Efficient Remote Sensing Image Compress Algorithm For Circuit Design

The development on remote data acquisition technology has not only deliveredmuch more detailed remote images to us but also demanded much higher requirementsto data compress technology. The characters of remote sensing images, which includeabundant edges, rich textures and various orientations, have determined theunacceptable performance of natural image coding software like JPEG2000. Thus, aspecial algorithm must be developed for this particular kind of image data. Our countryis strong in space technologies but not in related sensing technologies, among whichlays the remote image coding ability. In order to fully utilize the potentials of satellites,targeted remote image coding chips were imperative, so was the corresponding codingsoftware.This article summarizes the state of art technologies on image coding and proposesa special algorithm for coding detail rich remote images. Aim to adapt to the specialcharacter of remote image, we constructed a special trained KL transform to furtherexcavate redundancy in prediction residual, boosting the compress ability of theproposed algorithm. To adjust to hardware parallel design requirements, we conducted asuitable fast rate estimation method, avoiding abundant algorithm complexity andcircuit design difficulty. With the efficient entropy coding, our algorithm outperformJPEG2000on both objective and subjective evaluation, making good support for thedesign of special chip systems for remote image compression.