基于视频动态检测的报警算法和装置毕业设计.docVIP

基于视频动态检测的报警算法和装置设计 摘 要 本文针对已经初具规模的视频监控网络，尝试使用运动物体检测算法，设计一套能够稳定检测动态视频异常，并通过报警装置提示工作人员异常出现的系统。与过去常常采用的传感器或者传感器网络组成监控系统相比，视频监控系统有着直观、方便、信息内容详实等优点。加上运动物体检测算法后，基本能够达到传感器网络的报警灵敏度，同时多手段的监测也为事故后的取证提供了多方面的信息和凭证。能有效的减少现场工作人员，视频监控人员的工作强度，更重要的是提高报警效率及准确度。 本文通过编程仿真，比较了运动物体检测的三种主流算法的优缺点，最终选择背景差分作为动态监测的核心方法。通过matlab环境下的大量实验，确定了帧差分算法的阈值。另外，由于环境噪声、干扰等问题的存在，使用中值滤波和形态学的方法处理帧图像，使之能够通过运动检测算法将图像异常检测出来。最后，使整个系统能够良好的检测动态视频中的异常，并通过报警装置提醒监控人员或现场工作人员。 关键词：视频监控，报警算法，运动检测 Design of Alarming?Algorithm and?Device Based on Video?Motion Detection ABSTRACT With a certain scale of video surveillance network, this paper try to use the motion detection algorithm to design a stable system which can detect the anomalies in dynamic video and inform the staff of the anomalies through the alarm device. And compared with sensors or sensor network which are often used in monitoring system in the past, video surveillance system is convenient, intuitive and detailed informative, etc. Using motion detection algorithm, video surveillance system can mostly achieve alarm sensitivity of sensor networks, and also means of monitoring for the forensics may provide a various information and documents after the accident. The system can effectively reduce the on site working staff and video monitoring personnel working strength, most importantly to improve the efficiency and the accuracy of alarm. By the simulation programming, this paper compare the three main motion detection algorithm for the advantages and disadvantages, and ultimately select the background differential method as the core of dynamic monitoring. Through a large number of experiments in Matlab, determine the threshold of the frame difference method. In addition, due to environmental noise, interference and other problems, the use of median filtering and morphological approach to the frame make the changes can be better detected through the motion detection algorithm. Finally, the entire system can detect the motion of a good exception in video, and inf