As manufacturing expands into perception systems and edge computing, moviTHERM has launched the Irley IoT thermal camera—a practical example of integrating smart, connected sensors for improved safety and operational awareness.

The camera is designed to provide early detection of fire hazards, equipment malfunctions, and thermal anomalies in industrial and commercial environments. The device consolidates thermal imaging, continuous monitoring, and alarm notifications into a single platform aimed at supporting preventive maintenance and risk management efforts.

Irley supports Ethernet and LTE connectivity, enabling deployment both within facilities that have wired network infrastructure and in remote or hard-to-reach locations where cellular connectivity is required. This flexibility allows integrators to position the camera in critical zones without requiring additional network setup.

Vision Systems Design (VSD): What specific types of image processing or machine learning algorithms are implemented in this thermal camera to distinguish between normal heat variations and genuine early warnings for fire or equipment failure?

Markus Tarin (MT): moviTHERM has developed several proprietary algorithms, which makes Irley unique; it is not just another thermal smart camera. The smarts in the camera perform pixel-by-pixel analysis in real time at 30 fps. The camera can run up to 80 fps for specialty applications with fast-moving objects. Irley can analyze the rate of rise and vector-based hotspot detection with directional spread, to name one of the many algorithms. But almost more important than detecting a fire early is Irley's ability to figure out what not to trigger on: for example, moving vehicles with hot exhaust pipes or hot engine components. It is an all-in-one solution for busy areas where it needs to differentiate friend vs. foe.

VSD: How do you plan to manage software updates and incorporate ongoing improvements or new AI models in the system over time, especially for embedded hardware deployments?

MT: Our team has built in an over-the-cloud firmware update function. Irley comes with a built-in LTE cellular modem and is always reachable. We can push updates or new AI models to any Irley camera remotely with a push of a button.

VSD: Can you describe the security architecture used to protect the thermal data, alert signals, and device control commands transmitted over Ethernet or LTE networks?

MT: We thought this through very carefully, as we deal with critical infrastructure monitoring with stringent cyber security regulations. The data transfer uses a 2048-bit encryption. Our cloud server runs on a secure and dedicated AWS server. We manage our own SIM cards and data plans. Our SIMs run on a first-responder network with redundancies. We use a dual-core, multi-carrier SIM that automatically switches over to the next strongest provider, in case one goes down. We also have a secure and encrypted remote tunnel interface that allows us to support our customers anywhere in the world. This also allows our customers to configure the camera via the cloud interface—all through one and the same interface.

VSD: What calibration methods or environmental adaptation techniques are included to ensure consistent detection accuracy across diverse industrial settings that have variable temperatures, humidity, or potential infrared obstructions?

MT: The thermal camera has an automated self-calibration built in. It is a so-called NUC (non-uniformity correction). During a NUC operation, which just takes a second, a small flag drops in front of the thermal detector, and the camera electronic performs a pixel-by-pixel non-uniformity correction. This ensures that the camera stays within the accuracy specification. The NUC is triggered based on timing and environmental conditions. It can also be performed manually via the cloud interface. Like with any camera, if something is truly obstructing its view, there is nothing the camera can do to circumvent that.