CitrusFarm: Multimodal Agricultural Robotics Data from UC Riverside

1.3 TB of data collected over 1.7 hours and 7.5 km by a mobile robot in citrus fields. The dataset provides nine sensing modalities, including stereo RGB, depth, near-infrared, thermal, LiDAR, and GPS-RTK data. It was gathered by the Autonomous Robots and Control Systems Lab at the University of California Riverside in the summer of 2023.

Use Cases

• Developing visual-inertial odometry algorithms based on synchronized stereo RGB, IMU, and wheel odometry data.
• Training semantic segmentation models for crop and tree detection based on multispectral and thermal images.
• Creating high-precision field maps for autonomous navigation based on LiDAR point clouds and GPS-RTK data.
• Analyzing crop health and growth stages based on near-infrared and thermal imagery across different daylight conditions.

Strengths

• Offers nine synchronized sensing modalities, including stereo RGB, depth, monochrome, near-infrared, thermal, LiDAR, IMU, GPS-RTK, and wheel odometry.
• Covers 7.5 km of robot travel over 1.7 hours, constituting 1.3 TB of data across seven sequences.
• Includes data from three citrus fields with varying tree species, growth stages, planting patterns, and daylight conditions.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Autonomous Robots and Control Systems Lab, University of California Riverside.

Collection Method

Collected by a wheeled mobile robot in the Agricultural Experimental Station.

Time Range

Summer 2023.

Freshness

Data was collected in summer 2023; last platform update is unknown.

Geography

Agricultural Experimental Station, University of California Riverside.