A Multisensor Fusion-based High-Quality Depth Estimation Dataset for Dynamic Coal Flow

Depth estimation has become a key technology for improving efficiency in industrial applications such as intelligent sorting and robotic grasping. Limited installation space, stringent safety requirements, and the dynamic nature of observed objects restrict the application of spatial information sensors in industrial environments. To address these challenges, we develop a multisensor fusion-based high-quality depth estimation dataset to meet the specific needs of coal and non-ferrous metal mining in key processes such as extraction, transportation, and processing. This dataset integrates hardware devices, including wide-field-of-view visible light sensors, active depth sensors, trigger-signal generators and distributors, while employing a reprojection method based on a time-synchronized signal system to provide high-precision and spatiotemporally aligned data.