Optimal 3D surface metrology-localization of fiducial points

The authors have developed a method for optimal estimation of surface fiducial point locations based on optically sensed range maps obtained with a specially designed apparatus originally developed for portrait sculpture. The algorithm used is based on a Kalman filter, a recursive spatially variant optimal estimator. The results demonstrate that accurate localization of surface landmark points can be readily achieved by using this method. The algorithms developed have been implemented and simulations have been used to judge robustness, accuracy, and overall performance. The use of the covariance matrix to carry forward the errors in different parameters has resulted in stable and robust estimates. This technique has application in industrial photogrammetry and machine vision