2026-10-12T03:28:48Z https://www.scidb.cn/oai

10.57760/sciencedb.08354 2023-09-26T17:46:34Z

2023-09-26 Synthetic Digital City Dataset doi:10.57760/sciencedb.08354 en Synthetic Digital City Dataset (SDCD) is a large-scale, high-resolution monocular depth driving dataset, which consists of 930K high-resolution RGB images and corresponding dense depth maps. SDCD is generated and collected under 7 different weather driving conditions [Sunny, Overcast, Night, Rain, Snow, Sleet, Dust], and common adverse disturbances caused by the transmission of communication equipment [Noise, Blur, ColorQuant, Pixelate, JPEG] is imposed on RGB images of SDCD as well. SDCD provides adequate training data for investigate the generalization and robustness of depth estimation algorithms, and it serve to boost the development of the perception system in autonomous driving. Monocular depth estimation; Computer Vision; Automotive engineering Jihao Li PUBLIC https://creativecommons.org/licenses/by/4.0/ dataset Science Data Bank