Visual Realization Of Aircraft Maintenance Process And Its Key Technologies

Traditional aircraft maintenance training mainly relies on paper manuals and manual teaching,which is inefficient and of low quality,cannot meet the current trend of urgently introducing aircraft maintenance personnel.In the maintenance site,there is no real-time information support and guidance,and the difficult problems on the site can not be solved in time,resulting in low maintenance efficiency,but also easy to induce technical errors caused by human factors,such as the maintenance personnel carelessly missed a certain step of the maintenance process,resulting in many adverse consequences.Based on the current situation of civil aviation maintenance industry,an effective solution with the help of visualization technology is proposed.Aircraft maintenance visualization is proposed,which is a combination of Virtual Reality(VR)and Augmented Reality(AR)with aircraft maintenance process.It is mainly to digitize paper manual materials into text information,pictures,videos,3D models and other multimedia means,combined with human-computer interaction development,so as to provide aircraft maintenance personnel with intuitive,efficient maintenance training services and real-time reliable maintenance guidance services.Based on Unity3 D engine and with the help of HTC VIVE Pro and Microsoft Holo Lens,the Virtual Reality aircraft pre-flight inspection training system and the Augmented Reality induced aircraft maintenance system are developed respectively.Visual induced aircraft maintenance system pursues high-speed and high-precision process guidance.Image recognition is the key technology of visual inducement,the speed and accuracy of recognition determine the performance of the system.Corner detection in image recognition will be researched in this thesis.The traditional corner detection algorithm has the problems of slow speed and low accuracy.In view of the shortcomings of the traditional corner detection algorithm,this thesis takes advantage of the combination of Wiener filter and 8-neighborhood comparison method to filter the corner points for the first time,and finds the optimal corner quality level and the minimum distance limit of the adjacent corner points to filter the corner points for the second time in order to get the best corner set,the corner detection algorithm is optimized.The developed Virtual Reality pre-flight inspection training system and Augmented Reality induced aircraft maintenance system have met the functional requirements of the system and have certain applicability.In terms of the optimization of the key technology of visual induced maintenance,the simulation results show that the optimization algorithm has a significant improvement in speed and accuracy compared with the traditional algorithm,and the optimization results are conducive to the real-time matching between the sample image and the maintenance site.Furthermore,the optimization results can meet the performance requirements of visual induced maintenance system.