you have any question, do not hesitate to contact me on G+:https://plus.google.com/u/0/117209043337528009210

Discussion in the paper

2011-10-25T08:17:00.001-07:00

When you describe the meaning of your little bit of truth, do it simply. The simplest statements evokes the most wisdom; verbose language and fancy technical words are used to convey shallow thought.
That is part of the book《How to write and publish a scientific paper》 by Robert Day

Batch change the table name or column name in Access using C# ADOX

2011-09-20T21:58:00.000-07:00

A good design weighs much in Database management.

However, sometimes we have no choice, but to modify the database to meet the expansion of data.

In my case, I am using Access database, which is poorly designed previously.

And I have to change the table name and column name.

There are several approaches to realize this goal, such as follows:

Since JET SQL can't directly rename the table name or column name, we could use a series of operations

including create new column, copy column and delete column to rename a column.

The same operation could be applied to tables as well.

Of course these operation are implemented by sql inquiry.

However, that seems quite cumbersome to some extent.

Another way to fix this problem could be the ADOX, the ADO Extension libiary.

Firstly, we have to add the reference of ADOX into our project.

Then, connect to the Access datqbase using connection string.

Create the ADOX object to change the table name column name.

Sample codes as follows:

string accessConnectString = "Provider=Microsoft.Jet.OLEDB.4.0;Data Source=" + connectionString;

ADODB.Connection connection = new Connection();

connection.Open(accessConnectString, null, null, -1);

adoCat = new ADOX.CatalogClass();

adoCat.ActiveConnection = connection;

//Change column name

Table adoTable = adoCat.Tables[tableName];

adoTable.Columns[oldFieldName].Name = newFieldName;

//Change table name

adoCat.Tables[oldTableName].Name = newTableName;

Above is the code for C#, similar code in other language such as VB.net could be quite close.

For more examples, please refer to :

http://www.supidea.com/post/change-access-table-name.aspx

If you are using other type of database such as SQLServer or Oracle, things might be different, be cautious!

http://www.supidea.com/post/replace-sql-table-name.aspx

Matrix Arithmetic Operators Review

2011-09-07T02:10:00.000-07:00

Again, this is terrific if you want to keep all your work and research in the same line , especially in programming.

Matrix and array arithmetic

Syntax

A+B A-B A*B A.*B A/B A./B A\B A.\B A^B A.^B A' A.'

Description

MATLAB software has two different types of arithmetic operations. Matrix arithmetic operations are defined by the rules of linear algebra. Array arithmetic operations are carried out element by element, and can be used with multidimensional arrays. The period character (.) distinguishes the array operations from the matrix operations. However, since the matrix and array operations are the same for addition and subtraction, the character pairs .+ and .- are not used.

`+`	Addition or unary plus. `A+B` adds `A` and `B`. `A` and `B` must have the same size, unless one is a scalar. A scalar can be added to a matrix of any size.
`-`	Subtraction or unary minus. `A-B` subtracts `B` from `A`. `A` and `B` must have the same size, unless one is a scalar. A scalar can be subtracted from a matrix of any size.
`*`	Matrix multiplication. `C =` `A*B` is the linear algebraic product of the matrices `A` and `B`. More precisely, For nonscalar `A` and `B`, the number of columns of `A` must equal the number of rows of `B`. A scalar can multiply a matrix of any size.
`.*`	Array multiplication. `A.*B` is the element-by-element product of the arrays `A` and `B`. `A` and `B` must have the same size, unless one of them is a scalar.
`/`	Slash or matrix right division. `B/A` is roughly the same as `B*inv(A)`. More precisely, `B/A = (A'\B')'`. See the reference page for `mrdivide` for more information.
`./`	Array right division. `A./B` is the matrix with elements `A(i,j)/B(i,j)`. `A` and `B` must have the same size, unless one of them is a scalar.
`\`	Backslash or matrix left division. If `A` is a square matrix, `A\B` is roughly the same as `inv(A)B`, except it is computed in a different way. If `A` is an `n`-by-`n` matrix and `B` is a column vector with `n` components, or a matrix with several such columns, then `X = A\B` is the solution to the equation AX = B*. A warning message is displayed if `A` is badly scaled or nearly singular. See the reference page for `mldivide` for more information.
	If `A` is an `m`-by-`n` matrix with `m ~= n` and `B` is a column vector with `m` components, or a matrix with several such columns, then `X = A\B` is the solution in the least squares sense to the under- or overdetermined system of equations AX = B. The effective rank, `k`, of `A` is determined from the QR decomposition with pivoting. A solution `X` is computed that has at most `k` nonzero components per column. If `k < n`, this is usually not the same solution as `pinv(A)*B`, which is the least squares solution with the smallest norm .
`.\`	Array left division. `A.\B` is the matrix with elements `B(i,j)/A(i,j)`. `A` and `B` must have the same size, unless one of them is a scalar.
`^`	Matrix power. `X^p` is `X` to the power `p`, if `p` is a scalar. If `p` is an integer, the power is computed by repeated squaring. If the integer is negative, `X` is inverted first. For other values of `p`, the calculation involves eigenvalues and eigenvectors, such that if `[V,D] = eig(X)`, then `X^p = V*D.^p/V`.
	If `x` is a scalar and `P` is a matrix, `x^P` is `x` raised to the matrix power `P` using eigenvalues and eigenvectors. `X^P`, where `X` and `P` are both matrices, is an error.
`.^`	Array power. `A.^B` is the matrix with elements `A(i,j)` to the `B(i,j)` power. `A` and `B` must have the same size, unless one of them is a scalar.
`'`	Matrix transpose. `A'` is the linear algebraic transpose of `A`. For complex matrices, this is the complex conjugate transpose.
`.'`	Array transpose. `A.'` is the array transpose of `A`. For complex matrices, this does not involve conjugation.

Nondouble Data Type Support

This section describes the arithmetic operators' support for data types other than double.

Data Type single

You can apply any of the arithmetic operators to arrays of type single and MATLAB software returns an answer of type single. You can also combine an array of type double with an array of type single, and the result has type single.

Integer Data Types

You can apply most of the arithmetic operators to real arrays of the following integer data types:

int8 and uint8
int16 and uint16
int32 and uint32
int64 and uint64

All operands must have the same integer data type and MATLAB returns an answer of that type.

Note Except for the unary operators +A and A.', the arithmetic operators do not support operations on complex arrays of any integer data type.

For example,

x = int8(3) + int8(4);
class(x)

ans =

int8

The following table lists the binary arithmetic operators that you can apply to arrays of the same integer data type. In the table, A and B are arrays of the same integer data type and c is a scalar of type double or the same type as A and B.

Operation	Support when A and B Have Same Integer Type
`+A`, `-A`	Yes
`A+B`, `A+c`, `c+B`	Yes
`A-B`, `A-c`, `c-B`	Yes
`A.*B`	Yes
`Ac`, `cB`	Yes
`A*B`	No
`A/c`, `c/B`	Yes
`A.\B`, `A./B`	Yes
`A\B`, `A/B`	No
`A.^B`	Yes, if `B` has nonnegative integer values.
`c^k`	Yes, for a scalar `c` and a nonnegative scalar integer `k`, which have the same integer data type or one of which has type`double`
`A.'`, `A'`	Yes

Combining Integer Data Types with Type Double

For the operations that support integer data types, you can combine a scalar or array of an integer data type with a scalar, but not an array, of type double and the result has the same integer data type as the input of integer type. For example,

y = 5 + int32(7);
class(y)

ans =

int32

However, you cannot combine an array of an integer data type with either of the following:

A scalar or array of a different integer data type
A scalar or array of type single

The section Numeric Classes, under Classes (Data Types) in the MATLAB Programming Fundamentals documentation, provides more information about operations on nondouble data types.

Tips

The arithmetic operators have function equivalents, as shown here:

Binary addition	`A+B`	`plus(A,B)`
Unary plus	`+A`	`uplus(A)`
Binary subtraction	`A-B`	`minus(A,B)`
Unary minus	`-A`	`uminus(A)`
Matrix multiplication	`A*B`	`mtimes(A,B)`
Arraywise multiplication	`A.*B`	`times(A,B)`
Matrix right division	`A/B`	`mrdivide(A,B)`
Arraywise right division	`A./B`	`rdivide(A,B)`
Matrix left division	`A\B`	`mldivide(A,B)`
Arraywise left division	`A.\B`	`ldivide(A,B)`
Matrix power	`A^B`	`mpower(A,B)`
Arraywise power	`A.^B`	`power(A,B)`
Complex transpose	`A'`	`ctranspose(A)`
Matrix transpose	`A.'`	`transpose(A)`

Note For some toolboxes, the arithmetic operators are overloaded, that is, they perform differently in the context of that toolbox. To see the toolboxes that overload a given operator, type help followed by the operator name. For example, type help plus. The toolboxes that overload plus (+) are listed. For information about using the operator in that toolbox, see the documentation for the toolbox.

Examples

Here are two vectors, and the results of various matrix and array operations on them, printed with format rat.

Matrix Operations		Array Operations
`x`	`1` `2` `3`	`y`	`4` `5` `6`
`x'`	`1 2 3`	`y'`	`4 5 6`
`x+y`	`5` `7` `9`	`x-y`	`-3` `-3` `-3`
`x + 2`	`3` `4` `5`	`x-2`	`-1` `0` `1`
`x * y`	Error	`x.*y`	`4` `10` `18`
`x'*y`	`32`	`x'.*y`	Error
`x*y'`	`4 5 6` `8 10 12` `12 15 18`	`x.*y'`	Error
`x*2`	`2` `4` `6`	`x.*2`	`2` `4` `6`
`x\y`	`16/7`	`x.\y`	`4` `5/2` `2`
`2\x`	`1/2` `1` `3/2`	`2./x`	`2` `1` `2`/3
`x/y`	`0 0 1/6` `0 0 1/3` 0 0 1/2	`x./y`	`1/4` `2/5` `1/2`
`x/2`	`1/2` `1` `3/2`	`x./2`	`1/2` `1` `3/2`
`x^y`	Error	`x.^y`	`1` `32` `729`
`x^2`	Error	`x.^2`	`1` 4 `9`
`2^x`	Error	`2.^x`	`2` `4` `8`
`(x+i*y)'`	`1 - 4i 2 - 5i 3 - 6i`
`(x+i*y).'`	`1 + 4i 2 + 5i 3 + 6i`

Diagnostics

From matrix division, if a square A is singular,

Warning: Matrix is singular to working precision.

