A High-Speed, Multi-Channel Lossless Compression Algorithm for High-Resolution Video on FPGA

作     者：Du, Xingkai Xie, Longhua Huang, Hongchuan Chen, Huawei Guo, Xiaolong Zhao, Tingyu 

作者机构：Zhejiang Sci Tech Univ Dept Phys Zhejiang Key Lab Quantum State Control & Opt Field Hangzhou 310018 Peoples R China Hangzhou ToupTek Photon Co Ltd Hangzhou 310030 Peoples R China 

出 版 物：《IEEE ACCESS》 (IEEE Access)

年 卷 期：2025年第13卷

页      面：95911-95921页

核心收录：

主　　题：Field programmable gate arrays cameras image compression compression algorithms universal serial bus 

摘      要：With the increasing image resolution, data bit depth, and frame rate of industrial cameras, the signal output speed needs to be improved to keep pace with the rapidly developing image sensor technologies. The USB 3.0 interface has become a common transmission method for industrial cameras due to its high versatility. However, limited bandwidth often requires data compression techniques to address transmission bottlenecks. This paper proposes a four-channel high-speed lossless compression algorithm based on FPGA (Field-Programmable Gate Array). The algorithm uses the difference between adjacent pixels, which has fewer data bit widths, to replace the original pixels for transmission. This approach enhances the transmission frame rate while maintaining image quality. Additionally, a novel first-in, last-out data inversion technique is employed to solve the problem of real-time processing of four-channel data, significantly reducing FPGA hardware resource usage. The performance of the proposed algorithm is evaluated through on-board FPGA experiments. When transmitting 4096 x 4096 resolution video via USB 3.0, the frame rate reaches 18.3 fps with a data size reduced by up to 49.98% (corresponding to a compression ratio of 50.02%), with an encoding overhead delay of 20.76 mu s, resulting in a frame rate increase of approximately 71.03%. Additionally, decompression speeds for a large resolution image set (5568x3712) are improved by at least 40% compared to mainstream algorithms. The algorithm demonstrates low FPGA resource utilization and high processing speed, ensuring efficient compression and transmission even for complex image scenes. These results provide a viable high-speed lossless compression solution for industrial camera applications.