基于GPS车辆数据和图像配准的道路中心线提取  

Road Centerline Extraction Based on GPSVehicle Data and Image Registration

作　　者：王馨苑 周绍光[1] 胡屹群 

机构地区：[1]河海大学地球科学与工程学院,江苏南京210098

出　　处：《地理空间信息》2018年第1期48-50,64,共4页Geospatial Information

基　　金：国家自然科学基金资助项目(41271420/D010702)

摘　　要：GPS车辆数据可提供详细的道路网信息,而它可能存在数据误差和冗余问题,可采用形态学处理中的细化操作得到近似道路中心线。针对细化操作得到的曲线弯曲度过大且存在孤立数据点、数据冗余等情况,以得到的近似道路中心线为基准线,利用霍夫变换获取图像配准的起始点,并结合图像配准得到平滑无冗余的道路中心线。以马鞍山车辆数据为例,实验结果表明,与细化操作相比,该方法在道路中心线提取质量上有了显著提高。In view of the data errors and data redundancy of GPS vehicle data,this paper used the refinement operation of morphological processing to receive the approximate road centerline.There also existed some problems in the refinement operation like isolated data points,data redundancy and large curvature of the curve,etc.To solve this,based on the approximate road centerline,the paper used Hough transform to obtain the starting point of image registration,and combined the image registration to obtain the smooth road centerline without redundancy.Taking the Ma'anshan vehicle data for example,the result shows that this method has a significant improvement in the quality of the road centerline compared with the refinement method.

关 键 词：GPS车辆数据 细化 图像配准 霍夫变换 道路中心线 

分 类 号：P237[天文地球—摄影测量与遥感]