If the inverse was found, but is not reliable,

Warning: Matrix is close to singular or badly scaled.
    Results may be inaccurate.  RCOND = xxx

From matrix division, if a nonsquare A is rank deficient,
```
Warning: Rank deficient, rank = xxx tol = xxx
```

Matrix Operators in detail

2011-09-07T02:06:00.000-07:00

Matrix Operators

The following table gives a list of all arithmetic operators which work on matrices. They are divided into three categories:

The only new operator you are likely to see is the backslash or right divided operator. The expression a/b is equivalent to the operator b\a in Matlab.

Matrix Operations

Each of these operators works pairs of matrices with appropriate dimensions or matrices and scalars.

Operator	Comments	Help
+	Matrix addition	plus
-	Matrix subtraction	minus
*	Matrix multiplication	mtimes
^	Matrix exponentiation	mpower
/	Right matrix division	mrdivide
\	Left matrix division	mldivide
'	Hermitian transpose	ctranspose
.'	Transpose	transpose

Plus +

If both operands are matrices of the same dimensions, the matrix entires are added elemement wise.

>> [1 2; 3 4] + [2 3; 5 7]  ans =       3     5      8    11

If either operand is a scalar, that scalar is added to each entry element in the other matrix:

>> [1 2; 3 4] + 5  ans =       6     7      8     9

In order to add a scalar r to the diagonal elements of a matrix A, use A + r*eye(size(A)).

Minus -

If both operands are matrices of the same dimensions, the matrix entires are added elemement wise.

>> [1 2; 3 4] - [2 3; 5 7]  ans =      -1    -1     -2    -3

If the first operand is a scalar, each entry in the second matrix is subtracted from that scalar. If the second operand is a scalar, that scalar is subtracted from each element in the first matrix.

>> 5 - [1 2; 3 4]  ans =       4     3      2     1  >> [1 2; 3 4] - 5  ans =      -4    -3     -2    -1

In order to subtract a scalar r from the diagonal elements of a matrix A, use A - r*eye(size(A)).

Matrix Times *

You can take the prodcut of two matrices A and B if the column dimension of the first matrix equals the row dimension of the second. That is, size( A, 2 ) == size( B, 1 ).

Matrix multiplication is defined such that given a column vector v with length equal to the row dimension of B, then we define A*B so that A*(B*v) == (A*B)*v.

>> [1 2; 3 4] * [2 3; 5 7]  ans =      12    17     26    37  >> [1 2; 3 4] * [2 3]'  ans =       8     18

If either operand is a scalar, each entry in the other matrix is multiplied by that scalar.

Matrix Power ^

At least one operand must be a scalar, and any matrix must be square. I have not yet found a use for raising a scalar to a matrix.

>> [1 2; 3 4]^3  ans =      37    54     81   118  >> [1 2; 3 4]^1.1  ans =     1.2742 - 0.0793i   2.3245 + 0.0363i    3.4867 + 0.0544i   4.7609 - 0.0249i

Matrix Right Divide /

A/B is equivalent to (B'\A')'. The column dimensions must be equal, that is, size( A, 2 ) == size( B, 2 ).

>> [1 2; 3 4] / [2 3]  ans =      0.6154     1.3846  >> [1 2; 3 4] / [2 3; 5 7]  ans =      3.0000   -1.0000    -1.0000    1.0000

Matrix Left Divide \

A\B is equivalent to inv(A)*B. The row dimensions must be equal, that is, size( A, 1 ) == size( B, 1 ).

>> [1 2; 3 4] \ [2; 3]  ans =     -1.0000     1.5000  >> [1 2; 3 4] \ [2 3; 5 7]  ans =      1.0000    1.0000     0.5000    1.0000

Complex Transpose '

The complex (or Hermitian) transpose operator ' transposes the the matrix and takes the complex conjugate of each entry in the transposed matrix. A' is equivalent to conj(A.').

If the matrix is real, this is equivalent to the standard matrix transpose.

>> A = [1 2; 3 4; 5 6]  ans =       1     2      3     4      5     6  >> A'  ans =       1     3     5      2     4     6  >> A = rand(2) + rand(2)*i  ans =     0.9501 + 0.8913i   0.6068 + 0.4565i    0.2311 + 0.7621i   0.4860 + 0.0185i  >> A'  ans =     0.9501 - 0.8913i   0.2311 - 0.7621i    0.6068 - 0.4565i   0.4860 - 0.0185i

Transpose .'

The transpose operator .' takes an m × n matrix A and creates an n × m matrix B such that A(i, j) = B(j, i).

>> A = [1 2; 3 4; 5 6]  ans =       1     2      3     4      5     6  >> A.'  ans =       1     3     5      2     4     6

Element-wise Operations

Each of these operators works element wise. If either operand is a scalar r, that scalar is interpreted as the matrix r*ones(m, n) where m and n are the dimensions of the other operand.

Operator	Comments	Help
.*	Element-wise scalar multiplication	times
.^	Element-wise scalar exponentiation	power
./	Element-wise (right) scalar division	rdivide
.\	Element-wise left scalar division	ldivide

Times .*

If both operands are matrices of the same dimensions, the matrix entires are multiplied elemement wise.

>> [1 2; 3 4] .* [2 3; 5 7]  ans =       2     6     15    28

If either operand is a scalar, the operation is the same as for *.

This can be used, for example, in selection:

>> A = [1 2 3; 4 5 6; 7 8 9]; >> B = [0 1 0; 1 0 1; 0 1 0]; >> A .* B  ans =          0        2        0         4        0        6         0        8        0

Power .^

If both operands are matrices of the same dimensions, the elements in the first matrix are raised to the power of the corresponding elements in the second matrix.

>> [1 2; 3 4] .^ [2 3; 5 7]  ans =             1           8          243       16384  >> [1 2; 3 4] .^ 2  ans =       1     4      9    16  >> 2 .^ [2 3; 5 7]  ans =       4     8     32   128

Right Divide ./

If both operands are matrices of the same dimensions, the elements in the first matrix are divided by the corresponding elements in the second matrix.

>> [1 2; 3 4] ./ [2 3; 5 7]  ans =      0.5000    0.6667     0.6000    0.5714

If either operand is a scalar, the operation is the same as for /.

Left Divide .\

If both operands are matrices of the same dimensions, the elements in the second matrix are divided by the corresponding elements in the first matrix.

>> [1 2; 3 4] .\ [2 3; 5 7]  ans =      2.0000    1.5000     1.6667    1.7500

If either operand is a scalar, the operation is the same as for /.

Data Proceeding Plan for LAI from MODIS and AVHRR

2011-09-06T05:13:00.000-07:00

Theoretically, the LAI data could be fused based on MODIS and AVHRR products(By Quantitative Remote sensing).

However, the quality of product from MODIS could be potentially better compared with that from AVHRR.

Thus, an alternative approach which may correct the data product from AVHRR to MODIS, according to the
accuracy differences between the overlap period during 2000 to 2005.

Moreover, attempt of exacting ground control point value from spatial data may encounter Nan or likewise invalid value, which should be considered afterwards.

During this consideration, it's likely that to replace the invalid data using interpretation in temporal data record could be an efficient and qualified method.

Some thought about spatial data process

2011-04-28T02:25:00.000-07:00

These days, I have access and processed mass of remote sensing data including data from MODIS ,MTSAT and so on.

During this period, I have almost record ever single strategy or approach in process.

While I still can't avoid errors or some inappropriate actions.

I applied IDL/ENVI and ArcEngine/ArcGIS to operate on the data.

The main data process stream includes format conversion, projection,mosaic, extraction and so on.

Since there is some difference between ENVI and ArcGIS system, so data exchange should be extremely carefully concerned.

Most of time, the error occurs in generating new data,like mosaic images into a new one.

So we shall think before we leap.

The detailed information will comes soon.

Mosaic Multi-image Using GeoProcessor in ArcEngine

2011-04-26T04:21:00.000-07:00

As I ever mentioned and discussed in some similar function in ENVI/IDL.

But this time, another approach has been used---GP in ArcEngine.

GeoProcessor is powerful tool, seems I ever discussed in another article before.

This time, I have applied it in mosaic.

Compared with IDL, ArcEngine has exposed much more parameter to the developers.

And I setup some of them to guaranteeing the quality of the final result.

For further discussion,please send the feedback!

Time Schedule for Graduation Project Essay

2011-04-20T21:53:00.001-07:00

As time goes, the coming graduation project needs to be done in the next month meeting to the requirement of our department.

So according to my current status of progress of that, a schedule time table a made comes as following:

Finishing framework of paper by 23th, April;
Finishing all detail analysis work by 27th, April;
Finishing photos, tables, and so on by 28th, April;
Finishing references ,thanks note and so on by 30th, April;

Attention should be paid upon the logistical construction and reasonable descriptive explanation.

Apply Mask for image using shapefile in ENVI or IDL

2011-03-30T01:28:00.000-07:00

If your images are all of the same size I think that you can solve
your problem simply by using ENVI.

