Vector quantization of raw SAR data
Jean-Marie Moureaux, P. M. Gauthier, Michel Barlaud, Pascal Bellemain · 2002
Synthetic aperture radar (SAR) is a microwave imaging system which collects a lot of data to synthetize a radar image by numerical process. In order to reduce the data flow to be transmitted to the land-based receiver, the authors propose a data compression scheme using two vector quantization techniques: full search vector quantization using a codebook designed by the Linde-Buzo-Gray algorithm lattice vector quantization (LVQ). Conclusions are supported by a comparative study between vector quantization (LVQ) and block adaptive quantization (BAQ). All the results show that VQ outperforms BAQ at low bit rates. Furthermore, LVQ because of its low complexity seems very well-suited to SAR data compression.>