High Precision Underwater Three-dimensional Reconstruction Method Based On Binocular Stereo

With the development of human society,the consumption of oil,natural gas and other resources is increasing day by day,and the exploition of these resources in land has been basically completed.In the ocean,which account for 2/3 of the earth's surface area,a large number of resources have not yet been exploited.Using visual information to carry out the three-dimensional reconstruction of the surrounding environment has important application value on navigation and exploration activities of underwater robot,and it's also an important research direction in the field of computer vision.However,due to the complexity and particularity of the underwater environment,the quality of underwater photo is decreased,and the image will be distorted at the same time.This makes the traditional camera calibration and three-dimensional reconstruction algorithm can't be directly applied to underwater.In view of the subject content and the request of the algorithm,the three-dimensional reconstruction algorithm in air is researched deeply.The disparity map is obtained by multi resolution stereo matching and multiple constraints.Aiming at solving the problem of steps shadowed error in current method,the dissertation puts forward a new method to optimize the disparity map.The method uses anisotropic filter so that make the accuracy and robustness of the algorithm improved and the running time reduced.In view of the difference between the imaging model underwater and in air,calibrated the underwater camera system is calibrated in this dissertation.In order to study the underwater imaging process,the ideal underwater imaging process is modeled.At the same time an improved underwater camera calibration method with plane calibration plate is proposed.Refraction parameters are added to the calibration model and then underwater camera calibration is completed.By introducing the pixel offset error parameter,we design algorithm to compensate it and experiments are carried out to verify the compensation effect.After the compensation of the underwater images,they can be restored to the images in air.Then the binocular reconstruction algorithm in air can be used to get the three-dimensional model of underwater objects.In this dissertation,a set of experiments are designed to model three underwater objects.Comparing with the reconstruction results in air,the proposed algorithm has the advantages of high precision,high efficiency and good integrity so that can be applied to the actual underwater research work.