| With the rapid development of aerospace manufacturing,product assembly quality visualization technology has been widely used in aerospace industry,especially in highcomplexity,high-precision,high-quality product assembly.However,there are several problems which cannot be realized in the existing technologies,such as full point cloud,high precision,fast assembly quality evaluation.Therefore,we proposed a pre-evaluation technology for product assembly quality visualization to solve the problems using the combination of fast high-precision registration and multi-feature automatic constraint of point cloud.The structure of this research is as follows:Firstly,the trinocular vision measurement system and principle were studied while the measured point cloud data was obtained.The point cloud data of the connecting part of the wing and the fuselage was selected as the research object to obtain the point cloud data.Fast and high-precision registration and coordinate unification of the theoretical data and the data to be assembled were needed to achieve product assembly quality assessment.Thus,a new Iterative Closest Point(ICP)algorithm based on the combination of the similarity feature matching method and the weighted Euclidean distance threshold method was proposed.According to the similarity feature matching method,the initial position of the massive point cloud and the theoretical data was quickly optimized.Based on the point cloud data mentioned above,a new ICP algorithm based on the weighted Euclidean distance threshold method was studied.Fast and high-precision registration and coordinate system of the basic point cloud data were realized using the new ICP.Comparing with the classical ICP,the registration accuracy of the new ICP was improved by 37% while the registration time was halved.Secondly,a multi-feature constrained assembly quality assessment method based on "point cloud recognition-region extraction-feature matching-constraint" rules was proposed.The method was used to study the identification of various feature geometric elements based on Gaussian mapping principle.Feature fitting and model extraction of few point cloud data in the target area was done according to the characteristics of the identified features.The featured pairs were matched based on the angle and distance relationship between the extracted features.In order to accomplish fast assembly of point cloud data,the multiconstraint feature direct assembly method was applied under initial feature constraints.Finally,in order to evaluate the assembly quality,the quality evaluation coefficient was set to get the coaxial angle error of the upper and lower hole shaft assembly and the interference parameter distribution of the contact surface.A visual model was established,while the quality evaluation parameters were indexed with the three-channel color model of the computer system.The interference results and interference types were analyzed.The preevaluation of the assembly quality of the product based on the secondary development of CATIA was realized.The visual pre-assessment platform was established to evaluate the quality of docking assembly of large parts.Through the research on the visual pre-evaluation technology of product assembly quality for the fast registration of the fusion point cloud and the multi-feature constrained virtual assembly,the technology can be applied to assess the quality of parts such as product(the fuselage,the wing,the tail,the large aircraft,etc.)connections and assembly.In the existing technology,the assembly quality results cannot be obtained automatically in real time,which resulted in problems such as long production cycle,unpredictable assembly quality results,poor assembly accuracy,multiple rework and increased wear of the assembly.The technology developed in this research can be used to detect,reduce or eliminate the influences of parts errors and assembly errors on product assembly.Eventually,the product assembly cycle was shortened while the assembly efficiency was improved. |
PDF Full Text Request