Multimodal Mobile Robotic Dataset for a Typical Mediterranean Greenhouse: The GREENBOT Dataset.

Area of Science:

• Robotics and Automation
• Agricultural Technology
• Computer Vision

Background:

• Precise localization is vital in agricultural settings like greenhouses, but Global Navigation Satellite System (GNSS) signals can be unreliable due to structural interference and dense crops.
• Existing datasets often do not adequately represent the complex sensory challenges found in indoor farming environments.

Purpose of the Study:

• To introduce a novel dataset specifically curated for the unique challenges of agricultural environments, particularly greenhouses.
• To facilitate the development and testing of Simultaneous Localization and Mapping (SLAM) algorithms in indoor farming scenarios.
• To enable the creation of detailed 3D plant models for applications like precision agriculture and robotic operations.

Main Methods:

• A mobile robot platform equipped with standard mobile robot sensors collected data within a Mediterranean greenhouse containing tomato crops.
• The dataset captures sensor data along all greenhouse corridors, simulating real-world operational conditions.
• State-of-the-art SLAM algorithms were applied to the dataset to assess its suitability for localization and mapping tasks.

Main Results:

• The dataset provides a unique resource for developing and evaluating SLAM algorithms in challenging greenhouse environments.
• Initial tests demonstrate the feasibility of using the dataset to assess the performance of current SLAM techniques.
• The collected data supports the construction of detailed 3D plant models, crucial for advanced agricultural applications.

Conclusions:

• This dataset represents a significant contribution to agricultural robotics, offering a benchmark for SLAM research in greenhouses.
• It opens new avenues for developing autonomous systems for tasks such as robotized spraying and crop monitoring.
• The availability of this dataset online will accelerate innovation in precision agriculture and indoor farming technologies.