Motion compensation for three-dimensional measurements of macroscopic objects using fringe projection

Fringe projection is an established method to accurately measure the 3d structure of macroscopic objects. To receive a high accuracy and robustness at once there is need of several images with different projection patterns. Over the sequence of these images it is necessary that each 3d object point corresponds to the same image point at every time. This situation is no longer given for measurement objects moving under the process of projection and image acquisition. To solve this problem we propose a method to realign the captured images. Because of different view angles over the measurement process this have to be done in 3d space. Primarily our method is based on a kind of back projection of realigned sparse 3d point clouds in the image planes as well as the projection plane. Therefore you need a well calibrated measurement system and images without or only small motion blur. The 3d point clouds are results of a single shot analysis like fourier transformation and realigned by ICP. Back projection in the projection plane is necessary because of an unknown phase step between two realigned images each. Overall our method results in a motion compensated dense 3d point cloud.