Studies On 3D Reconstruction Algorithm Based On Monocular Vision

Three dimensional reconstruction plays an important role in the field of robot vision?intelligent transportation and so on.It is an important field of computer vision research.Because monocular vision based image and image acquisition equipment is simple,easy to install and operate,data information fusion is simple,and the popularity rate is relatively high.The 3D reconstruction based on monocular vision has been widely used as a hot topic in the field of computer stereo vision.Aiming at the problem of single camera vision calculation in 3D reconstruction efficiency and robustness,and study the application of 3D reconstruction technology based on monocular vision in automobile auxiliary driving.Application of 3D reconstruction in vehicle driving control platform.The thesis proposed a new 3D reconstruction method based on SURF operator.It mainly includes calibration of coordinate position,image acquisition and preprocessing,extraction of feature points,matching of three-dimensional information and recovery of 3D information.In this method,the principle of edge filtering is first used to preprocess the noise before the image processing,so as to achieve the purpose of obtaining better image information;Through the calculation and transformation of the image coordinate system in the horizon,a three-dimensional space pixel coordinate system is set up;By extracting the pixel coordinate point data in the spatial pixel database,the 3D reconstruction of the original image is realized to a certain extent.From the angle of combining theory with practice,the algorithm in this thesis overcomes the original algorithm in image reconstruction speed and reconstruction effect is not realistic.In addition,the running time of the algorithm is relatively reduced,and it can improve the reconstruction effect,3D reconstruction view can get high quality,meet the requirement of three-dimensional reconstruction,with the actual the effect being good.The 3D reconstruction based on monocular vision realizes the reconstruction of 3D coordinates based on monocular vision,and improves the real-time and accuracy of the algorithm.On the basis of existing algorithms,this study optimizes the preprocessing algorithm of graphic image on the basis of improving the existing algorithms,so that it can be better adapt to the requirements of the simulation experiment in the thesis,optimize the process of the acquisition of camera parameters,and put forward the number of pixels in the storage and transformation of the pixel sitting mark of the graphic image.According to the idea of the library,as a temporary resource pool,it meets the requirement of the transformation between the two-dimensional plane coordinate and the three-dimensional image coordinates,which greatly improves the time efficiency of the algorithm and saves the space complexity.After the simulation test,it can meet the requirements of the experiment.The thesis is supported by nature science funding?Research on accurate ranging method based on vehicle monocular vision?(Grant No.61741207),and as a subtopic to research.