Study Of Optical Micro-tomography Image Analysis And Recognition Method Of Grains On Diamond Grinding Wheel

As one of the main materials in grinding,diamond grain is an important part of grinding wheel.Its distribution characteristics directly affect the grinding results.The grinding performance of diamond wheel can be evaluated by studying the distribution of diamond grains,the height and shape of diamond,and by predicting the wear degree of grains.Therefore,the analysis and shape recognition of diamond grains is helpful to guide the grinding process and optimize the grinding performance.In this paper,based on the optical microscopy measurement technology,the grain tomography images of different depths were obtained by focusing synthesis in-site measuring instrument.The characteristic parameters of the grains extracted by analyzing the optical micro-tomography images of the grinding wheel,and the recognition and classification of grain shapes were realized.The main research work includes:(1)Grain image preprocessing.The grain tomography images of different depths were analyzed,and the different image preprocessing methods were studied for the degradation of grains caused by different factors.The image enhancement algorithm based on MSRCR and histogram equalization was used to enhance the grain region,and the better region information of grain was obtained.(2)Grain image segmentation.Based on the histogram threshold,an improved Otsu adaptive threshold segmentation algorithm with particle swarm optimization was proposed to segment the broken grains at different levels.The segmented grain regions could be merged by Hough line detection and the least squares fitting.(3)Grain feature extraction and recognition.Studying on the geometric feature extraction method of single grain,the characteristics of different grains were represented by wavelet descriptors and they were used as the characteristic input of the follow-up support vector machine to identify the shape of grains.Orientation information between multiple grains was analyzed;the characteristic parameters of the single grain at different depths were obtained.The support vector machineclassifier was used to realize the recognition and classification of different grains.The experimental results showed that the optical micro-tomography image analysis and recognition algorithms in this paper can effectively realize the feature extraction and classification of diamond grains.