Infrared spectrum identification method for different targets under spectral feature similarity condition

Abstract

The invention discloses an infrared spectrum identification method for different targets under a spectral feature similarity condition. The method comprises: processing the acquired different target infrared spectrum images without the significant spectrum characteristics; through radiation calibration, denoising, sample collection, dimension reduction with a principal component analysis method, clustering and the like, obtaining clustering centers of different target samples after spectrum dimensionality reduction, calculating a dimensionality reduction spectrum of a to-be-detected spectrum by using the same processing method during target identification, and comparing the dimensionality reduction spectrum with a known clustering center so as to realize identification of different targetswith similar spectrums. The method can be applied to spectrums of different targets with similar contours and no significant characteristic absorption. The method can extract dimension reduction spectrums with significant difference characteristic information, realize the distinguishing of different targets according to the difference between the dimension reduction spectrums, calculate the distances between the different targets by using the dimension reduction data obtained by the to-be-detected spectrums, and find out the category most similar to the to-be-detected spectrums, thereby achieving the purpose of target identification.

Description

technical field [0001] The invention belongs to the technical field of infrared detection and spectral image processing, and in particular relates to an infrared spectrum identification method for different targets under the condition of similar spectral features. Background technique [0002] With the rapid development of infrared detection technology, infrared spectral imaging technology has been widely used in target detection and recognition because of its characteristic of multi-mode detection of spectral information and spatial information. For example, in the fields of geological exploration, atmospheric remote sensing, identification of true and false targets and interference objects, etc., it has become effective and focuses on the development of cutting-edge technologies. [0003] In recent years, the use of infrared spectral imaging technology to detect and identify different targets has become a key technical means for the application of infrared imaging recognit...

AI Technical Summary

Problems solved by technology

[0008] However, in practical applications, different targets have low signal-to-noise ratios, similar spectral profiles, and no significant spectral feature information such as absorption peaks under strong backgrounds. Traditional identification methods based on spectrally significant features cannot be effective. to detect and identify different targets

Images