Multispectral image compression using hierarchical vector quantization

Abstract

With the increasing performances of remote sensing systems, the problem of image compression for efficient transmission and storage becomes more urgent. There are many high efficiently lossy image compression standards, but they are mainly targeted on the human visual system, and designed to compress three-band (RGB) color images. In this paper, we propose a hierarchical VQ (HVQ) encoder for multispectral remote sensing image, which realizes high coding efficiency and easy codebook specification. Simulation results show that the HVQ encoder outperforms the distortion about 25% than a normal VQ encoder for the same compression rate.