Egocentric Warehouse Video Dataset

First-person video of real warehouse operations — picking, packing, sorting, and navigation — captured across diverse fulfillment center layouts with logistics-specific annotations for training warehouse robotics and AMR systems.

Dataset at a Glance

85K+
Video clips
600+
Hours recorded
40+ warehouse layouts
Environments
6+
Annotation layers

Comparison with Public Datasets

How Claru's dataset compares to publicly available alternatives.

DatasetClipsHoursModalitiesEnvironmentsAnnotations
Ego4D (workplace subset)~15K~80RGB, IMUMixed workplaceNarrations, hands
IndustReal8K24RGB-DLab warehouseAssembly actions
Claru Warehouse85K+600+RGB, Depth40+ real warehousesActions, objects, paths, zones

Use Cases

Autonomous Mobile Robots

Navigation and obstacle avoidance in dynamic warehouse environments. Example models: Locus Robotics, 6 River Systems, Fetch Robotics.

Pick-and-Place Systems

Grasping diverse products from shelves and bins under operational conditions. Example models: Covariant, RightHand Robotics, Berkshire Grey.

Warehouse Safety Monitoring

Detecting unsafe worker behaviors and zone violations in real-time. Example models: Voxel51, Kinema Systems, Dexterity.

Key References

  1. [1]Duan et al.. A Survey on Autonomous Mobile Robot Navigation in Warehouses.” Robotics and Autonomous Systems 2023, 2023. Link
  2. [2]Mahler et al.. Learning Ambidextrous Robot Grasping Policies.” Science Robotics 2019, 2019. Link
  3. [3]Grauman et al.. Ego4D: Around the World in 3,000 Hours of Egocentric Video.” CVPR 2022, 2022. Link

How Claru Delivers This Data

Claru deploys collectors in active fulfillment centers and distribution warehouses across North America, capturing genuine picking, packing, and sorting workflows. Unlike synthetic warehouse simulations, Claru's data includes real product diversity, authentic worker movement patterns, and the visual complexity of operational warehouses.

Frequently Asked Questions

The dataset covers order picking (goods-to-person and person-to-goods), packing, inventory putaway, cycle counting, forklift operation, and general warehouse navigation across various rack and bin configurations.

Yes. All collection occurs in active warehouses and fulfillment centers during normal operations, capturing authentic product diversity, real worker movement patterns, and operational constraints.

Beyond temporal actions and object tracking, the dataset includes spatial path annotations (2D floor-plane trajectories), zone occupancy labels, pick/place event markers, and worker pose estimation for human-robot interaction modeling.

Request a Sample Pack

Get a curated sample of egocentric warehouse video data with full annotations to evaluate for your project.