Research On Technology Of 3D Surface Reconstruction Based On Binocular Vision

With the development of science and technology and the needs of modern industry,higher-precision measurement methods have become a new direction for the development of measurement science.In particular,the current manufacturing industry has strict requirements on specifications of products,and high-precision three-dimensional measurement technologies are increasingly important for monitoring quality of products.Due to the limitations of measurement speed,equipment cost and method,traditional 3D measuring instruments cannot meet the needs of modern industry.Nowadays,the technology of 3D measurement based on the machine vision has been gradually applied to the field of precision measurement.And the technology of three-dimensional measurement is a research direction combining machine vision and image processing,and it is also one of the most popular research topics in machine vision.Different from the traditional measuring equipment,the measuring system developed from the optical principle is essentially non-contact,which can avoid problems such as wear and deformation of the target's surface,and has the advantages of accurate measurement accuracy,flexibility,and high reliability.This paper studies and designs a binocular stereo vision measurement system combined with structured-light technology.It aims to solve the inherent problems of traditional vision systems such as the scarcity of feature points,difficulty in extracting feature points,difficulty in matching,and low measurement accuracy.The system uses Gray code encoding technology and cyclical fringe projection technology together,and builds a complete three-dimensional measurement system based on the binocular vision system.The main contents of this paper include the construction of camera model,the calibration of binocular stereo vision system,the study of structured light projection strategy,feature extraction,stereo matching and 3D reconstruction.In this paper,the measurement scheme of the system and the implementation process of the core algorithm are introduced and explained in detail,and an effective solution is proposed for some difficulties and key points in the algorithm.First of all,this article has carried on the thorough research to the camera model,and proposed a kind of simple and high-accuracy calibration method of binocular stereo vision system,which has laid a good foundation for the following three-dimensional reconstruction.Second,in the section of the coding image,in order to overcome many difficulties in the traditional visual measurement,this paper uses an effective strategy of structured light projection.What's more,in the section of the decoding image,this paper proposes an effective binarization algorithm based on the gray features of the encoded image.This method can obtain a more ideal results.And this paper also proposes a very effective algorithm to achieve correction of feature points' coded information.The algorithm can effectively complete the detection and correction of error coded information.Experiments results show that this correction algorithm is not only efficient but also effective.In addition,in the section of feature extraction,in order to improve the accuracy of boundary feature extraction,a high-precision algorithm of sub-pixel-level boundary detection is proposed in this paper.Finally,we made use of the system to perform multiple measurement experiments on different targets,so as to verify the practicability of the measurement system and the effectiveness of all algorithms.From the experimental results,we can find that the measurement system and the proposed algorithm are reliable and effective,it not only has a high measurement resolution,but also has a high measurement accuracy.