Simply put, there are many input devices in the conference room: cameras, DVDs, VCRs, physical booths, desktop computers, a large number of notebook signals, etc. However, there are few display terminals: projector, plasma, large screen display, and the role of matrix comes out. It can send any signal of the equipment providing signal source to any display terminal, so that the audio and video can be synchronized or unsynchronized, which is convenient and fast. Common interface types are classified (VGA, AV, RGB), and of course, there are mixed matrices, which are different interface types in devices and also classified according to the number of interfaces. For example, according to the level of the series, there is a telecom broadcast level: there is no flicker and snowflake when switching, which is very stable, and you can know it by watching CCTV programs, followed by the professional matrix, and there is a little black screen when switching. But it didn't flash, and then it was used for civilian use. This is used in most meeting rooms. There are flashing snowflakes and jitter at the moment of switching, but the picture is very stable after switching. With the rapid development of digital technology and the improvement of software and hardware level, the continuous application of high-performance DSP and high-speed bus makes the video matrix scheme based on digital technology realized. Hikvision will launch a new model: DS MD (Matrix Decoder Card) in the board product line in the near future, and based on this product, Hikvision proposes a digital video matrix solution. At the same time, Hikvision believes that digital video matrix will be an emerging hot spot in the security industry and will also be the future development trend. First, the basic concept of video matrix. Basic functions and requirements of video matrix As a video matrix, one of the most important functions is to switch and output the input video images. An accurate summary is to switch the video image from any input channel to any output channel for display. Generally speaking, M×N matrix means that it can support both M-channel image input and N-channel image output. What needs to be emphasized here is that it must be arbitrary, that is, any input and any output. In addition, a matrix system should usually include the following basic functions: character signal superposition; Control the decoder interface between the pan/tilt and the camera; Alarm interface; Control mainframe, audio control box, alarm interface box, control keyboard and other accessories. For domestic users, for the convenience of operators who don't understand English, the character superposition should be Chinese. Matrix systems also need to support cascading to achieve higher capacity. In order to meet the requirements of different users for the capacity of the matrix system, the matrix system should support modularization and PnP, and the combination of different capacities can be realized by adding or reducing video input and output cards. The development direction of matrix system is multi-function, large capacity, networking and remote exchange. Generally speaking, a matrix system with a capacity of × is a large-capacity matrix. If a larger capacity matrix system is needed, it can also be realized by cascading multiple matrix systems. The larger the matrix capacity, the higher the required technical level, and the more difficult the design and implementation. Classification of video matrix According to the different ways of video switching, video matrix can be divided into analog matrix and digital matrix. Analog matrix: video switching is completed in the analog video layer. Signal switching is mainly realized by using single chip microcomputer or more complex chip to control analog switch. Digital matrix video switching is completed at the digital video layer, which can be synchronous or asynchronous. The core of digital matrix is the processing of digital video. It is necessary to add AD conversion at the video input to convert analog signals into digital signals, and add DA conversion at the video output to convert digital signals into analog signals for output. The core part of video exchange will change from analog exchange of analog matrix to digital video processing and transmission. 2. Introduction of digital video matrix. Classification of digital video matrix According to the different implementation methods of digital video matrix, digital video matrix can be divided into bus type and packet switching type. Bus-type digital video matrix As the name implies, bus-type digital matrix means that data transmission and exchange are realized through a bus used by * * *, such as PCI bus. The most common bus matrices are PC-DVR and embedded DVR. For PC-DVR, its video output is VGA, and the image display is completed by PC graphics card. Usually, there is only one output (a graphics card), and the output (a graphics card) is rare; General video output of embedded DVR is a monitor, and some new embedded DVR can also support VGA display. The above two examples can both realize multi-channel video output (or split screen), and these two products can be regarded as a special case of video matrix, that is, only multi-channel video output. The bus-type digital matrix packet switching digital video matrix packet switching matrix composed of PC-DVR (PC+H card, HC card) realizes the transmission and exchange of image data through packet switching (usually IP packet). Packet switching matrix has been widely used at present, such as remote monitoring center, which compresses the image at the local video recording end and then sends the compressed code stream to the remote end through the network (which can be high-speed private network, Internet, local area network, etc.). ) and displayed on the big screen after remote decoding. At present, packet-switched digital matrix has two limitations: long delay and poor image quality. Because it is transmitted through the network, it will inevitably bring delay. At the same time, in order to reduce the occupation of bandwidth, it is often necessary to compress the image at the sending end and then decompress it at the receiving end. It is difficult to ensure good image quality after lossy compression, and the encoding and decoding process will increase the delay. Therefore, the current packet switching matrix is not suitable for occasions that require high real-time performance and image quality. Packet switched digital matrix III. Advantage analysis of digital video matrix cost advantage: video matrix and DVR are combined into one. Digital video matrix scheme can realize the functions of video matrix and DVR at the same time with only one device, which greatly saves the cost. Matrix control and DVR integration are convenient and flexible. If the analog matrix is used, at least one matrix host and one DVR host are needed, which is complicated to install and debug, and besides the cost of DVR, it also needs to pay a high cost for the analog matrix. In addition, for the control of analog matrix, it may be necessary to connect other devices, such as display devices, matrix controllers, matrix control keyboards, etc. Some complex functions even need a special PC to configure. Analog matrix scheme also needs video signal distribution