Overview of the plan: With the initial construction of the "Safe City", the amount of video data is showing an exponential growth trend. The basic functions of "watching", "controlling", "recording" and other functions of videos can no longer meet the practical needs of public security. Moreover, traditional video surveillance systems usually achieve security protection through personnel monitoring and recording, which cannot actively and effectively guarantee security, especially when there are too many monitoring points. Personnel monitoring cannot take into account all monitoring scenarios, and it is difficult for monitoring personnel's attention to ensure accurate and efficient monitoring of all monitoring scenarios 24 hours a day
With the initial construction of the "Safe City", the amount of video data is showing an exponential growth trend. The basic functions of "watching", "controlling", "recording" and other functions of videos can no longer meet the practical needs of public security. Moreover, traditional video surveillance systems usually achieve security protection through personnel monitoring and recording, and in fact cannot actively and effectively guarantee security, especially when there are too many monitoring points. Personnel monitoring cannot take into account all monitoring scenarios, and it is difficult for monitoring personnel's attention to ensure accurate and efficient monitoring of all monitoring scenarios 24 hours a day; In addition, "passive video recording" can usually only be replayed for evidence collection through video recording after the event has occurred. On the one hand, the loss has already been incurred and cannot be recovered; On the other hand, the efficiency of manually replaying video evidence is very low, while high-tech achievements rely on human wave tactics to generate value.
So how to use these video data more efficiently, how to conduct deep mining on video data, and how to provide more valuable videos that can be structurally described have become important challenges that need to be addressed at present. As a video practical application for future smart public security, intelligent mining of local video content has become the most urgent need for current video data.
Solution