DSI(Dam surface inspection) is a multi-class defects segmentation data set collected by a climbing robot on dam concrete surface. We have included the CSSC dataset in our dataset; to download CSSC dataset, please contact the authors of the [Paper].
We include CSSC dataset with crack and spalling defect in the left bottom of image. DSI dataset includes crack, patched area, erosion, spot, and climbing robot rope segmentation labels.
We include augmentation of multi-class data, some of the augmentations containing fewer pixel labels than the original labeling are ignored by our txt files in zip files. Download via Baidu Netdisk (code: pnx6, Extracting passwords：buildings-2105107)
If you find this useful please consider citing
@Article{buildings13020285,
AUTHOR = {Hong, Kunlong and Wang, Hongguang and Yuan, Bingbing and Wang, Tianfu},
TITLE = {Multiple Defects Inspection of Dam Spillway Surface Using Deep Learning and 3D Reconstruction Techniques},
JOURNAL = {Buildings},
VOLUME = {13},
YEAR = {2023},
NUMBER = {2},
ARTICLE-NUMBER = {285},
URL = {https://www.mdpi.com/2075-5309/13/2/285},
ISSN = {2075-5309},
DOI = {10.3390/buildings13020285}
}