Space-based infrared hyperspectral image intelligent compression method

Abstract

The invention discloses a space-based infrared hyperspectral image intelligent compression method. The method comprises the steps that infrared hyperspectral image cubes are determined; multiple narrow bands are selected according to infrared physical characteristics of a target; the target is detected in images corresponding to the selected narrow bands; suspected targets are preliminarily screened through target infrared feature fusion; suspected targets are further screened through multi-frame track association; a central spectrum of each suspected target is acquired; a spectrum on pixels of each suspected target and the central spectrum of the corresponding suspected target are compared to determine a target region, and other regions, except the target regions, are regarded as background regions; and nondestructive compression is performed on the hyperspectral image cubes in the target regions, and destructive degree-controllable compression is performed on the hyperspectral imagecubes in the background regions. According to the method, nondestructive compression is performed on the target regions, destructive compression is performed on the background regions, and therefore non-distortion of target information is guaranteed while a high compression rate is realized.

Description

technical field

[0001] The invention belongs to the intersection field of information processing and remote measurement and remote sensing, and more specifically relates to a space-based infrared hyperspectral image intelligent compression method. Background technique

[0002] Data acquired by infrared hyperspectral satellites for remote sensing of moving targets (such as rockets) or dynamically changing objects (such as fires) need to be downloaded in real time. Due to the limited download bandwidth, its data needs to be compressed at a high rate. Although traditional hyperspectral image lossless compression does not lose target information, its compression rate is low, while traditional hyperspectral image lossy compression has high compression rate but loses target information.

[0003] Therefore, traditional hyperspectral image compression methods cannot meet the data compression requirements of infrared hyperspectral satellites for remote sensing of moving targets or d...

Images