Reverse Three-Dimensional Reconstruction Of Mechanical Parts Based On Kinect

Non-standard parts are an important part of many industrial products,and its quality is also an important factor affecting the quality of related industrial products.Therefore,accurate and rapid completion of the production and testing of non-standard parts is the key to industrial intelligent manufacturing.However,most of the non-standard components on the existing equipment are three-dimensional geometric structures.According to the existing three-dimensional non-standard components,the product model is quickly designed and generated to solve the problem because the manual design and imitation are inefficient and the data model is easy.Loss and other issues are important goals for implementing the relevant strategies of “Industry 4.0”,“Industrial Automation” and “Made in China 2025”.Reverse engineering can map the entire shape of the product,reconstruct a three-dimensional geometric model of the surface of the product,and generate the digital files needed to make the product.With the introduction of new 3D scanning devices such as Microsoft Kinect and ASUS XTion,research on 3D reconstruction technology based on these devices has attracted more researchers' attention.In this paper,the Kinect equipment is used to obtain the depth image of non-standard parts by the advantages of insufficient illumination and color interference.An image enhancement algorithm is studied to solve the instability and existing problems of the original depth data,and the data accuracy is improved.Through in-depth analysis of the camera calibration principle,the camera parameters obtained from the calibration are used to calculate the surface coordinate information of the part under the current camera pose.Through the research on point cloud filtering,point cloud segmentation,point cloud registration,and point cloud surface reconstruction,the 3D reconstruction scheme for nonstandard components based on Kinect is found.Finally,edge fitting and feature extraction are performed on the surface information of the generated object to obtain more accurate surface feature information of the non-standard parts.The experimental results show that the 3D reconstruction scheme studied in this paper can obtain the 3D model of non-standard components well,and overcome the technical bottleneck of plane detection.It can quickly acquire the feature point information of parts and realize the detection of component feature defects.The whole reconstruction system is simple and has good robustness,and has broad application prospects in industrial intelligent manufacturing.