Prediction of surface roughness in machining process by machine vision system

Abstract

The machine vision process has a new emergence technique to solve various engineering problems, Especially in welding, medical industries, automatic machining process like surface roughness detection, Coating level detection. Here this work concentrated on the prediction of surface roughness value in the turning process, the machining process done by CNC machining based on the Specification of Computer Numerical Control Machine. Aluminum 6063 Preferred as workpiece materials because of it used in automobile industries, construction, and other engineering works. The Contact stylus Probe Equipment measured the surface roughness value of the Aluminium 6063. These roughness values used to compare with the vision measurement value to predict the error percentage. This work mainly concentrated on using various soft computing approaches to predict the surface roughness value. In the future we can modify the testing materials based on applications. Soft computing model development was the main part of this work. Here the Artificial neural Network, Adaptive Fuzzy Inference System and Random Forest Classifier models developed and used to predict the surface roughness values. Then finally the Genetic Algorithm used to optimize the best value between the ANFIS and Random Forest Classifier. The DSLR camera captures the images and the pictures are extracted at grayscale value for further work. In this work, we selected ANN only for identified the sampling surface roughness accuracy in Aluminium 6063. Real work carried with ANFIS and Random forest classification. In ANN speed, feed rate, depth of cut feed and Grayscale value feed as the input. newline