Supervised image segmentation method for hyperspectral image based migration dictionary learning

Abstract

The invention discloses a supervised image segmentation method for a hyperspectral image based migration dictionary learning, which mainly solves the problem of unbalance of classes in hyperspectral image segmentation. The implementation process of the method comprises the following steps of: (1) inputting a target image and an auxiliary image, and extracting features; (2) setting loop termination times, training a classifier by a dictionary learning method for a target domain labeled sample set; (3) calculating a migration sample set; (4) updating a minority class sample set in the target domain labeled sample set; (5) calculating the class labels and the classifier weight in a target domain unlabeled sample set in current loop; (6) calculating the class labels of a final target domain unlabeled sample set; (7) outputting the segmentation result of the target image by the obtained class labels of the final target domain unlabeled sample set and the labels of the target domain labeledsample set. The method has the advantage of being efficient in segmentation of the hyperspectral image with unbalanced classes, and can be used for detection and recognition of a radar target.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to the segmentation of hyperspectral images, which can be used for detection and identification of radar targets. Background technique [0002] Hyperspectral imaging is one of the cutting-edge technologies in the development of remote sensing technology. It is shifting from aeronautical remote sensing to a combination of aerospace remote sensing. effective technical means. Hyperspectral images have the advantages of rich ground detail information, multiple imaging bands, and high correlation between adjacent bands. Many spectral imagers at home and abroad can provide spectral bands with dozens or even hundreds of bands, and the amount of data is huge. For example, a The data volume of the scene image AVIRIS is about 140 megabytes. These imaging instruments make the research on hyperspectral images more reliable and practical. [0003] There are many kinds of ground object...

AI Technical Summary

Problems solved by technology

However, when using a dictionary for sparse representation, attention is paid to whether the distribution of data is similar, and less consideration is given to distance, which reduces the accuracy of supervised segmentation of hyperspectral images when there is a problem of category imbalance.

Images