A brief analysis of the relationship between high-definition video and intelligent technology

       As the resolution of high-definition video continues to improve, it will provide clearer and clearer data, further promoting the development of intelligent technology. At the same time, the continuous development and application of intelligent technology will also promote the further practical application of high-definition video, providing strong support for the promotion of high-definition video surveillance systems.

　 In the application process of high-speed rail video surveillance system, users' requirements for image resolution are constantly improving, and the demand for high-definition video is also increasing day by day. The overall demand is divided into two categories, one is security-related applications, and the other is service-related applications. In security-related applications, the relevant railway departments have paid special attention to the tracking of wanted criminals in passenger stations, the prevention of criminal cases in stations, and the activities of scalpers. For example, in some criminal cases that occurred in passenger stations, it is necessary to retrieve video recordings to view the physical characteristics of the suspects, but some of the currently installed analog cameras cannot provide effective image details. In addition, there are some key entrances and exits that require the video to provide sufficiently clear facial features, but the existing analog videos cannot do this while taking into account the scene.

　　Among the service-related applications, the most typical application is ticket queue control. Users hope that the video can cover a larger scene to achieve dynamic monitoring of the number of passengers purchasing tickets, so as to add ticket windows when there are dense crowds of people. At the same time, they hope that the video can provide some details of the ticket seller and passengers in the ticket purchasing process, so that it is easier to handle some disputes in combination with voice. At present, the relevant departments of the Ministry of Railways have launched the application test of high-definition video. In some passenger stations, high-definition video products and technologies of different manufacturers have been tested and verified first, and the test results are being further evaluated. In the latest "Technical Conditions for Railway Integrated Video Surveillance System", the definition and technical conditions of high-definition video have been clarified, laying a technical foundation for further application.

Problems Brought by High-Definition Video Surveillance Applications to High-Speed ​​Rail

　　While the high-definition video surveillance system brings high-definition clarity, it also brings a series of practical problems. The reason is that although the front-end camera equipment can meet the high-definition requirements, the "high-definition" of the front-end does not represent the "high-definition" of the entire video surveillance system. Due to the doubling of the amount of front-end video information, it will inevitably bring greater changes to the video encoding, transmission, and storage systems. Because the high-definition video surveillance system is a system solution, it cannot be represented by a high-definition camera. At the same time, it will also bring about an increase in demand and technological changes in computing, transmission, exchange, and storage in the entire industry chain. A true high-definition surveillance system should consider high definition from the perspective of the entire surveillance system, including the acquisition of front-end video sources, encoding and compression of video signals, transmission of video signals, video browsing, playback of video files, and other links to fully support high-definition surveillance.

　　For users, high definition is meaningful only when it includes front-end, platform, storage, browsing, display and other links. In the process of migrating to high definition, high-speed rail video surveillance systems are also facing similar problems. The larger the video resolution and the higher the clarity, the larger the bit rate will be when data is transmitted, which increases the pressure on network bandwidth. At the same time, the doubling of data traffic has increased the amount of data storage. Taking 720P as an example, its 25 frames/second video bit rate is between 500K-3Mbps, and the storage space per hour is about 1G-2G. The storage pressure is huge both in terms of cost and space. In addition, the requirements of high-definition video for central control, management software and hardware performance have also doubled, and the same is true for the workload of managing and post-analyzing recorded video images. These problems are all problems that must be solved in the promotion and application of high-definition video surveillance systems in high-speed rail.

　　
Intelligent application: saving storage costs to a certain extent

　　The emergence of intelligent analysis technology can solve the problems faced by high-definition video surveillance systems to a certain extent. In high-definition video surveillance systems, intelligent video analysis has a foundation because of high-quality video sources. In addition, many intelligent technology applications are only practical on the basis of high-definition video, such as face recognition and other applications. High-definition video technology can bring us high-definition, high-quality video sources containing rich and complete information. Without high-quality video sources, video intelligent analysis will affect the accuracy due to video resolution issues, and will also lose information due to scene issues.

　　First, intelligent video analysis technology can be used to further extract effective information for high-precision encoding, reducing the bandwidth occupied by invalid information. For example, the texture of the video scene can be segmented, and then the video information can be further converted into semantic information, and the image encoding can be converted into the understanding of the image. In this way, a scene can be expressed as a combination of objects such as buildings, roads and targets, thereby streamlining the amount of information.

　　Secondly, on the timeline of the video, the use of intelligent analysis technologies such as dynamic detection, event detection, and behavior analysis can also selectively store videos dynamically and store alarm events, thereby filtering out a large amount of redundant video information occupying storage space. At present, this mode is widely used in railway integrated video surveillance systems, that is, only when user-customized abnormal events occur in the scene will the clips be stored, which greatly saves the cost of storage devices.

　　Thirdly, within the spatial scope of the video, users can store and process only the areas of interest, which can also greatly reduce the problems caused by the increase in video resolution. For example, if the video scene contains some irrelevant areas such as the sky, these areas can be marked as not to be stored, so that each frame of the image can save a certain amount of storage space. Finally, using intelligent technologies such as video content analysis and video search, we can quickly locate event videos, realize labeled video data management and data query, and thus improve the data management efficiency of the entire video surveillance system. On the other hand, without video intelligent analysis, the high-definition video surveillance system will only remain at the primary stage of "viewing" and will not rise to the advanced stage of "using".

　　In terms of implementation, both front-end and back-end intelligent applications are very meaningful. In front-end applications, intelligent algorithms embedded in high-definition cameras can realize real-time analysis of video data and flexibly determine the transmission and storage strategies of video streams, thereby saving a large amount of resources consumed by redundant video information. In back-end applications, intelligent technology is applied to the back-end, combined with cloud computing and cloud storage technology, to realize intelligent retrieval and intelligent storage of the entire high-definition monitoring system.

　　Therefore, high-definition video and intelligent technology promote each other and serve as the basis for each other's development. In essence, high-definition video only provides data content, while intelligent technology is the first use and consumption of content, while generating secondary data and content. The broader application of high-definition video surveillance systems must be further data mining on this basis. Through deep information mining, rich and colorful application solutions will surely be generated in various industries.