The dataset consists of Sentinel-2 images at 10m spatial resolution along with their corresponding binary road masks acquired from seven diverse regions across India, covering urban, rural, and mountainous terrains. We also offer a set of enhanced version of the Sentinel-2 images, which incorporates techniques such as gamma correction and CLAHE, for improved road visibility. For ease of reuse in different deep learning frameworks, the dataset is made available in TIF as well as PNG format. The dataset contains a total of 5,634 high-quality samples, each of 256x256 pixels, obtained after manual refinement.

The dataset is divided into two major parts: Sentinel-2 images and their corresponding binary road masks. Both the masks and the images are available in PNG as well as TIF formats. We have also included enhanced versions of the Sentinel-2 images (images_enhanced_png.zip). The dataset is organised into five folders each containing 5634 images. All the images and masks in our dataset have a resolution of 256x256 pixels. Each image and its corresponding mask are labelled numerically (eg. ‘1.tif’, ‘2.tif’ or ‘1.png’, ‘2.png’), with every number corresponding to the same file across all five folders.