Hyperspectral imagery compression via SLMS filter and multiband lookup tables
Jinwei Song, Zhongwei Zhang, Xiaohong Chen · 2013
Hyperspectral Imagery contains massive spectral and spatial information of the ground radiance. In order to reduce the storage volume and bandwidth requirement of such imagery, a new hyperspectral imagery lossless compression algorithm which uses the Signed Least Mean Square (SLMS) Filter and Multiband Lookup Tables (MLUT) is proposed and implemented in this paper. The experimental results illustrate that SLMS filter produces high compression ratio. However, it is not sensitive to the calibration artifact in the calibrated imagery. The Multiband Lookup Tables algorithm can promote the result of SLMS filter by exploiting the calibration artifacts when compressing the calibrated data. MLUT is implemented and compared with other techniques that take advantage of the calibration artifacts. The results show that MLUT generates compatible performance in the lowest complexity. © 2013 by Binary Information Press.