Polsar image feature classification method based on dfic superpixels

Abstract

The invention discloses a PolSAR image ground object classification method based on DFIC superpixels. The implementation steps of the present invention are: (1) input polarization synthetic aperture radar PolSAR image; (2) carry out Claude target decomposition to PolSAR image; (3) initialize superpixel information; (4) calculate the feature between pixel points (5) update superpixels; (6) construct convolutional neural network; (7) classify PolSAR images; (8) iteratively cluster DFIC superpixels with decomposed features to optimize classification results. Compared with the superpixel segmentation method of the prior art, the present invention has the advantages of accurate segmentation boundary and compact interior of the superpixel, and has the advantages of high classification accuracy and no influence of noise points compared with the single pixel classification method of the prior art. The invention can be used for the classification of polarimetric synthetic aperture radar PolSAR images.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a polarimetric synthetic aperture radar PolSAR (Polarimetric Synthetic Aperture Radar) image feature based on decomposition feature iterative clustering DFIC (Decomposition Feature Iterative Clustering) superpixels in the field of image classification technology Classification. The invention can be used to classify the ground objects in the image obtained by the polarization synthetic aperture radar PolSAR. Background technique [0002] Polarimetric Synthetic Aperture Radar PolSAR has become one of the research hotspots in the field of remote sensing due to its advantages of all-weather, all-time, and unaffected by weather. Therefore, more and more experts and scholars are devoted to the research of PolSAR images of polarimetric synthetic aperture radar. Among them, the classification of ground objects of PolSAR images of polarimetric synthetic aperture radar is ...

AI Technical Summary

Problems solved by technology

Although this method has achieved good results due to the use of the polarization scattering similarity between pixels, the disadvantage of this method is that in PolSAR images with more speckle noise, Can not overcome the impact of speckle noise on object classification

The disadvantage of this method is that because the method uses pseudo-color images for superpixel segmentation, it fails to take advantage of the scattering characteristics of the image, so the superpixel segmentation effect in complex areas is poor, and the classification effect becomes poor. Difference

The disadvantage of this method is that when generating superpixels, this method does not use the decomposition features of PolSAR images, which makes the superpixel segmentation lines generated during superpixel segmentation inaccurate

Images