3D Modeling for Slope Angles and Volume of Underground Conduits Using RGB-D Camera

Abstract

This paper proposes a 3D modeling system to measure the slope angles and volume of underground electrical conduits installation. The proposed system consists of an RGB-D camera and a RTK device for data acquisition, and exploits AR markers when capturing 3D data. The RGB-D data at the consecutive frames are matched and reconstructed as 3D point clouds for the surfaces of underground conduit. The RTK data compensate for the scale errors in the 3D point cloud which geometric calibration has done with the RGB-D camera. AR markers indicate the starting and end cross-sections in the reconstructed point cloud data. To measure the slope angles of side surfaces and volume of underground conduit, first, the front direction of 3D point cloud data is accurately determined while minimizing the area of cross-sectional convex hull. Then, the piecewise slope angles and volumes are calculated and accumulated for the quantized y and z points. Histogram analysis is proposed to statistically determine the cross-sectional convex hull and point cloud on the conduit surfaces. According to tests with acquired synthetic and real data, the proposed system determines the slope angles and volume of the electrical conduits. For synthetic data, the proposed system shows errors within 1%, and for real data, it achieves errors within 3% compared to highly precise measurements. The proposed system demonstrates a capacity to measure underground conduits with consistency and precision. Consequently, this system is expected to be effectively applied to analogous construction environments.