The Research On Active Fortify Based On The Characteristic Object

Along with the developing of network technology and information technology, video surveillance technology has also experienced the process from analog surveillance to digital surveillance, local single machine surveillance to network surveillance based on Internet. The developing process main reflects in four steps: local analog video surveillance system, based on the local single machine video surveillance system, based on the DVR/DVS network video surveillance system and this paper studying the intelligent video surveillance system.In the intelligent video surveillance system, it mainly includes five key technologies: digital video/audio compression technology, digital video/audio transmission technology, and large number of high-performance video content storage and retrieval technology, motion detection and alarm monitoring technology as well as the scene target recognition technology in the monitoring scene.This paper mainly put the two key technologies: the digital image processing, analysis technology of the latter results, and the motion detection, alarming technology as the studying main line. The paper comparative analysis on the image processing and image analysis typical algorithm, and carrying out the experiments on the image recognition technology in the process of pre-image, moving target detection technology and image analysis technology. Finally, according to the requirement of grain reserves intelligent video surveillance, by constructing the image hue, brightness and its gradient space to preliminarily determine the region of interest, taking the same color, and combining the maximum gradient (threshold) of two-dimensional segmentation method to identify the object edge, then describing the edge in the Fourier describing operator, and comparing its Fourier descriptors and parameters of hue with the priori knowledge of multi-dimensional characteristics of object to achieve the target of a reliable and high-precision and characteristics object identification. In the realization of motion detection and alarming, this paper adopts the algorithm that combining based on background subtraction with time-difference algorithm, and taking this frame-difference algorithm to achieve the remote monitoring (motion detection and alarming). According the software standard development process requirement of video monitor and image analysis system survey, analysis, implement, to develop the Client / Server structure that take the embedded DVR / DVS as the core-based intelligent grain reserves surveillance system, to achieve video compression, storage, transmission, GIS scheduling, video surveillance equipment controlling and other important functions. Because of adopting the embedded DVR / DVS, the scale of video surveillance system can be applied adaptability in accordance with the actual requirements to deploy the dynamic, high-performance and flexible network.