Research Detection Algorithm Of Moving Objects On Video

Intelligent Video Surveillance System, which does not require human intervention in the circumstances, uses computer vision technology and video analysis methods to analyse image sequences automatically, to obtain the positioning of the target, identification and tracking, and analyzing and judging the target behavior, which happened under unusual circumstances to complete a timely response, such as day-to-day management, can be widely used in border defense, battlefield surveillance, intelligent transportation systems, intelligent buildings and digital home systems, forest fire protection systems, oil refineries, airports and other important places to monitor system. Therefore, the video detection technology not only has important application value and a huge market potential.For video detection of the status quo and problems, this paper, the video images of moving objects under the detection methods, including video detection system of its features and system requirements, in accordance with its system requirements, design the installation of video detection hardware acquisition programs and hardware architecture, and According to the system functional requirements, design software systems technical route; studied the pretreatment of video detection algorithms and image background model in order to overcome the external environment changes, such as image noise caused by image interference, caused by the reduced accuracy of video detection. The prospects for video images of isolated points, the impact of noise and other interference, as well as the image object region segmentation, mathematical morphology based on open computing algorithm, and then study the images under the prospects of sport objects watershed region segmentation algorithm provides fast, accurate video detection algorithm.For video detection system to obtain accurate system requirements, this article analyzes the video detection system characteristics and system requirements, in accordance with its system requirements, design a video detection hardware to install programs and hardware acquisition architecture, implementation of the video data collection and information access for the latter part of the deal laid the foundation for software; according to the system functional requirements, design a software system for technical route, which includes video capture, video source fault diagnosis, video decompression, detection of regional identification, image filtering, background model, background Differential, regional segmentation, shadow detection, object recognition, motion tracking, behavior analysis and other parts, according to system requirements, a detailed analysis of each part of the functional requirements, the software algorithms for the late design and implementation has provided a guarantee.Pretreatment of video detection algorithms and image background model is an important aspect of video detection, image background model are the prospects for effective access to an important prerequisite for moving objects with the foundation. Pretreatment methods including the effective detection region to identify, filter processing operation, the effective detection region to determine the use of prior knowledge required Detect the key regions, can be concerned about the invalid region, reducing the computational complexity, improve the computing speed; Image Filtering Contrast analysis of the merits and demerits of different filtering method, recommend the use of video detection filter Median filter, filter treatment to reduce the impact of noise and other interference; overcome the external environment in order to effectively change the background image changes caused by the video detection to reduce the impact of accuracy, using the adaptive Gaussian mixture background model to obtain the background and the background update, in order to rapidly calculate the parameters of Gaussian mixture model, using the K-means clustering algorithm, to estimate the Gaussian mixture model parameters. Thus, in accordance with the current video image frame and background image of the differential access to the prospect of objects, objects for the sports division of the latter creates the conditions.Against the prospect of image noise, interference and other effects of isolation, caused by the prospect of misuse of object segmentation using mathematical morphology open computing smoothing object boundaries and isolated points, the structural elements smaller than deburring filtered out, cutting off long lap The separation of the role played, to remove noise and isolated points as well as the impact of the border burr for image segmentation has laid a foundation; accurately obtain the object of each exercise regional implementation objects are an important basis for detection, this paper watershed image segmentation algorithm based on the future Thus, multiple access to each moving object region, implementation of the moving object detection.Experimental process, this paper has developed a VC 2005 Video Detection experimental prototype system and to make use of video detection sequence, to conduct experiments on video sequences to run automatically Detect. The experimental results show that the video detection algorithm is effective and can meet the needs of video detection. Therefore, this article could be for video detection technology to provide a sense of reference.