YTU-CH Dataset

YTU-CH (Yildiz Technical University - Cultural Heritage) Dataset was specifically developed for edge detection in cultural heritage structures. The dataset comprises a variety of images captured from the Internet, cameras, and low-cost UAVs, all featuring three bands (RGB) and an 8-bit radiometric resolution. In the dataset, contours and cracks have been manually digitized in all images. To standardize the dataset, images were divided into 512×512 pixel patches due to their variable sizes. Patches with an edge detail ratio of less than 2% were excluded, resulting in a final set of 5,380 labeled image patches. The dataset is then split into 90% (4880 image patches) and 10% (500 image patches) for training and testing, respectively. 

The creation of this dataset was funded by the Scientific and Technological Research Council of Türkiye (TUBITAK) 1001 program under Project no. 122Y017.