Postharvest Shanghai Bok Choy Dataset for Pre-Cutting Sorting
Published: 25 July 2026| Version 1 | DOI: 10.17632/84z7hs67fm.1
Contributors:
Yajie Yang, , , Description
This dataset contains RGB images of postharvest leafy greens collected in multiple batches under a conveyor-based robotic sorting environment. It includes 575 images and 909 annotated plant instances. Each instance is annotated in AnyLabeling JSON format with a bounding box and an instance segmentation polygon; 864 instances also include a directed line from the root end toward the leaf region for orientation estimation. The dataset provides a fixed split of 460 training images and 115 validation images. It can support research on object detection, instance segmentation, root-end localization, orientation estimation, and vision-guided robotic sorting of leafy vegetables.
Files
Institutions
- Tongji UniversityShanghai, Shanghai
Categories
Computer Vision, Robotics, Image Processing, Agricultural Engineering