Efficient implementation of FPGA based video compression engine using spatial predictive method for space application

Abstract

In manned mission application, there is a requirement of audio-video conferencing between Crew members and ground stations. Uncompressed graphics, audio, and video data require considerable storage capacity and transmission bandwidth. Since bandwidth is limited to a suitable video compression is to be selected and implemented. H264 is the choice because it is more efficient than other compression standards. There are no space grade chips available for this engine as of today. Thus a suitable space grade implementation is required that can withstand radiation/temperatures requirement in space. FPGA is targeted for the implementation so that along with the H264 implementation other data handling functionality can also be implemented there by reducing mass, power, and volume. Space grades FPGA with very large gate count, internal memories are available (Xilinx Virtex 5). CCSDS based image compression standard is used for compression of the image as well as video data, but high compression ratio is not achieved by this standard. So ISRO planned to use the real time video compression standard for space application. The implementation is also supported the variable aspect ratio of the frame, up to 1024x1024 and variable frame rate (frames per second). This project is targeted on the Vertex 5Q-XQ5VFX130T FPGA and successfully placed and routed FPGA. The full engine is working up to the frequency of 70 MHZ and individual module like DCT, Quantization is working up to 220 MHZ.