Research On The Technology Of 3D Model Segmentation For Laser Processing

Laser processing technology with its high precision and high adaptability is more and more used in high-end manufacturing field.China is the only country with all kinds of industries in the world.However,restricted by various factors,the key technologies and equipment of large-scale laser processing system in China are heavily dependent on imports,and some key technologies are blocked by foreign countries.As the brain of the laser processing system,the matching software technology will directly affect the processing accuracy.How to divide the 3D model of the object to be processed into multiple components to meet the processing conditions of the system,so as to improve the processing accuracy and efficiency,is an important problem to be solved in the research and development of the matching software of the laser processing system.Due to the shortcomings of the existing automatic segmentation technology,such as poor segmentation effect and poor universality,it is imperative to develop a 3D model segmentation technology in large laser processing scene.Based on the project of”research and development of laser precision manufacturing technology and equipment for aerospace typical complex patterns”in the national key R&D plan of the laboratory,this thesis studies the 3D model segmentation technology in laser processing scene.In this thesis,the 3D model segmentation technology is studied.Aiming at the shortcomings of the existing automatic segmentation algorithm,An interactive 3D model segmentation technology for laser processing is designed and implemented.Compared with the automatic segmentation algorithm,this segmentation technology adds manual selection operation in the segmentation process,which greatly improves the segmentation accuracy and universality.The main work of this thesis is as follows:1.This thesis summarizes the existing 3D model segmentation technology,and proposes a semi-automatic segmentation technology based on model simplification.Through the segmentation of simplified models with different magnification,the better component combination is selected and restored,and good segmentation results are obtained.2.This thesis proposes to use bounding box technology for manual intervention,and use bounding box to manually select the best part of segmentation quality in different simplification ratio model segmentation results,and then combine it on the original model.3.According to the characteristics of industrial model,a 3D model simplification algorithm for industrial processing scene is proposed.In the process of simplification,the single-layer boundary of 3D model is identified effectively to avoid boundary simplification.4.The segmentation technology in this thesis is compared with seven typical automatic segmentation algorithms,such as random cut,core extraction,shapediameter,normcuts,randwalks,fitprim,k-means and so on.The quantitative evaluation of the segmentation technology in this thesis is completed through the comparison of patch error rate and area error rate.The simplification efficiency and effect of the proposed algorithm are tested to evaluate whether it meets the segmentation requirements.Experiments show that the 3D model segmentation technology designed in this thesis can get good segmentation results,and the simplified algorithm can adapt to the characteristics of industrial model.At the same time,this segmentation technology not only has good segmentation effect for industrial models,but also can produce better segmentation results than the above seven segmentation algorithms for natural models,which shows its good universality.