论文部分内容阅读
机载激光雷达(LiDAR)技术的出现为地面汽车目标检测提供了新的途径。为了从机载LiDAR点云数据中提取汽车对象,根据不同地物的属性特征,提出了一种航空影像辅助下的城区机载LiDAR汽车目标检测方法。首先利用形态学开重建滤波完成地面和地物的分类,然后在地物点的基础上结合正射影像,通过归一化植被指数(NDVI)特征完成对植被和非植被地物的初步分类,最后在非植被地物的基础上,根据地物对象的形状特征及高程信息完成汽车和建筑物及阴影植被等非汽车对象的分类,从而完成汽车目标的提取工作。3个实验区的计算结果表明:该方法能有效从LiDAR点云中提取汽车目标,正确度和完整度的均值分别为95%和85%,满足实用性要求。
The emergence of Airborne Lidar (LiDAR) technology provides a new approach for the target detection of ground vehicles. In order to extract the car object from the on-board LiDAR point cloud data, an airborne image-assisted LiDAR vehicle target detection method is proposed based on the attributes of different features. Firstly, morphological reconstruction filtering is used to classify the ground surface and ground objects. Then, based on the features of the ground objects and the orthophotos, the classification of vegetation and non-vegetation ground objects is completed by the normalized vegetation index (NDVI) Finally, on the basis of non-vegetation features, the classification of non-vehicle objects such as cars and buildings and shadow vegetation are completed according to the shape and elevation information of objects, so as to complete the extraction of vehicle targets. The experimental results show that the proposed method can effectively extract the vehicle target from the LiDAR point cloud, with the mean of accuracy and completeness being 95% and 85%, respectively, which meets the practical requirements.