Research On 3D Object Segmentation Technology For Laser Etching

In the field of high-end manufacturing,high-precision laser processing and manufacturing technology is increasingly valued in China,however,due to the influence of various factors,there is still a certain gap between China and the world class level in the key equipment and key software technology of high-precision laser processing,therefore,in recent years,China has been continuously increasing its research support in this field in order to solve these many key problems.How to divide a 3D object into suitable parts to facilitate subsequent processing is one of the important problems to be solved in laser processing software.Under different application backgrounds,the requirements for segmentation results are often quite different,and traditional 3D object segmentation techniques are not completely applicable to the field of high-precision laser processing.Based on the national key R & D project undertaken by the laboratory,this paper studies the 3D object segmentation in the key software technology of laser etching processing.This paper mainly focuses on the 3D object segmentation method under the background of context of industrial processing such as laser etching,aiming at the 3D object model with a large amount of data,has noise interference or too detailed details,a segmentation method is studied and implemented,which can quickly obtain segmentation results in line with subsequent processing.The research work and contributions completed in this paper are as follows: 1.Based on the summary of a variety of existing 3D object segmentation algorithms,in view of the particularity of 3D object segmentation in laser etching processing scenes,an interactive multi-resolution hierarchical segmentation process is proposed,with the help of the 3D object simplification algorithm,the multi-resolution hierarchical representation of 3D objects is realized,and the enumeration effect is generated on the segmentation results,meanwhile,the error caused by automatic segmentation is further eliminated by combining the process of interactive selection and combination,and the segmentation quality is effectively improved.2.The 3D object simplification technology is researched.According to the specific requirements of 3D object segmentation processing in the laser processing scene,the simplified algorithm based on the quadric error metric is analyzed,and based on this,a new simplified method of quadric error measurement to improve efficiency is proposed,in the case that there is no obvious difference in the simplification quality,the simplification efficiency is significantly improved,and the multi-resolution hierarchical representation of 3D objects is realized with the help of this algorithm.3.Through the combination of 3D object simplification technology and 3D object segmentation process,the level of detail of the 3D object model with a large amount of data is effectively reduced,blurring noise and unnecessary details while preserving the main convexity features used for segmentation,not only has the segmentation speed of 3D objects been significantly improved,but the segmentation quality has also been improved to a certain extent.4.The related algorithms in this paper are implemented,and the simplified process and segmentation process are tested and evaluated respectively,which verifies the effectiveness of this algorithm in laser processing scenarios.Through experiments,it is proved that the algorithm proposed in the paper can effectively optimize the segmentation results of large data volume 3D objects in the field of laser processing,meet specific processing requirements,and achieved the desired quality and efficiency of segmentation.Meanwhile,the proposed method has a certain universality for all kinds of models.