基于类Haar特征和自适应提升算法的前车识别

Front Vehicle Identification Based on Haar-like Feature and Adaptive Boosting Algorithm

针对汽车高级驾驶辅助系统(ADAS)中前方车辆识别率低的问题,基于机器视觉原理研究了前方道路图像中的类Haar特征;并进行积分图计算。在提取类Haar特征基础上,采用自适应提升(Ada Boost)算法进行正负样本训练并级联,得到训练好的模型;进而检测和识别汽车行驶中前方车辆。最后基于Open CV计算机视觉库在Visual Studio开发环境中进行了算法实现和测试。结果表明,每帧视频图像识别时间小于40 ms,检测率准确可靠,满足多场景、多工况下的前方车辆实时识别。

Aiming at the low rate of front vehicle recognition in the advanced driver assistance system( ADAS),the haar-like features in the road ahead image are studied based on the principle of machine vision,and the integral graph calculation is performed,the adaptive boosting( AdaBoost) algorithm is used to train the positive and negative samples and cascade to obtain the trained model on the basis of extracting the haar-like features,then detect and identify the vehicle in front of the vehicle. Finally,the algorithm is implemented and tested in the visual studio development environment based on open source computer vision library. The results show that the recognition time of each frame of video image is less than 40 milliseconds,the detection rate is accurate and reliable,and it can meet the real-time identification of front ve hicle in multiple scenes and multiple working conditions.