Artificial Intelligence-based Recognition of Wear and Corrosion on Steel Surfaces

Based on the commonly used steel wear and tear identification method is limited by many factors and the accuracy is not high enough to be stable, this paper adopts a new artificial intelligence algorithm to improve the accuracy of steel wear and corrosion identification. In this paper, a total of six neural network models such as CNN, VGG, etc. are used to classify a total of 1,800 steel photos with six different degrees of wear and tear, and after the model construction and parameter debugging, we get the classification accuracy of CNN, InceptionResNetV2, and Xception models is higher, and MobileNetV2 model has higher sensitivity to the scratch category. The conclusion that the sensitivity of images is higher, using these emerging AI algorithms of deep learning is an effective way to improve the accuracy of steel wear and corrosion recognition.