CoPyro - Computational lensless two-colour one-sensor pyrometer for airborne quantitative temperature measurement

• Applicant: Dr.-Ing. Johannes Meyer
• Subject Area: Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 541018302

The possibility of airborne quantitative and spatially resolved temperature measurements enables various applications. For example, large and difficult-to-access structures, such as bridges or wind turbines, can be inspected for material defects on or beneath the surface using pulse thermography. In addition, contactless temperature measurement and hazard assessment can be carried out from the air in hostile situations (fires, disasters). However, related state-of-the-art sensor technology is only limitedly usable on flying platforms due to the weight, required space, and energy consumption of the necessary optical and electronic components. Therefore, the aim of the proposed project is to research and develop a lightweight, small, and energy-efficient sensor for quantitative temperature measurement. The approach pursued is based on the concept of a two-color pyrometer, where the infrared light of an object is detected by two sensors at two wavelengths to derive quantitative temperature measurements. To mitigate the weight and space requirements and the energy consumption typically required by measurement systems of this kind, the proposed project pursues the approach of a lensless optical system with only one sensor. The incoming light is also filtered for two spectral ranges via two optical paths but then the two optical paths are superimposed on only one sensor. Lightweight and small optical modulators such as diffractive optical elements (DOEs) or phase masks replace the lenses of the conventional system and allow an optical encoding of the amplitudes and phases of the two optical paths according to their local structures of the modulators. An algorithmic reconstruction calculates the desired temperature measurement from the recorded superposition of the two signals. In contrast to the classical approach in the design of such optical systems, the project follows the concept of computational imaging, where both the degrees of freedom of the optical system (e.g., structures of the DOEs) and the parameters of the image reconstruction (coefficients of numerical methods, weights of a neural network, etc.) are optimized holistically in a data-driven manner. This allows the overall system to be optimized with respect to various criteria (e.g., measurement accuracy, speed, energy and storage requirements). The entire formulation (forward model of image formation, reconstruction method) must be continuously differentiable, enabling the use of modern optimization methods based on gradient descent. Since the reliable use of the calculated measurement values requires an indication of the measurement uncertainty, methods for quantifying and propagating it through the reconstruction method are being investigated. The sensor concept will be prototypically implemented on a drone and tested based on the scenarios described initially.

• DFG Programme: Priority Programmes
• Subproject of: SPP 2433: Metrology on flying platforms
• Co-Investigator: Professor Dr.-Ing. Jürgen Beyerer