Phase error correction for a robust surface reflectivity-invariant 3-D scanning

Abstract

A phase error correction method is proposed to produce an absolute phase with robustness, specifically by utilizing phases generated at high frequency. The proposed algorithm integrates spatiotemporal phase unwrapping into a 3D scanning system by transmitting a unique phase signal through its projector and capturing the patterned surface of an object through cameras at two different perspectives. By projecting multiple fringe patterns on the object surface, the algorithm obtains and analyzes spatiotemporal information of an object. The algorithm looks for phases containing errors and rectifies them by redirecting the misled fringe orders to the correct values. As the algorithm corrects errors for both images obtained from the left and right cameras, it processes and reconstructs them into a 3-D model. The suggested method allows for 3-D scanning objects with varied surface qualities. In doing so, neither of the scanning setup nor the algorithm requires modification with respect to the object to be scanned. Also, the object does not need to undergo any surface treatment to put it in a favorable condition for scanning. With the error-correction algorithm, the noise in an image caused by a phase error can be effectively adjusted, thus yielding a robust 3-D scanning result.