Recognition device of the three-dimensional position and orientation based on the recognition method as well as the vision of the three-dimensional position and orientation based on the visual

摘要

PURPOSE: To enable recognizing the position and posture of a camera for photographing a known object body speedily, accurately and robustly. ;CONSTITUTION: Image feature points are calculated in an image processing part 7 using the signals from a camera 4 photographing a work. By obtaining a local optimum value as a minimum error calculated for between the image feature point and a predicted estimation image feature point against a plurality of search parameters of position and posture vector of the camera 4 with a calculation processing of each part of a quantized space searching part 10, a stepwise independent search is executed for each search parameter. By repeating this with multiple scale by turns for specified times, data showing the optimum position and attitude of the camera 4 are calculated and recognized. Thus, position and posture can be recognized speedily, accurately and robustly.;COPYRIGHT: (C)1995,JPO