| Recent advances in the field of remote sensing require us to compress abundance of remote sensing image data. Wavelets have been successfully used in compression in all kinds of image. The applications of Integer wavelet transform in remote sensing image compression are explored.Firstly, statistical properties of wavelet coefficients of remote sensing images are studiedby moment method of estimation and x2 test. A useful result about the relationship betweenthe characteristic of subsample and shape parameter is found. Through the x2 test, theestimated distribution functions are proved well fitting for the high frequency subband wavelet coefficients.Secondly, LIFTS, a list-free embedded coding algorithm is realized and revised. LIFTS nearly achieves the efficiency of SPIHT but requiring greatly reduced computational complexity and memory as used in the compression of remote sensing image.Finally, a scheme of the compression of multispectral image is proposed based on three dimension integer wavelet transform and three dimension LIFTS. It can realize lossy or lossless compression. Experimental results reveal the scheme performance well.
|
PDF Full Text Request