1: Open one of your images and load it in a display
2: Convert your shapefile to an ENVI vector file (EVF) by clicking
> Vector > Open Vector File > Shapefile
3: Load your vector on the display where you have loaded your
image (In the "available vectors list" select the vector, then click
on "load" and then on "Diplay# (your_display)"
4:In the "Vector Parameters" window that appears, select "File >
Export active layer to ROI"
5: Select "basic tools > masking > build mask" and select your
display
6: Select "Options > Import ROIS", then select the ROI you created
from your vector and click on "OK". Next, choose a name for the mask
file that you want to create and click on "Apply". Now, in the
"available bands list you should see a new image, which has value 0
outside the vector and 1 inside it.
7: Now, you can use this "0-1" image as a mask for your image. Go
to Basic tools > apply mask. Select one of the images that you want to
mask, then click on "Select mask band" and select the mask image.
Click on "OK".

If you have many images and you don't want to repeat n-times the
last step you can use an IDL procedure that automatically opens the
images and apply the mask on each one.
I think that something like this should work: it's a simple
modification of the example program that you can find on the ENVI
User's guide.

PRO multiple_mask

; Select input files (Select all your images)

files_list = DIALOG_PICKFILE(/READ,title ='Select input$

files',/multiple_files)

; Select the mask File

mask_file = DIALOG_PICKFILE(/read, title = 'Select Mask File')

ENVI_OPEN_FILE, mask_file, r_fid=m_fid

; Count the number of files

num_files = N_elements (files_list)

FOR count = 0, num_files-1 DO BEGIN

; Selects the n- file. Output file name is

"input_file_name"+"-masked"

in_file = files_list [count]

out_file =files_list [count]+'-masked'

ENVI_OPEN_FILE, in_file, r_fid=fid

IF (fid EQ -1 OR m_fid EQ -1) THEN RETURN

; get some useful information and set the output filename.

ENVI_FILE_QUERY, fid, ns=ns, nl=nl, nb=nb, bname=bname

; Set the keyword parameters

dims = [-1l, 0, ns-1, 0, nl-1]

pos = LINDGEN(nb)

m_pos = [0]

; Call the 'doit' to apply the mask

ENVI_MASK_APPLY_DOIT, fid=fid, pos=pos, dims=dims,m_fid=m_fid,$

m_pos=m_pos, value=0, out_name=out_file, in_memory=0, $

r_fid=r_fid

ENDFOR

END

how does no_copy work in IDL

2011-03-18T02:54:00.000-07:00

Uh, well, that's because what you are passing into and out of procedures and functions by arguments and keywords *IS* a pointer.That is to say, a variable in IDL is, among other things, a

C pointer. This is what is passed into a method like SetProperty:

myobject->SetProperty, Data=thisData

The variable thisData is, essentially, the pointer to the data. We say the data is "passed by reference", meaning that what the procedure received was the actual physical address of the data

in memory (I.e. the pointer to the data). If the data is copied before it is passed into the procedure, we say it is "passed by value". For example, to pass this data by value we could create \an expression. Expressions are passed by value:

myobject->SetProperty, Data=thisData * 1

(An IDL variable is actually a structure that contains information about the size and type of data, etc. as well as the actual C pointer to the memory location of the data. So it is a little more complicated than saying a variable is a pointer.)

A few comments, though: When an operation like

myobject->setproperty,data=thisdata

is performed, the setproperty method will usually end up making a copy of the data set, even though it receives the data by reference. There will be one copy inside the object, and one

copy in the calling routine, still available after the call.

The way to avoid this is by either using a /no_copy keyword,or (more generally available) to use the call

myobject->setproperty,data=temporary(thisdata)

After this call, "thisdata" will be undefined, and the data will not have been copied. Likewise, one could avoid a lot of copying in David's second example, by using

myobject->SetProperty, Data=temporary(thisData) * 1

bearing in mind, though, that thisData will be undefined after the call.

Where No_Copy is useful is when you are transferring some

information from one memory location to another. For example,

from a local variable in an event handler to the user value

of the top-level base, or from a local variable to an IDL

pointer (heap variable). These operations actually copy the data to another memory location, unless you tell IDL not to with the NO_COPY keyword. Then all IDL transfers is the

C pointer to the data that already exists in memory. Though it's simpler to use the temporary() function, as in:

out_data = temporary(internal_data)

local_data = temporary(*in_data) ;; Shorthand to avoid *'es

*in_data = temporary(local_data) ;; Put it back

or similar.

I've never had occasion to need or use NO_COPY, except where they are already supplied by IDL. Pass variables, or pass pointers to variables, and you will be fine.I guess all (?) instances of /no_copy could be replaced by using the temporary() function instead...?

比地震更可怕的人性

2011-03-12T07:29:00.001-08:00

晚上看电视，发现了奇怪的现象：CCTV在报导日本地震时实在是太给力，我怎么没见你报导汶川地震给力过啊？汶川地震的时候，你用语多么委婉，镜头就没什么恐怖的，尽是红歌，人民情绪稳定；人家日本地震了，镜头搞得2012一样，我眼瞎居然看到了毁灭这两个字，主持人都敢自己估计死亡人数超过一万了？
我很佩服中国的国民教育，面对这种自然灾害，对人家日本是个什么态度。不过从镜头和报导中你却可以看出人家用素质平静地展示一个民族的修养。
那些发电祝贺的朋友，你的人性哪里去了？
一个国家的强大，是全民族人民自身素质的提高。
中国教育的失败，就体现在这里了。你灌输给人民的是什么，是人性的泯灭？
不管是教育者还是受教育者，都应该好好想一想，国家强大的出路该怎么走？
历史要铭记，但不能误导。
比地震更可怕的人性的泯灭！

Some announcement and modification about IDL mosaic routine

2011-03-09T04:14:00.001-08:00

I have post an article about mosaic image in IDL recently, without fully validation progress.

Then some error was detected when I took an regular check of the resulting dataset.

I traced the code, and I found that the error source---mapinfo.

I have assumed that those spatial reference should be similar with that in ArcGIS, while it turns out to be not.

Basically, the mapinfo does not only includes the reference, but also includes the starting points coordinate information as well.

So when mosaic has been carried upon the image, the mapinfo should be re-calculate as a consequence,

I have searched the Google for that, fortunately I was guided into the official website of ITT, and they have provided a

solution to this problem. That is absolutely a great hint to us.

The routine is like this:

This is an example of how to call GEOREF_MOSAIC_SETUP and use the calculated parameters in MOSAIC_DOIT.

Example Code:

; program to estimate mosaic parameters

pro georef_mosaic_setup, fids=fids, dims=dims, out_ps=out_ps, $
  xsize=xsize, ysize=ysize, x0=x0, y0=y0, map_info=map_info
compile_opt strictarr, hidden

; some basic error checking
;
if keyword_set(dims) then $
  if n_elements(fids) ne n_elements(dims[0,*]) then dims=0
;
if n_elements(fids) lt 2 then begin
  xsize = -1
  ysize = -1
  x0 = -1
  y0 = -1
  return
endif

; if no DIMS passed in
;
nfiles = n_elements(fids)
if (keyword_set(dims) eq 0) then begin
  dims = fltarr(5, nfiles)
  for i=0, nfiles-1 do begin
    envi_file_query, fids[i], ns=ns, nl=nl
    dims[*,i] = [-1L, 0, ns-1, 0, nl-1]
  endfor
endif

; - compute the size of the output mosaic (xsize and ysize)
; - store the map coords of the UL corner of each image since you'll need it later
;
UL_corners_X = dblarr(nfiles)
UL_corners_Y = dblarr(nfiles)
east = -1e34
west = 1e34
north = -1e34
south = 1e34
for i=0,nfiles-1 do begin
  pts = [ [dims[1,i], dims[3,i]],   $  ; UL
          [dims[2,i], dims[3,i]],   $ ; UR
          [dims[1,i], dims[4,i]],   $ ; LL
          [dims[2,i], dims[4,i]] ]  ; LR
  envi_convert_file_coordinates, fids[i], pts[0,*], pts[1,*], xmap, ymap, /to_map
  UL_corners_X[i] = xmap[0]
  UL_corners_Y[i] = ymap[0]
  east  = east > max(xmap)
  west = west < min(xmap)
  north = north > max(ymap)
  south = south < min(ymap)
endfor
xsize = east - west
ysize = north - south
xsize_pix = round( xsize/out_ps[0] )
ysize_pix = round( ysize/out_ps[1] )

; to make things easy, create a temp image that's got a header
; that's the same as the output mosaic image
;
proj = envi_get_projection(fid=fids[0])
map_info = envi_map_info_create(proj=proj, mc=[0,0,west,north], ps=out_ps)
temp = bytarr(10,10)
envi_enter_data, temp, map_info=map_info, /no_realize, r_fid=tmp_fid

; find the x and y offsets for the images
;
x0 = lonarr(nfiles)
y0 = lonarr(nfiles)
for i=0,nfiles-1 do begin
  envi_convert_file_coordinates, tmp_fid, xpix, ypix, UL_corners_X[i], UL_corners_Y[i]
  x0[i] = xpix
  y0[i] = ypix
endfor

;print, 'fids = ', fids
;print, 'dims = ', dims
;print, 'out_ps = ', out_ps
;print, 'xsize = ', xsize
;print, 'ysize = ', ysize
;print, 'x0 = ', x0
;print, 'y0 = ', y0
;print, 'map_info = ', map_info

