Turning Surface Roughness Classification Detection Based on Machine Vision

In the field of machining, surface roughness is an important parameter for evaluating the surface quality of workpieces. This paper takes the resnet18-based network and proposes an improved residual neural network, which improves the accuracy rate to a certain extent by introducing the attention mechanism. Experiments show that in the classification and identification of turning surface roughness, the improved model avoids artificial design characteristics compared with artificial networks, which can effectively meet the classification and detection needs of turning surface roughness.