Pose Estimation in Automated Visual Inspection System using PSO

In this paper, we propose a Particle Swarm Optimization (PSO) based approach to determine the pose of a rigid body in Automated Visual Inspection having three degree of freedom. We have experienced the effect of noise at 20 dB SNR and 40 dB SNR and also observe mismatch getting from incorrect correspondence between object space points and image space points, on the estimation of three parameters. The maximum error in translation parameters (about x-axis and y-axis) is less than 0.1 cm and rotational error is less than 0.08 degree at 40 dB SNR and 0.45 cm and 0.6 degree at 20 dB SNR respectively. The error in parameter estimation is insignificant up to 5 pairs of mismatched points out of 16 points and results skyrockets when the mismatch occurs in more than 5 pair of points. These results have proved the robustness of PSO in determining the pose of object in automated visual inspection.