; delete the tmp file
;
envi_file_mng, id=tmp_fid, /remove, /no_warning

end

;============================
; example of mosaic_doit calling georef_mosaic_setup

pro example_georef_mosaic
;
;first restore all base save files
;
envi, /restore_base_save_files
;
;Initialize ENVI and send all errors and warnings to the file batch.txt
;
envi_batch_init, log_file='batch.txt'
;
;Open the input files  
;
envi_open_file, 'c:\infile1.img', r_fid=file1_fid
if (file1_fid eq -1) then begin
     envi_batch_exit
     return
endif
envi_open_file,'c:\infile2.img', r_fid=file2_fid
if (file2_fid eq -1) then begin
     envi_batch_exit
     return
endif
;
;Build the necessary keywords by querying the input images
;
envi_file_query, file1_fid, ns=file1_ns, nl=file1_nl, nb=file1_nb
envi_file_query, file2_fid, ns=file2_ns, nl=file2_nl, nb=file2_nb
;
;create fid array
fids = [file1_fid, file2_fid]
;
;create band arrays for POS keyword
pos1 = lindgen (file1_nb)
pos2 = lindgen (file2_nb)
pos = [[pos1], [pos2]]
;
;set output pixel size.  This could also be queried from the input image.
out_ps = [15., 15.]
;
;call georef_mosaic_setup to calculate the dims, xsize, ysize, x0, y0,
;and map_info structure. Pass the FIDs of the two input files and output 
;pixel size you want. If you want to use a spatial subset, 
;pass in the dims already filled out
;
georef_mosaic_setup, fids=fids, out_ps=out_ps, dims=dims, xsize=xsize, ysize=ysize,$
x0=x0, y0=y0, map_info=map_info
;
;setup the see through values and output bands names
use_see_through = [[1L], [1]]
see_through_val = [[0L], [0]]
out_name = 'outimage.img'
;
;Call the doit. Use a background value of 0 and set the output data type to byte.
;
envi_doit, 'mosaic_doit', fid=fids, pos=pos, dims=dims, out_name=out_name, $
r_fid=out_fid, xsize=xsize, ysize=ysize, x0=x0, y0=y0, georef=1, map_info=map_info, $
out_dt=1, pixel_size=out_ps, background=0, see_through_val=see_through_val, $
use_see_through=use_see_through
;
;exit ENVI
;
envi_batch_exit
end

So it has remarkably simplified the whole process, with appropriate accuracy as well

Try it!

What if life lie to me?

2011-03-08T07:32:00.001-08:00

Life has never show its identity to me.
It's just the God, always watching us from there, but never talk.
I have to admit God has been nice to me, I have parents,and now I have met the one for years.
But, what if life or God has lied to me?
Friends doubted, family did,as well.
Friends blessed, brothers did,as well.
But those are not the point.
I am searching, which I'm quite aware of, but i couldn't identify whether I found it or not.
Why I am still confused or feeling lost ?
Maybe there is no God, so there is no damn lying at all.
Maybe I should make voice of my own.

Mosaic Image using IDL/ENVI Routine

2011-03-07T04:16:00.000-08:00

I have ever realized some application based upon ArcEngine provided by ESRI, while I couldn't bear the efficiency.

Then I developed some routines to accelerate the process by IDL.

Before that, I have referred to some papers, and got some help from the forum as well. Those helped me a lot.

Here is some conclusion of my attempts and experiences.

First, according to the ENVI Help, the mosaic routine requires parameters as following:

ENVI_DOIT, 'MOSAIC_DOIT' [, BACKGROUND=integer], DIMS=array, FID=file ID [, /GEOREF] [, /IN_MEMORY] [, MAP_INFO=structure] [, OUT_BNAME=string array], OUT_DT={1 | 2 | 3 | 4 | 5 | 6 | 9 | 12 | 13 | 14 | 15}, OUT_NAME=string, PIXEL_SIZE=array, POS=array [, R_FID=variable] [, /SEE_THROUGH_VAL], USE_SEE_THROUGH=array, XSIZE=integer, X0=array, YSIZE=integer, Y0=array
Some of them are optional, some are not! So we have to prepare those connected with our data.

Basically,if all the images are georeferenced already(Since data without spatially reference would be quite unacceptable).

The background could be assigned directly like:

background=-9999

Dims refers to :

DIMS

The "dimensions" keyword is a five-element array of long integers that defines the spatial subset (of a file or array) to use for processing. Nearly every time you specify the keyword FID, you must also specify the spatial subset of the corresponding file (even if the entire file, with no spatial subsetting, is to be processed).

DIMS[0]: A pointer to an open ROI; use only in cases where ROIs define the spatial subset. Otherwise, set to -1L.
DIMS[1]: The starting sample number. The first x pixel is 0.
DIMS[2]: The ending sample number
DIMS[3]: The starting line number. The first y pixel is 0.
DIMS[4]: The ending line number

Usually, we could retrieve that by:

FOR f=0,SIZE(file,/n_elements)-1,1 DO BEGIN

ENVI_OPEN_FILE, file[f], r_fid=fid

IF (fid EQ -1) THEN BEGIN

RETURN

ENDIF ELSE BEGIN

in_fid[f]=fid

ENVI_FILE_QUERY, fid, ns=ns, nl=nl, nb=nb, DATA_TYPE=dtype,dims=dim

dims[*,f]=[-1,0, ns-1,0, nl-1]

pos[*,f]=0

use_see_through[f] = 1

see_through_val[f] = -9999

ENDELSE

ENDFOR

Of course, we have to initiate the variables like:

in_fid=LON64ARR(SIZE(file,/n_elements))

dims=LON64ARR(5,SIZE(file,/n_elements))

pos=LON64ARR(1,SIZE(file,/n_elements))

use_see_through=INTARR(SIZE(file,/n_elements))

see_through_val=FLTARR(SIZE(file,/n_elements))

The "file" refers to those image you would like to mosaic.

So meanwhile we retrieved fid, pos, use_see_through, and see through_val.

Then we have to calculate the x0,y0,xsize, ysize by:

ul_x=DBLARR(SIZE(file,/n_elements))

ul_y=DBLARR(SIZE(file,/n_elements))

lr_x=DBLARR(SIZE(file,/n_elements))

lr_y=DBLARR(SIZE(file,/n_elements))

FOR f=0,SIZE(file,/n_elements)-1,1 DO BEGIN

ENVI_CONVERT_FILE_COORDINATES, in_fid[f], [0,dims[2,f]], $

[0,dims[4,f]], XMap, YMap,/to_map

ENVI_CONVERT_FILE_COORDINATES, in_fid[0], xf, yf, XMap, YMap

ul_x[f]=xf[0]

ul_y[f]=yf[0]

lr_x[f]=xf[1]

lr_y[f]=yf[1]

ENDFOR

ulx=MIN(ul_x)

uly=MIN(ul_y)

x0=LON64ARR(SIZE(file,/n_elements))

y0=LON64ARR(SIZE(file,/n_elements))

FOR f=0,SIZE(file,/n_elements)-1,1 DO BEGIN

x0[f]=LONG64(FLOOR(ul_x[f]-ulx))

y0[f]=LONG64(FLOOR(ul_y[f]-uly))

ENDFOR

ulx=FLOOR(ulx)

uly=FLOOR(uly)

lrx=CEIL(MAX(lr_x))

lry=CEIL(MAX(lr_y))

ns_result=LONG64(lrx-ulx)+1

nl_result=LONG64(lry-uly)+1

outtiffname=out_m_d_path+'\'+year_str+'\'+outfileprefix+year_str+day_str+tifffileSuffix

pixel_size=[0.05,0.05]

xsize=ns_result*pixel_size[0]

ysize=nl_result*pixel_size[1]

pixel_size refers to your image resolution according to your situation.

Then, the georeference should be set as mapinfo or GEOREF parameter:

We could set the mapinfo from the first image as default georeference,

or, set the GEOREF parameter in the routine.

But it is strongly recommended that you make a test between the two approaches.

The way to retrieve the mapinfo is like:

map_info=ENVI_GET_MAP_INFO(fid = in_fid[0])

Finally we could call the routine:

ENVI_DOIT, 'mosaic_doit',DIMS=dims, FID=in_fid,$

OUT_DT=4,OUT_NAME=outtiffname,PIXEL_SIZE=pixel_size,$

POS=pos,background=-9999,$

XSIZE=xsize,YSIZE=ysize,X0=x0,Y0=y0,$

SEE_THROUGH_VAL=see_through_val,$

USE_SEE_THROUGH=use_see_through,$

MAP_INFO=map_info

So that is the mainstream we have to complete.

And there would be still more details, and I'll update that soon.

Enjoy it!

IDL memory optimization and management

2011-03-03T03:47:00.000-08:00

These are some tips of how to optimize the memory in IDL.

For the detailed explanations of each tip will be coming soon.

1. no_copy

IDL> b=30

IDL> b_ptr=ptr_new(b, /no_copy)

IDL> help,b

B UNDEFINED = <Undefined>

2. empty_ptr 空指针

IDL> empty_ptr=ptr_new(/allocate_heap)

IDL> print,ptr_valid(empty_ptr)

IDL> help,*empty_ptr

<PtrHeapVar3> UNDEFINED = <Undefined>

3. 多重指针引用

IDL> p1=ptr_new(20)

IDL> p2=p1

IDL> *p2=30

IDL> print,*p1

IDL> p1=ptr_new(20)

IDL> p2=p1

IDL> p2=ptr_new(30)

IDL> print,*p1

4. 如果不使用空指针需要检查指针是否有效

p=ptr_new()

...

if (ptr_valid(p)) then begin

*p=new_value

endif else begin

p=ptr_new(new_value)

endelse

下面不用检查

p=ptr_new(/allocate_heap)

*p=new_value

5. 通过释放结构体释放其中所有指针

IDL> state={a:{b:ptr_new(10),c:ptr_new(20),e:{d:ptr_new(30)}}}

IDL> help,/heap

Heap Variables:

# Pointer: 3

# Object : 0

INT = 10

INT = 20

INT = 30

IDL> heap_free,state

IDL> help,/heap

Heap Variables:

# Pointer: 0

# Object : 0

6. 考虑如何调用image

image = findgen(256,256)

state={data:ptr_new(image,/no_copy)}

pstate=ptr_new(state,/no_copy)

help,image[10,20] ;this is an error

% Variable is undefined: IMAGE.

% Execution halted at: $MAIN$

使用下面的命令

IDL> help,(*pstate).data

<Expression> POINTER = <PtrHeapVar324>

IDL> help,(*(*pstate).data)[10,20]

<Expression> FLOAT = 5130.00

7. 关于compile_opt

IDL> a=findgen(20)

IDL> print,a(2)

2.00000

IDL> print,a[2]

2.00000

IDL> compile_opt strictarr

IDL> print,a(2)

% Attempt to call undefined procedure/function: 'A'.

