当前的汽车安全辅助驾驶和无人驾驶汽车是图像领域的研究热点,针对汽车在启动或行驶时车前存在行人可能导致的安全问题,着重研究了基于双目视觉的车前行人检测方法.进行了双目相机的相机标定和立体标定;通过改进后半全局立体匹配算法获取深度图,确定车前行人所处位置的感兴趣区域(Region of Interest,ROI),剔除冗余的背景信息;分割并提取了图像的降维梯度直方图(Histogram of Gradients,HOG)特征信息;将特征输入到支持向量机(Support Vector Machine,SVM)分类器训练,检测并标记出车前的行人目标.实验证明,所提算法对车前场景下的动态行人可以更为有效地检测,具备更优的检率精度、时效性和鲁棒性.
Research on Pedestrian Detection Method in Front of Vehicle Based on Binocular Vision
At present,vehicle safety assisted driving and autonomous vehicle are the research hotspots in the field of image.For the safety problems that may be caused by pedestrians in front of the vehicle when the vehicle is starting or running,the method of pedestrian detection in front of the vehicle based on binocular vision is studied.Firstly,the camera calibration and stereo calibration of binocular camera are carried out.Then,the depth map is obtained by an improved semi-global stereo matching algorithm to determine the Region of Interest(ROI)of the position of pedestrians in front of the vehicle and eliminate redundant background information.The feature information of the dimension reduction Histogram of Gradients(HOG)is segmented and extracted.The features are input into the Support Vector Machine(SVM)classifier for training,and the pedestrian targets in front of the vehicle are finally detected and marked.Experiments show that the proposed algorithm can detect dynamic pedestrians in front of vehicles more effectively,with better detection accuracy,timeliness and robustness.
国家科技期刊平台
NSTL
万方数据
维普
