OSGeo Statement: EU Cyber Resilience Act

QGIS Planet

GeoAI: key developments & insights

It’s been a while since my post on geo and the AI hype in 2019. Back then, I didn’t use the term “GeoAI”, even though it has certainly been around for a while (including, e.g., with dedicated SIGSPATIAL workshops since 2017).

GeoAI isn’t one single thing. It’s an umbrella term, including: “AI for Geo” (using AI methods in Geo, e.g. deep learning for object recognition in remote sensing images) and “Geo for AI” (integrating geographic concepts into AI models, e.g. by building spatially explicit models). [Zhang 2020] [Li et al. 2024]

Today’s post is a collection of key GeoAI developments I’m aware of. If I missed anything you are excited about, please let me know here in the comments or over on Mastodon.

Background

A week ago, I had the pleasure to attend a “Specialist Meeting” on GeoAI here in Vienna, meeting over 40 researchers from around the world, from Master students to professor emeritus. Huge props to Jano (Prof. Krzysztof Janowicz) and his team at Uni Wien for bringing this awesome group of people together.

The elephant in the room: LLMs

Unsurprisingly, LLMs and the claims they make about geography are a mayor issue due to the mistakes they make and the biases behind them. An infamous example is AI’s issue with understanding topology:

Image source: Janowicz, K. (2023). Philosophical Foundations of GeoAI: Exploring Sustainability, Diversity, and Bias in GeoAI and Spatial Data Science. arXiv e-prints, arXiv-2304.

Even if recent versions of ChatGPT (such as GTP 4o) do a better job with this specific example, this doesn’t make their answers reliable. So between the trustworthiness, reproducibility, explainability, and sustainability issues … LLMs have a long way to go. And it’s not clear whether they are going in the right direction right now.

Geospatial foundation models

Prithvi, a model developed by NASA, IBM, et al. in 2023, is one of the first geospatial foundation models. Like much of GeoAI, Prithvi deals with remote sensing data. Specifically, it is trained on Landsat and Sentinel-2 (HLS) imagery, with applications in flood mapping and wildfire prediction. And maybe best of all: the model is open-source and publicly available.

Spatiotemporal machine learning model specifications

In the general AI community, model cards have become a common way to share information about models. However, identifying the right model for spatiotemporal tasks is hard since there are no standardized descriptions in existing model catalogs (e.g. Hugging Face, DLHub or MLFlow). To address this issue, [Charette-Migneault et al. 2024] have proposed the Machine Learning Model (MLM) extension for the SpatioTemporal Asset Catalogs (STAC). But, yet again, this development is targeting models trained with remote sensing imagery.

Similarly, the OGC Training Data Markup Language for Artificial Intelligence (TrainingDML-AI) and its ISO equivalent are limited to EO as well …

Spatial knowledge graphs

For those among us working mostly with vector data, the KnowWhereGraph is an interesting development. It’s the first geo-enriched knowledge graph [Janowicz et al. 2022] that helps answer geospatial questions by integrating a variety of spatial datasets through hierarchical grids, standard region boundaries and appropriate ontology and knowledge graph schema development. However, so far, the KnowWhereGraph is mostly limited to the United States.

Explainable AI (XAI) and geo

While answers from knowledge graphs are intrinsically explainable, many other (Geo)AI solutions are built on AI approaches that result in black box models.

Graph neural networks (GNNs) have become very popular in GeoAI (including in urban analytics and mobility [Jalali et al. 2023] [Liu et al. 2024]) but their black box nature limits their practical usefulness in domains where transparency and trustworthiness are crucial. To offer insights into how model predictions are made, [Liu et al. 2024] propose a spatially explicit GeoAI-based method that combines a graph convolutional network and a graph-based XAI method, called GNNExplainer to explore the correlation between urban objects.

Reproducibility et al.

The AI hype in geo is still going strong. Journals are being flooded with paper submissions and good reviewers are hard to come by. In many geo-related venues, it is still acceptable to present an AI paper without making code or model available. (We recently discussed this issue for mobility AI specifically [Graser et al. 2024].)

I’m convinced we can and should do better: quality over quantity, moving steadily, building and fixing things.

References

Learn More

Plugin Update – October, 2024

In the last month of October, 18 new plugins were published in the QGIS plugin repository.

Here follows the quick overview in reverse chronological order. If any of the names or short descriptions catches your attention, you can find the direct link to the plugin page in the table below:

