基于视频图像的交通参数检测  

Traffic Parameter Detection Based on Video Images

作　　者：杨玉珍[1] 贺晓锋[1] 陈阳舟[1] 鲍占阔[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100022

出　　处：《控制工程》2008年第3期349-352,共4页Control Engineering of China

基　　金：北京市交通工程重点实验室开放基金资助项目(2006)

摘　　要：针对交通图像的特点,提出了一种将灰度判断与边缘检测相结合的方法,来获得车流量、车速等相关交通参数。将先锋遗传算法(Vanguard Genetic Algorithms)应用到图像阈值分割中,利用先锋遗传算法寻求全局最优阈值,可以比较准确地将图像中不同灰度的车辆从背景中分离出来。分析了小波变换的特点,应用小波变换对交通图像进行车辆边缘检测。采用帧平均法来处理视频流,可以减小由于摄像头抖动或背景微小变化而产生的误差。实验结果表明,该方法简单、有效,能够获得较满意的检测结果。Considering the characteristics of traffi image,to get traffic parameters such as vehicle flow and speed, a measurement of tragic parameter detection based on the combination of gray value aud edge detection is proposed. It makes ttse of vanguard genetic algorithm in image segmentation, and uses vanguard genetic algorithm to get the optimal threshold. Using the method, different grey vehicles can be gotten from the background exactly. The feature of wavelet transform is analyzed, and the edge of vehicle is detected with wavelet transform. A frame averaging method is introduced to reduce the error brought by camera wobbling or background changing. Experimental results show that it is a simple and feasible method.

关 键 词：交通参数检测 先锋遗传算法 小波变换 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]