| Filename | Latest commit message | Latest commit date |
|---|---|---|
|
|
||
| example images | ||
| README.md | ||
SugarcaneWeedDataset
This repository provides access to SugarcaneWeedDataset, a realistic, in-the-field dataset of sugarcane and weed imagery captured under variable environmental conditions for the development and evaluation of vision-based weed detection models in perennial crops.
📌 Overview
SugarcaneWeedDataset consists of 2,139 high-resolution RGB images (4608×3072) captured in sugarcane fields, annotated for:
- 869
sugarcane-only - 855
weed-only - 415
mixed
This dataset addresses the real-world challenge of green-on-green weed detection in perennial crops, a scenario underrepresented in existing benchmarks.
📥 Download
You can download the dataset via the following link:
🔗 Download Dataset (Google Drive)
📂 Dataset Splits and Annotations
In addition to the raw images, we provide official dataset split definitions and annotation files for detection and segmentation tasks.
These resources include:
- Predefined train/val/test splits
- Detection annotations (bounding boxes)
- Segmentation annotations (pixel-wise masks)
🔗 Download Splits and Annotations (Google Drive)
All files are provided as compressed .zip archives. Please extract them and follow the folder structure described in the corresponding task directories.