% Execution halted at: $MAIN$

IDL> print,a[2]

2.00000

8. 代码优化——计算一条命令的运行时间

IDL> t0 = systime(/seconds) ;获得运行前的时间

IDL> a = bytarr(100)

IDL> a = a + 1

IDL> t1 = systime(/seconds) ;获得运行后的时间

IDL> print, t1 - t0

0.099999905

9. 代码优化——数组运算

a = fltarr(1000)

i) for a = 0L, n_elements(a) - 1L do a[i] = a[i] + 1; not a good way in IDL

ii) a = a + 1 ; better way

10. 代码优化——使用一维下标引用二维数组，速度更快

IDL> multi =findgen(5,3)

IDL> print,multi ;二维数组与一维数组存储方式一样

0.000000 1.00000 2.00000 3.00000 4.00000

5.00000 6.00000 7.00000 8.00000 9.00000

10.0000 11.0000 12.0000 13.0000 14.0000

IDL> print,multi[8]

8.00000

11. 代码优化——使用temporary

a = temporary(a) + (b > 0) ;快速！！原来的a成为未定义!

12. 代码优化——使用gt, lt等操作符

Example1:

IDL> x = findgen(360) * !dtor

IDL> y = sin(x)

IDL> plot, x, y

IDL> ind = where(y gt 0.5 and y lt 0.8, count)

IDL> print, count

IDL> check = where(y gt 1.0, count2)

IDL> print,count2

IDL> print, check

-1

IDL> plots, x[ind], y[ind], psym=1, color='FFFF00'x

Example2:

IDL> cutoff = 50

IDL> image = DIST(100, 100)

IDL> mask = BYTARR(100, 100)

IDL> for x=0, 99 do $

for y=0, 99 do $

if (image[x, y] gt cutoff) then $

mask[x, y] = 1

IDL> faster_mask = image gt cutoff

Example3:

IDL> for i=0, n_elements(a)-1 do $

If b[i] gt 0 then a[i] = a[i] + b[i] ;1.6msec

IDL> a = a + (b gt 0)*b ;0.14msec

IDL> a = a + (b > 0) ;0.11msec

IDL> a = temporary(a) + (b>0) ;0.09msec

13. 代码优化——注意写自己的算法

result = array * scalar1 * scalar2 循环两次

result = array *( scalar1 * scalar2) 循环一次

14. 代码优化——数据类型转换的问题

d=findgen(10)

result = d + 1 ;由于d是浮点型，运算前1先转为浮点型再加

result = d + 1.0 ;改为1.0，省去数据类型转换的时间

15. where的用法

Function replace_space, str

Compile_opt idl2

Space = (byte(‘ ‘))[0]

Underscore = (byte(‘_’))[0]

Bstr = byte(str)

Ind = where(bstr eq space, count)

If (count gt 0) then bstr[ind] = underscore

Return, string(bstr)

End

Groovesharkers pull together!

2011-02-28T02:43:00.001-08:00

What type of online music service you are obsessed with most?

Wait a second, whatever it is? Forget it, now we have Grooveshark!

Check it out!

www.listen.grooveshark.com

And it makes no differences with those previous services you ever loved?

Unfortunately you might be correct.

Indubitably, its appearance seems plain, but admirable.

I have ever been using Google Music for long, trust me, I really did.

And I even have share some playlists with friends.

The only thing that really matters is that Google Music, until this moment, have not provide a platform for users to exchange info and communicate.

Of course, the good news is that Google have purchased the copyright, so ever single piece of delicious music we enjoy in Google Music is genuine.

Bad news is that Google Music is unavailable if we are not in China.(Seem this is not the problem)

Anyway, Grooveshark fixs all the problem.

With a Grooveshark account, you can search your favorite musics, save them, share them, manage them, enjoy them…

Fresh words about the reschedule of the online music services, which indicate that those websites who provide illegal music download or play without copyright, would be shut down in the end.

Just like VeryCD, copyright would be protected gradually by the commercial industry.

Are you ready for the switch?

Chick me for new music trip!

近期关于GOES卫星的变更

2011-02-24T04:37:00.000-08:00

由于NASA的SWPC项目中的GOES 11号备用卫星出现问题，原计划于2011年九月更换的卫星将提前八个月，也就是在二月底更换。

替换GOES 11号卫星的是GOES目前正在服役的GOES 15号卫星。

由于两颗卫星的所在的经度区域不同，需要进行运作前的轨道校正等等工作，因此会带来一定时间段的特定数据无法收集和发布。

受到影响的数据集主要包括Proton通道等等。

不过NASA对于经度差异给出了解决方案：准备在不久的将来使用GOES 14号卫星继续在135度经度收集数据。

时间数据的存储格式NetCDF

2011-02-21T03:34:00.001-08:00

一篇不错的入门级别文章：

在ArcGIS中可以通过多种方式给空间数据库添加时间属性信息。例如对点、线、面等矢量数据结构的属性表，通过添加属性字段或者关联到时间表来存储时间属性；对于栅格数据来说，可以利用一个Table 存储不同时间获取的栅格图像，记录栅格图符号以及相应的获取时间；还有一种方式就是采用多维数组，将时间作为一维存储，包括NetCDF、HDF格式，主要用于地球、大气和海洋科学。把时间属性添加到空间数据库以后，就可以实现动画图表、时间动画以及追踪分析功能。

在ArcGIS 9.2中，添加了NetCDF这种多维数组来存储时间属性信息。NetCDF可以是N维数组: X维（例如：经度）、Y维（例如：纬度）、Z维（例如：海拔）、时间维、其它维等。NetCDF可以表现的地理现象有多种，包括：温度、湿度、气压、盐度等。NetCDF所表现的地理单元可以是规则格网、不规则格网、或者是点数据。

假设有2005年1月1日、2006年1月1日、2007年1月1日的三幅NDVI影像图。可以采用三维数组的方式存储到空间数据库中，包括X维、Y维和时间维。可以表达的地理现象是：不同的地理位置上，不同时间的NDVI值。“附件一”为利用三维和四维数组存储数据。

如果在三维数组的基础上添加了第四维，例如添加了海拔维，则可以表示的地理现象是：在不同的地理位置和不同的时间点，气压随海拔高度变化而产生的变化。
由于NetCDF是一种灵活的、自描述的、并能表达大量数组数据的格式，因此NetCDF在地球、海洋、大气科学中得到了广泛的应用。数据文件包含元数据信息，不用预读全部数据，可以直接读取某一数据子集，一个写用户和多个读用户可以同时使用一个数据文件。

NetCDF文件是二进制文件。在文件头是元数据信息，用来描述NetCDF的维数、变量和属性，方便用户获取NetCDF数据的含义。NetCDF文件的结构是通过CDL符号来描述，它是ASCII格式。“附件二”为NetCDF的文件格式。

   NetCDF不能直接通过添加数据按钮添加到ArcGIS中，需要利用Toolbox中的工具转化为Raster、Feature图层或者表格才能在ArcGIS中使用。还可以利用工具将GIS数据导出成NetCDF。

转化以后的NetCDF数据与普通GIS数据一样，可以实现很多GIS功能，包括：
   显示：NetCDF显示工具同普通栅格和矢量图层相同
  图表：同其它图表一样以表单为驱动
   动画：多维数据可以通过其中一维制作动画（例如时间、压力、高程等）
   分析：NetCDF图层和图表同普通图层一样工作 (例如对NetCDF点要素创建缓冲区，投影栅格数据、查询表单等等)

--

关于华为收购案的思考

2011-02-18T19:40:00.000-08:00

近期关于华为在美收购案受阻的新闻沸沸扬扬，200万美元的收购居然引起了奥巴马的注意。

涉及国家安全的问题究竟怎么确定界限？
个人认为这是一个非常捉摸不定的话题。

我以前一直担心万一中美关系恶化，Google要是和政府合作，是不是我的所有资料都会灰飞烟灭？
而美国举着国家安全的口号就可以让这么小的一个收购案夭折。
究竟是什么触犯了国家安全？

华为是一个什么样的公司？我不清楚，但是在美国看来，中国的所有企业都有中国政府的印记！
这一点我倒不怀疑，就像Google不敢在大陆安置服务器一样，只要是大陆的，就是政府的。

另外一个角度，其实无非是技术和数据的取舍。
在美国看来，华为可能在某些技术上，例如光纤领域，或者是资金上有一定优势。但是，你要获取我的敏感数据，那是不行的。你可以买我的国债，但是活不能让你干！

相比之下，美国的一些计算机领域公司例如Google就做的很聪明，Google从搜索引擎起家，专门做数据，如今已经是全球最具影响力的互联网企业。Google提供的其他的服务什么的，在我看来，就一个词——数据挖掘。

Goolge采用这种潜移默化的方式进入中国比起华为，还是要精明很多。即使Google中国从大陆推出，也基本没有改变Google前进的脚步。

数据应该来说是技术的结晶，往往比起技术或者硬件本身来讲更加宝贵。

Google的退出，一方面使得中国对美国互联网关闭了大门；另一方面也使得美国对中国所有企业关闭了大门。

这是谁的责任？

单位立体角示例图

2011-02-17T23:43:00.001-08:00

采用SketchUp描述模型中的几何关系是非常直观的，比起简单的平面图，其3D效果更佳逼真。
如下图是单位立体角的示意图：

IDL开发专题-灵活、多样的输入输出

2011-01-27T17:40:00.000-08:00

ESRI 北京看到的一片文章，关于IDL读写的。

1 输入与输出
   IDL中输入输出的基本原则是：“只要有数据，就可以将其读进IDL”，IDL对数据的格式没有要求，也就是任何数据格式都能自由的解析，这使得IDL成为目前功能最强、最灵活的科学可视化分析语言。
1.1 标准输入输出
1.1.1 输入输出函数
   IDL中读写ASCII码文本是很重要的一部分。标准输入是键盘输入，标准输出指的是屏幕输出显示。表1-1中列出了标准输入输出的几个函数。

