基于点云的建筑物表面损伤三维重建与属性提取方法

3D Reconstruction and Attribute Extraction of Building Superficial Defects Based on Point-cloud

传统的建筑物外表面损伤检测手段对于不规则表面的几何属性提取显得十分困难,这也对于损伤信息的准确采集造成很多阻碍。针对这些问题本文提出了基于GIS环境的点云融合建筑物表面损伤三维重建算法,在采集端使用Kinect提取场景和深度信息,可以应用于建筑表面损伤测量过程中。首先利用PCL算法对Kinect采集到的点云数据进行预处理和格式转化,将点云数据的坐标系转化为具体构件所对应的坐标系来对损伤定位。然后利用Meshlab进行点云切割和关键区域的信息提取,将处理后的点云数据在GIS环境中转化成深度图像,从而实现了对不规则损伤部位的量化属性(如体积、面积、深度等)提取和三维重建。实验表明,采用该方法进行损伤三维重建能有效减少计算工作量,精度较高。

It is difficult to extract the geometric attributes of irregular surfaces through traditional methods of detecting building superficial defects,which also causes obstacles in accurately gathering information on building defects. As for these issues, the paper proposes a three-dimensional reconstruction algorithm for building surficial defects based upon the point cloud in GIS environment.At the acquisition side,we use Kinect to select color information and depth information,which can be used in measuring the defect area of buildings. Firstly,it uses PCL algorithm to conduct preprocessing and format conversion on point cloud data collected by Kinect,and converts coordinate system of point cloud data into the coordinate system of corresponding building components to detect defects. Then,Meshlab is used to split point cloud data and extract information from crucial areas. It also transforms the processed point cloud data into range images in GIS environment so as to achieve extraction and conduct three-dimensional reconstruction for quantitative attributes（ such as volume,area and depth）of defect areas. The result shows that this method can simplify the computational process with higher accuracy for 3 D reconstruction.