Remote Sensing Image Compression Using Wavelet Support Vector Machine

The high ratio compression of high resolution remote sensing image is an important and urgent technology proplem of the most martial countries. It can infect the ability of using space information. The art of the state method can not fulfil the requirement of the high resolution remote sensing image compression. So, based on the character of high resolution remote sensing image, it becomes an urgent work that a self-adapted high ratio compression algorithm was studied.With the analysis of the characters of high resolution remote sensing image, a novel image compression algorithm that combines Wavelet Support Vector Machines (WSVM) regression and wavelet transform was presented in this paper. WSVM regression can learn dependency from training data and realizes compression by using fewer training point (Support Vectors) to represents the original data. So, we can compress wavelet coefficients based on this feature. Using wavelet transform, image is decomposed into subbands of different scales. The lowest subband is coded using DPCM for its great importance, and the other coarser subbands are compressed by WSVM. For the characteristic of the wavelet coefficients is different in different scales and directions, it is a key problem to design the appropriate organization method of the coefficients. Moreover, effective entropy coding technique is also studied to encode the support vectors and the corresponding weights. Experiment results demonstrate the coding efficiency of the proposed algorithm.The proposed scheme can not only be used in optical remote sensing image, but also be used in the fields of deep space exploring and avagation remote sensing, etc.