Research On Feature Enhancement Algorithm Of Assembly Surface Point Cloud Data

Virtual assembly is one of the key technologies in current virtual manufacturing,which is achieved through accurate measurement of assembly surfaces and computer-aided realization of virtual assembly,and point cloud-based assembly surface measurement and 3D reconstruction technology is an effective means.As important features such as threaded holes and step surfaces on the assembly surface are key details that represent the 3D model and are also key features that reflect the actual assembly surface,the study of point cloud feature enhancement methods for the assembly surface has some practical value.In order to protect the sharp features of the assembly surface,this paper enhances the point cloud features of the assembly surface in terms of denoising and streamlining,and achieves high quality reconstruction results and improves the assembly efficiency while maintaining the detail of the assembly surface.The main work of this paper is as follows:In response to the problems that traditional denoising algorithms cannot completely eliminate noise points,excessive denoising and too smooth denoising effect when dealing with models containing noisy point clouds,this paper studies a denoising algorithm for feature enhancement of assembly surface point clouds,which divides the noise into two types of noise according to the features of assembly surface point cloud data.The first type of noise is filtered out from the assembly surface point cloud data using the K-field search method,and the first type of noise is removed using the sampling point neighbourhood point count method;the second type of noise is denoised using the improved bilateral filtering method.The point cloud is then resampled to eliminate the noise from defects in the point cloud data while effectively maintaining the sharp features.In response to the problems caused by the loss of feature data,holes in the point cloud model,huge time consumption and excessive overhead of computer resources when processing the point cloud model by traditional streamlining algorithms,this paper researches a streamlining algorithm for feature enhancement of the assembly surface point cloud,using the point cloud streamlining theory mainly by using the information entropy hybrid sampling method,for the extraction and retention of the point cloud feature data points using local information entropy to achieve This paper then uses the K-Means clustering method to sub-classify the assembly surface point cloud data and determine the characteristic and non-characteristic regions according to the curvature of different point cloud data,so as to avoid the phenomenon of holes in the point cloud data caused by the extraction of characteristic points.Finally,a comparative analysis with different classical algorithms shows that the denoising algorithm in this paper can achieve the desired denoising effect while increasing the retention of assembly surface features,and can also effectively avoid the phenomenon of excessive smoothness produced by traditional denoising algorithms after denoising.