一种微纳遥感卫星实时视频压缩的实现与应用

The implementation and application of real-time video compression of micro-nano remote sensing satellite

为满足微纳遥感卫星高速视频数据快速压缩下传的需求,提出一种实时H.264视频压缩的轻量化实现方案.系统硬件主要由经过宇航质量保证的高性能可编程逻辑单元FPGA(Field Programmable Gate Array)和H.264视频编码单元DSP(Digital Signal Processor)异构组成,可在一块VPX标准6U单元模块上实现超高清4K分辨率视频压缩处理.可编程逻辑单元主要完成超高清分辨率视频数据的标准Camera Link接口协议解析、视频帧拆分和标准Video Port接口协议转换.H.264视频编码单元在实时多任务操作系统的调度下主要完成原始视频帧的拼接恢复和H.264实时视频压缩处理.视频压缩质量评价及卫星在轨应用情况结果表明,单个6U视频压缩单元的4K超高清视频实时压缩处理能力达到2.5Gbps(Giga bit per second),在50∶1,20∶1和10∶1三种压缩比下PSNR(Peak Signal Noise Ratio)均优于36.6dB.参照国际电联ITU-R BT-500-11建议书关于图像质量的五级划分,视频压缩质量达到"优"级.相关研究适合应用于微纳视频遥感监视卫星,且已在珠海一号卫星上得到成功应用.

In order to meet the requirement of real-time compression and transmission of the high speed video data derived by micro-nano remote sensing satellite,this paper proposes a small scale implementation method of real-time H.264 video encoding.The system hardware consists of the high performance programmable logic device FPGA（Field Programmable Gate Array）and H.264 encoding unit DSP（Digital Signal Processor）,both with space quality assurance,which can be implemented on a 6 U VPX standard module to support 4 K ultra high definition video compression.The programmable logic device mainly accomplishes the standard Camera Link interface protocol parsing,the video frame splitting,and the standard Video Port interface protocol conversion of the ultra high definition video data.Under the scheduling of real-time multitasking operating system,the H.264 encoding unit mainly accomplishes the video block mosaic and the real-time video compression based on the H.264 algorithm.Theresults of the video quality assessment and the in-orbit application indicate that the compression capacity of a single6 Uvideo compression unit is achieved up to 2.5 Gbps（Giga bit per second）,and the PSNR（Peak Signal Noise Ratio）result is clearly superior to 36.6 dB with the compression ratio of 50∶1,20∶1 and 10∶1.Refer to the five-grade quality scale in Recommendation ITU-R BT-500-11,the＂excellent＂video quality is achieved.This implementation is suitable to apply in micro-nano video remote sensing satellites,and it has been applied to the OVS-1 video observation satellites successfully.