Hyperspectral rapid target detection method and system

Abstract

The invention relates to a hyperspectral fast target detection method, comprising: preprocessing original spectral data; obtaining multiple sets of background spectral vectors of the original spectral data; spectrum segment, constitute an effective spectrum segment set; according to the above-mentioned effective spectrum segment set, extract hyperspectral data to form n small spectrum segment sets; calculate the spectral differential value of the above small spectrum segment set; calculate the background spectrum differential value, and According to the obtained spectral differential value of the small spectrum set and the background spectral differential value, the target position is roughly extracted; according to the matching between the roughly extracted target position and the background spectrum, the accurate target detection result is obtained, and the hyperspectral fast target detection is completed. The invention also relates to a hyperspectral fast target detection system. The invention can improve the target detection speed while ensuring high precision and low false alarm, and has a high degree of intelligence and good robustness.

Description

technical field [0001] The invention relates to a hyperspectral rapid target detection method and system. Background technique [0002] Hyperspectral technology has the characteristic of combining graphs and spectra. It can detect the shape characteristics and spatial distribution of objects, and can also obtain irradiance and spectral information. Complete accurate detection of targets in the natural background; it has a wide range of applications in the fields of military target reconnaissance and detection, camouflage and stealth, civilian environmental monitoring, and resource management. How to efficiently and quickly find targets in background environments (such as vegetation, deserts, oceans, and snow fields) has become a hot spot and difficulty in hyperspectral target detection. [0003] Traditional hyperspectral target detection methods include RX algorithm, orthogonal subspace projection OSP algorithm, constrained energy minimization CEM algorithm and adaptive cos...

AI Technical Summary

Problems solved by technology

The traditional hyperspectral target detection method has the following shortcomings: with the advancement of technology, hyperspectral images have more and more fine reflectivity spectral information, resulting in strong similarity and high redundancy between adjacent spectral segments. For hyperspectral images, the above algorithms cannot meet the real-time requirements; most of the traditional hyperspectral target detection algorithms require prior spectral information of known targets, and target spectral information is usually difficult to obtain, and factors such as lighting conditions will also affect the prior target The accuracy of the spectrum, therefore, it is difficult to obtain typical target spectral information in practical applications

Images