Reference temperature
A temperature difference is computed only when a valid reference exists — an identical component in the image, or an ambient value that was supplied. Without a reference, the temperature itself is shown, not a difference.
Thermal analysis
A thermal image is a measurement, not just colour
The difference between a radiometric file and a colour image is the difference between data you can measure and a visual representation only. That distinction governs everything the platform can — and cannot — say about an image.
Radiometric vs. visual
What is actually in the file
| Original radiometric file | Colour image / screenshot | |
|---|---|---|
| Per-pixel temperature data | Present | Not present |
| Camera measurement parameters | Stored in the file | Lost in conversion |
| Spot and region-of-interest measurement | Possible | Not possible |
| Difference against a reference value | Possible with a valid reference | Not possible |
| What can be displayed | Measurements with a documented source | Visual display only |
A file extension proves nothing about radiometry. A JPG file may carry embedded measurement data — or be an ordinary photograph. That is why the check is run on the file’s contents, not its name.
Measurement context
Every value has a source, and every source has a limit
Measurement parameters
Values such as emissivity and distance affect the result. The platform displays the parameters used for the calculation, so review can take them into account.
Regions of interest
A region of interest defines where a value was measured. Changing the region changes the result, so the region is stored together with the measurement.
Identifying a component is not a thermal finding
Identifying a component answers “what is this”. A thermal finding answers “what is anomalous here”. These are two separate layers, and not every identification leads to a finding.
Reading the result correctly
Three distinctions that prevent mistakes
Model confidence is not fault probability
Where a confidence value is shown, it refers to identifying the component in the image. It says nothing about the likelihood that a physical fault exists.
No findings does not mean “healthy”
When no findings are suggested, it means no region was raised for review under the current analysis settings. It is not a statement that the equipment is sound.
Missing context limits the conclusion
When information about the environment or capture conditions is missing, some calculations are unavailable. The platform shows the limitation instead of filling in a guess.
This result supports professional review and does not certify equipment condition.
File support
Support is tested, not declared
Initial support focuses on original FLIR radiometric files, from camera models and versions that have actually passed a support check. Additional models and manufacturers will be included only after they are tested against real files. Until then, a file outside the supported set receives an explicit message rather than a partial analysis that looks sound.
Not sure where this fits in your operation?
Tell us what equipment you inspect and what reporting looks like today, and we will come back with a concrete assessment.