Aircraft Skin Damage Detection Based On The Information Fusion Of Machine Vision And Ultrasonic Sensor

Aircraft is a key component of the safety of the aircraft flight and passenger. Due to the harshenvironment of the high altitude, long time and high capacity of the flight, the aircraft skin will appearcrack, corrosion, impact and other defects. These defects will cause damage to the structure integrityof aircraft skin, but the state of aircraft skin has directly influence on the safety of aircraft flight. Atpresent, the aircraft skin detection is mainly based on manual inspection and the experience and stateof the inspector decides the results. Therefore, the development of the aircraft skin automationdetection has good practical significance.The thesis was mainly based on the features of machine vision and ultrasonic inspection,as wellas the support vector machine with its approved algorithm was implied on intelligent fault diagnosisresearch which use the sensors put on the aircraft skin wall-climbing robot, and then transmit it to theground platform via wireless.The main contents are as follows:First of all, the development history and current research status of wall-climbing robot fordetecting damage of aircraft skin are reviewed and a new kind of wall-climbing robot is presented inthe paper. The causes and characteristics of three damage types which are normal, aircraft skin crack,and corrosion are analyzed. The aircraft skin crack has three different levels, that is normal, damagedand badly damaged.Second, The main research topics of multi-sensor information fusion such as informationrepresentation，sensor modeling and fusion level are studied. The classification of multi-sensor fusionalgorithms is also presented and the advantages and disadvantages of different algorithms arecompared.SVM is used as the fusion algorithm.Then, pre-process specifically at aircraft skin image is developed and characteristic parameterssuch as Gray level Concurrence Matrix of aircraft skin image are extracted. Meanwhile, the charactersof echo wave are extracted from the ultrasonic inspection. The support vector machine (SVM)algorithm is put forward to classifying the aircraft skin fatigues. The Genetic algorithm (GA) andParticle swarm optimization algorithm (PSO) are used to optimize the parameters and Fuzzy SVM(FSVM) is also used to classify the aircraft skin fatigues. Based on the experiment results, it can beconcluded that the FSVM method is a very effective method on the classification of aircraft skinfatigues.Lastly, the software to classify the aircraft skin fatigues with MATLAB and VC++is designed.The software fully plays the advantage of MATLAB and VC++which can classify the aircraft skin fatigues intuitively, quickly and accurately.This research is of great significance for improving the level of detection of aircraft skin damage,shortening maintenance time of aircraft, and ensuring the safety of flight.