| Being small in size, low in cost and easy to install, embedded video surveillance system is widely used in many fields. Accordingly, scholars and manufacturers have paid attention to it in the domestic and overseas.Firstly, on the base of requirement analysis, a design is given. The system is composed of two parts server and client. From the bottom to top, the server includes 4 layers: a) hardware layer, b) operating system layer, c) a variety of runtime libraries necessary to the application, d) application layer. The application captures video, encodes as H.264, analyses video content and transmits the video and the results of video content analysis to the client. On the other hand, the client receives, decodes, presents video and processes the results of analysis.Secondly, a video surveillance system is implemented in an integrated manner. It captures video data, via DSP built-in CPU encodes the data as H.264 stream. Utilizing protocols such as RTP, RTCP and RTSP, video data are packaged and sent the client.Thirdly, in the embedded video surveillance system, a method is designed and implemented for analyzing and locating vehicle license plate. The original image is preprocessed(such as histogram equalization and binarization) in RGB and HSV color space to get two binary images, and then the candidate regions that may contain the plate are extracted from the binary image of HSV space. Next, the candidate regions are positioned one by one, and the plate boundaries are determined. As a result, the plate is located. Further the more, the characters are segmented by projecting the pixels in the plate image.Finally, the system is tested. The results show that it can correctly capture, encode and transmit video data. The license plate location method can locate and extract the license region and accurately segment the characters in the image with complex background. The system has high flexibility and performance, and it is applicable in vehicle management. |
PDF Full Text Request