Sample dataset#

Stepping into the dynamic world of computer vision becomes a lot easier when you have the right data to work with. In this regard, the “Smart Warehouse” dataset serves as a significant asset. Comprising of unannotated images and videos, this dataset mirrors real-life warehouse scenarios, providing a relevant and practical context for a wide array of computer vision tasks.

Below, you will find a detailed description of the “Smart Warehouse” dataset contents.

Note

All datasets provided in this documentation are licensed under the Apache License 2.0. All files in the datasets have to be extracted and uploaded to the project using Upload media button, not through dataset import feature.

Folder Overview#

Each category in this dataset, represented as a folder, focuses on a specific perspective or aspect of the warehouse environment. The two folders are:

  • BoxesOnTheConveyorBelt

  • WarehouseOverview

BoxesOnTheConveyorBelt#

This folder contains three image sets showcasing:

  1. a collection of good packages

  2. one faulty package on the conveyor belt

  3. packages viewed from a top-down perspective

Placeholder

These image sets are useful for segmentation, detection (including normal and rotated bounding boxes), and task chains, which involve applying a sequence of different computer vision tasks on the same data.

You can download the dataset from this link:

WarehouseOverview#

This folder contains four videos, each showing the same warehouse from a different perspective. There is also a video demonstrating a situation involving a faulty AMR.

Placeholder Placeholder

These videos can be used for detection tasks (such as identifying objects in the video), segmentation tasks, and anomaly detection/segmentation tasks.

You can download the dataset from this link: