Open AccessAutomatic detection for calcium leaching of dam concrete based on DeepLabv3+
Author(s) -
Jun Zhang,
Mang Luo,
Changsen Li,
Hao Mei
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1651/1/012035
Subject(s) - leaching (pedology) , calcium , service life , environmental science , geotechnical engineering , engineering , materials science , metallurgy , soil science , reliability engineering , soil water
Long-termly resistance to upstream water, dam concrete is prone to calcium leaching. It will lead to lesions of the concrete and gradual attenuation of strength, impermeability and frost resistance, thus reducing the durability and service life of the structure and even threatening the safe operation of the dam. At present, the concrete leached calcium is mainly checked by the workforce, but it is time-consuming, inefficient, and hard to quantitatively evaluate, such as the area of leached calcium. A semantic segmentation method based on the DeepLabv3+ with ResNet50v2 backbone is proposed to identify the calcium leaching of dam concrete automatically. The calcium-leached concrete data set is established, including 94 high-resolution images in dam corridors to verify the method. The results indicated that the DeepLabv3+ model finally reaches 0.72 mIoU on the test set, which is a practical way to detect calcium leaching of dam concrete.