Study On Human Body Detection And Tracking In Intelligence Visual Surveillance

Intelligence visual surveillance is currently one of the most active research topics in computer vision. Visual surveillance automatically analyses the video sequences, and attempts to detect, track and recognize certain objects. Based on this, the visual surveillance system can understand and analyze the object's behaviors, in order to do daily management and reflect when unusual conditions occur. The model-based and area-based human body tracking in intelligence visual surveillance are studied in this paper.Firstly, as the base of model-based human body tracking method in this paper, a novel evolutionary algorithm called Probability Evolutionary Algorithm (PEA) is proposed. PEA is inspired by the Quantum computation and Quantum-inspired Evolutionary Algorithm (QEA). The individual in PEA is encoded by a probabilistic superposed bit. The observing step is used in PEA to obtain the observed individual, and the update method is used to evolve the population. The function optimization and 0-k knapsack problem experiments show that PEA has apparent superior in application area, searching capability and computation time compared with QEA and canonical genetic algorithm (CGA).Secondly, a model-based human body tracking method based on PEA is presented. In the PEA based human tracking framework, tracking is considered to be a function optimization problem, so the aim is to optimize the matching function between the model and the image observation. Then PEA is used to optimize the matching function. Experiments on synthetic and real image sequences of human motion demonstrate the tracking accuracy and computation efficiency of the proposed human tracking method compared with the tracking method based on particle filters.Thirdly, an area-based tracking system for multiple faces is presented based on Relevance Vector Machine (RVM) and Boosting learning. At the first frame, a face detector based on Boosting learning is used to detect faces, and the face motion models and face color models are created. In the tracking process, different tracking methods (RVM tracking, local search, giving up tracking) are used according to different states of the faces, and the states arechanged according to the tracking results. When the full image search condition is satisfied, a full image search is started in order to find new coming faces and former occluded faces. In the full image search and local search, the similarity matrix is introduced to help matching faces efficiently. Experimental results demonstrate that this system can (a) automatically find new coming faces;(b) recover from occlusion, for example, if the faces are occluded by others and reappear or leave the scene and return;(c) run with a high computation efficiency.