Unmanned aerial vehicle large-view-field hyperspectral image generation method and system

Abstract

The invention discloses an unmanned aerial vehicle large-field-of-view hyperspectral image generation method and system, and the method comprises the steps of converting a to-be-spliced hyperspectral image into the same coordinate system of a reference hyperspectral image in a waveband-by-waveband manner for the reference hyperspectral image and the to-be-spliced hyperspectral image collected by an input unmanned aerial vehicle, and determining an overlapping region of the reference hyperspectral image and the to-be-spliced hyperspectral image; performing spectrum uniformization correction on each wave band of the to-be-spliced hyperspectral image by using the overlapping region; and calculating an optimal suture line of an overlapping region of the reference hyperspectral image and the to-be-spliced hyperspectral image, and fusing the hyperspectral images band by band based on the optimal suture line by adopting a weight pyramid image fusion strategy to obtain a final seamlessly spliced large-view-field hyperspectral image. According to the invention, seamless splicing can be accurately completed in an actual scene in which the overlapping rate of adjacent hyperspectral stripe images is relatively low, a large-view-field hyperspectral image is generated, and the problem of spectrum inconsistency is effectively eliminated.

Description

technical field [0001] The invention relates to image processing technology, in particular to a method and system for generating hyperspectral images with a large field of view of an unmanned aerial vehicle. Background technique [0002] Hyperspectral imagery is a special multi-band image, and each ground object has its unique spectral curve in hyperspectral imagery. Taking advantage of this feature, hyperspectral imagery has been successfully applied to mineral mapping and crop index estimation, etc. Various applications. In most cases, hyperspectral images are acquired by hyperspectral imager satellites. However, remote sensing satellites are photographed above clouds, so most channels in satellite hyperspectral images are susceptible to cloud interference, and such interference is difficult to eliminate. In addition, the inherent characteristics of low spatial resolution and long revisit period make it difficult for remote sensing satellites to obtain high-quality, mult...

AI Technical Summary

Problems solved by technology

Direct stitching of source images will lead to inconsistencies in the spectra of stitching results, which will adversely affect subsequent applications

In addition, in order to improve efficiency, when UAVs collect hyperspectral data, the overlap rate is usually set below 50%, which increases the difficulty of stitching algorithms based on image registration.

Images