Research And Implementation Of JPEG Decoding IP

JPEG has provided the methods and grammars for compressing static picture. Its main application is to store and transmit static picture through the compressed form, Therefore it has a widespread application in digital picture, digital camera, embedded picture in web and many other domains. Along with the development in mobile termination, multimedia, Internet network, correspondence as well as picture scanning technology, real time processing of image data gets more and more important. Software based data processing usually has difficult to meet the timely requirements of the high-speed realtime systems. Therefore it has been inevitable tendency to realize picture processing algorithm with hardware. It is also hot research topic at present.Firstly, this thesis elaborated the JPEG coding and decoding principle in detail. It mainly includes following several processes, Huffman decoding, inverse quantification, inverse discrete cosine transformation (IDCT), color space transform. It provides the certain theorem foundation for JPEG decoding design.Secondly, a system model was designed with the system level language SystemC for JPEG decoding. This system model is helpful for speeding up the development of the system chip of JPEG coding and decoding.Thirdly, RTL code was designed with HDL language for JPEG decoding, and this model was verified with FPGA chip on the EP2S180 development board. Additionally, the logic synthesis of JPEG decoding was researched based on the SIMC0.35 library in this thesis.Fourthly, Huffman decoding was optimized in the RTL code. That will enhance Huffman decoding speed and reduce the memory consumption. Besides the processing speed of the post-processing module is improved further.