2026-10-11T14:18:05Z https://www.scidb.cn/oai

10.57760/sciencedb.011o2 2026-09-09T16:04:36Z

2026-09-09 remote sensing denoising dataset doi:10.57760/sciencedb.011o2 en This remote sensing denoising dataset consists of four subfolders, namely train, val, test, and example. The train, validation, and test sets contain a total of 20,590 high-quality clean remote sensing images with a uniform resolution of 512×512 pixels, all saved in PNG format. The example folder contains 150 manually screened real remote sensing noise samples, which can be fused with clean remote sensing images to synthesize authentic noisy remote sensing images that conform to real-scene imaging characteristics. This dataset is specially constructed and dedicated for the research, training, and evaluation of remote sensing image denoising algorithms. remote sensing; denoising; dataset mufanxing PUBLIC https://creativecommons.org/licenses/by/4.0/ dataset Science Data Bank