NTSC视频信号的数字解码方案研究与实现

The research and implementation on the digital decoding of NTSC video signal

当前数字电视发展速度很快,在个别地方甚至已经开始逐步普及,但是由于经济,技术等原因,在电视广播领域,模拟电视广播还将在很长一段时间内占据主导地位.在不改变现有模拟电视广播系统的条件下,如何有效的提高视频图像的质量已经成为一个热门的话题.将数字解码技术用在改善传统的模拟电视广播质量上,用数字方法提高模拟视频的图像质量,和传统的模拟视频解码相比,解码性能较高,并且为后端数字视频处理提供优质视频源.对于模拟视频信号的解码系统,亮色分离和色度解调是视频解码的两个关键点.对于NTSC(national television system committee)制式的视频解码,当前已有的几种优质的解码算法中,亮色分离均用3D自适应梳状滤波,色度解调采用数字锁相环,均能实现优质的视频解码效果,但其共同点为算法复杂,硬件实现难度大,成本高.本文在深入研究NTSC制式编解码的基础上提出一种精简高效的数字解码方法,将传统的NTSC制式彩色全电视信号,通过高速高分辨率的FPGA(field programmable gate array)硬件采集系统转化为数字视频流,提出一种改进型自适应2D梳状滤波器进行亮色分离,相比3D梳状滤波器,在满足视频解码质量需要的基础上大幅度简化了算法,利用相位估计法恢复色同步副载波,相比数字锁相环,除简化了算法外,还避免了数字锁相环偶尔失锁的可能,且对前端视频数字化采集的采样率无特殊要求.解码后的视频分量信号可在数字视频设备上直接显示,并可为后端的数字视频图像处理等提供优质的数字视频信号源.该解码算法在MATLAB上模拟,具有解码后的图像质量良好,运算速度快和解码系统精简等特点.

Digital TV technology is developing quickly and becoming all-pervading in some regions now. But, because of the economic and technological reasons, the analog TV broadcast will still be the most popular media in some fields, such as TV broadcasting, video surveillance and so on. And how to improve the video quality under the condition of using the analog composite video source is a popular subject. After the analog composite video is decoded into RGB video components, the digital technology can work on it to improve the digital video quality evidently. Compared with the analog technology, the digital decoder has many advantages, such as high--powered decoding, and providing high quality video source for the back end of digital video processor. Also, the software digital video decoding system can be used as the front end of analog composite video decoding IC. Y/C（Lumu/chroma） separation and the chroma demodulation are the most important points of the decoding of analog composite video. Refer to the arithmetic of the NTSC decoding mentioned in the reference, they usually use 3D adaptive comb-filter to separate the luminance and the chroma, DPLL（digital phase locked loop） to demodulate the ehroma. They perform well on the NTSC decoding, but they are too complicated, and hard to be realized on hardware. This thesis focuses on a smart and high-powered digital decoder of NTSC analog video, based on the research of the NTSC encoding and decoding. The NTSC composite video is sampled by the high-speed, high- resolution FPGA（field programmable gate array） data acquisition system. The digital decoder uses improved 2D adaptive comb-filter to separate the luminance and the chroma. Compared with the 3D adaptive comb-filter, it is smarter and perform better on the Y/C separation. The digital decoder also uses phase estimation arithmetic to recover the phase of color burst. Compared with the DPLL, it is smarter and has no risk of losing effect under some special conditions, and it also has no special request on the sample clock. The RGB video components after being decoded can be displayed on many digital monitors directly, and can be processed by the back-end digital video processor. The digital decoding arithmetic is verified by the MATLAB on PC（computer）. The image decoded from the analog NTSC video is nice, and the digital decoding system is effective, high powered and simple.