Design And Implement Of Airport Enhanced Ground Inspection System Based On Computer Vision

During airport’s Line Maintenance and Special Maintenance,ground crew will perform an overall inspection for airplanes to eliminate hidden security risks.Among the inspection,it is a crucial link to inspect injuries on the skin of airplanes.Visual inspection is a common method during skin’s check.In many cases,time reserved for ground crew’s inspection is short,and to some degree,inspection results are influenced by ground crew’s experience or other human factors,which will lead an unstable inspection result.Low inspection efficiency may result in flight delay,while inspection errors can cause serious accidents.As can be seen that,automated inspection of airplane skin injuries is important for both passengers and airlines.This paper researches the common methods used for airplane skin injury inspection during plane’s maintenance at home and abroad,and points out the problems of current inspection systems.Meanwhile,this paper mainly discusses the inspection method based on computer vision,and carries on the thorough research of the detection and recognition algorithm.This paper analyzes the demand of airplane skin injury inspection system,and completes the overall design and implementation work.Firstly,for the purpose of ground crew’s high inspection efficiency,Enhanced Ground Inspection system is divided into four parts:center management subsystem,inspection subsystem for ground crew,inspection subsystem for non-ground crew and inspection subsystem.This paper defines a kind of simple communication protocol between subsystems and completes image collection,image transmission and other information processing modules.Secondly,non-destructive testing(NDT)technology based on computer vision is applied for inspection system,where NDT is used for positioning and identifying the injuries on airplane skin.According to the features of airplane skin injuries,this paper proposes to utilize SUSAN edge detector to position the injuries.After positioning,the machine learning method is utilized for injury identification.Finally,this paper completes the design and implementation of the Enhanced Ground Inspection system based on actual needs of project,and uses the actual data for verification.