1.1.2 格式化输出
   数据的格式化输出是按照设计的格式进行输出。IDL支持“%”和“\”。
   标准的输入输出时可以进行格式的自由控制，格式化控制的格式为：[n]FC[+][-][width]，各个字段代表的含义见表1-2



如下面的

例子
  PRO Test_format
  COMPILE_OPT idl2
  ;
  PRINT, FORMAT = '(A6)', '123456789'
  PRINT, FORMAT = '(I2,$)', INDGEN(6)
  ;
  PRINT, FORMAT = '("First", 20X, "Last", T10, "Middle")'
  PRINT, FORMAT = '("First", 20X, "Last", T4, "Middle")'
  ;
  PRINT, FORMAT = '("First", 5X, "Last", TL2, "Middle")'
  ;
  PRINT, SYSTIME(/JULIAN)
  PRINT, FORMAT='(C())', SYSTIME(/JULIAN)
  PRINT, FORMAT='(C(CMOI,"/",CDI))',SYSTIME(/JULIAN)
  PRINT, FORMAT='(C(CHI2.2,":",CMI2.2,":",CSF05.2))',SYSTIME(/JULIAN)
  ;
  PRINT, FORMAT='(%"I have %d monkeys, %s,", " and ", I0, " parrots.")', 10, 'Scott', 5
  PRINT, format='(%"I have \xb1%d monkeys ")', 5
  ;
  print,'0F'X ; 十六进制转十进制
  print,"17 ; 八进制转十进制
  print, 119, format='(B)';十进制转二进制的
  print, 119, format='(O)';十进制转到八进制
  print, 119, format='(Z)';十进制转到十六进制
END

1.2 文件读写
1.2.1 打开文件
IDL从磁盘上的文件读写数据，必须首先把一个逻辑设备号连接到一个特定的文件，然后进行文件操作如打开、关闭和读取等。IDL中的逻辑设备号范围是-2—128，其中1--99是可以用户任意指定的，100-128是根据get_lun和free_lun来管理的，其他是特殊函数专用的。
0是常规标准输入流，一般是键盘。也就是说下面命令是一致的：

IDL> read,x
IDL> read,0,x

-1是常规标准输出流，一般是显示屏幕，下面命令的功能是一致的：

IDL> PRINTF, -1, X
IDL> PRINTF, -1, X

   -2是错误流，一般是显示屏幕。
   文件操作相关的函数见表1-4。

1.2.2 读写ASCII码文件
   IDL在读写该文件分为有两种格式化文件：自由文件格式和确定的文件格式。
   自由文件格式是用逗号或空白（tab键和空格键）分开文件中的每个元素的ASCII文件。确定的格式文件是用格式说明按照给定的规范进行编排的。二者相比后者比前者更正规一些。
1.2.2.1 读写自由文件格式
   IDL下用ReadF从文件中读入自由格式数据，Printf写入自由格式数据到文件中。IDL中读写自由格式文件遵循下面的几个规则：
   1) 如果读入到字符串变量中，那么，在当前行剩下的所有字符都将读入该变量中。
   2) 输入数据必须用逗号或空白分隔（空格键或tab键）。
   3) 输入通过数字变量完成。数组和结构都可作为数字变量的集合。
   4) 如果当前读入行是空的，并且还有变量要求输入，则读取另一行。
   5) 如果当前读入行不是空的，但是没有变量要求输入，则忽略此行剩下的数。
   6) 尽量将数据转换为变量所希望的数据类型。
   7) 复数数据必须有实数和虚数两部分，用逗号分隔，并用括号括起来。
1.2.2.2 读写确定的文件格式
   读写确定文件格式可同样用ReadF和PrintF命令，它们刚才已用于自由格式文件，但现在文件格式已由Format关键字明确声明。（在读写标准输入和输出时，也可将Format关键字用于Read 和Print命令）。
1.2.2.2.1 通用的格式说明符
   通用格式说明符Format关键字的语法参考1.1.2.
1.2.2.2.2 逗号分隔的确定格式
   有时数据文件必须用确定格式书写，以方便它们被其它软件读取。用逗号分隔的数据文件就是这类文件的典型代表。
   Format=’(F10.3, A1, 2x, F10.3, A1, 2x, F10.3)’
1.2.2.2.3 字符串中读取格式数据
   ReadS是一个有用的IDL命令，可以从字符串变量而不是从文件中为自由格式或确定格式读取数据。ReadS运用了和命令Read和ReadF相同的读取格式数据规则。也就是说，使用ReadS 就象从数据文件中读取一样，所不同的是所读的对象是一个字符串变量。
   当大量信息需从文件头部读取时，此命令特别有用。例如，假设ASCII 数据文件的第一行说明了数据文件的行数和列数，此文件头可以从文件中读取，并且可创建一个大小正确的数组来读取数据。

IDL> firstLine=' 10 24500 12 June 1996'
IDL> columns=0
IDL> rows=0
IDL> date=''
IDL> reads,firstline,columns,rows,date
10 24500 12 June 1996
IDL> print,columns;,rows,date
10
IDL> print,rows;,date
24500
IDL> print,date
12 June 1996

1.2.2.3 向导式读取函数
IDL中提供了ASCII_TEMPLATE，QUERY_ASCII和READ_ASCII函数。

asciiFile =currentPath+'\data\ascii.txt'
  ;依次读取
  OPENR,lun,asciifile,/get_lun

  IF lun EQ -1 THEN RETURN;
  tmp = ''
  WHILE(~EOF(lun)) DO BEGIN
READF,lun,tmp
  ENDWHILE

  FREE_LUN,lun
  ;分块读取
  tmp = STRARR(3)
  data = FLTARR(2,4)
  fdata = INTARR(20)
  OPENR,lun,asciifile,/get_lun

  READF,lun,tmp
  READF,lun,data

  FREE_LUN,lun

  ;自定义格式读取
  tmp  =''
  row =0
  columns =0
  type =0
  OPENR,lun,asciifile,/get_lun
  ;
  READF,lun,tmp
  READS,tmp,row,columns,type
  data =MAKE_ARRAY(row,columns,type = type)
  READF,lun,tmp
  READF,lun,tmp
  READF,lun,data

  FREE_LUN,lun

  template =Ascii_template(asciiFile)
  data = Read_ascii(asciiFile,TEMPLATE=template,count = yNum)
  help,data
  print,data.(0)

1.2.2.4 综合应用举例
针对上面介绍的三种读取方式，对同一个数据采用了不同的读取方式进行了读取，附代码。
1.2.3 读写二进制文件
二进制数据文件基本是以一长串的二进制字节存在文件中，这些字节的含义（也就是说，这些字节如何翻译成特定数据类型和结构的）很难描述的，除非刚开始就知道文件写入的是什么内容。二进制数据比格式化数据紧凑得多，经常用于存储大数据文件。有两种命令读写二进制数据，它们与前面用来读取格式数据文件的ReadF和 Print 命令等效。它们是ReadU和WriteU命令。
要读取非格式数据文件，简单定义变量，打开文件读取，并用ReadU命令将字节一个接一个地读入变量中。如果给定了变量的数据类型和组织结构，每个变量按其要求从文件中读出相应的字节数。
1.2.3.1 读写常规二进制文件

IDL> B = BYTARR(192, 192, /NOZERO)
IDL> ; Read the data into B:
IDL> READU, 1, B
IDL> ; Close the file:
IDL> CLOSE, 1
IDL> ; Display the image:
IDL> TV, B

1.2.3.2 读写带头文件的二进制文件
   当数据中含有特定的头文件时，读取的时候有时希望直接跳过，此时需要用Point_Lun函数，该函数支持对文件进行字节定位。
   调用格式为POINT_LUN, Unit, Position，其中Unit是文件打开后的lun，Position一般是需要跳转的字节数。
1.2.3.3 关联变量读写二进制文件
   IDL关联变量是把一个IDL数组或结构变量的组织结构映射到数据文件的内容上。文件被看作是这些重复单元的一个数组。第一个单元的索引号是0，第二个单元的索引号1等等。关联变量不象常规变量那样将整个数据组都存储在内存里。而是当一关联变量被引用时，IDL仅对需要的部分数据执行相关的输入或输出请求，这部分数据就是要读入内存的。
   有时候大型的二进制数据文件都由一系列的重复单元组成。一个IDL关联变量通常是处理这种数据形式的最好方式（有时候是唯一的方式）。
关联变量有下面的几个优点：
? 当该变量被用于表达式时，才产生文件的输入和输出动作。不需要单独的读或写命令。
? 数据集的大小不受内存容量的限制，因为有时它可处理大型的数据集。对于物理存储器来说是太大的数据，通过把此数据分成块就能很容易地处理。
? 不必提前声明用于映射该数据的数组或结构的数量。
? 关联变量是效率最高的I/O形式。
1.2.3.4 向导式读取函数
READ_BINARY和BINARY_TEMPLATE
Read_Binary是直接读取二进制文件的函数，用法可参考下面的语句：

vHeadData = READ_BINARY(FILEPATH('head.dat', $
SUBDIRECTORY=['examples', 'data']), DATA_DIMS=[80,100,57])
IVOLUME, vHeadData

1.2.3.5 例子
   读FY-3的9210AWX文件，FY-2C的卫星黑体亮度温度产品，下面的格式说明是参考国家卫星气候中心的《风云二号C卫星业务产品使用手册》。

