Author(s): |
Huijun Hu, State Key Lab of Software Engineering, Computer School, Wuhan University, Wuhan 430072, China;College of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan 430065, China Yuanxiang Li, State Key Lab of Software Engineering, Computer School, Wuhan University, Wuhan 430072, China Maofu Liu, College of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan 430065, China Wenhao Liang, Zhejiang Dahua Technology Co. Ltd, Hangzhou 310053, China |
Abstract: |
In this paper, we use back-propagation neural network to classify the defects in steel strip surface images. After image binarization, three types of image features, including geometric feature, grayscale feature and shape feature, are extracted by combining the defect target image and its corresponding binary image. For the classification model based on back-propagation neural network, we utilize hyperbolic tangent function as the activation function and determine the number of neurons in the hidden layer by experiments. Experiment results show that the back-propagation neural network makes good performance in the predication accuracy and the average predication time.
|