Abnormal target detection method and device and computer readable medium

Abstract

The invention relates to an abnormal target detection method and device and a computer readable medium, and the method comprises the steps: carrying out the operation of a to-be-detected hyperspectralimage, so as to obtain a correlation coefficient between any two wavebands in the hyperspectral image; determining the correlation coefficient smaller than a preset correlation coefficient thresholdas a target correlation coefficient; determining a target waveband corresponding to the target correlation coefficient, and determining a target hyperspectral image according to the target waveband; determining a spectral vector of each pixel point in the target hyperspectral image; performing background mean value operation and covariance matrix operation on all spectral vectors in the target hyperspectral image to obtain a background mean value and a covariance matrix; and for each pixel point in the target hyperspectral image, determining an abnormal target detection result according to thespectral vector, the background mean value and the covariance matrix of the current pixel point. According to the method, the detection precision of the abnormal target can be improved.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to an abnormal object detection method, device and computer-readable medium. Background technique [0002] Hyperspectral image is a combination of spectral information reflecting target radiation and image information reflecting two-dimensional space. It has a large number of bands, narrow band width, nanometer-level resolution, and strong correlation between spectra, so it can effectively detect The difference in the spectral characteristics of the target features. However, the improvement of spectral resolution leads to information redundancy, which increases the difficulty of later data processing and causes waste of storage and processing capacity. Therefore, data dimensionality reduction processing for high-dimensional and massive information can be considered to improve data processing speed. [0003] Abnormal target detection is a hot issue in the hyperspectral fi...

AI Technical Summary

Problems solved by technology

At present, the abnormal target detection method usually uses the data information of the hyperspectral image to estimate the statistical information of the target or the background, and the redundant information in the hyperspectral will lead to a high probability of false alarms in the abnormal target detection results, which will reduce the detection of abnormal targets. detection accuracy

Images