Analysis of IPC Video Coding Technology in Network Monitoring System

First, the development status of network cameras The digitalization, networking and integration of video surveillance have become a trend of development. As a video surveillance front-end camera, the camera will undoubtedly play an important role in this change. In recent years, the camera market has shifted from analog cameras to network cameras. According to data, network cameras have grown at a rate of no less than 50% per year in recent years, and the network camera market has ushered in a period of rapid development. From the first professional standard H.264 network camera chip solution to the 720P network camera solution to the latest third-generation monitoring solution, the network camera technology has achieved major breakthroughs and rapid application. As the core technology of the network camera, the video coding technology has been restricted by the network conditions and the coding and decoding technologies. Now it is gradually mature, and it has the characteristics of strong pertinence, multiple coding modes, superior performance, and high adaptability. This article mainly introduces the video coding technology of the network camera.

Second, the principle of network camera architecture Network camera is truly all-digital products, video signals from the image sensor (such as CCD, CMOS, etc.) acquisition and digitization, the subsequent processing of all the use of digital signals to ensure the clarity of the image, and the use of the network Pack and transmit these multimedia data (video, audio, text information, etc.).

1. Video Capture The core part of the video capture part - an image sensor, mainly CCD and CMOS, the level of sensor performance directly determines the level of the camera. CMOS image sensors have low bandwidth, inherent anti-halberd characteristics, while maintaining the advantages of small size, light weight, low power consumption, high integration, and low price. CCD image sensors have a wide range of applications because of their high sensitivity and low noise. With the advancement of CCD and CMOS sensor technology, the difference between the two will gradually decrease, and the two will coexist for a long period of time. At present, all kinds of CCDs and CMOSs ​​are competing and introduced. Each manufacturer has its own interface standard. Even if there are physical interfaces on the surface, the internal data is still different. Therefore, it is necessary to interface with the original sensor. In this regard, the need for coordination and cooperation between chip manufacturers and sensor manufacturers will have better results. This is also the network camera codec chip must do the homework, such as the Hi3515 and Hi3520 SONY's 720PCCD this interface that requires CPLD transfer directly defined to the chip inside, by the implementation of the CPLD need SONY factory to help, This directly helps some customers solve the original problem that cannot be obtained. Coding Algorithms With the development of video coding and processing algorithms, multiple image and video coding compression standards have been released successively for different application requirements. For example, MPEG for multimedia storage media and digital television applications developed MPEG-1, MPEG-2, MPEG-4; for low bit rate real-time video communications; ITU-T applications developed H.261, H.263, H. 263+, H.264; In the face of current and future security applications, AVS-S, the advanced digital audio and video encoding and decoding technology standard developed by the AVS (AudioVideocoding Standard) working group in China, has been established. It can be said that there are many options for video surveillance coding standards. However, the current standards used by network cameras are MJPEG, MPEG4, and H.264. M-JPEG is a motion still image compression technique, but M-JPEG only compresses the intra-frame spatial redundancy and does not compress the time redundancy between frames, so the compression efficiency is not high. The focus of MPEG-4 is to solve the problem of audio and video transmission on low bandwidth, making it one of the important trends developed by current network product manufacturers. H.264 is currently the mainstream solution with high compression ratio and high image quality. Good network adaptability has a wide range of applications and can adapt to different occasions. However, the complexity of the algorithm is high and it requires strong hardware support.

3. Coding and Network Currently, for the sake of bandwidth and cost, users are required to use higher bitstreams to achieve the current cost. This imposes higher requirements on the coding efficiency of the network cameras. The network camera transmits data through the network, and the implementation mode on the network transmission is not the same. However, if the data packet is too large, the network congestion will be caused. Therefore, the internal control of the encoder is very strict, and it is required that the image quality cannot be influenced. With the continuous development and changes in the actual application requirements of network cameras, their requirements for chip encoding are becoming more and more demanding, and a more realistic scenario is achieved on the premise of a higher compression ratio. For example, Hass provides a compression solution that can achieve a compression rate of 720P@2Mbps, which is 3-4Mbps faster than the average product. This increases the minimum compression ratio by at least 50%. The 2Mbps code rate has another meaning: the user can not Increase the network and storage costs to upgrade to the 720P experience. In the future, with the rapid application of 3G networks, the encoding methods supporting various networks have also become an important development direction.

