TorchGeo 0.8 includes 28 new pre-trained model weights and a number of improvements required for better time series support, including a complete rewrite of all GeoDataset and GeoSampler internals, encompassing 8 months of hard work by 23 contributors from around the world.

You may have noticed that https://github.com/microsoft/torchgeo is now https://github.com/torchgeo/torchgeo. This is not an accident!

Note

TorchGeo now belongs to YOU, please join our monthly Technical Steering Committee meetings!

Dados raster estão entre os datasets mais pesados do mundo GIS. Ortomosaicos, imagens de satélite e modelos digitais de elevação frequentemente possuem dezenas ou até centenas de gigabytes.

Historicamente, trabalhar com esses arquivos sempre foi um desafio para profissionais de geotecnologia. Entre os principais problemas estão:

• Leitura lenta de arquivos grandes
• Alto consumo de disco
• Necessidade de armazenamento local
• Dificuldade de uso em ambientes cloud
• Baixa escalabilidade em servidores GIS

Foi nesse cenário que surgiu o Cloud Optimized GeoTIFF (COG).

Hoje o COG é considerado um dos formatos mais importantes para infraestruturas modernas de dados geoespaciais, permitindo trabalhar com rasters gigantes de forma muito mais eficiente.

Neste artigo vamos entender:

Quando trabalhamos com ortomosaicos ou rasters muito grandes, um dos principais desafios é como armazenar e publicar esses dados com boa performance, sem sobrecarregar o servidor GIS.

Uma arquitetura moderna que vem sendo cada vez mais utilizada é baseada em:

• Cloud Optimized GeoTIFF (COG)
• MinIO ou Google Cloud Storage ou Amazon S3 (Object Storage)
• GeoServer

Essa combinação permite que o GeoServer leia diretamente rasters armazenados em object storage, sem precisar copiá-los para o servidor.

Neste post vou mostrar um passo a passo simples e prático para implementar essa arquitetura:

Hoy tengo el placer de impartir el webinar “Geomática hoy: del levantamiento al gemelo digital con datos abiertos, IA y estándares”, organizado por la Delegación Territorial de la Comunidad Valenciana y Región de Murcia del Ilustre Colegio Oficial de Ingeniería Geomática y Topográfica (COIGT).

La idea de la sesión es compartir una reflexión sobre el momento tan interesante que vive la geomática. Nunca antes habíamos tenido a nuestra disposición tantos datos, tanta capacidad de procesamiento y tantas herramientas para transformar la información geoespacial en conocimiento útil para la toma de decisiones.

In recent days, the Roadmap to Accelerate Digital Sovereignty in Spain has been presented, a document that reflects something that has been becoming increasingly clear for some time now: digital sovereignty has ceased to be an abstract concept and has become a strategic issue. We are no longer talking only about innovation, but about decision-making capacity, resilience, technological autonomy, and control over critical infrastructures. Technological dependence is a structural vulnerability.

En los últimos días se ha presentado la Hoja de Ruta para acelerar la soberanía digital en España, un documento que refleja algo que desde hace tiempo resulta cada vez más evidente: la soberanía digital ha dejado de ser un concepto abstracto para convertirse en una cuestión estratégica. Ya no hablamos solo de innovación, sino de capacidad de decisión, resiliencia, autonomía tecnológica y control sobre infraestructuras críticas. La dependencia tecnológica es una vulnerabilidad estructural.

La nueva versión gvSIG Desktop 2.7 incluye la herramienta de Swipe o Cortinilla. Con esta herramienta es posible comparar cartografía diferente de forma sencilla, como por ejemplo dos ortofotos de años diferentes.

El funcionamiento de esta herramienta es a partir de dos vistas diferentes, donde en cada una de ellas se deberá tener una de las capas a comparar. El usuario podrá elegir la opción de realizar cortinilla de forma vertical u horizontal, y se aplicará sobre las capas que se tenga visibles en ese momento en las dos vistas. Una vez seleccionadas ya solo deberá desplazar la barra central horizontalmente o verticalmente para poder ver las diferencias.

En este vídeo te mostramos el funcionamiento:

• Softwares
  • Examind Suite
  • Apache-SIS
• Lire la suite →

The call for papers is now open! We welcome proposals for talks and workshops from all levels of expertise, end users, technical developers, academics, and community contributors alike.

Deadline: 12 April 2026 at 23:59 (Europe/Zurich)

Submit your proposal: https://conference.qgis.org/presenting/

Submissions can cover any topic relevant to the QGIS community, for example:

My previous blog post reviewed the concept of the Hausdorff distance (which more descriptively could be called farthest distance.) Despite its usefulness in matching geometric data, there are surprisingly few open-source implementations, and seemingly no efficient ones for linear and polygonal data.  This even includes CGAL and GRASS, which are usually reliable for provising a wide spectrum of geospatial operations.

