Speaker
Description
Various IR camera suppliers sell their products, regardless of their intended field of use and according to specifications that primarily address temperature measurement applications. For NDT applications by Infrared Thermography, users often select their cameras based on rather arbitrary criteria (wavelength range, acquisition frequency, temperature range, field of view) without taking into account the targeted application. Furthermore, cameras are generally delivered with temperature calibrations that are rarely utilized in NDT.
IR camera monitoring over time usually involves returning the device to the supplier for maintenance, where a few component checks are performed and NUC tables, bad pixel lists, and calibrations are updated according to the supplier’s own standards.
The rising use of IR cameras in production and maintenance inspections has highlighted the necessity to define calibration procedures that truly reflect industrial needs. This work presents Safran's ongoing efforts to monitor IR cameras, deploy an internal methodology, and eventually establish a dedicated standard.
Currently, specification requirements for NDT mainly rely on existing standards and user experience. However, the few available standards in Infrared Thermography often differ significantly in definitions and do not address operational conditions for camera characterization and validation. As a result, there is a need to clarify how to characterize an IR camera, define its application domain, and monitor its key parameters over time.
This study proposes a measurement methodology based on building SITF (Signal Intensity Transfer Function) curves using a HGH blackbody, allowing the estimation of critical features such as sensitivity, gain, NETD (Noise Equivalent Temperature Difference), measurement range, sensor correction, and bad pixel identification.
Even though the calculation of gain, offset looks straightforward, the process is influenced by different acquisitions conditions and processing parameters. Its even more tricky for sensitivity and NETD estimation since they are strongly linked to the setup of the camera, especially the integration time. Because existing standards lack guidance on operational conditions, camera characteristics are evaluated according to supplier recommendations.
Recognizing that NDT applications typically display intensity in Digital Levels (DL) instead of temperature, the paper introduces and discuss a new term, NEDLD (Noise Equivalent Digital Level Difference), more suitable than NETD.
Finally, the author will discuss in this paper how to determine this dynamic range and how to correct bad pixels in relation to specific applications.
