A Subspace-Based Blind Extraction and Detection Method for Hyperspectral Sub-pixel Targets

Abstract

A subspace-based hyperspectral sub-pixel target blind extraction and detection method belongs to the field of hyperspectral remote sensing. The method includes: obtaining the spectral signal of the target mineral; extracting the suspected signal by using a blind source extraction method based on the mean square co-prediction error; performing orthogonal processing on the suspected signal, obtaining an orthogonal base, constructing an orthogonal subspace, and then extracting the target The signal is projected into the subspace to obtain an orthogonal subspace projection operator; the detection value of each pixel is calculated through the projection operator, and the pixel with the highest detection value is determined as the target. The present invention uses the subspace-based blind source extraction method to detect and extract the target object, eliminates the influence of the background after extraction, highlights the target, and solves the problem in the prior art that the use of images for mineral surveys is not effective, and furthermore The hyperspectral mineral target detection effect is improved.

Description

technical field [0001] The invention belongs to the field of hyperspectral remote sensing, and in particular relates to a method for blindly extracting and detecting hyperspectral sub-pixel targets based on subspace. Background technique [0002] Hyperspectral imaging technology plays a pivotal role in the detection of material information and plays an important role in the development of remote sensing. Hyperspectral target detection is a key technology for identifying ground objects from hyperspectral images. Its essence is to judge whether the target is Included in hyperspectral pixels. In the process of sensor imaging, when the target fills the entire pixel, the target is a pure pixel target, which can be easily detected by classical detection methods; when the target size is smaller than a pixel, it is a sub-image The form of pixels exists, and the detection at this time becomes the problem of sub-pixel target detection. [0003] Common models in the development proce...

AI Technical Summary

Problems solved by technology

[0004] The present invention proposes a subspace-based hyperspectral sub-pixel target blind extraction detection method to solve the problem in the prior art that the use of images for mineral surveying is not effective, and then realizes the improvement of the hyperspectral mineral target detection effect the goal of

Images