Please login first
Classifying UAVSAR PolSAR imagery using target decomposition features
* 1 , 2 , 1
1  Department of GIS/RS, Science and Research branch, Islamic Azad University, Tehran, Iran.
2  School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran.


The changes in the earth's surface significantly increase the natural disasters, resulting in severe damage to man-made objects, such as roads, buildings, bridges, and so on. Radar techniques have advantages such as lack of sensitivity to weather conditions, night and day, and cloud cover conditions which can be used to identify, alert, and mitigate these damages. Land use classification due to the importance of these areas and the need to care for them is one of the important applications of remote sensing. Therefore, using polarimetric synthetic aperture radar (PolSAR) images have many capabilities due having the scattering information on the four polarized of HH, HV, VH and VV, and consequently their dependence on the shape and structure of the environment. In this study, UAVSAR image is used. Meanwhile, the support vector machine (SVM) model is one of the well-known classification methods, in addition to being able to run on different types of features from different kinds and in large numbers, which can also distinguish classes those are not linearly separable. On the other hand, it is possible to use data mining method to facilitate data analysis like classification application. In this regards, it is recommended to use random forest (RF) technique. The RF is one of the useful methods for data classification which uses a tree structure for decision making. This method uses strategies to enhance the probability of reaching the goals with conditional probability. In this study, by incorporating a variety of target decomposition methods in PolSAR images, producing the land cover types are generated. Then, using the set of analysis and classification of characteristics, 70 features were obtained by applying SVM, RF, and KNN classification methods. In order to estimate accuracy, the output of these methods was evaluated by reference data.

Keywords: Polarimetric Synthetic Aperture Radar, Classification, support vector machine, random forest, K-nearest neighbors