An image dataset created by the Brazilian Agricultural Research Corporation (Embrapa) for studying object detection and instance segmentation in viticulture. It provides images and annotations of five different grape varieties, capturing variance in pose, illumination, focus, and genetic and phenological traits. The dataset was created by Thiago Teixeira Santos.
Use Cases
- Training instance segmentation models based on annotated grape instances.
- Performing grape variety identification based on genetic and phenological variations.
- Relaxing the task to object detection to locate grapes within images.
- Relaxing the task to semantic segmentation to identify grape pixels.
- Relaxing the task to image classification to determine grape presence.
Strengths
- Images capture real-world variance in grape pose, illumination, and focus.
- Includes instances from five different grape varieties, showing genetic and phenological variations.
- Annotations support multiple computer vision tasks: instance segmentation, detection, and classification.
Limitations
- Row count and dataset size are unknown, which may limit suitability assessment.
- Column-level documentation is absent; field semantics must be inferred after download.
- Last update date is unknown; freshness unverified.
Provenance
- Source
- Brazilian Agricultural Research Corporation (Embrapa)
- Collection Method
- Images taken on field.