建筑物的空间信息可以为城市规划、土地管理等提供重要参考.近年来,利用遥感方法提取建筑物引起了许多学者的关注,其核心是提取能够准确描述建筑物的特征,将其与其他物体区分开来.机载激光雷达获取的点云可以准确描述物体的三维特征,但具有离散性;航空图像具有丰富而连续的纹理信息,但缺乏对空间位置的描述.此外,现有特征之间可能存在冗余,从而降低了效率和准确性.为此,本文提出一种基于机载LiDAR(light detection and ranging)点云与航空影像融合的建筑物提取方法.首先,将两类数据源配准融合,实现更丰富的特征提取;其次,使用特征选择方法,降低特征维度,减少计算消耗的同时提高提取精度.实验结果表明,所提方法的有效性,优于生产中常用的LiDAR-Suite和Terrasolid软件.
A Building Extraction Method Based on LiDAR Point Cloud and Image
The spatial information of buildings can provide im-portant references for urban planning,land management,etc. The extraction of buildings with remote sensing methods has attracted the attention of many scholars in recent years. The core is to extract features that can accurately describe the buildings,and then distinguish them from others objects. The point cloud acquired by airborne LiDAR can accurately de-scribe the three-dimensional features of objects,but it is dis-crete. Aerial images have rich and continuous texture informa-tion but lack the depiction of the spatial location. In addition,there may be redundancy among existing features,reducing the efficiency and accuracy. In this paper,a building extrac-tion method for the fusion of the LiDAR point cloud and the aerial image is proposed. The registration and fusion of these data can achieve richer feature extraction. Then,the feature selection method is employed to reduce the dimension,and lessen the calculation consumption and improve the accuracy. Experimental results show the effectiveness of the proposed method. It is superior to the LiDARSuite and Terrasolid soft-ware that are commonly used in production.
building extractionairborne LiDARaerial imagefeature selection
万方数据
维普
