Research On Multi Moving Human Tracking In An Indoor Environment

In computer vision domain, Research on Multi human tracking in an indoor environmentis currently one of the most active topics. The technology mainly applies in such sensitivefields as security surveillance system, video tracking, and so on, and has provided thefoundation support for the computer intellectualization development. The core problem ofmulti human tracking is to detect the objects from image sequences with computer visiontechnology, and then, according to features relevance, to realize classified tracking. The article,based on analyzing images' features and sequences relation, aims at detecting moving objects,analyzing moving features, and finally realizing the recognition and tracking for multi movinghuman.According to image characteristics, the paper mainly introduces three parts: motiondetection, results analysis, and multi-human tracking. In motion detection part, we realizes aadaptive method suitable for the object detection having the static background, but alsoproposed one method suitable for the object detection having the movement background. Inthe process of results analyzing, through analysis on connected-through field algorithm, acontour following algorithm is proposed to attain and revise the location and size of thetracking window. In motion tracking part, through improving the tracking algorithm for asingle object, on the basis of connected component analyzing, a tracking algorithm isproposed which applies in multi moving human tracking. The experiments prove that thealgorithm has the following features: auto detecting moving objects under static background,adjusting tracking window adaptively, Forecasting the centroid movement and reducingreduces the wrong track probability.The results indicate that, these methods complement one another during processing,enhance the efficiency and the precision of the multi moving human tracking in an indoorenvironment.