结合Steger方法和影像分类的道路中心线提取  

Road Centerline Extraction Based on Steger Method and Image Classification

作　　者：邓巧 周绍光[1] 胡屹群 王馨苑 

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2期64-67,共4页Geospatial Information

基　　金：国家自然科学基金资助项目(41271420/D010702)

摘　　要：针对直接利用监督分类结果提取道路中心线过程过于复杂的问题,研究一种综合利用Steger方法和影像分类结果的道路中心线提取新策略。以影像分类结果从Steger方法获取的条带中心线中初步筛选出候选道路中心线段,根据道路的连续性和方向的一致性进行合理延伸、连接,即可实现道路中心线提取。实验证明,提取的道路中心线在完整率、准确率和提取质量方面精度较高,并在一定程度上解决了道路断裂问题。Extracting road centerline through image classification directly was complicated. To solve this problem, a new strategy which combined the Steger method strip centerline result with classification result was studied to get the road centerline in this paper. Firstly, the strip centerlines acquired by Steger method were preliminarily selected by the classification result directly. Then the road centerline result was extracted by extending and connecting the selected result reasonably based on the continuity and consistence in direction of the road. The experimental result indicates that the accuracies of completeness, correctness and quality are efficient, and this method partly solves the road fracture problem.

关 键 词：Steger方法 影像分类 道路中心线提取 

分 类 号：P237[天文地球—摄影测量与遥感]