基于植被指数限制分水岭算法的机载激光点云建筑物提取

发布时间：2017-12-28 21:10

本文关键词：基于植被指数限制分水岭算法的机载激光点云建筑物提取　出处：《光学学报》2016年10期 　论文类型：期刊论文

【摘要】：建筑物提取在建筑物重建和城市管理中起着重要的作用。利用基于植被指数限制的分水岭算法分割机载激光雷达点云,并利用一定的规则识别建筑物区域。对激光点云进行内插生成网格数据;利用植被指数限制的分水岭分割算法分割激光点云生成的数字表面模型数据,在分水岭淹没过程中引入植被指数可以较好地区分建筑物和植被区域;在区域相邻关系的基础上,利用一些准则(高程差值、尺寸和植被指数)识别建筑物区域。利用国际摄影测量与遥感学会基准数据中法伊英根测试区域对建筑提取结果进行评价,在像元级别,平均完整度、正确度和质量分别为89.2%、94.3%和84.7%;在对象级别,平均完整度、正确度和质量分别为81.8%、93.1%和76.9%;在物体面积大于50m2的对象级别,平均完整度、正确度和质量可以达到99.1%、100%和99.1%。
[Abstract]:Building extraction plays an important role in building reconstruction and city management. A watershed algorithm based on the Vegetation Index Limited is used to segment the airborne laser radar point cloud, and a certain rule is used to identify the building area. The laser point cloud data interpolation mesh generation; digital surface model data segmentation algorithm of laser point cloud generated by watershed vegetation index limit, introduced in the process of watershed submerged vegetation index can distinguish the buildings and vegetation areas; based on region adjacent relation, using some criteria (size, height difference the building area recognition and vegetation index). The extraction results to evaluate the building using the ISPRS benchmark data in the test area of Sino French Iran yngen, pixel level, average integrity, accuracy and quality were 89.2%, 94.3% and 84.7%; the average at the object level, integrity, accuracy and quality were 81.8%, 93.1% and 76.9% in the area of the object; more than 50m2 of the object level, average integrity, accuracy and quality can reach 99.1%, 100% and 99.1%.
【作者单位】：武汉大学遥感信息工程学院;
【基金】：国家自然科学基金(41571434,41171292)
【分类号】：P237
【正文快照】： 建筑物区域提取是建筑物重建的必要前提,在城市规划和管理中起着重要作用[1-2]。机载激光雷达(LiDAR)是一种非常有效的遥感手段,可以获取坐标和高程三维信息[3-4]。可以利用获取的三维点云生成数字高程模型(DEM)数据[5-7]、识别线特征[8-10]、提取植被区域[11-12]和建筑物区域

【相似文献】