安防监控视频编码技术

　　视频编码技术基本由IS0/IEC制定的 MPEG - x 和 ITU - T 制定的 H .26x两大系列视频编标准推出。从 H .261视频编码建议，到 H .262/3、 MPEG -1/2/4等都有一个共同的不断追求的目标，即在尽可能低的码率(或存储容量)下获得尽可能好的图像质量。而随着市场对图像传输需求的增加，如何适应不同信道传输特性的问题也日益显现出来。于是 IEO / IEC 和 ITU - T 两大国际标准化组织联手制定了视频新标准 H .264来解决这些问题。

　　H .261是最早出现的视频编码建议，目的是规范 ISDN 上的会议电视和可视电话应用中的视颍编码技术。它采用的算法结合了可减少时间冗余的帧间预测和可减少空间冗余的 DCT 变换的混合编码方法，和 ISDN 信道相匹配，其输出码率是p* 64kbps。 p 取值较小时，只能传清晰度不太高的图像，适合于面对面的电视电话; p 取值较大时(如 p >6)，可以传输清晰度较好的会议电视图像。 H .263建议的是低码率图像压缩标准，在技术上是 H .261的改进和扩充，支持码率小于64 kbps 的应用。但实质上 H .263以及后来的 H .263+和 H .263++已发展成支持全码率应用的建议，从它支持众多的图像格式这一点就可以看出，如 Sub - QCIF 、 QCIF 、 CIF 、4CIF甚至16CIF等格式。

　　MPEG -1标准的码率为1.2 Mbps 左右，可提供30帧 CIF (352 x 288)质量的图像，是为 CD - ROM 光盘的视频存储和播放所制定的。 MPEG -1标准视频编码部分的基本算法与 H .261/H.263相似，也采用运动补偿的帧间预测、二维 DCT 、 VLC 游程编码等措施。此外还引人了帧内帧(1)、预测帧( P )、双向预测帧( B )和直流帧( D )等概念，进一步提高输妈效率。在 MPEG -1的基础上， MPEG -2标准在提高图像分辨率、兼容数字电视等方面做一些改进，例如它的运动矢量的精度为半像素;在编码运算中(如运动估计和 DCT )区分“帧”和“场”;引人了编码的可分级性技术，如空间可分级性、时间可分级性和信噪比可分级性等。近年推出的MPEC4标准引人了基于视听对象( AVO , Audio - Visual Object )的编妈，大大提高了视频通信的交互能力和编码效率。MPEG4中还采用了一些新的技术，烟ル状编仍、自适应 DCT 任意形状视频对象编码等。但是MPEG4的基本视频编码器还是属于和 H .263相似的一类混合编码器。

　　总之， H .261建议是视频编码的经典之作， H .263是其发展，并将逐步在实际上取而代，主要应用于通信方面，但 H .263众多的选项往往令使用者无所适从。 MPEG 系列标准从例仔储媒体的应用发展到适应传输媒体的应用，其核心视频编码的基本框架是和 H .261玫的，其中引人注目的MPEC4的“基于对象的编码”部分由于尚有技术障碍，目前还以百遍应用。因此在此基础上发展起来的新的视频编码建议 H .264克服了两者的弱点，化合编码的框架下引人了新的编码方式，提高了编码效率，面向实际应用。同时，它是两大国际标准化组织共同制定的，其应用前景不言而喻。

　　The video coding technology is basically launched by the two major series of video coding standards, MPEG x developed by IS0/IEC and H.26x developed by ITU T. From H. 261 video encoding recommendations to H. 262-3, MPEG-1/2/4, and others, there is a common goal that is constantly pursued, which is to obtain the best image quality possible at the lowest possible bit rate (or storage capacity). With the increasing demand for image transmission in the market, the issue of how to adapt to different channel transmission characteristics is also becoming increasingly apparent. So two major international standardization organizations, IEO/IEC and ITU T, jointly developed the new video standard H.264 to address these issues.

　　H.261 is the earliest video encoding recommendation to standardize the video address encoding technology in conference television and videophone applications over ISDN. It adopts a hybrid encoding method that combines inter frame prediction that can reduce time redundancy and DCT transform that can reduce spatial redundancy, matched with the ISDN channel, with an output bit rate of p * 64kbps. When the value of p is small, it can only transmit images with low clarity, making it suitable for face-to-face television and phone calls; When the value of p is large (such as p>6), it can transmit conference television images with good clarity. H. 263 proposes a low bit rate image compression standard, which is an improvement and expansion of H. 261 in technology and supports applications with bit rates less than 64 kbps. However, in essence, H.263 and later H.263+and H.263++have developed into recommendations that support full rate applications, as evidenced by its support for numerous image formats such as Sub QCIF, QCIF, CIF, 4CIF, and even 16CIF.

　　The MPEG-1 standard has a bit rate of around 1.2 Mbps and can provide 30 frames of CIF (352 x 288) quality images. It is designed for video storage and playback on CD ROM discs. The basic algorithm of the MPEG-1 standard video encoding part is similar to H.261/H.263, and also adopts motion compensated inter frame prediction, two-dimensional DCT, VLC run length encoding, and other measures. In addition, concepts such as intra frame (1), predicted frame (P), bidirectional predicted frame (B), and direct current frame (D) were introduced to further improve the transmission efficiency. On the basis of MPEG-1, the MPEG-2 standard has made some improvements in improving image resolution and compatibility with digital television, such as its motion vector accuracy of half a pixel; Distinguish between "frame" and "field" in encoding operations such as motion estimation and DCT; Introduced scalability techniques for encoding, such as spatial scalability, temporal scalability, and signal-to-noise ratio scalability. The MPEC4 standard launched in recent years has attracted attention based on audiovisual objects (AVO, Audio Visual Object