TorchGeo 0.7 adds 26 new pre-trained model weights, 33 new datasets, and more powerful trainers, encompassing 7 months of hard work by 20 contributors from around the world.

Note

The following model and dataset descriptions were generated by an imperfect human, not by an LLM. If there are any inaccuracies or anything else you would like to highlight, feel free to reach out to @adamjstewart.

The Hausdorff Distance is a useful spatial function which can appear slightly mysterious. Partly this is due to the name.  It honours Felix Hausdorff, one of the founding fathers of topology, and a polymath who was creative in music and literature as well as mathematics.    

“You can’t always get what you want
But if you try sometimes, well, you might find
You get what you need” (Rolling Stones, You Can’t Always Get What You Want, Album: Let it Bleed, 1969)

I had an idea for a map – where did all of the non native plants in Britain come from? I thought there might be something interesting there if I could find the data and map it.

I called up my new friend Claude and asked for some help with the data “is there a dataset that shows common flower and shrub species found in great britain and their country of origin (possibly with some historical narrative of how they arrived in the uk)?” That took me to the Royal Botanical Gardens Kew’s “BIFloraExplorer” with over 3000 species in Britain and the Botanical Society of Britain and Ireland’s “Plant Atlas 2020” – loads of time saved, I had some authoritative and relevant data in minutes.

Unusually mild weather helped make week five productive. I can feel the benefits as I write this, a week later.

• 11 hours, 37 minutes all training

• 19.3 miles running

• 1,890 ft D+ running

Tuesday I did hill sprints on the "Wallenberg Wall" in my neighborhood. My cadence has increased, and I was a second faster on average. 23 seconds instead of 24. Getting faster is one of my goals, and I'm making measurable progress.

This new version is compatible with QGIS 4 (based on Qt 6 framework). 

Until QGIS 4 is officially released, in order to try the new Semi-Automatic Classification Plugin you can install the prerelease (QGIS 3.99 master).

The NDFF (Nationale Databank Flora & Fauna) holds over 200 million verified observations of plant and animal species from the Netherlands. These data are contributed by volunteers and professionals from hundreds of organizations. Together they contain a wealth of information about the occurrence of plant and animal species in the Netherlands.  Since 2025, the database is accessible to everyone free of charge. That means that one can download all data for 5 × 5 km grid cells (certain restrictions apply). 

Observations are provided as polygons. This post provides an example of how to quickly create summary raster maps showing the total number of observed species. Non-native species are excluded to focus on native biodiversity patterns. Moreover, each record is treated as a single observations regardless of the reported number of individuals. 

What does it mean to speak about European digital infrastructure? It means:

• Use of open standards (such as those promoted by the Open Geospatial Consortium).
• Compliance with regulatory frameworks such as INSPIRE.
• A real possibility of self-hosting.
• Independence from unilateral changes in licenses or terms of use.
• The ability to audit, evolve, and adapt the system over the long term.

Europe has clearly defined its commitment to the digital commons and technological sovereignty.
But that strategy is not materialized through declarations. It is materialized through real infrastructures.

¿Qué significa hablar de infraestructura digital europea? Significa:

• Uso de estándares abiertos (como los promovidos por el Open Geospatial Consortium).
• Cumplimiento de marcos normativos como la directiva INSPIRE.
• Posibilidad real de autohospedaje.
• Independencia frente a cambios unilaterales de licencias o condiciones de uso.
• Capacidad de auditar, evolucionar y adaptar el sistema a largo plazo.

Europa ha definido claramente su apuesta por el bien común digital y la soberanía tecnológica. Pero esa estrategia no se concreta en declaraciones. Se concreta en infraestructuras reales.

We are happy to announce that GeoServer 3 is approaching general availability with a target release date of 15th of April 2026.

This major upgrade modernises the platform’s foundation with the migration to Spring 7 and JDK 17, brings a refreshed user experience and replaces legacy image-processing components with ImageN to deliver significantly improved raster performance and maintainability. The release aligns GeoServer with current Java ecosystems, strengthens security and vulnerability management, and simplifies cloud-native deployments. You can read more about the GeoServer 3 initiative on this page.

No maps in this post!

I mentor a couple of people and I was about to start a new project this week. I was going to knock up a Google Sheet or Doc to track goals, talks and meetings and share with the mentee and my co-mentor. It occurred to me that I could vibe code an app to track the mentee’s progress and to provide structure to our regular meetings.

With a lot of help from Claude I’ve got something running that does that for a few mentors and mentees, it runs on desktop and mobile and quite a lot of the input is voice to text which is particularly useful in mobile. It took about 3 hours, to build test and deploy.

