FPGA Design Of Two-Dimensional DWT Based On The Lifting Scheme

With the rapid development of remote-sensing technologies and the increasing demand of better Image quality, the resolution and sample rate has become higher and higher, thus, result in the conflict of data volume growing far beyond the communicating capacity between satellites and the earth. Since the available bandwidth of the satellite communication sub-system is limited, there has been a great demand of an effective approach to compress the data before transferring and preserve the information in the high resolution images. FPGA device supplies a perfect hardware platform for image compress because of its high integration and powerfull function. Users could implement their designs by coding with HDL. We can do our designs and do the simulation with PC, we could debuge the design with specific software , these make the designing much easier and reduce the designning time. Implement arithmetic with hardware will enhance the speed .for the satellite image compress system , it's quite advantageous. Discret Wavelet transform(DWT) feeds the need of multiresolution and progression transmission for modern image compression. Especially in a great compression rate ,the DWT's performance is much better than the DCT's, so that the JPEG 2000 standard select the DWT as its core algorithm. The lifting scheme for the wavelet transform is very simple ,and is easy to be implemented with hardware,so the lifting algorithm is a good choise when we use the FPGA device to implement the wavelet transform. This paper introduces the wavelet arithmetic that be used in data compress system.especially introduces the lifting scheme and integer to integer wavelet(IWT) arithmetic.According to the requirement of hardware design,the key technologies in VLSI design of the lifting scheme,including the boundary extension algorithm,filter architecture,IWT,fix point operation and inplace operation are discussed. Then, the design of the two dimension wavelet transform is putted forward with verilog. the structure is fit for the real_time processing of remote image.