Quantify pixel-level detection of dam surface crack using deep learning

Abstract An increasing number of detection methods based on computer vision are applied to detect cracks in water conservancy infrastructure. However, most studies directly use existing feature extraction networks to extract crack information, which are proposed for open-source datasets. As the crack distribution and pixel features are different from these data, the extracted crack information is incomplete. In this paper, a deep learning-based network for dam surface crack detection is proposed, which mainly addresses the semantic segmentation of cracks on the dam surface. Particularly, we design a shallow encoding network to extract features of crack images based on the statistical analysis of cracks. Further, to enhance the relevance of contextual information, we introduce an attention module into the decoding network. During the training, we use the sum of cross-entropy and Dice loss as the loss function to overcome data imbalance. The quantitative crack information is extracted by the imaging principle after using morphological algorithms to extract the morphological features of the predicted result. We built a manual annotation dataset containing 1577 images to verify the effectiveness of the proposed method. This method achieves state-of-the-art performance on our dataset. Specifically, the precision, recall, Intersection of Union (IoU), F1_measure, and accuracy are 90.81%, 81.54%, 75.23%, 85.93%, 99.76%, respectively, and the quantification error of cracks is less than 4%.