A weak small target detection method based on a hyperspectral image

Abstract

The invention discloses a weak small target detection method based on a hyperspectral image, and the specific scheme is as follows: the method comprises the following steps: 1, carrying out target detectability analysis by utilizing a signal-to-clutter ratio, and preferably selecting a plurality of spectrum segments with differences between a target and a background from large to small by calculating the signal-to-clutter ratios of different spectrum segment targets relative to neighborhood backgrounds of the targets; 2, weak and small target extraction is carried out on the spectral band withthe largest difference between the optimized target and the background, the background is restrained through a multi-structure element mathematical morphology method, and a plurality of suspected targets are obtained through self-adaptive threshold segmentation; and step 3, by using the optimal detection spectrum band spectrum information, taking the aliasing spectrums of the target and the background in different scenes as standard spectrums when the target is located in different scenes, and calculating the similarity between the aliasing spectrums of the suspected target and the backgroundand the standard spectrums based on a spectrum angle matching principle to confirm the weak and small target. The invention belongs to the technical field of target detection and recognition, and canrealize efficient confirmation of long-distance weak targets in a complex environment background.

Description

technical field [0001] The invention belongs to the technical field of target detection and recognition, and in particular relates to a method for detecting weak and small targets based on hyperspectral images. Background technique [0002] High-probability detection of weak and small targets in complex cloud backgrounds has always been a key technology in the field of target detection and recognition. However, for the actual air target detection process, long-distance detection is usually required. After the target is imaged by the detection system, it generally lacks geometric shape, texture and other information. The environmental background of the non-target area, especially the complex and changeable cloud background will increase the difficulty of target detection, and with the development of aircraft stealth technology, the radiation intensity of the target itself is greatly reduced, making the target signal easily detected by complex clouds. The background drowns ou...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problems of low detection efficiency of traditional target detection algorithms and difficulty in dealing with target spectral variation caused by spectral aliasing. The present invention provides a weak target detection method based on hyperspectral images

Images