An Automatic Laser Scanning System for Objects with Unknown Model

Abstract

Laser scanning has been widely used in the industrial manufacturing, especially in reverse engineering, quality control and surface defect detection. Achieving automatic scanning can succeed in reducing production costs and increasing production efficiency. This paper proposes an automatic intelligent scanning system based on ROS, including robot arm, RGBD camera and line laser scanner. The system is designed to scan objects with unknown model and we use high precision calibration methods to improve the accuracy of the system. A novel scanning trajectory planning strategy is developed, including three steps. Firstly we plan preliminary trajectory based on the RGBD camera data. Then the robot arm pose is adjusted with PID controller in real time to optimize the scanning results. Finally, according to the distribution of point cloud density, the portion whose density is lower than the threshold will be re-scanned. The experiment we illustrate shows the feasibility and high accuracy of our system.