RIT-18 Remote Sensing Image Semantic Segmentation Dataset

RIT-18 is a dataset for semantic segmentation of remote sensing images. The dataset was collected by a Tetracam Micro-MCA6 multispectral imaging sensor mounted on a DJI-1000 octorotor.

The characteristics of this dataset include

• Ultra-high-resolution multispectral imagery from drones
• Six spectral VNIR bands
• Contains a severely imbalanced class distribution, involving 18 object classes (plus background).