Research On Detection Method Of Abandoned Object In Complex Environment

Today, people’s security awareness is increasing. So ensuring public safety and national stability has been become the focus attention. Now, the main means of terrorist attacks are used by leaving objects on waiting room, subway, exhibition halls, airports, stadiums and so on. The illegal parked acts in specific area of outdoor public places have seriously affected the order of public traffic. So an intelligent video surveillance system based on real-time monitoring is needed. Detection of abandoned objects is key parts of the intelligent video surveillance system. Some domestic and foreign scholars have done a lot of research of abandoned objects detection. However, there are some shortcomings in the existing video surveillance system. This paper has been improved some methods for these shortcomings.In this paper, we proposed a detection system of abandoned objects in complex environment. Preprocessing was done in the first step so that we could restrain the ghost generated by the first frame and built a correct background. Secondly, Gaussian mixture modeling with two different update rates was used to detect foreground masks; according to pixel classification for the two foreground masks, we could obtain the temporary static targets; shadow detection was done to restrain the influence of illumination. In addition, this paper presented a classification method based on region growing. This method is a better solution to solve the problem of classification between abandoned objects and removed objects. After detecting removed objects, we proposed updating background method to restrain ghost appear again. Then, this paper proposed the combination method between HOG feature and SVM classifier to remove the false detection due to stationary people and to solve the false detection when people and objects are adhesion. Finally, this paper presented a region matching method for matching temporary static area; according to residence time, the relative objects were identified as abandoned objects and alarming.This paper propose, a detection system of abandoned objects and their subjects based on MFC, describes and analyzes the realization of each module, series of relevant experiments were done in simulative environment. Experimental results show that the method used in this paper can effectively detect the abandoned objects with good accuracy and robustness by comparing the Ying’s method.