I am sure that I’ll find some problems or feature gaps in the future but that doesn’t matter in my use case. I don’t plan to build this into a business or to scale it up to thousands of users it’s just a tool that makes record keeping for me and my mentees easier.

IOSACal 0.7 was released yesterday. Here is a quick summary of what’s new.

This long cycle was mostly about documentation improvements and some maintenance tasks, the boring but essential work that keeps the project going.

Version 0.7 is already available in PyPI and conda-forge. There is an updated version record at Zenodo.

All changes were contributed by Stefano Costa.

Die FOSSGIS-Konferenz 2026 findet vom 25.-28. März 2026 in Göttingen und Online statt. Es sind nur noch wenige Wochen bis zur Konferenz. Die Vorfreude wächst stetig und die Vorbereitungen laufen auf Hochtouren!

Die Konferenz wird vom gemeinnützigen FOSSGIS e.V, der OpenStreetMap Community in Kooperation mit dem Geographischen Institut der Georg-August-Universität Göttingen organisiert und findet auf dem Campus der Uni Göttingen statt.

Auch in diesem Jahr zeichnet sich ein großes Interesse an der Konferenz ab. Die Anmeldungen steigen von Woche zu Woche. Zum Glück bietet das Zentrale Hörsaalgebäude der Uni Göttingen ausreichend Platz, so dass es die bisher größte FOSSGIS-Konferenz werden könnte.

• Added optional dependency Pandas for performance improvement in tabular data.
• In tool “Band classification” added option for using PyTorch pretrained model. In case a pretrained model is selected, and additional algorithm is selected for classification, using the same parameters of the named algorithm (e.g. random forest); after executing the pretrained model, the additional algorithm is executed on the embeddings for classification. Currently, it works with models pretrained by the Allen Institute for Artificial Intelligence (SatlasPretrain: https://satlas-pretrain.allen.ai) in particular, Sentinel-2 swin-v2-base single-image multispectral and swin-v...

TorchGeo 0.9 includes 13 new datasets and a number of improvements required for better time series support, encompassing 3 months of hard work by 15 contributors from around the world. We are now trying to make more frequent releases to get exciting new features out to users as quickly as possible!

Throughout the first month of 2026, our ninjas added a nifty set of improvements to three useful plugins that we love: the GeoMapFish Search, OpenStreetMap Nominatim Search, and OSRM Routing plugins. These plugins all provide genuinely useful functionalities as well as being great showcases of how easy it is to integrate QField with online REST endpoints.

All three plugins have been updated to ship with useful endpoint presets out of the box, and users also have the option to configure their own custom endpoints for more flexibility. To configure endpoints, open QField’s settings panel, navigate to the plugin manager, and click the settings button next to the desired plugin.

Se você trabalha com GIS/QGIS e ainda depende de cliques, processos manuais e retrabalho, chegou a hora de mudar isso.

Estão abertas as inscrições para o Curso Python com GIS do Zero, uma formação técnica e prática para quem quer sair do modo operacional e começar a programar o geoprocessamento de verdade.

Aqui você aprende Python aplicado ao GIS, não Python genérico.

O que você vai dominar:

Python do zero com foco técnico
Pandas e Geopandas aplicados ao GIS
Processamento de vetores e rasters (GDAL/OGR)
Análises espaciais e estatísticas zonais
SQL Espacial moderno com DuckDB Spatial
Automação de processos no QGIS
Introdução ao desenvolvimento de plugins

Tudo aplicado a problemas reais, como acontece no mercado.

Para quem é este curso?

Click on the image above to view Dead Reckoning v2

A couple of weeks ago I finished my Dead Reckoning V1 app and I was pretty pleased with it, all the gigs that the Dead had played summarised and displayed along with set lists, sound tracks and more. I released it just before Geomob and got some good feedback so a few days later I decided to publicise it to the Grateful Dead community on Reddit, then the storm broke!

La nueva versión 2.7 de gvSIG Desktop incluye una nueva calculadora de campos, cuyas principales ventajas son, por un lado, que se pueden rellenar varios campos a la vez, y por otro, que ha mejorado significativamente el rendimiento cuando se pretenden rellenar muchos registros a la vez, al no cargarse en memoria todos ellos sino por bloques.

La nueva calculadora se basa en la herramienta de actualizar tabla que se había incluido en la versión anterior, donde se había mejorado la parte de rellenado de registros. Ahora dispone de una pestaña “Opciones” donde se puede seleccionar el número de registros en los que se quiere que termine edición y reinicie. De esa forma no carga todos los datos en memoria y no bloquea el equipo, evitando así también posibles pérdidas de datos.