| 摘要: |
| 针对现有地下井室病害探测与维护方法的不足，提出了一种基于Kinect三维重建的地下井室可视化方法，以实现其三维可视化探测与维护。还提出了一种基于多项式曲面拟合的Kinect深度测量误差修正方法，利用联合双边滤波算法对深度图像数据进行预处理；结合SIFT特征匹配和改进的RANSAC算法获取相邻点云间的初始位姿，并利用基于邻域特征的ICP算法进一步实现不同视角点云的精确配准，从而获取全局一致的稠密三维点云。最后，在三维稠密点云的基础上进行曲面重建和纹理贴图，以实现地下井室真实三维重建。实验结果表明：所提方法可有效修正Kinect深度相机的深度测量误差，在0.5~4.5m的测距范围内，其三维重建精度可达2cm；在4.5~7m的测距范围内，精度也可以保持在4.5cm以内。所提重建方法可实现地下井室真实场景的三维可视化，为地下井室的探测和维护提供了技术支持。 |
| 关键词: Kinect 三维重建 地下井室 深度误差 点云 |
| DOI： |
|
| 基金项目: |
|
| Research on Visualization Method of Underground Well Chamber Based on Kinect 3D Reconstruction |
| YU Zhao-kai,PENG Xiao-feng,QIU Chang-jie |
| (PowerChina Guizhou Electric Power Engineering Co., Ltd., Guiyang 550003, China) |
| Abstract: |
| To address the shortcomings of existing underground wells disease detection and maintenance methods, a visualization method of underground wells based on Kinect three-dimensional reconstruction is proposed to realize its three-dimensional visualization detection and maintenance. A Kinect depth measurement error correction method based on polynomial surface fitting is proposed, and the joint bilateral filtering algorithm is used to preprocess the depth image data. SIFT feature matching and the improved RANSAC algorithm are used to obtain the initial pose between adjacent point clouds. The ICP algorithm based on neighborhood features is used to further realize the precise registration of point clouds from different perspectives, and to obtain a globally consistent dense 3D point cloud. Finally, surface reconstruction and texture mapping are per-formed on the basis of the 3D dense point cloud to realize the reality of underground wells 3D reconstruction. The experimental results show that the proposed method in this paper can effectively correct the depth measurement error of the Kinect depth camera, and its 3D reconstruction accuracy can reach 2cm in the range of 0.5~4.5m. In the range of 4.5~7m, it can also be kept within 4.5cm. The proposed reconstruction method can realize the 3D visualization of the real scene of underground wells, and provides technical support for the detection and maintenance of underground wells. |
| Key words: Kinect 3D reconstruction Underground wells Depth error Point cloud |
