On MNCP in two-stage technique for 3D camera calibration for machine vision metrology

The two-stage technique for three dimensional (3D) camera calibration for machine vision metrology developed by Tsai [1987] is reviewed. The technique has many advantages in terms of accuracy, speed, and versatility over others. However, the technique, which needs a lot of the calibration points (CP), has been limited in real-time video signal processing. The minimum number of calibration points (MNCP) needed in the technique is studied in the paper. The investigation results indicate that there exists the MNCP which is less than the number of CP needed in Tsai. To prove that there exists a MNCP, several lemmas are put forward and proved in detail. The computer simulation and experiment results are given