Research On Key Techniques Of 3D Reconstruction For Wind Turbine Blade Based On Images

3D reconstruction technology is an import method to obtain object models and 3D scenes,which is widely used in industrial detection,medical treatment,virtual reality and many other field.Traditional 3D reconstruction methods cannot obtain the texture information of target and are deeply influenced by environment.As a comparison,3D reconstruction method based on image feature points has the advantages of low-cost,and can obtain precise and realistic models with few restrictions on sence.In this paper,aiming at the problems of high cost,low efficiency and high labor intensity of traditional wind turbine blade defect detection methods,the 3D reconstruction technology is applied to the wind turbine blade defect detection system.By inputting the images of the blade from different angles,the real 3D model of the blade can be reconstructed,which makes the defect detection of the blade more efficient,safe and accurate.The main research contents are follows:Aiming at the problem of extracting too few feature points caused by the sparse texture of blade,an enhancement algorithm based on SIFT algorithm is designed.Firstly,the feature points can be obtained by the way of using the FAST corner detection algorithm in multi-scale space,Then the gray contrast information between the feature points and their multi-ring neighborhoods is calculated,and then the feature point descriptor is obtained by fusing it with the local gradient information of the sampling area.Finally,the matching results are purified and optimized according to the Euclidean minimum vector distance.Through simulation experiments,the effectiveness of the algorithm in this paper is verified.According to the analysis about the characteristics of 3D reconstruction of wind turbine blade,the 3D point cloud is reconstructed based on the combination of purified and optimized matching point pairs.Firstly,the 3D space point coordinates of the wind turbine blade model are calculated through the multi-view geometric constraints of matching point pairs,Then,the sparse point cloud model of wind turbine can be obtained by using Bundle Adjustment algorithm to optimize the coordinate,Finally,the sparse point cloud was clustered and classified to reduce the amount of redundant data,and the sparse point cloud was expanded and filtered to obtain the dense point cloud model.Aiming at the problem of low precision caused by high noise of dense point cloud model,an impoved algorithm based on Poisson surface reconstruction algorithm is designed.Based on the original Poisson surface reconstruction algorithm,a shielding factor is introduced,and the sample points are calculated using linear interpolation.The process of surface reconstruction is transformed to solve the shielding Poission equation.At the same time,the problem of overfitting the original algorithm to noise is solved effectively.Finally,the parameters are compared with the existing algorithm to verify the feasibility and superiority of this algorithm.