面向对象的机载LiDAR数据建筑物提取

发布时间：2018-07-01 07:52

本文选题：机载LiDAR + 建筑物提取　；参考：《应用科学学报》2016年01期

【摘要】：基于Li DAR数据,提出一种由粗到细的面向对象的建筑物自动提取方法.首先通过机载Li DAR数据构建出归一化数字表面模型(normalized digital surface model,n DSM),利用首尾两次回波高程计算出归一化差值(normalized difference,ND),并采用形态学运算消除边缘特殊回波点.基于n DSM和ND数据,依据建筑物的高程及穿透性信息,用阈值分割法进行建筑物粗提取.结合n DSM和ND数据以及强度信息,对粗提取得到的备选建筑物采取多尺度分割,合并亮度值相差较小的邻近分割结果对象,达到对分割结果的优化处理.最后利用目标对象的亮度、形状、面积和空间关系等特征,完成建筑物的精提取.实验结果表明,该方法可得到较高精度的建筑物信息,是基于机载Li DAR数据提取建筑物的新思路.
[Abstract]:Based on Li Dar data, an object-oriented automatic building extraction method from coarse to fine is proposed. Firstly, a normalized digital surface model (normalized digital surface modeln DSM) is constructed from airborne Li Dar data, and the normalized difference (normalized difference ND) is calculated by using the first and last echo heights, and the edge special echo points are eliminated by morphological operation. Based on the data of n DSM and ND, the rough extraction of buildings is carried out by threshold segmentation method according to the height and penetration information of buildings. Based on the data of n DSM and ND and intensity information, multi-scale segmentation is adopted for the rough extracted alternative buildings, and the adjacent segmentation objects with small difference in luminance are combined to achieve the optimal processing of the segmentation results. Finally, using the brightness, shape, area and spatial relationship of the target object, the fine extraction of the building is completed. The experimental results show that this method can obtain high accuracy building information and is a new idea for building extraction based on airborne Li Dar data.
【作者单位】：中国矿业大学环境与测绘学院;
【基金】：国家自然科学基金(No.41331175)资助
【分类号】：TU198;TN958.98

【相似文献】