Blade30 is a comprehensive dataset for multiple blade-related tasks, including blade stitching, segmentation, defect detection, classification and deduplication, contamination detection and classification, and more. Blade30 was collected during the real drone-based wind turbine inspection in various environments. It contains 1,302 real drone-captured images covering 30 full blades captured under various conditions (both on- and off-shore), accompanied by a rich set of annotations such as defects and contaminations, etc. Thus, Blade30 leads in both quality and quantity.

paper: https://www.sciencedirect.com/science/article/pii/S0960148122018481?via%3Dihub=

1. Baidu Disc:
   • part1 (blade1-15): https://pan.baidu.com/s/17kv5Xadz1QcSrvoG58WtBw code：1234
   • part2 (blade16-30): https://pan.baidu.com/s/1hzcwdc6sBXOeja3nkfartg code：1234
2. OneDrive:
   • full dataset: https://1drv.ms/u/s!AoXJBmXKVWu5tmtUzCJULhrtYuIP?e=KYOtlo
3. GoogleDrive:
   • blade_1_15_with_annotation：https://drive.google.com/file/d/1HbB4t9xV2oCgSSxR9hMEOU6v9qDfetmR/view?usp=sharing
   • blade_16_30_with_annotation：https://drive.google.com/file/d/1SwRdMzA7zCkNVlHuWvk8uK6eDToM0mUV/view?usp=sharing

If you use these works in your research, please cite:

@article{Yang2023Blade30,
	author = {Cong Yang and Xun Liu and Hua Zhou and Yan Ke and John See},
	title = {Towards accurate image stitching for drone-based wind turbine blade inspection},
	journal = {Renewable Energy},
        volume = {203},
        pages = {267-279},
        year = {2023}
}

Data Structure:

• Blade_1
  • mask: ground truth of blade region segmentation
    • 10_d78543cb-5fb1-4677-b92e-2ece550642c9.jpg: blade image
    • 10_d78543cb-5fb1-4677-b92e-2ece550642c9.json: ground truth of defects and contaminations