Dataset: Image-Based Method for Automatic Spillway Ruler Reading

This dataset contains the training and test images used in the work detailed in the article: Schviderski, A., Nassu, B. T. Image-Based Method for Automatic Spillway Ruler Reading (10.1109/ACCESS.2026.3688802). It can be freely used for academic and non-commercial purposes. Commercial use is not permitted unless explicitly authorized. The dataset is provided "as is" without warranties of any kind, and the creators assume no responsibility for any consequences arising from its use. If you use the dataset, please cite our paper! A. Schviderski and B. T. Nassu, "Image-Based Method for Automatic Spillway Ruler Reading," in IEEE Access, vol. 14, pp. 68910-68919, 2026, doi: 10.1109/ACCESS.2026.3688802. @ARTICLE{11499408, author={Schviderski, Alexandre and Nassu, Bogdan Tomoyuki}, journal={IEEE Access}, title={Image-Based Method for Automatic Spillway Ruler Reading}, year={2026}, volume={14}, number={}, pages={68910-68919}, doi={10.1109/ACCESS.2026.3688802},} Contains about 27,000 images extracted from videos captured by various cameras while in maneuvers of the spillway gates in a variety of scenarios, in three different dams (A, B, C) each with a ruler model. They are separated as train and test to do a fine-tuning process of a deep learning detection model, as described in the article.Each subfolder inside the train and test folders is related to a unique camera from a specific gate of a dam to avoid data leakage, and each camera has images and labels from digits (0 to 9) and pointer locations, which are used to indicate the aperture value of the gate in meters, or in meters and decimeters.