Aircraft Skin Defects Machine Vision Monitoring Technology Research

With the development of economy and technology, machine vision technology has deep into the all aspects of society, instead of human eyes completely changed people’s living environment. Machine vision detection and automation technology, a combination of machine vision is widely used in manufacturing the product defect detection of the industry, such as product assembly process detection and positioning product packaging, product appearance quality detection of the goods, the logistics or fruit sorting, machine vision can replace manual job done quickly and accurately. Beacause,it is not easy to detect the old aircraft skin defects, this paper analyses the advantages and disadvantages of existing detection technology, using machine vision technology, DSP technology, recognition technology, on the robot platform completed structure of aging aircraft skin defect detection system, this system can collect realime, dynamic skin damage information to health monitoring patform by the wireless transmission technology to, and the ground can be obtained online defect detection results, realized the aircraft skin defect image classification and rivet connection key parts corrosion image classification.Aging aircraft skin defect monitoring system mainly includ image acquisition.wireless communication modules, image storage modules, image processing modules, feature extraction modules, pattern recognition modules. According to the requirement of the system, complete the hardware design of image acquisition module, wireless communication module, image storage and complete the software design of the image processing module, the feature extraction module, the module pattern recognition. For aircraft skin defect feature extraction, the establishment of the aircraft skin image sample libraries, adopted gray characteristic value of matrix method to extract the aircraft skin defect images, the value of precision of characteristic meet the system requirements; For the rivet connection corrosion image feature extraction, presents an improved corrosion positioning algorithm to determine the center of the rivet center and radius of the improved algorithm improved the rivet center to determine the accuracy, and improve the precision of the rivet connection parts corrosion image characteristic value. In pattern recognition module, this paper expounds classification principle of the general linear support vector machine (SVM), general nonlinear support vector machine (SVM) and fuzzy support vector machine (FSVM), based on the sample center distance FSVM method,this system chose the fuzzy support vector machines for pattern recognition, completed aircraft skin image and rivet connection parts corrsion grade classification by the FSVM is given based on the sample spacing method, through the compared simulation experiments,find the algorithm has a recognition classification efficiency improvence to the aircraft skin defect images and rivet corrosion rate of skin image.Compared with current detection device,monitoring system based on machine vision of the aircraft skin defects can finish detection in front of the computer without professional operation personnel,it has good scalability,application prospect and great significance to improve the reliability of aging aircraft.