To address the problem that a deep neural network needs a sufficient number of training samples to have a good prediction performance, this paper firstly used the Z-Map algorithm to generate a simulated profile of the milling surface and construct an optical simulation model of surface imaging to supplement the training sample size of the neural network. Then the Deep CORAL model was used to match the textures of the simulated samples and the actual samples across domains to solve the problem that the simulated samples were not in the same domain as the actual milling samples. Experimental results have shown that high texture matching could be achieved between optical simulation images and actual images, laying the foundation for expanding the actual milled workpiece images with the simulation images. The deep convolutional neural model Xception was used to predict the classification of six classes of data sets with the inclusion of simulation images, and the accuracy was improved from 86.48% to 92.79% compared with the model without the inclusion of simulation images. The proposed method solves the problem of the need for a large number of samples for deep neural networks and lays the foundation for similar methods to predict surface roughness for different machining processes.

Current vision-based roughness measurement methods are classified into two main types: index design and deep learning. Among them, the computation procedure for constructing a roughness correlation index based on image data is relatively difficult, and the imaging environment criteria are stringent and not universally applicable. The roughness measurement method based on deep learning takes a long time to train the model, which is not conducive to achieving rapid online roughness measurement. To tackle with the problems mentioned above, a visual measurement method for surface roughness of milling workpieces based on broad learning system was proposed in this paper. The process began by capturing photos of the milling workpiece using a CCD camera in a normal lighting setting. Then, the train set was augmented with additional data to lower the quantity of data required by the model. Finally, the broad learning system was utilized to achieve the classification prediction of roughness. The experimental results showed that the roughness measurement method in this paper not only had a training speed incomparable to deep learning models, but also could automatically extract features and exhibited high recognition accuracy.