Plant Semantic Segmentation Dataset

Description:

We are pleased to announce the release of an open-access dataset containing detailed semantic segmentation masks for plant images. This dataset, developed by the Computer Vision and Biosystems Signal Processing Group at the Department of Electrical and Computer Engineering, Aarhus University, has been carefully annotated by Humans in the Loop. The precise delineation of plant structures ensures high accuracy.

Dataset Details:

• Images: The dataset contains 144 high-resolution images of plant seedlings from three containers. Moreover, these images were captured at various time intervals over two months, showcasing different stages of plant growth. By taking images at various times, we aimed to capture the diversity in plant growth over this period.
• Annotation: Each image includes highly precise segmentation masks that distinguish between two classes: plant and background (which encompasses soil and container). These masks are incredibly accurate, meticulously tracing the complex leaf structures of each plant.
• Bounding Boxes: Additionally, each individual plant within the containers has been marked with bounding boxes, enhancing visibility and facilitating further analysis.

Potential Applications:

This dataset is invaluable for research and development in plant biology, agricultural automation, and computer vision. Researchers and practitioners can leverage these annotated images and masks for tasks such as plant phenotyping, disease detection, and growth monitoring.