|
In recent years, domestic security technology rapidly develops. Video monitoring system, as an important part of security technology, is also widely applied. But as for traditional monitoring system, its stationary cameras and limited monitor target environment, often lead to low monitor efficiency. Therefore, through increasing equipments, changing control modes of the monitoring system, and utilization of the latest image processing technology, monitoring system has certain intelligence and the efficiency of monitoring can be greatly improved.This thesis mainly studied control methods of Yuntai, video monitoring methods, face detection methods, and the interaction between detected face information and control model of Yuntai.Besides using the traditional manual control, some innovative control methods such as setting path control and intelligent control are used in Yuntai control module. For the method of setting path control, firstly key points in the path are set by workers, and then through some algorithm the paths are formed. Finally according to the calculated path, Yuntai periodic automatic moves. For the methods of intelligent control, image processing algorithms are used to detect face positions, movement direction and other information. And then according to these movement, objects are tracked by Yuntai. Thus it realizes intelligent control.For the video monitoring module, DirectShow SDK is used to realize some functions such as video preview, video playback, grasping frames etc. For Face detection module, Adaboost and Harr algorithm is used to make classifier, then basic image processing function of OpenCV and face detection functions are used to detect human faces. Finally, face positions will be provide to Yuntai control module so as to realize Yuntai intelligent control.Overall, after the tests of hardware and software, basic intelligent functions of video surveillance system are achieved.
|