基于dm642的视频图像压缩系统实现及其应用研究信息与通信工程专业论文.docxVIP

哈尔滨工业大学工学硕士学位论文 哈尔滨工业大学工学硕士学位论文 - - II - Abstract Video reconnaissance is an important recon method in the modern war field . Because of the video signal’s high data rate and disadvantage in storage, the research of the video compression system becomes a research focus in the field of video recon, especially on the Near-Space platform. With the requirement of war field video reconstruction, the way to enhance information missed during video compression becomes another key technology. In this thesis, the frameworks and main techniques of H.264 video compression standard, such as the module of Intra or Inter prediction, DCT transform and quantization, and entropy coding. The difference and advantage between H.264 and MPEG are also analysed. A video compression hardware system based on the TMS320DM642 is designed and the H.264 video coding program is transplanted and realized. According to the TMS320DM642’s functions and features, the video compression algorithm is optimized by the steps of system optimization and hot function optimization with linear assembly. To reconstruct high-resolution(HR) images from a sequence of low-resolution(LR) compressed images of H.264 standard, a novel algorithm based on the integer DCT transform-domain quantization noise is proposed. Models of compressed video and integer DCT transform-domain quantization noise are surveyed. Then the reconstruction algorithm under the POCS theory is proposed with experiments and analysis. The video compression system based on TMS320DM642 can work stably, which can achieve a compression rate of more than 20 and a PSNR more than 30dB. The speed of video compression algorithm is improved significantly by the optimization. The SR algorithm has a great improvement in subjective visual quality and low computation complexity, in which PSNR can reach 30dB and iterations are less than 5 times. Keywords: video compression, DM642, super resolution - - PAGE IV - 目