Research And Implementation Of The Technique For Extracting Moving Human-Body From Images

In recent years, technique for extracting the information of human-body becomes increasingly attractive. The technique for extraction is a composition of sensor, signal processing and pattern recognition, and has a great practicability in practice, such as the new human-computer interaction system that works in the way of human-human interaction, the human-body recognition and human motion identification system. In the thesis, we work on a branch of the technique for extracting the information of human-body: the technique for extracting the information of moving human-body from images and give some implementation based on the branch.The technique for extracting moving human-body from images is mainly based on image processing and pattern recognition, mainly include the following different parts: (1) technique for moving object segmentation, (2) technique for moving object tracking, (3) technique for representation and description, (4) pattern recognition.After explain the theory of the technique for extracting moving human-body from images, we give two implementation based on it: (1) a module for human tracking in Videoconference System, the module is a implementation of the segmentation and tracking technique just mentioned above, and is based on a new tracking arithmetic which is based on the difference between video frames, and is first gave out by the thesis.(2)a real-time system for counting the passing people using a single camera, which comes from Dalian Everspry Image Tech, Co, Ltd, the system takes advantage of all the sub-technique of the technique for extracting moving human-body. In the system, we use a single camera to get the video data of the being-watched region, then extract the moving objects from the background and identify the moving-objects to see if it is a human-body, if it is human-body we will recognize and track it to give the count and the moving direction of the human in the moving objects.