利用局部梯度最大值的半自动道路提取  

An algorithm for semi-automatic road extraction based on local gradient maximum

作　　者：岳贵杰[1] 杜黎明[2] 刘凤德[3] 项琳[3] 张刚[3] 李健[3] 

机构地区：[1]首都师范大学资环学院,北京100048 [2]河南城建学院测绘工程学院,河南平顶山467036 [3]中国测绘科学研究院,北京100039

出　　处：《测绘科学》2015年第5期89-91,105,共4页Science of Surveying and Mapping

基　　金：国家科技支撑计划课题项目(2011BAB01B05);国家自然科学基金项目(41101435);地理空间信息工程国家测绘地理信息局重点实验室开放研究基金课题(201328);对地观测技术国家测绘地理信息局重点实验室开放课题(K201412)

摘　　要：针对目前测图中半自动道路提取的现状和不足,文章提出了一种基于局部梯度最大值的道路测图算法:根据人工输入的初始条件,将影像上局部梯度最大值点作为道路边缘兴趣点,利用基于局部梯度最大值点的边缘扩展算法建立道路跟踪模型,并将该算法应用到资源三号卫星(ZY3)应用系统三线阵数字线划图生产软件中半自动道路识别。针对资源三号卫星影像乡村道路实验表明,该方法操作简便、符合作业习惯,能实时进行半自动道路测图,降低了道路测图的工作强度,提高了道路边缘的定位精度和测图效率,具有较好的应用价值。Aiming to the current status and the insufficiency of road extraction in mapping, a semi-au- tomatic road mapping algorithm based on local absolute maximum gradient was introduced and realized in the paper. The method constructs the road edge extraction model according to the edge growing algorithm based on local absolute maximum gradient and manual seeds initialization. Finally, the extraction algo- rithm was applied to the semi-automatic road extraction for digital liner graph software of satellite mapping system of ZY-3. Results for ZY-3 rural road extraction showed that the model could extract road edge easi- ly with accuracy and meet human＇s operation habits, which would be of good application value in reducing human work and improving locating accuracy and efficiency.

关 键 词：半自动道路提取 资源三号卫星 梯度 局部梯度最大值点 

分 类 号：P283[天文地球—地图制图学与地理信息工程]