Vision Location Method for Key Features of the Metallurgical Machine with a Robot 3D Measurement System

Visually-guided installation of the long nozzle with an industrial robot is a novel technique, which is urgently needed for molten steel casting. The main problem is to automatically find the location of the outlet for the installation of the long nozzle. In order to realize the process, a robot 3D (three-dimensional) measurement system is set up. Based on the system, this paper proposes a new calibration and location approach based on the criteria spherical array composed of three criteria spheres. By creating the local coordinate system of the criteria spherical array, the objective function of the hand-eye calibration is built and solved. And then, the feature location is established. The proposed approach is efficient and reliable for the location of the metallurgical machine. In order to verify the effectiveness and reliability of this method, simulation experiments are carried out. The mean location error is 1.3491mm, which can meet the accuracy requirements.