PRO test_ReadBinAWX
;
currentPath = sourceroot()
file = currentPath+'data\FY2C_TBB_IR1_OTG_20061130_AOAD.AWX'
;
OpenR, file_lun, file ,/Get_Lun
  ;定位到信息部分
  point_lun,file_lun,20
  HeadLine =indgen(3)
  readu,file_lun,HeadLine
  ;HeadLine[0]-数据的头文件长度 $
  ;HeadLine[]-文件头文件记录数 $
  ;HeadLine[2]-数据的记录数
  point_lun,file_lun,58
  ;定位到信息部分
  BeginDate=indgen(5) ;依次为年月日时分
  EndDate =indgen(5) ;依次为年月日时分
  LatLong=indgen(4) ;依次为左上角纬度经度右下角经度纬度
  readu,file_lun,BeginDate
  readu,file_lun,EndDate
  readu,file_lun,LatLong
  ;
  data = bytarr(HeadLine[2],(HeadLine[0]))
;定位到数据部分
  point_lun,file_lun,HeadLine[0]*HeadLine[1]
  readu,file_lun,data

Free_Lun,file_lun
END

1.3 读写通用格式文件
   IDL可以读写许多常用文件格式，这些文件格式已在表1-5中列出。


1.3.1 图像文件信息查询
   IDL中提供了常用图像文件格式的查询程序。这些程序允许在没有真正读取其数据的情况下，就可以查询图像文件，并可以存取随着图像数据文件一起存储在文件里的元数据（关于数据的一些信息）。
   下面是新的图像查询程序列表：
? Query_BMP
? Query_DICOM
? Query_GIF
? Query_JPEG
? Query_PICT
? Query_PNG
? Query_PPM
? Query_SRF
? Query_TIFF
   所有这些查询命令都是以同样的方式工作。它们都是返回0或1的函数，通过返回值确定是否成功地（返回值为1）读取了图像文件里的元数据。如果它们成功地读取了文件，将保存文件信息的IDL结构变量作为输出命令返回给用户。用户通过存取这个结构里面的字段从而获取文件的有关信息。
1.3.2 JPEG文件操作
1.3.2.1 JPEG读操作
   用Read_JPEG命令就可以读取并显示一个JPEG文件。
   假设正在8位显示器上显示24位的图像，应该使用TV命令里的关键字True：
IDL> TV, thisimage, true=1
1.3.2.2 JPEG写操作
   对图形窗口进行拍照可以用TVRD函数。
IDL> image3D=TVRD(true=1)
   最后，用Write_JPEG命令，将图像用输出为JPEG文件，运行命令：
IDL> Write_JPEG, ‘test.jpg’, image3D, true=1, quality=75
1.3.2.3 举例
   对一个jpg文件进行查询及读取操作。

PRO test_CommonImg
  currentPath =file_dirname(routine_file_path(“test_CommonImg”))
  imgFile =currentPath+'\data\idl.jpg'
  ;文件信息查询
  queryStatus = QUERY_IMAGE(imgFile, imageInfo,type = type)
  HELP,imageInfo,/str
  IF queryStatus EQ 1 THEN BEGIN
CASE type OF
   'BMP': BEGIN
   END
   'JPEG': BEGIN
      jpgStatus=QUERY_JPEG(imgFile, fileinfo)
      IF jpgStatus THEN BEGIN
      HELP,fileInfo,/str
      READ_JPEG,imgFile,data
      TV,data,/true
         image=TVRD(0,0,127,287,/true)
      Write_Jpeg,currentPath+'\data\result.jpg', image,true=1, quality=75
      ENDIF
   END
   ELSE: BEGIN
   END
ENDCASE
  ENDIF
END

1.3.3 HDF数据读入
   HDF有两种基本对象—组（Group）和数据集（Dataset），同时也有一些其他的辅助对象，如数据类型（Datatype）、数据空间（Dataspace）、属性（Attribute）。HDF组织结构有些类似于Windows 中的文件结构，组相当于一个文件夹，数据集相当于一个数据文件，文件夹里面可以包含数据文件，也可以包含其他的组对象，这样嵌套下去，就形成了一个复杂的数据对象，这点正好的对应于HDF逻辑结构的层次性。
HDF文件的操作读取-编辑
   HDF_OPEN: 这个过程首先要打开HDF文件，通过create关键字来确定是读取文件还是创建一个新的文件。
   HDF_DFAN_ADDFDS:  添加文件描述；Optionally， add a file description.
   HDF_DFAN_ADDFID:  添加文件Optionally， add a file annotation.
   对HDF 文件添加数据用到的过程：
   添加8位图像，用HDF_DFR8_ADDIMAGE 或DFR8_PUTIMAGE；
   添加颜色表，用HDF_DFP_ADDPAL或HDF_DFP_PUTPAL；
   添加24位图像，用HDF_DF24_ADDIMAGE 或HDF_DF24_PUTIMAGE；
   添加多数据时，需要用下面的命令：
   HDF_SD_CREATE 或 HDF_SD_SELECT 建立一个SDS或选择一个现有的SDS； HDF_SD_DIMSET 设置维数信息；
   HDF_SD_ATTRSET 设置属性信息；
   HDF_SD_SETINFO 添加一个数据信息；
   HDF_SD_ADDDATA 添加数据；
   HDF_SD_SETEXTFILE移动数据到外部存储文件；
   HDF_SD_ENDACCESS 结束读取SDS；
   添加数据表（vData）时需要用下面的命令：
   HDF_VD_ATTACH 获取一个数据表的标识；to get a Vdata identifier.
   HDF_VD_SETINFO写入数据表的信息；
   HDF_VD_FDEFINE to prepare non-trivial fields (optional).
   HDF_VD_WRITE 写入数据表；
   将数据表(Vdata)添加到数据组合(Vgroup)中时，用下面的命令：
   HDF_VG_ATTACH to get a Vgroup identifier.
   HDF_VG_SETINFO to set the Vgroup name and class (optional).
   HDF_VG_INSERT to add the Vdata to a Vgroup.
   HDF_VG_DETACH to close the Vgroup.
   HDF_CLOSE to close the file

MTSAT分发格式介绍

2011-01-24T21:55:00.000-08:00

MTSAT satellite based augmentation system（MSAS）是由日本民航局为民用航空应用而开发的对GPS进行区域性星基增强的系统。MSAS系统原理与美国WAAS相似，并将与WAAS兼容。MSAS从1996年开始实施，系统建成时将包括两个空间转发器（MTSAT-1和MTSAT-2卫星），2个基准站和2个主控站。覆盖范围为日本飞行服务区，也可以逐步扩展到亚太地区。MTSAT是多功能传送卫星（Multi-functional Transport Satellite）的缩写，MTSAT将为空中飞机提供通信和导航服务，并可为亚太地区的机动用户发送气象数据。按照MSAS计划，1999年发第一颗MTSAT-1多功能卫星，2000年进入初始运行阶段，2004年发MTSAT-2，2005年进入全面运行阶段。但是，1999年11月15日的第一颗MTSAT发射失败，预计2002年才能达到初始运行阶段。

在上图的第二个绿色的方框，代表了裸数据。裸数据是最原始的地面接收数据，它含有满足CCSDS标准（CCSDS 102.0-B-4）的数据包(CADU)。

CADU由两个部分组成：同步字和编码的VCDU。编码的VCDU又分为VCDU和R-S编码。

下面，我着重介绍VCDU的具体格式。

一、 VCDU数据包由两部分组成，数据头和数据区，如下图：

1、 VCDU数据头格式如下：

version number：表示版本号；

VCDU-ID是标识码，S/C ID表示的是航天器ID VC ID表示的是虚拟信道ID；

VCDU counter 表示虚拟信道单元计数；

signalling field 是信号字段，replay flag 是回放标示，spare是备用字段，刚好组成8位；

2、VCDU数据区格式如下：

M_PDU格式头文件格式如下：

M_PDU的数据区是由众多CCSDS包组成的，下面将介绍CCSDS包格式。

CCSDS包包含头记录和数据区两部分，分别介绍如下。

2.1 Source Packet Header

Packet Identification 包识别码；

Version NO 版本号；

Type 类型 1标示常规 0标示检验；

Secondary Header Flag 副导头标记，1标示有副导头，0标示没有；

APID 应用处理程序标识；

Packet Sequence Control 包序控制；

Sequence Flags 顺序标记 01表示第一个，10表示第二个；

Packet Sequence Count 包计数；

2.2 Packet Data field

Application data field 应用数据块

Packet Error Contorl 数据包错误控制

The ever most excellent application of Wiki in GRE preparation

2011-01-17T18:43:00.001-08:00

When I was trying to prepare the GRE test, I found the ever most exciting tools in assimilating knowledge in English.

That is Wikipedia,absolutely!

I will post the coming content about how it may helps us to take GRE down.

Cloud into consideration

2011-01-15T05:04:00.000-08:00

Since the SW concerns about cloud extremely, I have to take the cloud into consideration in my paper.

It covers type, and base height, and dynamic change and some detailed classified of condition.

More attention shall be paid upon this, especially when cloud has taken greater and greater effect in global climate change.

Cloud and Radiation

2011-01-13T21:35:00.000-08:00

The study of clouds, where they occur, and their characteristics, play a key role in the understanding of climate change. Low, thick clouds primarily reflect solar radiation and cool the surface of the Earth. High, thin clouds primarily transmit incoming solar radiation; at the same time, they trap some of the outgoing infrared radiation emitted by the Earth and radiate it back downward, thereby warming the surface of the Earth. Whether a given cloud will heat or cool the surface depends on several factors, including the cloud's altitude, its size, and the make-up of the particles that form the cloud. The balance between the cooling and warming actions of clouds is very close although, overall, averaging the effects of all the clouds around the globe, cooling predominates.

The Earth's climate system constantly adjusts in a way that tends toward maintaining a balance between the energy that reaches the Earth from the sun and the energy that goes from Earth back out to space. Scientists refer to this as Earth's "radiation budget." The components of the Earth system that are important to the radiation budget are the planet's surface, atmosphere, and clouds. The energy coming from the sun to the Earth's surface is called solar energy. Most of it is in the form of radiation from the "visible" wavelengths, i.e., those responsible for the light detected by our eyes. Visible radiation and radiation with shorter wavelengths, such as ultraviolet radiation are labeled "shortwave." Both the amount of energy and the wavelengths at which energy is emitted by any system are controlled by the average temperature of the system's radiating surfaces, plus the emission properties. The temperature of the sun's radiating surface, or photosphere, is more than 5500°C (9900°F). However, not all of the sun's energy comes to Earth. The sun's energy is emitted in all directions, with only a small fraction being in the direction of the Earth.

