Amazon Bin Image Dataset: Over 500,000 Images from Operating Fulfillment Centers

Amazon Bin Image Dataset contains over 500,000 images and metadata captured from bins within pods in an operating Amazon Fulfillment Center. The images were taken by robot units as they carried pods during normal warehouse operations. The dataset is provided by Amazon under a CC-BY-NC-SA-3.0 license.

Use Cases

• Train object detection models based on images of items within storage bins.
• Develop inventory counting and auditing systems based on bin contents.
• Research robotic perception and grasping in warehouse environments based on the operational context.
• Benchmark computer vision algorithms for cluttered scene understanding based on real-world fulfillment center imagery.

Strengths

• Over 500,000 images provide a substantial volume for model training.
• Images are captured from real-world, operational Amazon Fulfillment Centers, offering authentic industrial data.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Amazon

Collection Method

Images captured by robot units carrying pods during normal Amazon Fulfillment Center operations.

Time Range

null

Freshness

Last updated is unknown.

Geography

null