Design And Implementation Of Video-Based Traffic Monitoring System For Indoor

With the rapid development of computer technology and digital signal processing technology, computer vision technology has become a hot research direction in recent years. People use video cameras and other devices to obtain information, and then converted by the computer into a digital signal analysis and processing, in order to better understanding of the use of these information. With the increasingly large size of digital information, classified information requirements are also increasing, and the development of artificial intelligence to find a new breakthrough for this problem. The idea of??using artificial intelligence to design appropriate classification, classification implemented through machine learning, classification of information and screening, can better achieve the application requirements.SVM( Support Vector Machine) has been considered one of the best readily available classification algorithm, HOG(Histogram of Oriented Gradient) feature will be widely used in human testing field since 2005, was proposed by HOG features plus SVM classifier to detect the human body has become a recent the more popular methods.HOG-based features and SVM classifier human recognition technologies include feature detection and feature extraction steps, the sample data extraction HOG feature, in the form of feature vectors represent and use appropriate classifier training is completed for a particular target detection and recognition task. This system is designed for indoor traffic through statistical analysis to help decision-makers to grasp the flow dynamics, the corresponding decisions.Video-based traffic monitoring system for indoor use plan view taken advantage of the human head is not covered by the mutual combination of HOG features and SVM classifier a good solution to the problem of indoor human head detection, for detection of the target, was detected by track guide The thought of using the logical analysis of simple and effective target tracking, saving system cost, achieve the target application under more complex scenarios. On this basis, in order to deal with more complex situations may cometo study particle filter based on Kalman filtering system complements, enhances the robustness of the system, expanding the scope of the system.