基于图像差分的背景差分算法改进及实验数据分析

Improvement and Experimental Data Analysis of Background Difference Algorithm Based on Image Difference

针对交通视频中车辆移动检测问题,结合交通视频镜头相对固定、实时性强的特点,研究了基于图像差分运动目标检测算法中的背景差分算法,在真实的环境中对运动的目标区域完成实验检测,然后对检测数据进行分析,在背景估计和背景更新算法上做了相应的改进。通过对实验数据的观察和分析,背景差分法在交通视频车辆移动目标检测中具有较好的适用性和通用性,能较好地对交通视频车辆移动目标进行检测。

Abstraet：The detection of moving target based on the video is critical in traffic parameter detection system,and the detection data of moving target must be accurate and efficient,so as to make full use of the traffic video information to lay the foundation for intelligent detection algorithm and moving target tracking at a later stage.This paper focuses on moving vehicles detection in traffic video.The traffic video is considered with a fixed and real-time lens.The paper mainly researches the background difference algorithm based on image difference of the moving target detection algorithm,completes the experiment detection for the moving target region in a real environment,and analyzes the test data to make corresponding improvements in the background estimation and the background update algorithm.Under the observation and analysis of experimental data,the paper draws a conclusion that the background difference is more applicable and universal in detecting moving target in the traffic video,and can well detect the moving target in traffic video.