Abstract: The invention relates to the calibration of a three-dimensional computer vision system. The method and system are implemented using a polygonal plate-like calibration piece (21) made of a material whose properties do not change significantly with changes in conditions or with time. The edges of the calibration piece (21) used have different lengths and the piece is provided with circular focusing marks (22) placed at its vertices.

Abstract: The present invention concerns the discrimination of good measurement results measured by a computer vision system from bad results. By the method of the invention, values calculated from the observation vectors of several measuring devices can be discriminated into good and bad values by using a quality factor. The quality factor is calculated from the parameters of an error ellipsoid. The error ellipsoid is formed using the calculated value and the observation vectors. If the value calculated from the error ellipsoid parameters exceeds a threshold value that has been input, then the measurement result is rejected.

Abstract: The invention relates to the calibration of a three-dimensional computer vision system. The method and system are implemented using a polygonal plate-like calibration piece (21) made of a material whose properties do not change significantly with changes in conditions or with time. The edges of the calibration piece (21) used have different lengths and the piece is provided with circular focusing marks (22) placed at its vertices.

Abstract: The present invention concerns the discrimination of good measurement results measured by a computer vision system from bad results. By the method of the invention, values calculated from the observation vectors of several measuring devices can be discriminated into good and bad values by using a quality factor. The quality factor is calculated from the parameters of an error ellipsoid. The error ellipsoid is formed using the calculated value and the observation vectors. If the value calculated from the error ellipsoid parameters exceeds a threshold value that has been input, then the measurement result is rejected.

Abstract: In order to register the space, there are installed two video cameras at certain angles with respect to each other. In this object space there are arranged control points, the object coordinates (X.sub.k, Y.sub.k, Z.sub.k ; k=1, 2, 3 . . . ) whereof are measured and the imagewise image coordinates (x.sub.ki, y.sub.ki ; x.sub.kj, y.sub.kj) are determined from the corresponding video cameras, whereafter the control points can be removed from the object space. On the basis of the image coordinate values and the object space coordinate values of the control points, there are calculated the orientation parameters (a.sub.11 . . . a.sub.33) of projective resection, whereafter the unknown object space coordinates (X, Y, Z) of the observed object points can be solved in a real-time process on the basis of the image coordinates (x, y) registered by the video cameras and by making use of projective intersection.