A Method for Target Detection and Tracking of the Recycling Robot in Nuclear Power Plant

The detector assembly recovery robot can replace the manual recovery of the detector assembly in the core of the nuclear power plant. To achieve the visual servo control for the recycling robot, this paper adopts a target detection and tracking method for the recycling robot in a nuclear power plant. Through the image processing algorithm, the target area is segmented to greatly reduce the computational amount in the recognition process and obtain more accurate target pixels. The compression tracking algorithm is used to track the detector components in real time to ensure the real-time information acquisition. The positioning accuracy of the visual servo control system was tested by repeated localization experiments, which show that the positioning accuracy is better than 2mm.