Our paper, Machine Learning Model Specification for Cataloging Spatio-Temporal Models, is now available online. It will be presented at the ACM SIGSPATIAL 2024 (GeoSearch'24) conference, which starts tomorrow in Atlanta, Georgia, USA. #ml #GeoAI #geochat
Link to paper: https://doi.org/10.1145/3681769.3698586
The accessibility of satellite data has catalyzed the hosting of these data sets in cloud computing environments, leading to several Earth Observation (EO) cloud platforms with analytical capabilities. However, transitioning between these platforms remains a challenge. The openEO specification addresses this by unifying backend services to be accessed via REST protocol using R, Javascript, and Python clients, yet it lacks a stable machine learning (ML) specification. This paper proposes extending the openEO API by developing a ML Application Programming Interface (API) specification for EO data cubes to facilitate ML models' reproducibility, reusability, and interoperability in multiple backend services and federated cloud platforms. Our efforts focus on establishing protocols for data pre-processing, model training, model parameter tuning, model prediction, model saving, and loading of existing models, facilitating the use of classical ML and deep learning (DL) algorithms for critical applications such as environmental monitoring and disaster response. Future efforts will focus on implementing this ML specification across multiple EO backend services and cataloging spatio-temporal ML models.
In recent decades, Earth Observation (EO) systems have seen remarkable technological advancements, leading to a surge in Earth-orbiting satellites capturing EO data. Cloud-based storage solutions have been adopted to manage the increasing data volume. Although numerous EO data management and analysis platforms have emerged to accommodate this growth, many suffer from limitations like closed-source software, leading to platform lock-in and restricted functionalities, restricting the scientific community from conducting open and reproducible research. To tackle these issues, we present OpenEOcubes, a lightweight EO data cubes analysis service that embraces open-source tools, standardized APIs, and containerized deployment, we demonstrate the service’s capabilities in two user scenarios: performing vegetation analysis in Amazonia, Brazil for one year, and detecting changes in a forested area in Brandenburg, Germany based on five years of EO data.OpenEOcubes is an easy-to-deploy service that empowers the scientific community to reproduce small and medium-sized EO scientific analysis while aggregating over a potentially huge amount of data. It enables the extension of functionalities and validation of analysis carried out on different EO data processing platforms.
We're pleased to share our latest work, published in Earth Science Informatics: "OpenEOcubes: an open-source and lightweight R-based RESTful web service for analyzing earth observation data cubes". This work is now accessible as an open-access document
code: https://github.com/PondiB/openeocubes
For a long time, a big constituency of users of PostGIS has been people with large data analytics problems that crush their desktop GIS systems. Or people who similarly find that their geospatial problems are too large to run in R. Or Python. These are data scientists or adjacent people. And when they ran into those problems, the first course of action would be to move the data and parts of th...
OpenEOcubes by Brian Pondi integrates STAC API (using Rstac package), the OpenEO standardized API, and data cubes concepts (using gdalcubes R package) for satellite image analysis.
🌍🛰️📡As we are heading into the holidays, we are looking back at 2023 and our proudest accomplishments:
➡️ https://openeo.cloud/2023/12/19/looking-back-on-2023/
Happy holidays!🎄🎅
Edzer Pebesma
Matthias Mohr
Barry Rowlingson
Planet PostgreSQL (Unofficial)
Jakub Nowosad
openEO_Platform