Recognition and Positioning Method for Large Component Docking Based on Visual Measurement Field

Measurement and positioning is a key issue in the docking of large components such as aircraft, ships, and launch vehicles. Existing measurement approach mostly rely on high precision and large-scale single/multi-point measurement equipment, which can't satisfy the requirement of measuring curved surfaces, and the high hardware cost also limits their promotion and application. This paper proposed a measurement and positioning method based on visual measurement field for the docking process of large components. This method has the advantages of high measurement and positioning accuracy, large measurement range, and the ability to measure curved surfaces. First, a large-scale visual measurement field is constructed by multiple groups of visual cameras, and the pose calibration problem between multiple cameras is studied. Then, a 6D pose recognition and positioning algorithm based on the template matching principle was proposed, which solved the problem of target feature recognition and positioning in the visual measurement field. Subsequently, a multi-camera data fusion method is studied to improve the accuracy and reliability of measurement and positioning. Finally, the effectiveness of the proposed method is verified through the docking experiment of rocket launcher segments.