Hyperspectral anomaly detection method based on collaborative representation and anomaly elimination

Abstract

The invention discloses a hyperspectral anomaly detection method based on collaborative representation and anomaly elimination. The method is characterized by comprising the following steps of: removing abnormal pixels in a local background set by adopting an elimination method of fusing global abnormality and local abnormality; and calculating a linear representation coefficient and solving a residual image based on a hyperspectral anomaly detection algorithm of collaborative representation, and finally marking pixels of which the detection values are greater than a threshold value as abnormal points by setting the threshold value to obtain an abnormal target detection result. Compared with the prior art, the method has the advantages that the precision and detection precision of linear representation are improved, and the problems of window selection in a hyperspectral anomaly detection algorithm based on collaborative representation and background set pollution caused by abnormal pixels of double windows are effectively solved.

Description

Technical field [0001] The invention relates to the technical field of hyperspectral anomaly detection, in particular to a hyperspectral anomaly detection method based on window adaptive collaborative representation and abnormal elimination. Background technique [0002] With the development of hyperspectral imaging spectrometers, the rapid development of spectral information of remote sensing images from panchromatic, multispectral to hyperspectral has brought remote sensing technology into a new stage. Hyperspectral images have high spectral resolution, and the obtained image pixels have hundreds of bands of spectral information. Based on the different spectral characteristics of various features, hyperspectral remote sensing has been used in feature classification, quantitative inversion, target detection and ecological environment Monitoring and other aspects are widely used. Hyperspectral image target detection can be divided into two categories according to whether the tar...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to design a hyperspectral anomaly detection method for collaborative representation and anomaly removal in view of the deficiencies of the prior art. It adopts the method of merging global anomalies and local anomalies, and obtains the optimal window and The abnormal elimination of the background set helps to improve the accuracy of the linear representation, thereby improving the detection accuracy. This method first calculates the background saliency to obtain the optimal window size of each pixel, and then the abnormal image in the local background set in the double window Meta-elimination, using the method of fusion of global anomaly and local anomaly to remove the abnormal pixels in the local background set, and then based on the hyperspectral anomaly detection algorithm based on collaborative representation, calculate the linear representation coefficient and obtain the residual image, and finally set the threshold , mark the pixels whose detection value is greater than the threshold as abnormal points, and obtain the abnormal target detection results, effectively solve the problem of window selection in the hyperspectral anomaly detection algorithm based on collaborative representation, and the problem of dual-window abnormal pixels polluting the background set, the method is simple , with high detection accuracy, which further promotes the wide application of anomaly detection based on collaborative representation

Images