Fast Trifocal Tensor Estimation Using Point Projecting Devices

This paper present a convenient method for estimating the trifocal tensor corresponding to three images acquired by 3 cameras and 10-15 point projecting devices. A set of virtual 3D points are made by the laser pointers which are projected onto the working volume by turns. Its images are corresponding points naturally, found with sub-pixel precision. Then the trifocal tensor is computed via rank-4 factorization from these corresponding points and the Euclidean stratification is done by imposing geometric constraints. Sample experimental results demonstrate that the method performs considerably faster compared to a state of the art method with more accuracy and without any other calibrating devices.