Research On Near-Lossless Image Compression And Decompression Based On VLSI Architecture

Compression is an effective way to solve the difficulty of storing and transporting the largeamount of digital image data. Near-lossless compression can achieve high image fidelity withrespect to lossy compression and can gain the compression ratio with respect to losslesscompression. More importantly, near-lossless compression can output constant bit-rate consistentwith the communication channel which can ensure the reliable transmission of image data. Whilelarge amount of computation and real-time digital video are getting more simple and convenient byusing the VLSI technology, it has become a trend in the field of video processing.In this paper, the basic principle and current situation of compression are introduced briefly, andthe international video and image compression standard is studied. Take improving the compressionration with guaranteeing the image quality as a starting point, a near-lossless compression anddecompression algorithm are presented. Firstly, the algorithm analyze the image redundancy withthe parser, and then encode the image information. Video encoder is designed after fully research ofbit-plane coding and Huffman coding methods. At the same time high-bit code table and low-bitcode table are proposed based on actual project. Graphic encoder is mainly used in the occasion inwhich the texture is complex but with few color types. The algorithm was modeled with C language,and the results show that the compression ratio and image quality achieve the predetermined effect.Then the hardware architecture is designed and implemented with RTL code according to thealgorithms. The RTL code was simulated and coverage was collected by building coverage drivenverification platform. The results show that the efficiency and area achieve the desired effect andfunction is consistent with the C model, at the same time coverage also meet the requirements.