Research On Microscope Image Identification And Surface Morphology Reconstruction Method Of Ferrographic Particles

Ferrography analysis is an important tool in oil analysis technology.By extracting information about wear particles generated during the running of the equipment,the technique is used to determine the wear status of the equipment.However,there are limitations such as the low efficiency and subjective error of manual ferrography analysis,and the ferrography image processing can only extract2 D characteristic information of wear particles.Therefore,based on image processing and 3D reconstruction techniques,this thesis explores a microscopic image recognition and surface morphology reconstruction method for identifying wear particles in ferrogram to improve the automaticity and accuracy of ferrography analysis and reduce the occurrence of missed and misjudged iron spectral analysis.On the basis of exploring the wear mechanism of the surface parts of machines and equipment and summarising the rules of the number and size distribution of wear particles at different wear stages in the machines life cycle,the relationship between the wear mechanism and the characteristics of wear particles is explored based on the typical morphological characteristics of wear particles.By constructing a mathematical model of the colour,size,shape and surface morphological characteristics of wear particles,and combining relevant technologies such as ferrography analysis,the wear mechanism of machines,computer image processing and 3D reconstruction,a method for microscopic image recognition of wear particles and reconstruction of surface morphology was proposed.By combining the rules of rotary ferrogram wear particle deposition,the dual-light microscope and CCD camera were selected to develop the abrasive particles microscope image acquisition system.Based on the dual-light microscope,the structure of the multi-view observation electric platform was designed to apply the abrasive particle microscope image recognition and surface topography reconstruction method.According to the control requirements of rotating ferrogram during microscopic observation,the matching hardware such as motion control card,stepping motor and driver are selected.By applying the basic motion control function of the control card,the control software is designed,the hardware of the ferrogram automatic analysis system is finally completed,realising the automatic control of rotating and translating the ferrogram and retracing the position of the abrasive grains.Based on the parameters required for the identification of abrasive particles,the algorithms of image enhancement,image segmentation,morphological processing and edge detection are designed to extract the edge contours of each abrasive particles in the field.The 2D characteristics parameters such as the colour,size and shape of the particles are calculated,and the 2D identification of the particles of the spectrum is achieved by combining typical particle morphological characteristics and wear mechanisms,allowing the identification and position marking of complex particles such as severe sliding particles and fatigue particles.By using the depth-of-field extension algorithm based on multi-focus image fusion,a spectral flip mechanism is used to acquire full-focus images of complex abrasive grains from different views.The SIFT feature point detection algorithm is used to detect and match feature points of wear particles in the image,then a stereo vision coordinate model is constructed for the microscope,the 3D coordinates of the feature points are finally calculated based on the pair of polar constraints and reprojection errors of images,and the reconstruction of the surface morphology of the abrasive particles is completed.by using the motorised stage to retrace the position of all complex abrasive particles in the field,the automatic reconstruction and analysis of the surface morphology of all the complex abrasive particles is finally achieved.By identifying the 2D characteristic Parameters and surface morphology type of the abrasive particles on the rotaryferrogram,the result shows that the method can automatically analyse the 3D morphology of abrasive particles in an efficient and accurate manner,and can achieve an effective identification of the wear status and health condition of the equipment.This paper contains 98 figures,15 tables,and 89 references.