Speeding up methane detection with AI
Telops/Exosens brings Mondrian in to cut the manual analysis load on its hyperspectral imagery — without compromising detection.
Telops/Exosens is a global leader in hyperspectral imaging systems and infrared cameras, with its airborne hyperspectral camera used to detect methane leaks at industrial sites, among other applications. Every mission generates thousands of images to review — a volume that was becoming unmanageable against fast-growing client demand. Mondrian built an AI tool that automatically sorts these images, cutting the manual analysis workload before a technician ever gets involved. A working proof of concept was ready in six weeks; five months later, Telops had a deployment-ready tool.
A pioneer in environmental detection since 2000
Infrared & hyperspectral imaging
+100 employees, 40 in R&D
Québec, Canada, with a strong international presence
Telops/Exosens designs and manufactures hyperspectral imaging systems and high-performance infrared cameras for defense, industry, and academic research. Its flagship product, the Hyper-Cam, is suitable for numerous applications, such as the detection and quantification of gas leaks from the air in its airborne version. The tool has become indispensable for large-scale industrial environmental inspections.
Thousands of images per mission and a bottleneck that's holding back growth
After each inspection flight, the Telops teams were left with thousands of images to review. Despite existing filtering tools, hundreds still required an expert’s eye, and demand continued to grow.
This bottleneck directly limited Telops' ability to take on new contracts. Growth meant hiring on a large scale and training highly specialized staff—a challenge to manage in a rapidly expanding market.
The images produced by the Hyper-Cam are very different from ordinary images: analyzing them requires specialized scientific expertise. The real challenge was to design a tool capable of mimicking this expert judgment without ever missing a genuine leak.
A smart filter that mimics expert judgment
Mondrian developed an AI model that automatically analyzes each image and determines whether it warrants a technician’s attention. Images that pose no risk are filtered out immediately; only those that might contain a leak end up on an expert’s desk. The solution was delivered as a turnkey system, complete with everything Telops needs to operate and improve it independently.
From the very beginning, Telops’ experts were at the heart of the project. Their practical knowledge directly guided the technical decisions, which allowed us to move quickly and deliver something useful within the first few weeks, rather than a prototype disconnected from reality.
| Component | Role in the solution |
|---|---|
| Computer vision model | Automated image triage |
| Training on real-world data | Designed to meet the specific requirements of Telops cameras |
Less manual analysis. Same reliability. Greater capacity.
| Indicator | Result | Business impact |
|---|---|---|
| Manual analysis volume | Up to 30% of images automatically screened out | Fewer specialized labor hours per mission |
| Reliability | No actual leaks were found in the best datasets | Service quality preserved |
Facing a similar challenge in your industry?
A 30-minute call is enough to find out where AI can create value for you.
Book a call