Complex Under The Conditions Of The Head Detection And Counting Technique And Realization

The head detection and counting technical studies under complex the conditions are based on the current automated analysis of surveillance data and effective data extraction problem. Actually more complex environmental monitoring, surveillance cameras tilt or look down under the circumstances, it is difficult to obtain in the integrity of the video Face detection information to complete the detection. According to this, this paper uses the methods of activities edge and the methods of Camshift to complete the head and head position and count.In this paper, the activities edge and the Hough transform use to detecte the looking down head, Camshift tracking method use to detecte the side of head, and set a virtual line method to calculate the number of heads. The main research content of this article are as follows:First, this article analyzes several active edge detection methods, based on 10 adjacent frame edge differential combined with the activities of the current edge detection methods, effective extracte the target edge of the current frame.Second, this article analyzes the characteristics of image of the looking down head and Hough transform circle detection methods, realizate of the use of rapid detection of the edge of the head with random Hough transform.Third, this article analyzes the characteristics of the image of the side of the head and the MeanShift, Camshift algorithm, give an automatic side face detection tracking method based on Camshift.Fourth, this article analyzes the largest neighbor matching algorithm, propose an method based on the movement information to establish the goal of chain, set up a virtual line in specific locations of the image, determine the relationship between the virtual line and the target chain to achieve the target head count.In this paper, the above-mentioned methods achieve the look down head and the side head position and count, experimental results show that the methods obtain good results, it achieves the goal of accurate detection of head and counting under complex conditions, has some practical value.