摘要
目的提出一种基于图像处理的车辆排队长度自动检测方法。方法首先,提出基于固定背景区域窗口最大相似度配准法来校正实时采集图像因抖动的干扰所引起车道区域的偏差;接着,将交通图像转换为简约纹理单元的纹理图像,突出有车辆车道与空车道的纹理差异,结合最大方差阈值分割算法,分离出车辆与空车道;最后,实现一个基于队尾空车道区域特征的排队长度自动检测算法。结果实现了车辆排队长度的自动检测。结论试验结果说明该方法具有较理想检测效果,能有效地克服抖动与光照干扰。
Objective To put forward an automatic extraction and analysis method based on image processing in crossing vehicle queue length. Methods Firstly, put up anti-jamming method. Secondly, according to segmentation arithmetic in maximal square difference, separate vehicles from roadway. Lastly, carry out an automatic extraction algorithm of vehicle queue length. Results Automatic extraction and analysis of vehicle queue length is achieved. Conclusion The result tested indicates that the approach is effective.
出处
《中国体视学与图像分析》
2006年第3期216-221,共6页
Chinese Journal of Stereology and Image Analysis
基金
福建省自然科学基金项目(编号:A0510010)
关键词
图像处理
排队长度
自动检测
智能交通系统
image processing
vehicle queue length
automatic extraction
intelligent traffic system
