场信号预处理的扩展视频编码器设计

Design of extended video encoder with pre-processing of interlace signal

虽然视频标准中给出了场信号的一些编码方法,但由于宏块划分方式比较多造成运动估计非常耗时,很难得到实际应用.为此,设计了具有运动自适应去隔行预处理功能的扩展视频编码器.该编码器先将场信号处理成逐行信号,之后采用固定帧方式进行编码.在将场信号处理成帧信号时,首先对相同奇、偶性的四场做运动检测,然后对具有不同运动特点的区域采用不同的插值方法得到待插值像素,最后插值场与当前场交织成逐行帧.测试结果表明,该编码器在相同压缩比下能取得与其它编码方法相同的峰值信噪比,编码一帧时间为宏块级帧/场自适应方法的1/3.此外,该编码器能有效地解决场信号在逐行扫描设备上的显示问题.

Video coding stands gives out several coding methods for interlace signal. However, many kinds of macroblock division methods for motion estimation, which makes these methods very time-consuming and can＇t be used actually. So an extended video encoder was designed, motion-adaptive de-interlacing was used to convert interlace signal into frame signal, then encoding frame signal with the fixed frame method. There are three steps for signal converting. Firstly, motion detection by utilizing four field signals with the same parity to get motion status of pixel, secondly, according to the motion status of pixel, use different filter to interpolate the pixel, thirdly, interlace interpolation field and current field to get progressive frame. The experimental result shows that it can get the same peak signal noise rate as MBAFF, but the encoding time is about 1/3 of other methods. What＇s more, when displayed on progressive screen, it can solve the problem such as line crawling etc.