Single Spectrum Driven Dual Class Sparse Representation Hyperspectral Image Object Detection Method

Abstract

The invention discloses a hyperspectral image target detection method driven by a single spectrum with dual-category sparse representation, including constructing a target dictionary and a background dictionary, using a given target spectrum as prior information when constructing the target dictionary; when constructing the background dictionary, Classify and select pixels with high representation frequency in each category as background training samples to obtain a global over-complete background dictionary; for the target dictionary and background dictionary, use the dual-category sparse representation model to sparsely represent the pixels to obtain the target sparse vector and the background sparse vector, the pixels to be detected are detected one by one, and the target detection result of the hyperspectral remote sensing image X is extracted. The background dictionary of the present invention is a global background dictionary, so that all background object categories on the global image are well represented; and the dual-category sparse representation method separates the target class and the background class for sparse reconstruction, and efficiently realizes the target in the hyperspectral image. Separation from the background to detect objects of interest.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a hyperspectral image target detection method driven by a single spectrum with a dual-category sparse representation. Background technique [0002] The rapid development of remote sensing earth observation technology and its application has changed the world mode of human cognition to a large extent, and it has become an important technical means to obtain surface information (Griffith, J..(1979). Remote sensing and image interpretation. John Wiley & Sons .). The value of remote sensing earth observation technology in practical applications depends to a large extent on whether the images obtained by satellite remote sensing can provide detailed and rich surface information. Compared with multispectral images, hyperspectral remote sensing images have the characteristics of large number of bands and extremely high spectral resolution. The spect...

AI Technical Summary

Problems solved by technology

This method will make the two sparse vectors interact with each other, and the recovered sparse vector will be less effective, thus affecting the final detection effect

Images