Based on the analysis of the actual processing environment, this paper first builds a fast detection system based on binocular vision for the detection of the machining accuracy of ship curved plates. The visual sensors and lenses in the system are analyzed and selected. After selection, the binocular vision is calibrated using Zhengyou Zhang’s calibration method, and visual distortion is corrected through matrix transformation. Then, in order to complete the calculation of the curvature of the ship’s curved board, a curvature calculation model for the ship’s curved board was constructed. The model can calculate the curvature information and depth information of the curved board, and then solve the curvature information and depth information models separately using the least squares method. Finally, to verify the accuracy of the visual system, a simulation experimental platform based on software and hardware was built. After simulation experiments, the curvature error probability and measurement time of the curved board were obtained, and the results met expectations.
本站僅提供期刊文獻檢索。
