The issue is straightforward. A machine learning model trained to classify agricultural crops using Sentinel-2 satellite imagery will perform well over farmland. But apply that same model to an area of open water, and it will generate nonsensical outputs — confidently labelling ocean pixels as wheat fields or vineyards. For an experienced GIS analyst, this kind of error is easy to spot and correct. But in automated workflows where no human is reviewing intermediate results, these failures can propagate silently through entire analysis pipelines.

This is not a theoretical concern. Research has shown that even well-regarded models like BigEarthNet, trained on Sentinel-1 and Sentinel-2 data, can swing from over 85 percent accuracy in optimal conditions to as low as 20 percent in unfavourable scenarios. The gap between best-case and worst-case performance is enormous, and most users have no way of knowing which end of that spectrum they are operating at for any given query.

The documentation problem

Compounding this reliability issue is a documentation gap. Models published on platforms like HuggingFace and Kaggle are often poorly documented — at least not in a machine-readable format that a processing platform could use to automatically validate whether a model is appropriate for a given dataset and region. In practice, this means users need to manually inspect model specifications, preprocess input data with custom Python scripts, and make judgement calls about applicability. That effectively limits the use of these models to specialists with both domain expertise and programming skills.

For geospatial AI to scale beyond expert users, platforms need to handle this validation automatically.

Model fencing as a solution

A research collaboration between Constructor University and rasdaman GmbH in Bremen, Germany, funded by the EU’s EFRE programme, is working on exactly this problem. The project, called FAIRgeo, introduces a concept called “model fencing” — automatically restricting AI model inference to the spatial, temporal, and thematic contexts where reliable results can be expected.

The approach works by enriching model metadata with machine-readable information about where and when a model is valid. When a user submits a query, the platform checks parameters automatically before execution: correct satellite source, correct spectral bands, appropriate patch size, and geographic applicability. If the model is being asked to operate outside its validated comfort zone, the system can flag the issue or prevent execution entirely.

Early results are promising on the usability front as well. What typically requires over a hundred lines of Python code can be reduced to a two-line datacube query, with the platform handling data selection, preparation, and tiling automatically. Performance benchmarks also show the integrated approach running faster than traditional Python implementations in most cases.

Why this matters beyond research

As AI becomes embedded in operational geospatial workflows — from agricultural monitoring to urban planning to climate risk assessment — the consequences of unreliable model outputs grow more serious. Decisions about land use, disaster response, and infrastructure investment increasingly depend on automated analysis of satellite data.

The geospatial industry needs standardised approaches to model validation and applicability metadata. Efforts like FAIRgeo, which is contributing its findings to OGC working groups on data quality and coverage standards, point toward a future where AI on Earth observation data is not just more powerful, but meaningfully safer.

The post Why AI Models Need Guardrails When Applied to Earth Observation Data appeared first on Spatial Tech.

The post Why AI Models Need Guardrails When Applied to Earth Observation Data appeared first on Spatial Tech.

Enterprise campus networks are under pressure. The combination of growing bandwidth demands, increasing device density, operational technology integration, and sustainability targets is forcing organisations to rethink how their physical network infrastructure is built. Nokia and Alcatel-Lucent Enterprise are responding with a deepened strategic alliance that combines optical fibre technology with secure campus networking — targeting the kinds of environments where connectivity failures have real operational consequences.

The partnership, now in its fifth year, integrates Nokia’s Optical LAN fibre infrastructure with ALE’s enterprise networking solutions for in-building and campus connectivity. The result is a converged fibre-based architecture capable of carrying multi-gigabit data speeds across complex facilities while reducing energy consumption and total cost of ownership compared to traditional copper-based network designs.

Why fibre is gaining ground in campus environments

Most enterprise campus networks still run on copper cabling for the last segment of connectivity — from switch closets to end devices. This architecture has served well for decades, but it is increasingly strained by the demands of modern campus operations. IoT sensor networks, high-density WiFi, CCTV systems, building management platforms, and operational technology applications all compete for bandwidth on infrastructure that was designed for a simpler era.

Fibre-to-the-edge architectures eliminate many of these constraints. Optical fibre supports significantly higher bandwidth over longer distances, requires fewer intermediate network layers, and consumes less energy than equivalent copper deployments. For large campus environments — hospitals, resorts, logistics facilities, transport hubs — the reduction in physical infrastructure also translates to meaningful space savings.

The Nokia-ALE approach consolidates what would traditionally be separate network layers into a single fibre backbone. At Ikos Resorts in Greece, for example, the combined solution runs guest WiFi, CCTV, voice communications, and building safety sensors through one converged high-availability architecture — replacing what would previously have required multiple parallel network infrastructures.

The operational technology angle

What makes this partnership particularly relevant for critical infrastructure is the operational technology integration layer. Modern logistics facilities, manufacturing plants, and transport networks increasingly depend on automated systems that require deterministic, low-latency connectivity. Automated warehouse systems, robotic material handling, real-time asset tracking, and industrial control systems all need network infrastructure that is not just fast but reliably available.

ALE’s contribution to the partnership includes automated device onboarding, asset discovery and classification, virtual network segmentation, and continuous monitoring — capabilities designed to handle the complexity of environments where hundreds or thousands of connected devices need to be securely managed without manual intervention. Virtual segmentation is particularly important in mixed-use environments where IT traffic and operational technology traffic need to coexist on the same physical infrastructure without interfering with each other.

Deployment track record

The partnership has now been deployed across more than 100 enterprises globally, spanning hospitality, healthcare, transport, and logistics. Notable projects include Grand Paris Express, Montreal Railways, Pantai Jerudong Hospital in Brunei, and Wembley Park in the UK — all environments where network reliability is not optional and where the consequences of downtime extend beyond inconvenience into safety and operational risk.

What this signals for enterprise networking

The broader trend is clear: enterprise campus networks are converging. The era of separate infrastructure for IT, OT, building management, and security systems is giving way to unified fibre-based architectures that carry everything on a single physical layer. This convergence is driven by economics — fewer network layers means lower cost — but also by operational necessity. Managing five separate network infrastructures across a large campus is unsustainable as device counts and bandwidth demands continue to grow.

For organisations in logistics, healthcare, manufacturing, and transport — sectors where both connectivity and physical infrastructure intersect — the Nokia-ALE model represents the direction enterprise networking is heading: fewer layers, more bandwidth, lower energy consumption, and a single converged platform capable of supporting both IT and operational technology workloads.

The post Nokia and Alcatel-Lucent Enterprise Push Fibre-Based Networks Into Critical Infrastructure appeared first on Spatial Tech.

The post Nokia and Alcatel-Lucent Enterprise Push Fibre-Based Networks Into Critical Infrastructure appeared first on Spatial Tech.