and multiplexing equipment to realize the recording function of DVR, while digital matrix scheme only needs to add a simple matrix module on the basis of DVR, which has relatively low cost, high integration and stability of digital matrix and video recording system, and also reduces the cost of future maintenance. Functional advantages: flexible configuration, powerful function and convenient use. In the analog matrix +DVR scheme, the matrix and DVR are independent and need to be controlled separately. The simulation matrix provides complex operation mode, poor operability and single function. If you want to realize complex functions, you need very complex operating procedures; Using digital matrix, the switching matrix and DVR can be controlled simultaneously through the control platform. The operation interface can be freely developed by secondary developers under Windows or Linux, and the applications and functions can be customized according to the needs of their own customers. The built system depends entirely on the developer's own software. In the digital matrix, image-based digital processing: video switching can be realized, and at the same time, many images can be processed, such as overlapping characters, overlapping images, covering areas and so on. These are the common functions of DVR at present, but for analog matrix, because its core is based on analog signal processing, it seems to be inadequate when facing these functions. Here, taking the character superposition function as an example, the simulation matrix often needs an external character superposition chip to realize, and usually only ASCII code, that is, English characters, can be superimposed, while the simulation that can realize Chinese character superposition is rare, let alone support simplified, traditional and even Japanese at the same time. As for the functions such as image superposition, it is basically impossible to realize in the analog signal layer. Digital matrix can provide richer image display methods. The traditional analog matrix can only perform the simplest: image output; On this basis, the digital matrix can also realize the display modes of N→ (multiple images can be displayed in one window through image scaling) and →N (an input image can be displayed at multiple output terminals at the same time), and even realize advanced functions such as picture in picture. Finally, the system stability, digital matrix ++DVR scheme, high system integration, low power consumption and high stability; However, using the analog matrix scheme, the possibility of problems is greatly increased because of the need for multiple devices. Potential: At present, the analog matrix control system with huge development space is very mature, and the structure and function of its products have not changed much in recent years, even more than ten years, and the potential to be tapped is very limited. The number matrix is completely different. At present, the development of digital technology is changing with each passing day, and the prospect is bright. First of all, with the improvement of hardware performance, in terms of high-speed bus: M's PCI bus has been very mature and popular, such as PCI-E or other high-speed serial buses are constantly proposed; In chip technology: there have been high-performance DSP of M, M and even GHz. It can be said that thanks to the continuous improvement of the performance of the hardware platform, the functions of the digital matrix will continue to improve and develop to the high end. At the same time, the progress of software can not be ignored. New image compression and processing algorithms are constantly put forward, the efficiency of image compression is constantly improving, and more complex and intelligent image processing algorithms are constantly being applied, such as intelligent motion detection and intelligent recognition technology (face recognition, fingerprint recognition, license plate recognition, signature recognition). These advanced image processing technologies can be applied by using the current hardware platform. Therefore, with the rapid improvement of software and hardware, we have reason to believe that the development space of digital matrix will be very broad, and it will definitely surpass analog matrix in performance and function. The secondary development is simple and convenient, which is different from the previous H card and D card using their own SDK. The new SDK will support H card, HC card and MD card simultaneously. Only one SDK is needed to realize coding, decoding and matrix control simultaneously. The codec part in the new SDK is compatible with the original SDK, and users only need to add matrix control part, which greatly reduces the complexity of users' secondary development. At the same time: H card, HC card and MD card can be mixed to facilitate the maintenance and expansion of existing projects. Four. Hikvision's digital video matrix solution Our digital video matrix solution is based on Hikvision's HC series compression card and the upcoming MD series matrix decoding card. HC card is responsible for the video recording, preview and network transmission functions of the system, and this part of the application has been very mature. MD card is the key to realize video matrix. In addition to all the functions of the original decoding card, MD card adds matrix output to realize the video matrix function. At the same time, MD card can be independent of HC card, only used as hardware decoding card, and connected to remote video server through network, which can be Hikvision board or embedded device, forming a network video matrix, which is the same as the original decoding card, but compared with D card in function, the new MD card will be greatly expanded, for example, the number of decoding paths will be greatly improved, and the original D card DSP can only solve paths, and now it can at least solve paths. At the same time, it has the function of multi-window picture segmentation, and users can output decoded images in any combination. The new synchronization function can realize the synchronization between decoders. The main parameters of. The operating system of dsmd card supports Windows /XP and Linux. Used as matrix: each card supports the output of road matrix, which can support roads stably. Using HC card, the video matrix of × can be realized, and the real-time compression of the channel can be guaranteed at the same time. Used for decoding: the decoding function is completely compatible with the original decoding card. Each card supports channel decoding (real-time decoding channel CIF, channel CIF or channel CIF) and channel analog output, and can support channel decoding and channel analog output at most. . The typical scheme of DS MD (1) is equipped with local video matrix and real-time video recording system. In this system, the road DVR system is composed of HC cards, which is exactly the same as the current board card scheme. In addition, an MD card is added to complete the road simulation output and realize the video matrix function. A× digital video matrix+real-time video recording system ⑵ forms a network matrix. The network matrix consists of blocks DS MD, which also supports the network monitoring center for network decoding and video output.