International Conference on Computer and Knowledge Engineering

Home / 12th International Conference on Computer and Knowledge Engineering

An Automated Visual Defect Segmentation for Flat Steel Surface Using Deep Neural Networks

Authors :

Dorna Nourbakhsh Sabet¹ Mohammad Reza Zarifi² Javad Khoramdel³ Yasamin Borhani⁴ Esmaeil Najafi⁵

1- Faculty of Mechanical Engineering, K. N. Toosi University of Technology, Tehran, Iran 2- Faculty of Mechanical Engineering, K. N. Toosi University of Technology, Tehran, Iran 3- Faculty of Mechanical Engineering, Tarbiat Modares University, Tehran, Iran 4- Faculty of Mechanical Engineering, K. N. Toosi University of Technology, Tehran, Iran 5- Faculty of Mechanical Engineering, K. N. Toosi University of Technology, Tehran, Iran

Keywords :

Steel surface،Visual detection،Defect segmentation،Deep neural networks

Abstract :

Defect detection on the metal sheets is an importance for the steel production industry. This paper proposes an automated visual defect segmentation for flat steel surface using two different deep neural network methods, namely U-Net and ResNet34, MobilNetV2, EfficientNetB0, InceptionV3, and VGG16 have been implemented as the U-Net’s encoder. In addition, FCN-8 model has been investigated as a different architecture for the SEVERESTAL dataset. Due to the imbalanced dataset, various techniques such as dice coefficient as loss function and data augmentation have been used. Moreover, no-defect steel surfaces have been considered as a separate class. As a result, a promising dice score is achieved on the validation set for the described four-class approach.