That distinction decides how much you can do with the file weeks after the inspection is over. This guide explains, in plain terms, what the file actually holds and what it lets you measure.
A picture you can see, and data you cannot
Open a radiometric JPEG in any ordinary image viewer and it behaves like a photo. That visible layer is a rendering: the camera took the measured scene, mapped it onto a color palette, and saved a viewable preview so the file is easy to share and recognize.
Behind that preview sits a second layer that a normal viewer never shows. The camera embeds the raw measured values for the sensor, plus a block of settings that describe how the shot was taken. The colors you see are a presentation choice. The measured layer is the evidence.
The measurement behind every pixel
The core of a radiometric file is a grid of measured values, one per pixel, captured by the infrared sensor. On its own, a raw sensor value is not a temperature yet. Turning it into degrees depends on several parameters, and the camera stores those parameters inside the same file so the conversion can be reproduced later.
The usual parameters include:
- Emissivity, how efficiently the surface radiates heat. A dull painted bushing and a shiny metal clamp at the same real temperature give off infrared differently.
- Reflected apparent temperature, to account for heat reflecting off the surface from the surroundings.
- Distance, and atmospheric values such as air temperature and humidity, which affect how much signal reaches the lens.
Because these are saved with the image, a reviewer can read the temperature at any point in the scene after the fact, not only at the single spot the operator happened to tap in the field.
What a normal thermal image loses
Many thermal pictures that circulate are just colored images: the temperature-to-color mapping has been baked in and the measured layer is gone. They are fine to look at and almost useless to analyze. You cannot ask "how warm is that connector" because the number no longer exists in the file, only its color.
You also cannot correct a wrong assumption. If the surface emissivity was set wrong, a flat picture is stuck with it. A radiometric file keeps the measured layer, so the parameters can be revisited and the reading recalculated on the same original data.
What you can actually measure from it
Because the measurement survives, a radiometric JPEG supports real review:
- Read the temperature at any point or region, not just one preset marker.
- Compare two points in the same scene, for example a warm contact against a similar contact nearby.
- Re-examine the shot with corrected parameters if the field settings were off.
- Keep the original as evidence that can be looked at again later, instead of a screenshot.
This is where thermal analysis in HealthyGrid works from. The platform reads the original radiometric file, surfaces the measured values, and helps the inspector line up components and regions of interest. Today the platform calculates with the emissivity and reflected temperature the camera stored in the file and shows them next to the result; changing them and recalculating is not available in the platform yet. It is important to be precise about roles here: identifying a component in the image is not the same as diagnosing a fault, and the tool presents measurements and context as evidence. The conclusion, and the responsibility for it, stay with the inspector.
Why the original file matters
The practical takeaway for anyone building an inspection record: keep the original radiometric file, not a flattened export. Once the measured layer is discarded, no software can bring it back. The original is what lets a second reviewer, an auditor, or a future you reopen the scene and check the numbers rather than trust a color.
