The Research And Implement Of 3D Point Cloud Processing For Blade Section Detection

The geometric precision and surface quality of blade blanks have great influence on product design,manufacture and assembly and assembly of the whole machine as the key core components of large power plant such as steam turbine,gas turbine and aeroengine.In order to guarantee the design,quality and cost of the products in the manufacturing system,the requirements of the detection technology are embodied in the fast,high precision,feature recognition and in-situ detection.How to carry out the precise processing of the test data has been an important research topic in the entire manufacturing cycle of the engine blades from rough forging to finished parts.With the rapid development of computer-aided technology,digital model reconstruction technology in the field of virtual measurement has been more and more in-depth application.In this paper,the rapid detection of the dimension parameters of the blade surface is established,and the technical path of the pretreatment and matching of the point cloud from the blade blank surface to the contour section detection is established.The surface data quality of die forging blank is improved.The function module based on the whole denoising technology is developed in the integrated development environment,and an effective denoising scheme for the blade surface is formed.The horizontal contrast experiment is designed,and the model matching degree is taken as the optimization target.The superiority of the solution and the function of the software in the whole denoising effect is demonstrated.The model matching of die forging blank is realized.the matching technology between the leaf model and the theoretical model is realized.The points of interest of the measurement module and the theoretical module are extracted,and the corresponding relationship between the measurement model and the theoretical model is estimated by using the feature descriptor;the initial matching is performed with the sampling consistency,and the subsequent matching is done by the nearest point iterative method;the matching visualization result is achieved.Aiming at the accurate segmentation of the point cloud in front and back of the blade,an algorithm for evaluating the front and back edge based on the inflection point detection is proposed.The accurate segmentation of the front and rear edge of the blade under the condition of high precision measurement is realized by rotating,intercepting,clockwise sorting and turning point detection.Finally,a comparative experiment is designed to verify that the proposed algorithm can segment the front and back edges of the blade with a low level on the line contour error.It is proved that in the point-by-point extension ellipse fitting process error.