Polarized SAR (Synthetic Aperture Radar) image classifying method

Abstract

The invention discloses a polarized SAR (Synthetic Aperture Radar) image classifying method. The polarized SAR image classifying method comprises the following steps: S1, extracting characteristics of a polarized SAR image as follows: scattering entropy H, anti-entropy A and a scattering angle alpha, wherein an obtained characteristic set (H, A, alpha) is taken as a first characteristic set; S2, after decomposing the polarized SAR image into two sub-aperture images, respectively extracting characteristics of the two sub-aperture images as follows: scattering entropies H, anti-entropies A and scattering angles alpha, thus obtaining two sub characteristic sets (H1, A1, alpha 1) and (H2, A2, alpha 2); S3, subtracting each corresponding characteristic value in the two sub characteristic sets to obtain a set (Delta H, Delta A, Delta alpha) of a difference value of each corresponding characteristic as a second characteristic set; and S4, inputting the first characteristic set and the second characteristic set into a decision-making tree classifying model to obtain the classified result of the polarized SAR image. The polarized SAR image classifying method disclosed by the invention is used, so that precision of classified result can be improved.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a polarization SAR image classification method. Background technique [0002] Due to the different scattering mechanisms of different ground objects on electromagnetic waves, more abundant ground object information can be obtained by analyzing the backscattering characteristics of fully polarized radar, so it has been widely used in ground target detection, ground object recognition, classification and extraction. . At present, in the process of studying the scattering characteristics of ground objects, the analysis of the scattering characteristics of ground objects is mainly realized by decomposing the polarization target of the echo signal containing the electromagnetic scattering characteristics of the target objects. The characteristics reflecting the scattering type of ground objects can be obtained by decomposing the polarization target, and then these characteristics can ...

AI Technical Summary

Problems solved by technology

[0005] 1. Classifiers are based on the establishment of rules for a small number of features, and it is difficult to fully reflect the characteristics of SAR images, so it is difficult to improve the classification accuracy;

[0006] 2. Usually, the maximum likelihood estimation method is used to classify the target, which makes these methods require samples to conform to a specific statistical distribution model, and the generated classification rules are difficult to interpret;

[0007] 3. Generally, it needs to be decomposed into multiple sub-apertures, and features such as the mean value and coherence of multiple sub-aperture images are extracted for classification, and the number of sub-apertures is inversely proportional to the spatial resolution of the ground. Therefore, although the number of sub-apertures increases, the extracted Features (such as mean, coherence) are more accurate, but the ground spatial resolution is reduced due to the number of sub-apertures, which inevitably limits the final classification accuracy

Images