ECLAIR: Emission CompiLation for AIR quality
This plugin compiles emission data for air quality. Data can be imported, edited and exported.
GRD_Loader
Load GRD Format Rasters.
Random Point on Lines…
The Random Points on Lines… is a simple interface and a user-friendly QGIS plugin that enables to generate a specified number of random points along selected line layers within QGIS. Users can control the layer selection and the number of points generated.
BGT Loader
A processing tool to download Dutch BGT data for a specific polygon area.
geonorge-tegneregelassistent
En plugin for å implementere stiler/tegneregler basert på norske standarder som finnes i Geonorge.
EvapoGIS
Evapotranspiração.
SkyDeck Plugin
Seamlessly Integrate and Manage SkyDeck Geospatial Data within QGIS.
Curva de Nivel
Cria curvas de nivel no territorio brasileiro.
Reservoir & Basin Analysis
This plugin offers some analysis tools for reservoirs and basins.
Feature Transfer Tool
Feature Transfer Tool provides a seamless way to copy and paste features between layers.
Parameter History
A better processing history plugin.
GeOSPR
GeOSPR (Consulta, Validación y estandarización).
AutonomousGIS-SpatialAnalysisAgent
The Spatial Analysis Agent is a user-friendly plugin that serves as a “Copilot” in QGIS software. This Copilot allows users to perform geospatial analysis directly within QGIS using natural language queries, making it accessible for both experts and beginners. The plugin leverages the full potential of over 600 QGIS processing tools, and other external tools such as Python libraries (e.g., Geopandas, seaborn, etc.). Whether working with vector data, raster analysis, the Spatial Analysis Agent offers a flexible, AI-driven approach to enhance and automate GIS workflows.
Aerodrome Utilities
Fetches OSM Data and processes it for aerodroms with various algorithms.
Relation Manager
This plugin helps in the management of 1:N project relations.
Easy Filter and Selection
Plugin gives easy selecting and filter feature for users that don’t want to write complicated SQL for simple problem solution.
GeoPF Altimétrie
Warning: France only! <br/> This plugin allows to call IGN Geoplateforme API directly from elevation profile tool.
Pian Exporter
The plugin exports the vector layer in WKT format for PIAN.

Learn More

How to contribute to GRASS GIS development

How to contribute to GRASS GIS development: Guidance for new developers in the GRASS GIS Project.

The post How to contribute to GRASS GIS development appeared first on Markus Neteler Consulting.

Learn More

[Changelog] Measurement tool is now available in the app

We’ve added measurement tools to Mergin Maps, allowing you to measure both distance and area directly within the app. The new "Measure" action is accessible from the "more" menu, where you can easily add points to calculate lengths and areas in your projects. The process is similar to creating a line feature, with a live display of the measured distance and options to undo, add points, or complete the measurement. Measurements are following your QGIS distance and area project units.

Read more about the measurement tool in our documentation: https://merginmaps.com/docs/field/measure/

Learn More

3.4.4 - Ebo

What's Changed

  • Improve handling of captured photo's ratio and resolution
  • Improve feature attribute pasting
  • Optimize addition of child feature in a parent feature being edited
  • Guard from potential crash when loading a project that requires authentication credentials
  • Fix large files constantly being re-downloaded when synchronizing cloud projects
Learn More

[Changelog] Feature identification in Maps [Beta]

You can now preview your project data directly from the project details page—no need to open QGIS or the mobile app!

Maps in dashboard are currently in beta, and we’re more than happy to hear your feedback. Share your thoughts and help us improve it here: https://wishlist.merginmaps.com/p/let-s-get-maps-in-dashboard-out-of-beta.

Learn More

(Fr) Du nouveau pour [CityBuilder] CityForge

Sorry, this entry is only available in French.

Learn More

New release for QField : 3.4 &#8220;Ebo&#8221;

Oslandia is the main partner of OPENGIS.ch around QField. We are proud today to forward the announcement of the new QField release 3.4 “Ebo”.

Main highlights

A new geofencing framework has landed, enabling users to configure QField behaviors in relation to geofenced areas and user positioning. Geofenced areas are defined at the project-level and shaped by polygons from a chosen vector layer. The three available geofencing behaviours in this new release are:

  • Alert user when inside an area polygon;
  • Alert user when outside all defined area polygons and
  • Inform the user when entering and leaving an area polygons.

In addition to being alerted or informed, users can also prevent digitizing of features when being alerted by the first or second behaviour. The configuration of this functionality is done in QGIS using QFieldSync.

Pro tip: geofencing settings are embedded within projects, which means it is easy to deploy these constraints to a team of field workers through QFieldCloud. Thanks Terrex Seismic for sponsoring this functionality.

QField now offers users access to a brand new processing toolbox containing over a dozen algorithms for manipulating digitized geometries directly in the field. As with many parts of QField, this feature relies on QGIS’ core library, namely its processing framework and the numerous, well-maintained algorithms it comes with.

The algorithms exposed in QField unlock many useful functionalities for refining geometries, including orthogonalization, smoothing, buffering, rotation, affine transformation, etc. As users configure algorithms’ parameters, a grey preview of the output will be visible as an overlay on top of the map canvas.

To reach the processing toolbox in QField, select one or more features by long-pressing on them in the features list, open the 3-dot menu and click on the process selected feature(s) action. Are you excited about this one? Send your thanks to the National Land Survey of Finland, who’s support made this a reality.

QField’s camera has gained support for customized ratio and resolution of photos, as well as the ability to stamp details – date and time as well as location details – onto captured photos. In fact, QField’s own camera has received so much attention in the last few releases that it was decided to make it the default one. On supported platforms, users can switch to their OS camera by disabling the native camera option found at the bottom of the QField settings’ general tab.

Wait, there’s more

There are plenty more improvements packed into this release from project variables editing using a revamped variables editor through to integration of QField documentation help in the search bar and the ability to search cloud project lists. Read the full 3.4 changelog to know more, and enjoy the release!

