Polarization SAR ground object classification method combining rotational domain polarization coherent characteristics

Abstract

The invention belongs to the field of complete polarization synthetic aperture radar imaging remote sensing technology and relates to a polarization SAR ground object classification method combining rotational domain polarization coherent characteristics. The method comprises the steps that (S1) a polarization SAR image to be classified is selected; (S2) coherent speckle filtering is performed on the polarization SAR image, and a polarization scattering matrix of all pixel points in the polarization SAR image obtained after filtering is obtained; (S3) polarization characteristic parameters are extracted, wherein the polarization characteristic parameters corresponding to all the pixel points in the polarization SAR image are extracted; (S4) the polarization characteristic parameters are normalized; (S5) a training sample and a test sample are selected; (S6) an SVM classifier is trained; (S7) classification processing is performed, and respective classification results of all the pixel points in the polarization SAR image are obtained; and (S8) classification precision is calculated. The method is easy to implement, has high robustness for polarization SAR data acquired through different sensors, and is used for performing high-precision ground object classification processing on the polarization SAR image.

Description

technical field [0001] The invention belongs to the technical field of full-polarization synthetic aperture radar (Synthetic Aperture Radar, SAR) imaging remote sensing, and relates to a polarization SAR ground object classification method combined with rotation domain polarization coherent features. Background technique [0002] Polarization SAR has all-weather and almost all-weather working capabilities, and obtains the full polarization scattering information of the target by sending and receiving electromagnetic waves with orthogonal polarization states. Land object classification is a common basic problem in application fields such as crop growth monitoring, rural and urban land census, and environmental monitoring. It is also an important application direction for polarimetric SAR image understanding and interpretation. The purpose of polarimetric SAR surface object classification is to use the polarimetric measurement data obtained by airborne or spaceborne SAR sensor...

AI Technical Summary

Problems solved by technology

This phenomenon causes a lot of inconvenience to the processing and application of polarization information of full-polarization radar, and is one of the technical bottlenecks in the fine interpretation and quantitative application of the polarization scattering mechanism of current full-polarization radar targets.

Images