Author:
Zhang Jun,Yan Zhaohui,Ma Xinlong,Cai Xingguo
Abstract
Abstract
Leakage is an important factor affecting the safety of the dam. In the past, manual inspection is a significant way to monitor leakage risk. However, it is time-consuming, inefficient and difficult to quantitative evaluate such as the leakage area. A semantic segmentation method based on the fully convolutional network is proposed to replace the manual inspection for the dam leakage automatic detection. Thirty-eight high-resolution images of dam leakage are collected. FCN-8s and VGG16 backbone are adopted. The results indicated that the FCN-8s achieves the mIoU to 0.59 on the test set, which proves to be an efficient way to detect the dam leakage.
Reference6 articles.
1. Research on seepage field of concrete dam foundation based on artificial neural network;Zhang,2020
2. Analysis of dam Leakage hazards and Treatment Measures;Tan;Heilongjiang Science and Technology Information,2008
3. A new method for dynamically estimating long-term seepage failure frequency for high concrete faced rockfill dams;Deng,2020
4. Study on leakage characteristics of reservoir dams;Tan;Dam and Safety,2019
5. Fully convolutional networks for semantic segmentation;Long,2015
Cited by
2 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献