Human Movement Detection Based On Haar Feature

The figure of the human body detection will be found in many applications likeintelligent surveillance, gaming and movie animation field. Human Movement detection inthe field of intelligent monitoring applications such as shopping malls, banks, railway stations,plazas, subway flow of people could gather, and the safety factor higher place, we can usehuman motion detection to do intelligent monitoring role. Intelligent monitoring system, witha lack of supervision, automatically identify and respond to events, alarm time, and lowconsumption of resources, will occupy the mainstream of the future monitoring equipment. Asthe game industry in recent years has grown in strength, the type of game content has changeda lot, but without exception, they are manually operated type. In order to increase the level ofparticipation of the people of the game, we can own images via video detection and a nalysisof human movement, directly with the body to control the game, his movement elementsadded to the game, so you can change the game mode of participation, The game reallyimmersive, so that the human motion detection in the game industry with commercial. HumanMovement motion detection for film and animation industry also have a positive impact. Thus,the moving body detection both have research value, but also has the potential market value.Detection system based on video image analysis of the human movement is a majorcomponent of the intelligent monitoring system. Video-based human detection technology isin the case of the absence of human intervention, the use of computer vision and videoanalysis, real-time automatic detection and analysis of image sequences. Therefore, thetechnology in the intelligent monitoring, intelligent transportation, driver assistance, motionanalysis, traffic statistics, and many other areas has broad application prospects. At the sametime, this topic is also research hotspot in recent years, computer vision, computer imageprocessing, pattern recognition and machine learning fields. Block the interaction betweennon-rigid characteristics of the human movement body, the human self-occlusion andobjectives, the impact of the complex and volatile context, changes in dress and light videohuman movement detection research still faces enormous challenges.Realized video histogram human detection algorithm based on silhouette featuresalgorithm using Haar features to describe the movement of human movement characteristics, and use AdaBoost algorithm to train the cascade of human movement classifiers, experimentsshow that the proposed algorithm can achieve higher accuracy in images and video the rate ofthe human body detection.The main work specific:1. Study the detection difficulties and focus of the current movement of humanmovement, and the first frame difference method and the moving averagemethod combined with relatively solve part of the problems faced by themovement of the human body detection.2. Study the current popular movement of human detection feature extractionalgorithm, Haar-Like, edgelet characteristics, Shapelet feature, and HOG featureextraction algorithm based on Haar features;3. Analysis of the the AdaBoost classifier principles and training process, explainthe relationship of the the haar characteristics of weak classifiers weak classifiertraining methods and processes, cascade classifier composing principle andideas4. Research and compare the moving target detection algorithm commonly used inthe field of Human Movement detection, including single Gaussian backgroundmodeling algorithm in the background subtraction method, codebook modelingmethod, moving average modeling method, and the proposed combination ofmorphological processing, asked the connected domain analysis, so that themovement of in a video sequence precise and fast positio ning of the humanmovement area5. One of the Human Movement detection system based on haar characteristics. Bya large number of experiments show the algorithms and the various processescan get a good set of test pictures and video test application, higher correctdetection rate and processing speed to get good results in terms of real-timeapplications outlook is expectations.