Research On Pose Estimation Technology Of Scattered Electronic Components Based On Point Cloud Registration

With the continuous improvement of production automation and intelligence,intelligent manufacturing technology has been widely used in the field of industrial production,which has the advantages of non-contact,high precision,high efficiency and real-time.At present,in industrial production,the insertion of electronic components is mainly based on manual insertion methods.This method has low insertion efficiency and high manual labor intensity,which greatly reduces the intelligent level of electronic component insertion.This project applies the point cloud registration technology to the pose estimation of scattered electronic components,and conducts in-depth research on the pose estimation method of scattered electronic components based on point cloud registration.The main work and conclusions of the paper are as follows:Firstly,the creation of the binocular vision model is discussed,mainly in four aspects: the establishment of the coordinate system,camera distortion correction,binocular vision stereo calibration and stereo correction;combined with the actual engineering environment,gray code combined with phase shifting is selected.Method for image encoding and decoding,and constructing a structured light encoding point cloud reconstruction system;using the calibration board image collected in the experimental environment,the binocular vision calibration and correction in the MATLAB software,the use of the projector to project the encoded image to complete the disorder Point cloud reconstruction of electronic components.Secondly,use background point removal,filter denoising and point cloud simplification related technologies to remove the background points,noise points and redundant points of the point cloud,highlighting the target information of the scattered electronic component point cloud;in order to complete the pose of each electronic component Information estimation,based on the detailed information and overall distribution characteristics of the scattered electronic components,an improved adaptive Euclidean clustering segmentation algorithm is proposed,which uses the normal angle and smoothness to distinguish edge points from non-edge points.The adaptive distance threshold is used to segment edge points and non-edge points to achieve effective segmentation of the point cloud of scattered electronic components.Experiments have verified that the proposed adaptive Euclidean clustering segmentation algorithm effectively improves the segmentation accuracy of scattered electronic components point cloud.Then,the ISS method is used to obtain the key points of the scattered electronic component point cloud,and improve the compactness of the scattered electronic component point cloud information expression;in order to improve the accuracy of the point cloud pose estimation,an improved EB?SHOT feature descriptor is proposed,which combines the SHOT feature descriptor with the spatial position information,and at the same time Use binary coding rules to encode the descriptors to realize the feature extraction of EB?SHOT feature descriptor;Then point cloud registration was carried out based on EB?SHOT feature descriptor.SAC-IA algorithm and ICP algorithm were used to carry out the two-step registration method of initial point cloud registration and precise point cloud registration to calculate the precise pose information of the electronic components point cloud.Experiments show that the proposed EB?SHOT feature descriptor can effectively reduce the error of pose estimation and improve the accuracy of pose estimation.Finally,completed the design and implementation of the pose estimation system for scattered electronic components,including the selection of the main hardware,the construction of the pose estimation platform,the design of software function modules and the system implementation;the experiment proves that the system has a great improvement in accuracy,error and speed,and meets the requirements of engineering.In summary,this subject has studied the position and pose estimation technology of scattered electronic components based on point cloud registration,and conducts an in-depth discussion on the point cloud segmentation method and point cloud registration method of scattered electronic components.A pose estimation algorithm with high speed and accuracy is proposed,and a pose estimation system of scattered electronic components is established,which realizes the intelligent pose estimation of scattered electronic components and improves the intelligent level of electronic component insertion.