Three-dimensional Measurement Of Colored Objects Based On Phase-shifting Amount

The three-dimensional measurement technology obtains the threedimensional(3D)space coordinates of the object and reconstructs the 3D shape of the object through the imaging sensing device.However,the existing researches on fringe structure light paid little attention to the influence of the surface color of the object.In practical applications,the decoding robustness on the colored object comes at the expense of increased additional patterns in the existing time-based absolute phase retrieval methods.More severely,multiple exposures are required in some extreme cases,which increases the time cost of the acquisition part.Therefore,it is urgent to introduce a new coding method that is insensitive to the surface reflectivity of the object and keeps the number of additional patterns same with solid-colored objects.This dissertation has carried out in-depth research on the above-mentioned problems,and proposed a new code element,several coding methods,and a new 360°measurement method.The details are given as follows:Firstly,this dissertation presents the theories of combining fringe structure light and stereo vision,the calibration method of the system and spatial geometry reconstruction in details.It provides a guarantee for the smooth development of the follow-up work.Secondly,the existing absolute phase retrieval methods are vulnerable to the color of objects.Thus,this dissertation targets to find a new carrier to code fringe orders and proposes the concept of phase-shifting amount sequence(PSAS).In PSAS,the fringe order is encoded in the phase-shifting amout sequence of additional patterns and directly derived from phase difference.The absolute phase and 3D point clouds can be obatained from that PSAS.The hole filling algorithm used in decoding can resist the influence of camera random noise on 3D measurement.Moreover,a statistic filter and a self-correction filter are designed to correct the errors.The time cost of correction algorithm is greatly reduced by accelerating them with GPU based on CUDA.Thirdly,a phase-shifting amount element(PSAE)is proposed to break the constraint that the existing image gray-scale and phase encoding methods in the colored objects measurement.For the first time,to the best of our knowledge,the phase related code element can be retrieved from one additional pattern.PSAE effectively reduces the amount of image data required to obtain symbol information while ensuring the robustness to the surface color of the object and the nonlinearity of the system,and further improves the universality of the coding and decoding method based on the phase-shifting amount.Fourthly,a kind of new N-ary Gray-code methods is proposed according to the PSAE which breaks the constraint that the traditional Gray-code methods cannot balance the amount of code elements and the application range.In 5-ary Gray code,125 fringe orders can be encoded in only 3 frames of additional patterns.It efficiently completes the measurement of the black plush toy with the fringe modulation degree close to 1.On the other hand,3-PSA-codes has the ability to obtain the fringe order from one additional pattern,which greatly reduces the time cost of the acquisition part in the colored object measurement.Furthermore,a special decoding method based on a maximum probability method is proposed to increase the decoding robustness.Finally,a method for the 360° 3D measurement of colored object is proposed.Its effectiveness has been verified by a system composed of a commercial projector,two mono cameras and an ordinary rotation stage.The initial transformation parameters are calculated from the proposed feature and phase constraint points and then refined with nonlinear optimization with a closed loop.The color of the object is obtained with a mono camera by projecting different channels light with a projector.