Computer vision-based surface roughness measurement using artificial neural network

Product quality is defined by the product lifespan and the product finish. The lower wear rate leads to good quality, especially in the machining components, so low-wear components have excellent surface roughness. Al7075 (Aluminum Alloy) is used in automotive industries for its predominant machining characteristics. The paper aims to predict the accuracy of surface roughness in the aluminum cylindrical shaft by the computer vision system. A feed-forward algorithm selected in the neural network for training and testing practice, Training Practice Speed, Depth of Cut, Feed Rate, Grayscale Values, defined as the input parameters, surface roughness is the output parameters. The accuracy is calculated by the difference between the vision measurement and stylus probe value. The accuracy level attains 95%. The method can make sustainable for the surface roughness measurement in the machining process.Copyright (c) 2022 Elsevier Ltd. All rights reserved.Selection and peer-review under responsibility of the scientific committee of the International Conference on Latest Developments in Materials & Manufacturing