Third, the application of the coding standard to determine and analyze the current industry mainstream coding algorithms MPEG-4 and H.264, MPEG-4 and H.264 who will dominate the network camera video coding? MPEG4 compression technology can obtain higher-definition video images, and can flexibly set the video resolution and the number of compressed frames for each channel, and the compressed image can be arbitrarily edited. However, its defects are also very obvious: First, there are serious drop frames and poor real-time performance. Under the premise that each path must be high-definition, it is difficult to complete real-time compression. Second, the compression efficiency is low, and the transmission bandwidth and storage space occupy a large area. . H.264 introduces an encoding mechanism for IP packets, which is conducive to packet transmission in the network, supports streaming media transmission of video in the network, and supports hierarchical encoding transmission under different network resources, thereby obtaining smooth image quality. H.264 can achieve broadcast-quality HD video resolutions of 720P and 1080i/Pd at lower bandwidths. At present, network cameras basically use H.264 encoding chips. However, this does not mean that any H.264 algorithms from manufacturers' network cameras are good. Therefore, it is necessary to have a clear understanding and understanding of the network camera coding algorithm so as not to enter into errors. If you only care about whether H.264 is used or not, it is considered that both manufacturers' MPEG encoding modules implement H.264, and they will be equally good. The fact is that the H.264 standard contains a large collection of algorithms, such as MPEG-1, MPEG-2, and MPEG-4. To achieve a large amount of calculations for all algorithms, no one chip can implement all the algorithms. The process of designing an H.264 encoding chip is actually a process of selecting a subset of algorithms in a set of H.264 algorithms according to the computing power of the chip. Therefore, we can use the subjective evaluation method in the image domain to discriminate between the coding chip and the coding algorithm. For example, by adjusting the code rate (such as several tens of kbps or one hundred or two hundred Kbps), adjust the motion amount of the image and check whether the image has any. Loss of frames, there is no mosaic, etc., according to the extent of these phenomena to determine the quality of the chip.

IV. Application Stability and security are two very important features of the video surveillance system. Therefore, users should pay special attention to the stability and security of the surveillance system in practical applications. In the past, monitoring systems used analog cameras and hard disk recorders to monitor, collect, and manage data. This mode is usually applied to close-range deployments. If natural disasters or other catastrophic events occur, monitoring data will likely be damaged. The appearance of the network camera has completely changed the previous monitoring mode. This model uses the network as a link to separate the data encoding and storage, and uses special data storage devices to achieve long-distance data acquisition, storage and management. Extremely high stability and safety. For example, the “Moat” checkpoint at the Suzhou World Expo security site uses a large number of network cameras. These network cameras are installed at all important checkpoints in Suzhou and Shanghai. The monitoring centers are located in Shanghai and Suzhou, and their data storage can be realized in Beijing. In places such as Shanghai and Suzhou, the prevention of data loss and damage has greatly facilitated the viewing, commanding, and scheduling of surveillance systems.

At present, network cameras are widely used in transportation and finance. Power, ports and important places.

V. Development trend With the subdivision of the security industry and the development of video surveillance in the direction of networking, intelligence and integration, the development of network cameras has gradually become more subdivided. High-definition and smart have become the main development direction of network cameras.

1. High-definition HD has always been one of the goals of network cameras. At present, most network cameras can reach the megapixel level, but not all network cameras can achieve high-definition applications in practical applications. This requires network cameras in addition to their own high-definition In addition to performance, it is necessary to support and match access-related devices for high-definition applications. As we all know, in the SD era, the cascading of the chips is implemented through PCI. A PCI can transmit 2 to 3 DI uncompressed streams, but it is not enough in the high-definition era. High-definition decoded data is transmitted via PCI. To the display chip, so many programs will scale and output the high-definition image, so that the final effect can not reflect the true high-definition effect. For example, Hass introduced a high-definition cascading channel solution that can achieve zero-delay multi-chip transmission effects, can be matched with the chip's integrated high-definition VGA output, and with DHCP technology, you can easily achieve high-definition.

2. Intelligent Currently, intelligent network cameras have become a hot spot for application promotion. It is the use of intelligent identification technology for automatic analysis of collected voice, video and other data and extract key information in the data source, so as to make timely and accurate for the monitor. Judgment basis is provided, among which smart cameras such as face recognition, behavior analysis, and panoramic tracking have been widely used in the surveillance field.

VI. Conclusion In summary, along with the common development of market demand and security technology, current network cameras present a flourishing situation. Various types of network cameras are widely used in all corners of the security field, including HD cameras and smart cameras. Become the mainstream product of video surveillance industry applications. This has benefited from the improvement of the core technologies of network cameras and the continuous improvement of coding technology. In recent years, network cameras have made great progress in image acquisition, A/D processing and encoding compression, network transmission and service and other technologies. Can support and match more other network equipment products, meet the actual application requirements of video surveillance systems, in order to achieve the purpose of security.