In this blog I share list a of journals, conferences and workshops where you can find or publish Machine Learning for Earth Observation (ML4EO) and Geospatial Artificial Intelligence (GeoAI) papers.
ML vs EO venues
ML4EO and GeoAI are very interdisciplinary fields, inheriting from computer science and Earth Sciences. However, we still don’t have “our own journal” yet. As an ML4EO researcher, you therefore have two options:
Publish in a computer science venue
Publish in an Earth Observation venue
It’s important to know that computer scientists often publish at conferences or workshops. Most computer science conferences and some workshops have proceedings that are published with regular publishers like Springer, ACM, and IEEE. Conference papers have short page limits (8-15 pages). Sometimes, people publish two versions of the same work: first a short conference paper, and later a longer journal paper with extended validation and experiments. In most other fields, including Earth Observation, you publish in journals and only submit abstracts to conferences and workshops, and when you’re accepted you get a talk at that conference. Computer scientists still give talks at conferences, but it is rare that it’s not tied to a publication!
A note on computer science conferences
Paper structure
Paper-writing as well as research design conventions vary between the fields:
Computer science papers focus more on methodology and experimental details, Earth Observation papers more on the data and application analysis.
Computer science papers are more likely to use existing benchmarks like BigEarthNet. Earth Observation papers often gather their own datasets.
Code and data availability: almost always shared in CS, not always available with EO papers
Finally, ML and EO papers often have different structures. Here are some short descriptions.
Computer science paper structure
List contributions in a bulleted list at the end of the introduction
Only cite enough related works in the introduction to give a brief background and highlight the knowledge gap. Remaining related works are discussed in a separate section directly after the introduction.
Before or at the start of the methodology, there is often a problem statement that describes the problem and objective in formal /mathematical terms.
There is a detailed experimental details section that describes the complete evaluation set up, including details like how many runs per model, which baseline algorithms you used, and even which hardware you used. The data is often discussed as part of the evaluation section (because data is used to validate the model).
There are no separate results and discussion sections. Results are interpreted where they are presented. The results section is organised per main finding, and focuses on direct implications of each finding.
Earth Observation paper structure
Contributions are mentioned in the text of the introduction, not listed as bullet points.
The related work is part of the introduction, there is no separate section.
The methods and data can be strongly intertwined and described in a single section.
There is not always an experimental details section (though most papers with ML are adopting that)
The results/discussion section can be significantly longer than in computer science papers, for example because the model is applied to a real-life use case and analysed in-depth.
List of venues
I’ve compiled a list of venues from:
Papers I saved in Zotero
Papers I cite in my papers
List of papers students can use for research projects in the university course I teach (Urban Computing)
Known ML4EO researchers publish there (e.g.: Gustau Camps-Valls, Begum Demir, Damien Borth, ESA Phi-Lab members)
For each venue, I give details like any limits on pages, the publishing model and I summarise the scope based on papers I’ve seen.
Page limit: None, but you need to pay 230$ per page for every page over 6 (including references and appendices)
Scope: this is one of the most popular venues for ML4EO papers. It’s a technical journal that borrows many CS papers like the structure and usually data and code are shared.
Scope: Technical journal on machine learning, focusing on methodological contributions. Has some examples of GeoAI papers with general scopes, including datasets.
Scope: A journal on applications of remote sensing, with some technical papers but mostly an application focus. There are some good GeoAI papers in here, but you may not always find the experimental details you expect in CS papers. I use this journal to get an idea of the tasks people are working on.
Scope: This is an Earth Observation conference where people present a lot of small projects or intermediate work, including Machine Learning. You’ll find more applications here than in ML venues.
Scope: very technical conference on computer vision. Focuses on methodological contributions that generalise to other datasets/cases. Applications can be accepted if they are very interesting/specific. This is where you’ll find a lot of geospatial foundation models and general vision models. Competitive and hard to publish at!
Scope: technical conference on deep learning. You can find general/broad-scope EO models here, like foundation models. Competitive and hard to publish at!
Scope: workshop at CVPR focused on computer vision for EO
WACV often has a workshop on GeoAI: for instance, CV4EO in 2025
Final tips
The journals and conferences I listed above are great places to search for methods and datasets for your specific ML4EO or GeoAI problem. However, many papers are technical and focus on specific tasks or applications – which is pretty overwhelming if you’re just
starting. So if you want a more gentle introduction, there is one category I didn’t mention: high-level and vision or position papers on EO. These can be found in prestigious venues like Nature journals, IEEE Geoscience and Remote Sensing Letters and IEEE Geoscience and Remote Sensing Magazine. It’s either very difficult to publish at such journals, or you cannot even submit yourself and have to be invited. However, these are great sources for papers written by experts like Begum Demir, Devis Tuia, Xiaoxiang Zhu and Gustau Camps-Valls. These papers do not give a lot of methodological details, but can help you find relevant research problems and are great examples of how to write papers :).
Finally, please reach out if you have any corrections or venues to add.