Research On The Application Of Endoscopic Image Feature Extraction Technology In Aeroengine Diagnosis

At present, endoscopy is one of the important non-destructive testing (NDT) methods. It plays an important role in fault diagnosis and maintenance of aeroengine. In its application, the methods of endoscopic data processing and fault diagnosis need to be improved urgently.So applying computer technology and digital image processing technology to endoscopy has become a significant direction in endoscopy development.The application of endoscopic image feature extraction technology in aeroengine fault diagnosis is researched in this paper.A fast fault diagnosis system is designed. The main contents can be summarized as follows:1. On the basis of the comprehensive and deep investigation, the constitution and working principles of the aeroengine are introduced. The main fault modes, mechanism, expressive features, effect, decision means and maintenance experience are concluded and analyzed.2. The application research of image pre-processing in endoscopy feature extraction is carried out. The methods of image enhancement, image segmentation and contours extraction are presented and tested. Aiming at the traits of the above-mentioned methods, some problems that should be paid attention to are put forward.3. Semiautomatic measurement and automatic measurement are brought forward, which is based on the analysis of traditional measurement. The automatic measurement based on 3-order spline interpolation and the automatic measurement based on feature point extraction are mainly expounded. The experimental results show that the methods can figure out the defection size accurately and quickly.4. The fast fault diagnosis system of aeroengine is designed. The actual application results show that this system can detect many kinds of faults and damage states effectively and accurately. The validity and feasibility of the endoscopic image feature extraction technology are validated.