Huangtu Dataset 1.0: Very High-Resolution Remote Sensing Dataset for Rural Waste Management

Huangtu Dataset 1.0 is a very high-resolution remote sensing dataset designed to support rural waste management optimization, land cover classification, waste assessment, and collection route planning.

The dataset accompanies the paper “Optimizing Rural Waste Management: Leveraging High-Resolution Remote Sensing and GIS for Efficient Collection and Routing”. It focuses on Huangtu Town, Sichuan Province, Southwest China, a rural area with complex geographical features and relatively low population density.

The source imagery was obtained from Google Earth Engine in 2020 at 18-level zoom with RGB bands. The images were processed through stitching, cropping, bit-depth adjustment, reprojection to UTM, and resampling to a spatial resolution of 1 meter.

The dataset contains 30 very high-resolution image sample blocks, each with a size of 1000 × 1000 pixels. It includes image samples and two types of pixel-level annotation labels:

VHR Images: 30 RGB remote sensing image patches.

Complete Land Cover Labels: semantic segmentation masks with six classes.

Farmland Boundary Labels: binary boundary masks for farmland boundaries.

The complete land cover labels contain the following classes:

0: Background1: Buildings2: Roads3: Forests / green spaces4: Farmland5: Water bodies

This dataset can be used for rural land cover mapping, semantic segmentation, farmland boundary extraction, waste management planning, rural infrastructure analysis, and GIS-based route optimization research.

Acknowledgements: This project was supported by the Graduate Quality Engineering Construction Funding Program of Chengdu University of Technology (2024YAL016).