基于智能检测车位信息的车辆识别算法

Vehicle Recognition Algorithm Based on the Intelligent Detecting Parking Information

摘要: 针对停车场车位图像信息进行边缘检测，智能标定车位坐标信息，提取汽车特征，通过训练分类器，加载训练好的汽车特征库，然后对感兴趣车位区域图像进行特征匹配，检测是否存在车辆，若存在车辆，则通过比对车辆停放位置坐标和车位坐标信息来判定车辆停放位置是否正确，结果表明，该算法对于室内停车场识别率高，处理速度较快。该算法具有全智能化的特点，满足实时性，准确性，鲁棒性的要求。

Abstract: The algorithm uses Edge Detection to calibrate intelligently coordinate information of parking spaces image information, then extracts the characteristics of automotive, trains classifiers and loads trained feature library. The paper aims at interested region of parking space to detect whether there is vehicle by feature matching, then judge whether the vehicle’s parked position is correct through comparing vehicle’s parking location and parking space coordinate information if vehicle exists. The results show that the algorithm is suitable for bright indoor carpark conditions, with a high recognition rate of vehicle and faster processing. The algorithm has the full intelligent features, and meets the real-time, accuracy, robustness requirements.