基于汽车前照灯的夜间车辆定位方法研究

全天候车辆视频检测白天和黑夜车辆时其检测条件差异很大,要选择不同的检测方法。在夜间无补光光源环境下,将摄像机获得的彩色视频图像进行灰度处理,夜间图像中汽车前照灯具有很强的特征,因此对灰度图像进行二值化处理和灰度统计来提取前照灯的特征,根据汽车前照灯在画面中的形态特征设计了相应的定位算法,实现了夜间车辆的定位检测。实验结果表明,该方法实现夜间车辆定位的突出特点是定位准确,且定位时间短,满足了视频交通系统实时性的要求。

Vehicle Positioning Method at Night Based on Headlamp

As the detection conditions of vehicle detection at day are much different from those at night,it is necessary to choose different detection methods at different time.At night without supplementary light,the color image of headlamp should be conducted by gray-scale processing.The image has a strong characteristic,which should be conducted by the binarization processing and grayscale statistics of the gray-scale image to extract the characteristics of the headlamps.According to the morphological characteristics of the headlamo image,an algorithm is advanced.Thus,the vehicle positioning at night is realized.Experimental results show that the method is characterized by accurately positioning and needing less time,which meets the requirements of the real-time video transport system.