Self-adaptive spectral unmixing optical implementation method and system based on fully-constrained least square method

Abstract

In order to solve the technical problem of huge processing data volume of the traditional hyperspectral unmixing algorithm, the invention provides a self-adaptive spectral unmixing optical implementation method and a self-adaptive spectral unmixing optical implementation system based on a fully-constrained least square method. The self-adaptive spectral unmixing optical implementation method and the self-adaptive spectral unmixing optical implementation system utilize a compression coding principle in a computational spectral imaging technology, combine a spectral unmixing algorithm-fully-constrained least square method with a spectral splitting and combining system, regard a spatial target scene as an analysis processing object, and can directly output an unmixing result without carryingout a reconstruction process of a spectral data cube, thereby avoiding the problem of large data volume, significantly reducing the burden of data storage, transmission and processing of the system, and providing an effective technical support for real-time application of spectral unmixing processing.

Description

technical field [0001] The invention belongs to the field of optical technology, and relates to an optical implementation method and system for hyperspectral unmixing, in particular to an optical implementation method and system for directly performing spectral unmixing based on a spectral imager. Background technique [0002] Hyperspectral imaging technology is a multi-dimensional information acquisition technology that combines imaging technology and spectral technology. Its appearance continues the image information chain of optical images from black and white panchromatic images via multispectral to hyperspectral. Hyperspectral imaging technology can detect the two-dimensional geometric space and one-dimensional spectral information of the target at the same time, which has the characteristics of "integration of map and spectrum"; and has achieved a breakthrough improvement in spectral resolution, making it undetectable in broadband multispectral remote sensing. ground o...

AI Technical Summary

Problems solved by technology

Such a huge amount of data not only puts great pressure on data processing, transmission, and storage systems, but also makes it impossible for hyperspectral imagers to achieve real-time or near-real-time applications.

Images