amos3 is a Python 3 client for the Archive of Many Outdoor Scenes (AMOS) that includes:

• methods to retrieve lists of cameras and related metadata
• methods to retrieve lists of images by camera
• methods to retrieve raw images by camera/timestamp
• methods to retrieve monthly ZIP archives of images by camera
• methods to construct datasets of one or more cameras with corresponding metadata and save locally or on AWS S3
• comprehensive unit tests with near-complete coverage

AMOS is a project of Media and Machines Lab at the Washington University in St. Louis. In their own words,

AMOS is a collection of long-term timelapse imagery from publicly accessible outdoor webcams around the world. We explore how to use these images to learn about the world around us, with a focus on understanding changes in natural environments and understanding how people use public spaces.

AMOS contains over 1 billion images from nearly 30K cameras over more than 15 years (as of June 2018).

This repository and its author are not affiliated with AMOS or WashU.

This initial release supports core functionality required to create simple datasets from AMOS on modern Python distributions. Integration with PIL/Pillow is also being considered to simplify the generation of standardized inputs for machine learning/computer vision applications.

• GitHub repository: https://github.com/mjbommar/amos3/
• Documentation: In Progress
• Contact: GitHub Issues

amos3 is distributed under the standard OSI MIT License (Expat) as documented in the repository LICENSE file.