Welcome to the documentation of FarmVibes.AI, a platform for developing multi-modal geospatial machine learning (ML) models for agriculture and sustainability. The goal of the platform is to help users obtain insights by fusing multiple geospatial and spatiotemporal datasets.

The platform includes a set of data ingestion and pre-processing workflows that help users prepare data for fusion models tailored towards agriculture. The platform comes with many dataset downloaders, including satellite imagery from Sentinel 1 and 2, US Cropland Data, USGS Elevation maps, NAIP imagery, NOAA weather data, and private weather data from Ambient Weather. Users can also bring in any rasterized datasets they want to make them fusion-ready for FarmVibes.AI, such as drone imagery or other satellite imagery. We also provide multiple data processing workflows that can be used on top of downloaded data, such as, index computation and data summarization.

The project provides instructions and user guides to set up local and remote FarmVibes.AI clusters. Both instances use pre-built operators and workflows, with the difference that the local instance runs and persists data locally, while the remote instance employs a remote Azure Kubernetes Service (AKS) cluster. In both scenarios, the user can interact with the FarmVibes.AI cluster via a REST API or a Python client (inside a Jupyter Notebook, for example).

This documentation is a work in progress and will be updated regularly. If you have any questions or find any issues while running the code, please make sure to open an issue or a discussion on our GitHub repository.

Quickstart & Useful Resources

For a first contact with FarmVibes.AI and its capabilities, please refer to:

  • Quickstart guide for information on how to install and get started using the platform locally.

  • AKS setup guide, if you prefer to setup a remote Azure Kubernetes Service (AKS) cluster to run FarmVibes.AI.

  • VM setup guide, if you prefer to setup a dedicated Azure Virtual Machine to run FarmVibes.AI.

  • Jupyter notebooks for practical examples on how to use the platform.

Additionally, the following user guides and links may be helpful:

Indices and tables

You can also look for classes and methods by name using the following indices: