Roughness Classification of End Milling Based on Machine Vision

At present, the sample comparison method is often used in the industrial field to classify the roughness of end milling, which has some problems such as high requirements to inspectors and subjective inspection results. So this paper proposes a classification method of end milling roughness based on machine vision. Firstly, the image acquisition device combined by a mobile phone camera and a miniature microscope is used to capture surface images of the end milling sample. Secondly, the image dataset is constructed by expanding the image sample size and preprocessing image. Then the classification results of the improved LeNet-5 and AlexNet are compared to determine the more appropriate structure. Finally, particle swarm optimization (PSO) is used to optimize the model. The experimental results prove that the classification accuracy of the improved PSO-AlexNet is higher than the improved LeNet-5 and AlexNet, and can meet the roughness classification requirements. So this method can eliminate the influence of human factors and evaluate the classification results of end milling roughness objectively and accurately.