|
Videometrics can measure the displacement, distortion of the objects, and get the movement parameters
by several views. As an optical measurement method, there are several advantages such as the high accurate
results, no need contact with the objects, wide measurement range and so on. In general, the definition criteria of
videometrics is evaluated by the following two ways: the first one is to get the image reprojection error which can
be calculated by comparing the image coordinates of the cooperation targets with the reprojection image
coordinates of the calculated 3D positions of the cooperation targets after camera calibration, the second one is to
get the 3D reprojection error which calculate the space distances of the calculated 3D positions of the cooperation
targets and the projection lines which pass the optical center and the image point of the corresponding targets after
camera calibration. The two methods actually obtain some kinds of calibration errors, the real measurement errors
are not the same. A measurement error evalution method is proposed by using the classical adjustment method.
Firstly, the calibration process of the videometric is deduced from the differential form of the collinear function
which regards the 3D positions of the cooperation targets and its coordinates in the image plane as the known
values and supposes that the mean-squared deviations of them are known. Secondly, according to the classical
adjustment method, the normal function and its weight matrix are known; the parameters and the mean-squared errors of the camera such as focus length, principle point, and etc. can be calculated. Thirdly, According to the values from the second step, and transform the differential function in the first step, the errors of the measurement can be calculated. The evaluation of the videometrics measurement errors is deduced from the three steps above and can be calculated when the cameras are ready to measure the objects and the cooperative targets are known. The simulation and experiment shows the proposed method is reasonable and efficient.
|
