Hyperspectral camera for condition assessment of painted metallic towers

Why this project

Metallic towers, one of the main components of overhead lines, rust over time. Despite applying paint roughly every 17 years to inhibit the process, corrosion continues and therefore impacts the stability of the towers. It is critical to smooth the replacement by detecting the level of corrosion of the towers as one of the main deterioration factors. Classifying corrosion is not trivial and requires detailed climbing inspection.

Approach

Back in 2021 we started technology market research and were looking to answer the below question:

Is there any technology out there that can detect rust behind the painting in a non-destructive way? 

We were particularly keen to find a tool that would be able to detect corrosion that had developed in metallic towers including corrosion that might be hidden by a layer of paint. Furthermore, we wanted to investigate whether there might be some form of technology that could accurately classify the severity of rust patches into different types (e.g. superficial corrosion, medium or severe corrosion).

Investigations led us to focus on hyperspectral camera which captures reflections from across the electromagnetic spectrum, using this information to identify and classify the chemical properties of the material or object under investigation. This contrasts with red-green-blue (RGB) cameras, which only capture the size, shape and visible colors of objects.

Hyperspectral cameras have already been explored and successfully used in several sectors, including in agriculture, the oil and gas industry, the shipping industry and marine archaeology (to name a few). It therefore looked promising to our teams.

Results

Following the initial laboratory research, the project evolved through multiple stages to progressively evaluate the applicability of hyperspectral imaging for the condition assessment of painted metallic lattice towers in increasingly realistic scenarios.

Between 2023 and 2025, Work Package 3 (WP3) focused on extending the approach towards a fully integrated, drone based inspection workflow and validating it on metallic lattice towers in outdoor conditions. This phase addressed both hardware integration and the development of an automated software pipeline, covering data acquisition, image alignment, 3D reconstruction, and corrosion detection and classification. WP3 demonstrated that hyperspectral imaging can detect surface corrosion and can help reduce false positives compared to conventional RGB imaging, particularly by leveraging spectral information in the NIR range. At the same time, it highlighted major challenges related to complex tower geometry, varying illumination conditions, shadows, limited spatial resolution, and the lack of reliable white referencing in outdoor drone operations. Despite substantial improvements to the processing pipeline, the target performance level—specifically an accuracy of 86% for severe corrosion detection—could not be consistently achieved, limiting the robustness and repeatability of the results in real world conditions.

Building on the learnings from WP3, Elia subsequently launched an extended feasibility study in 2025–2026 to further assess the operational applicability of the technology. This study was initiated following a dedicated workshop bringing together internal experts and external stakeholders, which confirmed the relevance of further exploration while also clarifying the key technical uncertainties to be addressed. To accelerate development and ensure a broad and objective evaluation, Elia deliberately involved multiple external partners with complementary expertise. The hardware track was developed in collaboration with imec and Quicksand, focusing on camera integration, embedded processing, and suitability for potential drone based deployment. In parallel, the software track was carried out together with APIXA and imec, addressing data processing pipelines, image alignment, spectral analysis, and robustness under dynamic outdoor conditions. This multi partner approach enabled rapid iteration, cross validation of results, and a comprehensive assessment of both the potential and the limitations of the technology.

The feasibility study confirmed that, while hyperspectral imaging shows promising capabilities in controlled environments and as a complementary source of information, its performance and repeatability remain challenging in real world outdoor conditions. In particular, sensitivity to environmental factors and the difficulty of achieving stable data normalisation currently prevent its deployment as a standalone solution for large scale operational inspections.

Based on these conclusions, Elia has decided to conclude the active development of this specific application. The knowledge and experience gained through the project will be retained, and technological developments in this domain will continue to be monitored by Elia’s Innovation department. In parallel, Elia will refocus on clearly defining the core inspection and decision making needs for lattice towers, in order to assess how remote inspection technologies can best contribute to a long term reduction of physical climbing inspections. Elia’s ambition to reduce climbing inspections through the use of remote inspection techniques remains unchanged and is formalised in an OKR targeting an 80% reduction in physical climbing inspections by 2029.

Partners