Automatic crack detection of dam concrete structures based on deep learning

Zongjie Lv, Jinzhang Tian, Yantao Zhu and Yangtao Li

Abstract

Crack detection is an essential method to ensure the safety of dam concrete structures. Low-quality crack images of dam concrete structures limit the application of neural network methods in crack detection. This research proposes a modified attentional mechanism model to reduce the disturbance caused by uneven light, shadow, and water spots in crack images. Also, the focal loss function solves the small ratio of crack information. The dataset collects from the network, laboratory and actual inspection dataset of dam concrete structures. This research proposes a novel method for crack detection of dam concrete structures based on the U-Net neural network, namely AF-UNet. A mutual comparison of OTSU, Canny, region growing, DeepLab V3+, SegFormer, U-Net, and AF-UNet (proposed) verified the detection accuracy. A binocular camera detects cracks in the experimental scene. The smallest measurement width of the system is 0.27 mm. The potential goal is to achieve real-time detection and localization of cracks in dam concrete structures.

Key Words

attention mechanism; crack detection; dam concrete structures; deep learning; focal loss; U-Net

Address

Zongjie Lv and Yangtao Li: 1) The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing 210024, China, 2) College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing, 210024, China Jinzhang Tian: 1) National Dam Safety Research Center, Wuhan, Hubei 430010, China, 2) Changjiang Survey, Planning, Design and Research Co.,Ltd., Wuhan 430010,China Yantao Zhu: 1) The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing 210024, China, 2) College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing, 210024, China, 3) National Dam Safety Research Center, Wuhan, Hubei 430010, China

← Back to Volume 32, No. 6