Detection Of The Human Body And In Complex Environments

As an important research direction in computer vision, human body detection andtracking merge two important technology which is detection and tracking together, andhave great application prospect in many fields,such as video monitoring, militarydetection reconnaissance and so on.But because of the variety and complexity of practical application environment,occlusion in the process of tracking human body, scale change and other difficulties,the human body target detection and tracking still has a great distance to application,and has a broad study space. How to solve the difficult problem above in order to makehuman body target tracking has high robustness, is of great significance to practicalapplication system. This article, in view of problem of the human body target detectionand tracking, first proposes a method to calculation local block, which divides thehuman body into several local pieces, calculates the change trend of target according tothe movement of local pieces, buid up calculation model calculates the whole objectocclusion degrees using the local similarity, and incorporates the model into the targettracking algorithm process based on detection of the human body. First the humanbody is divided into more than one local target pieces, histogram of orientedgradient(HOG) and histogram of color(HC) of each piece are extracted as features toanalyze. At the same time, the head of human body target is analyzed independentbased on the human body outline，it is proved that even in condition that it is in slowmotion of human head and appear illumination change in background and foreigninterference possess good testing ability.Taking two kinds of models and calculation method above into consideration,occlusion, interference and scale changes of the environment existing in the process oftracking are handled effectively, so the robustness of the tracking algorithm isultimately guaranteed。...