Distill Particle Feature Basing On Dividing Into Fractal And The Nerve Network Technology

Ferrography technology is a important mean for detecting Large-scale diesel engine trouble. At present wear particle identification work by artificial to finish through the microscope mainly. Combining traditional image processing method and neural network and divide into fractal, etc. modern computing technology, auto distill particle feature, analyze wear particle image. Thus differentiate the nature, form of the wear particle, size, quantity, etc., offer the strong ferrography data for the thing that the trouble of the diesel engine is diagnosed.The paper lucubrates to using fractal geometry and the neural network technology research in wear particle recognition processing some difficult questions. First introduced the image division and the characteristic withdraws in the tradition imagery processing method; Then in withdraws the wear particle outline image and in the gradation image foundation, using fractal geometry compute wear particle image edge outline and superficial gradation fractal dimension, the final wear particle image recognition has provided the shape characteristic; Finally using the neural network method to extract the wear particle image waveform characteristic and realize automatically distinguishing the wear particle image.