Research on image processing technology for online oil monitoring system

Accurate wear debris classifying using on-line oil monitoring system plays an important role in aero-engine condition monitoring and fault diagnostics. However, there still exist imperfections in on-line oil monitoring system such as: motion blur of abrasive particles in the acquired images, low efficiency of the traditional image segmentation and recognition methods. Aiming to improve this situation, improved image restoration and segmentation algorithms are developed and different kinds of classifiers are designed in our work. Practical experimental results indicate that better recognition results of the abrasive particle images can be achieved via the improved methods, which offer an analysis basis for the fault diagnosis and detection of aircraft engines.