learnrobotics.ai
Glossary← Back to feed
PERCEPTIONCURRENT2026-06-17

Viking Hill Dataset: A Lidar-Radar-Camera Dataset for Detection and Segmentation in Forest Scenes

Vladimír Kubelka, Oleksandr Kotlyar, Unal Artan, Martin Magnusson

DATASET
Multi-session forest dataset with 3D cuboid annotations across three modalities

This Robot LearningDatasetA collection of training or evaluation data. gives you co-registered Perception & SensingLidarA sensor that measures distance using laser light, often used in mapping and navigation., radar, and camera data for Robot LearningTrainingThe process of fitting a model using data or experience. tree detection and Perception & SensingSegmentationDividing an image into meaningful regions or object masks. models that work under forest canopy where vision fails. You can now Simulation & Sim-to-RealBenchmarkA standard test used to compare methods fairly. multi-modal fusion and radar-based Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. for autonomous forestry robots, with ground-truth tree diameters and seasonal variations included.

THE PROBLEM

This paper focuses on Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world.. This Robot LearningDatasetA collection of training or evaluation data. gives you co-registered Perception & SensingLidarA sensor that measures distance using laser light, often used in mapping and navigation., radar, and camera data for Robot LearningTrainingThe process of fitting a model using data or experience. tree detection and Perception & SensingSegmentationDividing an image into meaningful regions or object masks. models that work under forest canopy where vision fails. You can now Simulation & Sim-to-RealBenchmarkA standard test used to compare methods fairly. multi-modal fusion and radar-based Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. for autonomous forestry robots, with ground-truth tree diameters and seasonal variations included. Read the paper by tracking the Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. definition, the Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or data assumptions, and the evidence that supports the claimed improvement.

HOW IT WORKS

1

Task framing

The paper frames the work as Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world.. Start here because it defines what success means and which assumptions the rest of the method inherits.

2

Core method

This Robot LearningDatasetA collection of training or evaluation data. gives you co-registered Perception & SensingLidarA sensor that measures distance using laser light, often used in mapping and navigation., radar, and camera data for Robot LearningTrainingThe process of fitting a model using data or experience. tree detection and Perception & SensingSegmentationDividing an image into meaningful regions or object masks. models that work under forest canopy where vision fails. You can now Simulation & Sim-to-RealBenchmarkA standard test used to compare methods fairly. multi-modal fusion and radar-based Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. for autonomous forestry robots, with ground-truth tree diameters and seasonal variations included. When reading the method section, identify the inputs, the learned or engineered representation, and the Core ConceptsActionA command the robot sends to its motors, controller, or low-level system. or prediction produced by the system.

3

Data and supervision

The reported data scale is Multi-session forest Robot LearningDatasetA collection of training or evaluation data. with 3D cuboid annotations across three modalities. For robotics work, the data story is part of the method: check whether the system depends on Imitation & Reinforcement LearningTeleoperation (teleop)A human remotely controlling the robot, often to collect demonstrations., Simulation & Sim-to-RealSimulationA virtual environment where robots can be trained or tested., internet video, human labels, or Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. rollouts.

4

Evaluation evidence

The paper should be judged through its Simulation & Sim-to-RealEvaluationMeasuring how well a robot system performs. protocol: what data is used, what Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or simulator is tested, and which Evaluation & ResearchBaselineA reference method used for comparison. comparisons support the claim. Look for the gap between the headline result and the Simulation & Sim-to-RealDeploymentPutting the trained system on a real robot. setting you would actually care about.

KEY RESULTS

Main contributionConceptual contribution

This Robot LearningDatasetA collection of training or evaluation data. gives you co-registered Perception & SensingLidarA sensor that measures distance using laser light, often used in mapping and navigation., radar, and camera data for Robot LearningTrainingThe process of fitting a model using data or experience. tree detection and Perception & SensingSegmentationDividing an image into meaningful regions or object masks. models that work under forest canopy where vision fails. You can now Simulation & Sim-to-RealBenchmarkA standard test used to compare methods fairly. multi-modal fusion and radar-based Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. for autonomous forestry robots, with ground-truth tree diameters and seasonal variations included.

Data scaleMulti-session forest dataset with 3D cuboid annotations across three modalities

The reported data scale matters because Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. systems often fail when the Data, Distributions & Training IssuesTraining distributionThe kinds of examples the model saw during training. is too narrow.

WHY DEVELOPERS SHOULD CARE

This Robot LearningDatasetA collection of training or evaluation data. gives you co-registered Perception & SensingLidarA sensor that measures distance using laser light, often used in mapping and navigation., radar, and camera data for Robot LearningTrainingThe process of fitting a model using data or experience. tree detection and Perception & SensingSegmentationDividing an image into meaningful regions or object masks. models that work under forest canopy where vision fails. You can now Simulation & Sim-to-RealBenchmarkA standard test used to compare methods fairly. multi-modal fusion and radar-based Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. for autonomous forestry robots, with ground-truth tree diameters and seasonal variations included.

LIMITATIONS

The main limitation to check is whether the claimed behavior holds outside the paper's reported setup. That means testing across different Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. embodiments, scenes, objects, and data distributions.

WHAT COMES NEXT

The practical next step is independent reproduction with clear baselines, ablations, and stress tests. For a developer, the useful follow-up is to map the paper's Perception & SensingPerceptionThe process of turning raw sensor data into useful understanding of the world. assumptions onto a concrete Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. stack, then test the smallest version of the method that could run end to end.

Read on arxiv →HTML source →

RELATED PAPERS

COMPUTER-VISION

Diffusion Policy: Visuomotor Policy Learning via Action Diffusion

COMPUTER-VISION

RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control

IMITATION-LEARNING

Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation

IMITATION-LEARNING

Octo: An Open-Source Generalist Robot Policy

DEPTH-ESTIMATION

Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics

VLA

MolmoAct2: Action Reasoning Models for Real-world Deployment