Research On Key Technology Of High Precision 3D Reconstruction Based On Spatial Coded Structured Light

Coded structured light 3D reconstruction technology is a kind of active 3D reconstruction technology.Compared with the passive 3D reconstruction technology,this technique has the advantages of high matching precision,fast measurement speed and good robustness,which has been widely used in the field of industrial measurement,medical engineering and cultural relic repair.Currently,the research of coded structured light 3D reconstruction technology mainly focuses on two aspects:(1)improve the accuracy and speed of 3D reconstruction,(2)address the reconstruction problem of solid surface.In this paper,aiming at improving the accuracy of 3D reconstruction of spatial coded structured light,we mainly study structured light spatial coding algorithm,structured light decoding algorithm and system parameter calibration algorithm.The main work of this paper is as follows:(1)We study the structure light coding technology.Focus on the problem that spatial resolution of spatial encoded pattern is relatively low,we propose an adaptive spatial structured light coding algorithm based on surface geometric features of the object.Based on the feature of fringe density proposed in this paper,the sparse and dense fringes of structur light image are determined.Based on the feedback adjustment information of the geometric projection model of object surface,the fringe spacing of structured light encoding pattern is adaptively adjusted to obtain adaptive spatial structured light coding patterns with high spatial resolution rate,realize high accuracy three-dimensional reconstruction of specific objects.The experimental results verify the effectiveness of the proposed method.(2)We study the structure of structural light decoding technology.Aiming at the problem that the texture of the object surface and the influence of light on the structured light image lead to low decoding accuracy,we propose a color structured light decoding algorithm based on the technology of color migration.Firstly,using the color distribution of structured light coded images as prior knowledge,the structured light images captured by the camera are enhanced and de-illuminated by color migration,and the color information of structured light image stripes is restored.Secondly,a two-step accurate positioning of the color structure of light stripe center pixel sub-pixel extraction algorithm,which can accurately extract the center coordinates of the stripe,to achieve high-precision decoding.Compared to the experimental results,the proposed algorithm improves the matching accuracy by about 5%on average compared with the Retinex algorithm.By contrasting with the feature point detection algorithm,the error rate of the proposed algorithm is about 4%lower than that of Rongqiang Yang algorithm.The experimental results verify the effectiveness of the proposed method.(3)We study the structure light system calibration and structured light 3D reconstruction technology.Aiming at the problem of imprecise calibration of system parameters,the system calibration algorithm based on local homography matrices is proposed and a three-dimensional reconstruction system and calibration system based on spatial coding structure light are designed and implemented.Firstly,the structured light encoding pattern is projected on the calibration plate,and the global homography matrix between the projection pattern plane and the camera imaging plane is obtained by decoding.Then,the local homography matrix of each feature corner on the calibration plate is calculated,the local homography matrices are used to solve the exact pixel coordinates of the feature corners on the projection plane.Finally,the calibration system and the 3D reconstruction method based on structured light are implemented with high accuracy.