改进三帧差分算法在移动物体检测中的应用

Moving Object Detection Method Based on Three Frame Difference and Edge Information Extraction

移动物体监控系统利用检测算法识别监控区域的移动物体,并进行实时异常信息存储,检测算法中的传统三帧差分法中的阈值是固定的,因此重叠部分无法准确检测出来,存在空洞现象,可能发生误判;针对这些问题,对已有的三帧差分法进行改进,结合图像边缘提取和自适应的迭代阈值计算方法来提高移动物体检测的准确性,并对异常信息进行选择性存储,由实验结果可知,应用平台采用改进后的移动物体检测算法,较好的提高了移动物体检测的灵敏度,增强了检测系统的实时性和准确性,若仅存储异常信息,可节省视频存储空间,并在定位异常动态信息时节省查找时间。

Mobile object monitoring system using the algorithm can detect moving target recognition of regional monitoring, and Real- time Anomaly Detection Algorithm in information storage, the traditional three frame difference method in the threshold is fixed, so the over- lap cannot be accurately detected, there is a void phenomenon, can be a miscarriage of justice. To solve these problems, the existing three frames the difference in the improved method, calculation method to improve the accuracy of the moving object detection with iterative threshold image edge extraction and adaptive~ and selective storage of the abnormal information, experiment results show that the improved moving target detection algorithm in the application platform, improve the sensitivity of moving target detection algorithm is better, enhances the real-time and the accuracy of the detection system, and save the storage space and time to find information.