开源软件名称（OpenSource Name）：

开源软件地址(OpenSource Url)：

开源编程语言(OpenSource Language)：

开源软件介绍(OpenSource Introduction)：

MATLAB Code for the Research in Graduate School

What is Synthetic-aperture radar (SAR)?

Synthetic-aperture radar (SAR) is a radar that create two-dimensional images of landscapes under any weather circumstance.

And, what is polarimetric SAR (PolSAR)? Note that objects reflect radar waves with different intensities with respect to the polarization of the incident waves. So, we can obtain the information of the objects by analyzing the characteristic of reflect waves.

SAR Wiki: https://en.wikipedia.org/wiki/Synthetic-aperture_radar

PolSAR Wiki: https://en.wikipedia.org/wiki/Polarimetry

Folder - VelocitySAR

The first research topic in graduate school. The research is aimed to detect the velocity of moving target in the radar image. It is known that the moving target makes images blur in synthetic aperture radar (SAR) image. To fix blurring in image, we need the information of velocity of the target.

Folder - PolSAR

The source code of radar image toolbox for polarimetric synthetic aperture radar (PolSAR) was including in this folder. They are inside the folders PolSAR/src/@PolSAR_AnalyzeTool and PolSAR/src/@SeaIce, respectively. The toolbox was implemented to generate the experiment results for the master thesis.

PolSAR_AnalyzeTool

PolSAR_AnalyzeToolis the class for analyzing the characteristic of PolSAR images. It wraps the radar images, classification results as data member, and implements the member function for analyzing PolSAR images.

1. The conventional radar image analysis algorithms were implemented as class member function, including eigenDecomposition(), fourComponentDecomposition() and pauliDecomposition(). The conventional radar image analysis algorithms were used to classified the objects in the image.

2. We can assess polarimetric characteristic of the object in the image by using the member functions para<name>(), including paraCoPolCorrelation(), paraMoisture() and paraRoughness1(), etc. For example, we can find the level of roughness of the ground by calling paraRoughness1().

3. The machine learning algorithms, including support vector machine (SVM), fuzzy c-means (FCM) and k-means were wrapped in the member functions, mySVM(), myFCM() and myKmeans(), respectively. By calling these function, the supervised model, SVM, will trained on image and then show the classified results; the unsupervised model, FCM and k-means will show the classified results.

Note: The deep learning model, convolutional neural network (CNN), for PolSAR image was implemented by python which is in repository.

How to Use

• Initialization a PolSAR_AnalyzeTool class object as

img_a = PolSAR_AnalyzeTool()

• To apply "four-component decomposition", a kind of radar image analysis, on the radar image stored in img_a, call the member function fourComponentDecomposition() as

img_a.fourComponentDecomposition()

• To find the usage of a member function, type

> help img_a.fourComponentDecomposition

in the console. Then, the usage of fourComponentDecomposition is shown as

> help img_a.fourComponentDecomposition
> --- help for PolSAR_AnalyzeTool/fourComponentDecomposition ---

  fourComponentDecomposition implements the original 4-component decomposition.
 
  syntax:
  fourComponentDecomposition(varargin)
  Specifies function properties using one or more Name,Value pair arguments. 
  Name-value pair settings apply to all the lines plotted. 
 
  Inputs:
  Name-Value Pair Arguments:
     ('vv', s) - The pixel above SNR threshold is marked. If s is ture, 
                 function will show the figure.
     ('hh', s) - The pixel above SNR threshold is marked. If s is ture, 
                 function will show the figure.
     ('hv', s) - The pixel above SNR threshold is marked. If s is ture, 
                 function will show the figure.
 
  Example: 
     1.
 
  Other m-files required: plot_para
  Subfunctions: none
  MAT-files required: none
 
 ------------- <<<<< >>>>>--------------
  Author: K.S. Yang
  email: [email protected]
 ------------- <<<<< >>>>>--------------

SeaIce

SeaIce was the derived class from PolSAR_AnalyzeTool. This class was designed for sea-ice classification. The usage is the same as PolSAR_AnalyzeTool, and the member functions, which were specific for image classification, were implemented.

How To Use

• Initialization a class object as

img_s = SeaIce()

Folder - Homework

The MATLAB code used in course homework.