 

Contact us !

A question concerning QField ? Interested in QField deployment ? Do not hesitate to contact Oslandia to discuss your project !

 

Learn More

QField 3.4 “Ebo”: Geofencing and processing out of the box

QField 3.4 is out, and it won’t disappoint. It has tons of new features that continue to push the limits of what users can do in the field.

Main highlights


A new geofencing framework has landed, enabling users to configure QField behaviors in relation to geofenced areas and user positioning. Geofenced areas are defined at the project-level and shaped by polygons from a chosen vector layer. The three available geofencing behaviours in this new release are:

  • Alert user when inside an area polygon;
  • Alert user when outside all defined area polygons and
  • Inform the user when entering and leaving an area polygons.

In addition to being alerted or informed, users can also prevent digitizing of features when being alerted by the first or second behaviour. The configuration of this functionality is done in QGIS using QFieldSync.

Pro tip: geofencing settings are embedded within projects, which means it is easy to deploy these constraints to a team of field workers through QFieldCloud. Thanks Terrex Seismic for sponsoring this functionality.

QField now offers users access to a brand new processing toolbox containing over a dozen algorithms for manipulating digitized geometries directly in the field. As with many parts of QField, this feature relies on QGIS’ core library, namely its processing framework and the numerous, well-maintained algorithms it comes with.

The algorithms exposed in QField unlock many useful functionalities for refining geometries, including orthogonalization, smoothing, buffering, rotation, affine transformation, etc. As users configure algorithms’ parameters, a grey preview of the output will be visible as an overlay on top of the map canvas.

To reach the processing toolbox in QField, select one or more features by long-pressing on them in the features list, open the 3-dot menu and click on the process selected feature(s) action. Are you excited about this one? Send your thanks to the National Land Survey of Finland, who’s support made this a reality.

QField’s camera has gained support for customized ratio and resolution of photos, as well as the ability to stamp details – date and time as well as location details – onto captured photos. In fact, QField’s own camera has received so much attention in the last few releases that we have decided to make it the default one. On supported platforms, users can switch to their OS camera by disabling the native camera option found at the bottom of the QField settings’ general tab.

Wait, there’s more

There are plenty more improvements packed into this release from project variables editing using a revamped variables editor through to integration of QField documentation help in the search bar and the ability to search cloud project lists. Read the full 3.4 changelog to know more, and enjoy the release!

Learn More

Plugin Update – September, 2024

In September a total of 20 new plugins were published in the QGIS plugin repository.

Highlight

In the last month some AI-related plugins became available for users, namely IntelliGeo and TreeEyed, which in addition to the increasing number of tools, greatly contribute for the adaptation of QGIS to current and future needs, showcasing it as one of the best options for beginners and experts alike to conduct a number of geospatial analyses.

As stated by their authors, with IntelliGeo there’s a chat interface where users can detail their requests, and the output is either a PyQGIS code or a graphical processing model, which can in turn be executed directly in QGIS.

As for the TreeEyed plugin, its main objective is the monitoring of trees by generating vector and raster datasets from high resolution RGB imagery.

Overview

Here follows the quick overview in reverse chronological order. If any of the names or short descriptions catches your attention, you can find the direct link to the plugin page in the table below:

Project Setup
Sets up a QGIS project to my personal specs.
Continuous Network Analysis (Processing)
Processing plugin that adds several scripts to assist in decision making and validation of line-type vector networks by generating inconsistencies, further expanding the “Network Analysis” tool.
Oslandia
Official plugin for Oslandia customers.
Reach
Enables the use of real transit time as a spatial predicate for selects and joins.
Data Clock
Polar plot of seasonal data.
QGIS Light
QGIS made simple – a light user interface for core GIS functions.
Attribute Searcher
A minimalistic plugin to search for values in attributes quick and easy.
Topo Maps
微地形図の生成。Generate Topographic Maps.
IntelliGeo
IntelliGeo is QGIS plugin that facilitates interaction with Large Language Models in QGIS environment.
CartAGen
Cartographic generalization.
RiverBankErosionAndAccretion
This plugin calculates the erosion and accretion along a river’s course.
VectorStats
Plugin for descriptive and statistical analysis of vectors, with chart generation.
Jilin1Info(2023)
2023年全国50cm吉林一号影像拍摄日期查询
Historique Parcelle
Historique des parcelles (cadastre français).
String Writer
Writes QGIS layers to Surpac string file format.
TreeEyed
TreeEyed is a QGIS plugin for tree monitoring using AI.
SHP Buddy
Quickly create shapefiles for breeding experiments.
TEKSI Wastewater
TEKSI Wastewater plugin to manage wastewater networks.
KGR Finder
By simply drawing polygons or clicking on existing polygons, this extension makes it possible to download data from OpenStreetMap (OSM) or the iDAI.gazetteer and display it on the map, including all existing attributes. The plugin is designed so that other services can also be easily integrated in the future.
layer_style_master
This QGIS plugin copies symbology, labels, and rendering settings from one layer to multiple other layers.

Learn More