Demands of ATM surveillance market and the overall situations
Source: Nanjing Yuantuo Tech
Date: 2010/08/03
With the development of video processing, analyzing and transmission technology, surveillance system has developed from pure analog system to system combines analog line with figure line and pure IP surveillance. Intelligent analysis of surveillance system emerges. Requirements on intelligent surveillance technology of the security surveillance industry come from various kinds of actual surveillance needs with specific industrial characteristics. Different industries’ surveillance may need certain types of intelligent surveillance technology.
Personalized needs of the segmentation market determine that we must pay attention to the particularity of different systems. Yuantuo Tech is market-oriented and different departments such as sales department, market department, technology department, research and development department and product department corporate with each other to provide customers with customized products and service.
For example, the traditional major customer of the surveillance industry is the finance and banking industry. At present, the most urgent problem of banking surveillance seems to be the definition of images. However, as a matter of fact, the final purpose is to achieve face recognition and restoration of on-site situations. Therefore, it is extended to intelligent video analysis and we can see that bank surveillance system needs stable and reliable face recognition function. As to the ATM, banks not only needs to know the depositor’s identity but also protect the expensive teller machine as well as avoid some frauds. Face shielded recognition, special actions recognition (tracking, disguised withdrawal with a mask, etc) functions are very suitable for ATMs.
Another application is the security surveillance. It focuses on the places with a dense crowd. “Two robberies” should be combated to prevent vicious incidents. In some monitoring points such as squares and stations, surveillance system should always keep an eye on crowd’s movements, analyze and report sudden gathering of the crowds. In the road surveillance system, people’s sudden speed change analysis is added. It can judge the unexpected situations in time which is an effective preventative measure to “two robberies”.
With the development of the surveillance industry, application range of intelligent surveillance products is certain to seep into every industry. Meanwhile, each industry will combine with their own characteristics to make specific requirements on intelligent surveillance. Therefore, we should make intensive studies to meet customers’ requirements of different industries. Marketing staff corporate with research staff of our company have completed the specific solution that can be applicable to different systems. They have cleared way for our company to exploit new markets.
Intelligent surveillance functions in different application areas and technical details
Intelligent video analysis technology has been used in the surveillance industry for over 10 years. However, due to different requirements of different industries, the technology still needs further development and is thinning.
At present, intelligent surveillance technologies mainly contain 3 categories: 1. human and item recognition； 2. human and item’s motion track recognition； 3. judgment and compensation of environmental influence on video.
Human and item recognition: it mainly recognizes the contents that the surveillance system cares about, including face recognition, license plate number recognition, car type recognition and traffic light recognition. The key requirement of intelligent surveillance technology on recognition is the accuracy. For instance, license plate number recognition technology is the best in the market. Its recognition rate is 95[[]%] or even higher than 98[[]%] and thus it can meet the demands of customers of road monitoring industry. If recognition rate is lower than 90[[]%], it will bring great redouble to the management staff. Recognition technology is usually applied in road monitoring, banks and channel management industries. It provides basis for recognition record and hierarchical management.
Human and item’s motion track recognition and measurements: there is much segmentation, including virtual cordon, virtual alert area, intelligent tracking, population statistics, traffic statistics, appearance and disappearance of an item, sudden running and sudden gathering of people, etc. Except for the statistics, it usually judges certain process. Once there is an exceptional situation, for example, someone enters the alert area or there are people gathering in the northeast corner of the square, it will emit alarm information to remind staff on duty to pay attention to the corresponding area. As to the statistic technology, its key technology is to find exceptional situations and do quantity statistics. Therefore, the statistics is required to be accurate and try to reduce the false rate. The actual surveillance application area has much influence on motion track recognition processing technology. The key point of the technology is to find the exceptional situations as soon as possible and try to avoid any omissions to improve forecast accuracy. This function is mainly applied in safe city construction and industrial surveillance.
Judgment and compensation of environmental influence on video: environmental influence includes rain, snow and fog and other bad weather, low illumination at night, shielded camera or camera offset and camera dithering. Intelligent surveillance technology can achieve normal surveillance under very bad conditions. When it is influenced by the environment and video is unclear, it will try to find human in the image or raise the alarm after judging the offset of the camera. The key technology point of this function is that it can output intelligent analyzing information steadily in any application area and try to reduce environment influence on the video surveillance. This function is of wide suitability. Over 80[[]%] monitoring points have the potential needs of adding this function.