The Research Of Intelligent Video Surveillance System Based On OpenCV

With the rapid development of society, the improvement of material and the present situation of large-scale population flow in China, safety consciousness was gradually on the agenda. Video surveillance as a powerful means of modern security is gaining more and more attention and the video surveillance system is also beginning to be widely used up, gradually spread in ordinary people’s life. Intelligent and automated video surveillance technology will be a trend of the development of video monitoring field. It is needed to study the method to make the video monitoring more accurate and automatic identification of the abnormal situation, and the characteristics of the intelligent video surveillance system which can be identified by machine.A small intelligent video monitoring system is presented in this paper, based on the system application, through the analysis of the process of building system and the technology used in the system, to explore the characteristics of intelligent video monitoring system should have, the requirements between modules and the image processing algorithms to make video surveillance more intelligent, including intrusion detection, face recognition, gesture recognition and color recognition. Through the analysis of the advantages and disadvantages of these technologies, it is hoped that they can produce a more diversified role in the field of video surveillance. In face recognition, the recognition method based on graphics features combined with color identification, to avoid some images with facial features but not a real face, increasing the recognition accuracy. In the image processing, the operation of the histogram equalization has been used, in order to enhance the brightness and contrast of the image and focus on the outline of the face, for classifier filtering the image easily and increasing the accuracy of detection.This system is based on the OpenCV which is a open source library, using the client and server side of the framework model. The client is responsible for the collection of data and sending to the server. The server side used a certain strategy and the corresponding image processing method to determine whether to alarm, and through the AT commandcontroled the communications equipment to send alarm messages to the user’s mobile phone. The function of the system is clear, and based on the realization of the remote monitoring. The system through the intrusion detection decided whether authentication is required. Authentication used the PCA face recognition and gesture recognition method,once the target has not passed the authentication, then the SMS alert measures will be taken. System can automatically and accurately judge the abnormal situation caused by the invasion, and can quickly realize the SMS alarm, improving the accuracy and efficiency of the monitoring.