基于OpenCV实现多特征融合的移动车牌定位算法

IMPLEMENTING MOVING VEHICLE LICENSE PLATE LOCATING ALGORITHM WITH MULTI-FEATURE FUSION BASED ON OPENCV

针对移动车牌的实时定位,研究更加快速准确的定位算法具有很强的现实意义。采用背景差分法对车辆进行识别,用特征匹配法实现车辆跟踪,对包含有车辆的帧进行灰度化、均值滤波、形态学操作,使用车牌多特征融合精确定位出车牌区域。该移动车牌定位算法在VS2008平台上利用OpenCV和C++编程实现。实验结果表明该方法速度快、准确度高,可满足实时监控的需求。

It＇s of great practical significance to study a faster and more accurate locating algorithm aiming at the real-time localisation of moving vehicle license plate（ VLP）. We use background subtraction approach to recognise vehicles,and use feature matching method to realise vehicle tracking. For the frames containing vehicles,we carry out greying,mean filtering and morphological operations. We use VLP multi-feature fusion method to precisely locate the VLP region. The vehicle license plate localisation algorithm is implemented by using OpenCV and C ＋ ＋ programming on platform of VS2008. Experimental result demonstrates that this algorithm has high speed and accuracy,it can meet the demand of real-time traffic monitoring.