Study On High-precision Structured Light 3D Measurement

In this paper,aiming to improve the measurement accuracy,three-dimensional structured light measurement technique is studied based on sinusoidal grating projection.As an important research direction in the field of computer vision,three-dimensional measurement technology performs three-dimensional reconstruction to realize the recognition and detection of the target through computer using the three-dimensional structure surface measurement of the target object obtained through some tools.Nowadays,with the requirements of product quality and manufacturing precision improving constantly,the status and function of 3D measurement technology are also increasing.At home and abroad,huge amount of resources have been devoted to the research concerning this topic,and various new methods,theories,equipment,and algorithms emerge in an endless stream.Therefore,it is of great practical value to study the high precision measurement method of 3D objects with complex structure under the premise of low cost and digitization.We study the high-precision structured three-dimensional visual measurement technique based on sinusoidal raster coding in time coding.In actual measurement process,since the projection sine grating and the capture of the modulated image are distorted or disturbed for various reasons,the captured object-modulated sinusoidal raster image contains noise and interference other than the contour information containing the object,which reduces the measurement accuracy of the system.Therefore,this paper focuses on the sine grating gamma distortion correction algorithm of 3D visual measurement system based on structured light to suppresses the noise interference of captured images and realize high-precision object contour measurement.The main contents of this paper are given as follows:Firstly,the gamma nonlinear distortion is studied.Because of the design reasons,the camera and the projector itself will be with a gray-scale non-linear distortion of the characteristics.It will make the raster stripes does not meet the good sinusoidal distribution and cause phase error,which is how the gamma distortion come from.Although the Gamma distortion is due to the noise of the measurement system equipment,the direct replacement of some of the equipment is relatively expensive,so the way to solve the problem of gamma distortion is software method.Secondly,the gamma error compensation algorithm based on Hilbert transform is studied.We introduce the Hilbert transform(HT),which can produce the amount of phase shift of / 2 while maintaining the amplitude of the sine function,and also filter out the DC component.To be exactly,after extracting the inclusion of the gamma error wrap,then we establish the phase error model based on the spatial domain and the HT domain.Based on the fact that phase error of the two domains is the characteristic of the same amplitude and the opposite direction,average the phase in this field can compensate for the phase error and get the exact phase value.Following that,the gamma pre-coding method based on genetic algorithm is studied.We first project two groups of grey image processing,and then use the camera to capture its distortion caused by gamma distortion after the deformation of the grating.After that,we use the genetic algorithm to calculate the value,and then pre-encoded raster stripes projected onto the surface of the target object.Then we can directly measure use phase shift method to obtain the ideal phase without needing of phase compensationThen,the phase compensation method based on Hilbert transform and the pre-coding based on genetic algorithm are combined with four-step phase shift method and three-step phase shift method to correct the gamma phase error,respectively.The corrected phase Errors are compared with the uncorrected phase errors to prove that effectiveness of the proposed two methods in improving the accuracy of the measurement system.Finally,the three-dimensional reconstruction of the target object is carried out,and the phase of the package is expanded into a continuous absolute phase.Through the phase-depth correspondence,with the help of obtained the absolute phase,the depth information of the surface of the target object surface can be obtained.We realize the object of three-dimensional reconstruction using the eigenvalue matching method through point cloud technology.