Research On Key Technologies Of Fringe Projection 3D Measurement

Phase measurement profilometry,as a popular structured light 3D reconstruction technology,has been widely applied in many fields including military,agricultural,entertainment,medicine and scientific research etc.Obviously,its applications will be more extensive in the future.At present,many scholars focus on the research of fast phase unwrapping algorithm to adapt to high-speed dynamic measurement,and begin to explore 3D reconstruction for more scenes(defocus,panorama,etc.).However,there are few studies on the system calibration and phase unwrapping errors,which determine the system accuracy.Moreover,it indeed faces many problems and challenges in various applications of high precision and high speed.Therefore,we systematic study several key technologies of the system based on the fringe peojection,including camera calibration,projector calibration,phase unwrapping algorithm and 3D reconstruction under different conditions,and ensure high accuracy of the phase measurement profilometry.The main contents are summarized as follows:1.Camera calibration:The existing camera calibration methods are difficult to calibrate the camera in the defocused scene.Based on the characteristics of phase-shifting algorithm,this paper proposes three calibration methods for defocused camera.All these three proposed methods simplified the calibration process and obtained the accurate calibration results with blurred images.Meanwhile,color phase-shift wedge grating arraies provided a solution to calibrate a color defocus camera.Simulations and experiments showed the effectiveness of these methods.2.Projector Calibration:In the popular projector calibration method,the projector is used as a pseudo camera.The coordinates of "feature point" of projector image are solved by the constraint relationship between feature point coordinates of the camera image and the phase distribution of the projector image.Howover,the phase error caused by the non-linearity of the projector and camera has not been considered in the calibration.We used the Hilbert transform to compensate the nonlinear error during the projector calibration.The calibration accuracy has therefore been improved.3.Phase unwrapping algorithm:The phase error caused by environment random noise or equipment defocusing in time phase unwrapping algorithms can result in phase unwrapping failure or localized discontinuation.The gray-scale coding method based on half period correction and the phase coding method optimized by K-means algorithm are proposed in this paper.The two methods can effectively remove the phase error in the phase unwrapping.Based on the coding sequence,an optimized shift-phase method is proposed.Only four images are needed in the measurement process,which is suitable for high-speed dynamic 3D reconstruction and increase the fringe frequency.4.3D reconstruction under different conditions:Regarding the fringe projection system as a binocular vision model,the relationship between the absolute phase and 3D coordinates is deduced.The image coordiantes can be transformed into 3D space.Then 3D point cloud reconstruction of the target object is realized.Experiments of various objects in different scenes are carried out,such as static objects and dynamic objects measurement with a focus projector,and static objects measurement with a defocused projector.Experimental results validate the effectiveness of proposed method and also show that objects can be effectively recovered by the constructed SL measurement system.