POLSAR image unsupervised classification method based on target scattering identification

Abstract

The invention provides a POLSAR image unsupervised classification method based on target scattering identification. The POLSAR image unsupervised classification method comprises the steps of: 1, calculating a POLSAR image polarization scattering entropy and similarity parameters of surface scattering, even time scattering and volume scattering and initially dividing classes of POLSAR images by using the parameters; 2, selecting the minimum antenna receiving power characteristic polarization of surface features mainly referring to surface scattering, and with the minimum antenna receiving power characteristic polarization as an antenna polarization state, calculating an antenna receiving power of each pixel; 3, calculating the center of each class; 4, calculating a polarization scattering difference measurement of each pixel, judging the pixel as a class with the minimum difference measurement; and 5, checking whether the ending condition is reached, and if not, returning the step 4. According to the POLSAR image unsupervised classification method based on target scattering identification, the scattering of the ground features can be more accurately described and can well correspond to the actual scattering condition of the ground features, thus the operation time of class regulation is reduced.

Description

technical field [0001] The invention relates to the technical field of image classification, in particular to an unsupervised classification method for POLSAR images based on target scattering identification. Background technique [0002] Polarization Synthetic Aperture Radar (POLSAR) has become the most promising remote sensing device today because of its high-resolution, all-weather, wide-range imaging capabilities, and the ability to obtain complete target information characterization data. Nowadays, with the continuous emergence of various spaceborne and airborne POLSAR systems, the number of POLSAR images is increasing rapidly. How to quickly and accurately interpret these image data will be a hot issue of general concern in the field of remote sensing in the future. [0003] POLSAR image classification is an important research content of POLSAR image interpretation. According to whether there is manual intervention in the classification process, POLSAR image classific...

AI Technical Summary

Problems solved by technology

However, the premise of this method is that the POLSAR image obeys the Wishart distribution, which is usually unreasonable due to the complexity of the actual scene

At the same time, in the process of category adjustment based on this distance, frequent use of matrix inversion operations and logarithmic operations will result in a large amount of calculations and reduce the practicability of the algorithm

Images