Optimised calibration of machine vision system for close range photogrammetry based on machine learning

Real-time inspection of large mechanical parts manufacturing using camera-based scanning systems are increasingly adopted in industry 4.0. It leads to take preventive actions during the process and then fabricate right-first-time with respect specified tolerances. Therefore, use scanners requests a preliminary calibration process. consists on estimating intrinsic extrinsic parameters required relate 3D world point its projection image plane. Since selection grid poses affect quality, one approach-based machine learning (ML-approach) is proposed including polynomial approximation reprojection errors function 6 degree freedom (DoF) combined particle swarm optimization (PSO). Synthetic experimental evaluations have been performed while assessing performance ML-approach. The synthetic evaluation reveals better convergence comparison recent published methods by Wizard (CW-method) Rojtberg (R-method). ML-approach shows an average error RE < 12 µm sub-micrometre repeatability, which confirm benefit vision-based for volume real time.