Still image and spatial-temporal tomato data enabling detection, segmentation, tracking, and video-instance segmentation using strong and weak labels

2026-07-16 · Source: Computer Vision and Pattern Recognition · Field: Agriculture & Food Systems — Precision Agriculture & Smart Farming, Crop Science & Plant Technology, Artificial Intelligence & Machine Learning · Depth: Expert, quick

Summary

Researchers have released two new datasets, BUTom21 and BUTom-ST21, designed for visual sensing of tomato plants in commercial-like environments using robotic acquisition. BUTom21 comprises still images with detailed manual pixel-level annotations for fruit ripeness. Complementing this, BUTom-ST21 offers video-based spatial-temporal data, featuring semi-automated pseudo-labels for pixel-level ripeness, along with individual fruit count and ripeness information over time. These datasets aim to provide the research community with challenging, real-world imagery to advance methods for sensing and estimating the state of tomato plants and their fruit. BUTom-ST21, in particular, supports pushing the boundaries of field-based phenotyping through its unique spatial-temporal data.

Key takeaway

For Computer Vision Engineers or Research Scientists developing agricultural AI, these new BUTom datasets offer critical resources. If you are building systems for automated crop monitoring or field-based phenotyping, you should integrate BUTom21 and BUTom-ST21 to benchmark and refine your models. The spatial-temporal data, in particular, provides a unique opportunity to advance video-instance segmentation and fruit tracking capabilities, directly impacting the accuracy of yield estimation and ripeness assessment in real-world settings.

Key insights

New BUTom21 and BUTom-ST21 datasets offer challenging real-world tomato plant imagery with detailed ripeness labels for advanced phenotyping research.

Principles

• Real-world, challenging datasets drive phenotyping innovation.
• Spatial-temporal data enhances fruit count and ripeness tracking.

In practice

• Develop models for tomato fruit detection and segmentation.
• Explore video-instance segmentation for fruit tracking.
• Advance field-based phenotyping using spatial-temporal data.

Topics

• Computer Vision
• Agricultural Robotics
• Tomato Phenotyping
• Video-Instance Segmentation
• Dataset Annotation
• Spatial-Temporal Data

Best for: AI Scientist, Computer Vision Engineer, Research Scientist

Related on AIssential

• Tomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping — TomatoMAP dataset enables accurate, reproducible fine-grained tomato phenotyping using multi-angle imagery and AI models.
• SAM 3.1: Faster and More Accessible Real-Time Video Detection and Tracking With Multiplexing and Global Reasoning - AI at Meta — SAM 3.1 enhances real-time video object tracking via multiplexing, doubling processing speed and reducing GPU demands.
• Pixel-Precise Explainable Stress Indexing: A Semantic Segmentation Framework for Disease Severity Quantification in Field Crops — A deep learning pipeline accurately quantifies plant disease severity using semantic segmentation for real-time agricultural monitoring.
• USU-Corn-WeedDB: A UAV RGB Image Dataset for Multi-Species Weed Detection in Forage Corn — A new UAV RGB image dataset addresses data scarcity for multi-species weed detection in forage corn.
• Scene-Centric Unsupervised Video Panoptic Segmentation — VideoCUPS enables unsupervised video panoptic segmentation by generating pseudo-labels from scene-centric videos using depth, motion, and visual cues.
• The Turning Point of 3D Plant Phenotyping: 3D Foundation Models Enable Minute-to-Second Cross-Crop Reconstruction and Beyond — 3D Foundation Models drastically accelerate 3D plant phenotyping, reducing reconstruction time from minutes to seconds while maintaining accuracy.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computer Vision and Pattern Recognition.