Energy goes back to space from the Earth system in two ways: reflection and emission. Part of the solar energy that comes to Earth is reflected back out to space in the same, short wavelengths in which it came to Earth. The fraction of solar energy that is reflected back to space is called the albedo. Different parts of the Earth have different albedos. For example, ocean surfaces and rain forests have low albedos, which means that they reflect only a small portion of the sun's energy. Deserts, ice, and clouds, however, have high albedos; they reflect a large portion of the sun's energy. Over the whole surface of the Earth, about 30 percent of incoming solar energy is reflected back to space. Because a cloud usually has a higher albedo than the surface beneath it, the cloud reflects more shortwave radiation back to space than the surface would in the absence of the cloud, thus leaving less solar energy available to heat the surface and atmosphere. Hence, this "cloud albedo forcing," taken by itself, tends to cause a cooling or "negative forcing" of the Earth's climate.

Another part of the energy going to space from the Earth is the electromagnetic radiation emitted by the Earth. The solar radiation absorbed by the Earth causes the planet to heat up until it is emitting as much energy back into space as it absorbs from the sun. Because the Earth is absorbing only a tiny fraction of the sun's energy, it remains cooler than the sun, and therefore emits much less radiation. Most of this emitted radiation is at longer wavelengths than solar radiation. Unlike solar radiation, which is mostly at wavelengths visible to the human eye, the Earth's longwave radiation is mostly at infrared wavelengths, which are invisible to the human eye. When a cloud absorbs longwave radiation emitted by the Earth's surface, the cloud reemits a portion of the energy to outer space and a portion back toward the surface. The intensity of the emission from a cloud varies directly as its temperature and also depends upon several other factors, such as the cloud's thickness and the makeup of the particles that form the cloud. The top of the cloud is usually colder than the Earth's surface. Hence, if a cloud is introduced into a previously clear sky, the cold cloud top will reduce the longwave emission to space, and (disregarding the cloud albedo forcing for the moment) energy will be trapped beneath the cloud top. This trapped energy will increase the temperature of the Earth's surface and atmosphere until the longwave emission to space once again balances the incoming absorbed shortwave radiation. This process is called "cloud greenhouse forcing" and, taken by itself, tends to cause a heating or "positive forcing" of the Earth's climate. Usually, the higher a cloud is in the atmosphere, the colder is its upper surface and the greater is its cloud greenhouse forcing.

If the Earth had no atmosphere, a surface temperature far below freezing would produce enough emitted radiation to balance the absorbed solar energy. But the atmosphere warms the planet and makes Earth more livable. Clear air is largely transparent to incoming shortwave solar radiation and, hence, transmits it to the Earth's surface. However, a significant fraction of the longwave radiation emitted by the surface is absorbed by trace gases in the air. This heats the air and causes it to radiate energy both out to space and back toward the Earth's surface. The energy emitted back to the surface causes it to heat up more, which then results in greater emission from the surface. This heating effect of air on the surface, called the atmospheric greenhouse effect, is due mainly to water vapor in the air, but also is enhanced by carbon dioxide, methane, and other infrared-absorbing trace gases.

In addition to the warming effect of clear air, clouds in the atmosphere help to moderate the Earth's temperature. The balance of the opposing cloud albedo and cloud greenhouse forcings determines whether a certain cloud type will add to the air's natural warming of the Earth's surface or produce a cooling effect. As explained below, the high thin cirrus clouds tend to enhance the heating effect, and low thick stratocumulus clouds have the opposite effect, while deep convective clouds are neutral. The overall effect of all clouds together is that the Earth's surface is cooler than it would be if the atmosphere had no clouds.

Calculating the Ångstrom turbidity coefficient and other quantities related to aerosol optical thickness

2011-01-03T03:57:00.000-08:00

Optical thickness is a measure of the amount of direct sunlight reaching a detector that responds (theoretically) to a single wavelength of light. (In practice, all detectors respond to a range of wavelengths.) Optical depth is another commonly used name for the same measure -- these two terms are interchangeable.

Optical thickness (or optical depth) is affected by molecular (Rayleigh) scattering, gaseous absorption, and absorption or (mostly) scattering by aerosols. The portion of optical thickness due to aerosols is called aerosol optical thickness or aerosol optical depth, τ.

τ can easily be related to percent transmission of direct sunlight, T, which may be conceptually easier to understand than τ itself:

(1) T = 100•exp(-τ)

Wavelength, optical thickness, and atmospheric turbidity (haziness) are related through Ångstrom's turbidity formula:

(2) τ = β•λ^-α

where β is Ångstrom's turbidity coefficient, λ is wavelength in microns, and α is the Ångstrom exponent. α and β are independent of wavelength, and can be used to describe the size distribution of aerosol particles and the general haziness of the atmosphere. For two different wavelengths,

(3a) τ₁ = β•λ₁^-α
τ₂ = β•λ₂^-α

from which

(3b) τ₁/(λ₁^-α) = τ₂/(λ₂^-α)

Solving for α:

(3c) α = ln(τ₁/τ₂)/ln(λ₂/λ₁)

A typical range for α is 0.5-2.5, with an average for natural atmospheres of around 1.3±0.5. Larger values of α, when the τ value for the larger wavelength is much smaller than the τ value for the smaller wavelength, imply a relatively high ratio of small particles to large (r > 0.5 µ) particles. As τ for the larger wavelength approaches the τ for the smaller wavelength, larger particles dominate the distribution and α gets smaller. It is not physically reasonable for the τ value of the larger wavelength to equal or exceed the τ value of the smaller wavelength.

Now calculate β from either wavelength:

(4) β = τ₁•λ₁^α = τ₂•λ₂^α

where λ must be expressed in microns (500 nm = 0.500 μ). β values of less than 0.1 are associated with a relatively clear atmosphere, and values greater than 0.2 are associated with a relatively hazy atmosphere.

Given τ at two different wavelengths, the τ at a third wavelength can be inferred for the same atmospheric conditions. Rewrite (4) and solve for τ₃ using either the first or second wavelength:

(5) ln(λ₃/λ₁)α = ln(τ₁/τ₃) = ln(τ₁) - ln(τ₃)
ln(τ₃) = ln(τ₁) - ln(λ₃/λ₁)α
τ₃ = exp[ln(τ₁) - ln(λ₃/λ₁)α]

This calculation is useful when τ values determined with one instrument must be compared to values from another instrument that uses different wavelengths.

Here's a worked-out example for wavelengths used in the GLOBE sun photometer:

λ₁ = 505 nm, λ₂ = 625 nm
τ₁ = 0.185, τ₂ = 0.155
α = ln(0.185/0.155)/ln(625/505) = 0.8299
Using the first wavelength, β = 0.185•0.505^0.8299 = 0.1049
Find τ for a wavelength of 550 nm:
τ₅₅₀ = exp[ln(0.185) - ln(550/505)•0.8299] = 0.1723

References

(author unknown) WMO Background Air Pollution MONitoring (BAPMON) Network Information Manual, TD-9789, September, 1990.

Iqbal, Muhammad. An Introduction to Solar Radiation. Academic Press, Toronto, 1983.

OMI数据注意事项

2011-01-02T21:43:00.000-08:00

Important Information for OMI Data Users

Row Anomalies

Several row anomalies have occurred in the recent past. These anomalies affect the quality of the Level 1B and Level 2 data products. Please read this information carefully prior to using OMI data. Please respect the dates mentioned as the anomalies have occurred recently.

Anomaly 1: Since June 25th, 2007, cross-track scenes 53-54 (0-based).
Anomaly 2: Since May 11th, 2008, cross-track scenes 37-44 (0-based)
Anomaly 3: Since January 24th, 2009, cross-track scenes 27-44 (0-based).

Row Anomaly Behavior

Since "anomaly 3" the row anomaly exhibits a more dynamic behavior than before. Which rows are affected and to what extent varies with time. Please visit the "Overview" section of ourdetailed technical information page for the most up to date information and details.

Row Anomaly Corrections

The anomalies are known to the OMI team and are currently under investigation to examine whether corrections for the effects can be implemented in the Level 1b data. Please visit this information page regularly for updates on the status of corrections implemented and visit ourdetailed technical information page.

Row Anomaly Flagging

The Level 1B data are partially flagged for the anomalies listed above. We expect to provide appropriate flagging for the anomalies in the near future. Please visit our detailed technical information page for details on the current flagging status of the Level 1B and Level 2 products.

Recommendations To Users

At the moment no corrections have been implemented in the operational Level 1B and Level 2 data. It is recommended not to use the affected cross-track scenes. Please respect the dates mentioned above. All other OMI data, meaning other cross track scenes and earlier observations, is of optimal quality. Level-3 products are being produced after filtering for the cross track scenes mentioned per anomaly.

MODIS数据索引相关

2011-01-02T21:40:00.001-08:00

MODIS数据的投影和分块

MODLAND Sinusoidal Grid

10 degree Tiles

There are 460 non-fill 10 deg. by 10 deg. tiles in the grid.
The tile coordinate system starts at (0,0) (horizontal tile number, vertical tile number) in the upper left corner and proceeds rightward (horizontal) and downward (vertical). The tile in the bottom left corner is (35, 17).
Dark blue tiles contain only water (no land).
Tile system in black and white.
The boundaries of tiles outside of the grid's image area are not shown.
Table of Tile Bounding Coordinates (10 deg tiles).
Table of Tile G-ring Coordinates (10 deg tiles).
Projection parameters for the General Cartographic Transformation Package (GCTP).