2.3 Identifying and Reducing Measurement Errors

2.3 Identifying and Reducing Measurement Errors

Quick Answer: Quantitative IR errors usually come from wrong emissivity, ignored reflections, targets that are too small for the spot/IFOV, uncorrected atmosphere, soft focus, or steep viewing angles. Reduce them with correct camera parameters, proper geometry, environmental awareness, and contact or reference-emitter checks when safe.

A thermal image can look convincing while the temperature numbers are wrong. Level II thermography is the discipline of finding the error before the client finds it—and of stating residual uncertainty when physics will not allow a hard number. This section catalogs the dominant error sources and gives a field checklist you can apply on every quantitative job.

Error Source 1: Emissivity Error

Emissivity (ε) is the fraction of blackbody radiation a real surface emits at a given temperature and wavelength band. Cameras need a user-supplied (or assumed) ε to convert radiance to temperature.

What goes wrong: If true ε is lower than the camera setting, the camera often under-reads target temperature (it assumes more of the collected radiance is emitted by a cooler body than is true—exact bias also depends on reflected radiance). If ε is set too low on a high-ε surface, temperatures can over-read. On shiny bus bars, a 0.1 error in ε can swing indicated temperature by many degrees.

Mitigations:

• Use manufacturer or measured ε tables as a starting point, not gospel.
• Apply known high-ε tape or paint (when permitted and safe) and measure on that patch.
• Solve for ε using a contact reference on accessible equipment.
• Avoid reporting false-precision absolute temperatures on bare polished metal without a reference strategy.
• Match ε to the camera waveband and surface condition (oxidation, oil, dust change ε).

Error Source 2: Reflections (Reflected Apparent Temperature)

Real surfaces reflect radiation from the surroundings. The camera must know the reflected apparent temperature (RAT)—sometimes called reflected temperature or background temperature—to separate emitted from reflected radiance.

What goes wrong:

• Hot boilers, sunlit walls, open sky, or a thermographer’s own reflection on metal create false hot or cold spots.
• Using indoor ambient air temperature as RAT while viewing outdoor sky reflections biases building and roof work.
• Low-ε metals are reflection mirrors in the infrared.

Mitigations:

• Measure RAT with an accepted method (e.g., crumpled low-ε foil method for a diffuse estimate of background radiance) and enter it into the camera.
• Change viewing angle to deflect specular reflections away from hot sources when diagnosing metals.
• Use high-ε targets or coatings so emission dominates reflection.
• Mentally classify anomalies: true thermal features usually persist with angle/load changes; pure reflections often move or vanish when you move.

Error Source 3: Size-of-Source / IFOV / Spot Size

Every radiometric measurement area has a finite size on the target. For cameras, the instantaneous field of view (IFOV) and measurement spot tools define how much target area contributes. For pyrometers, the distance-to-spot (D:S) ratio defines the spot diameter at distance D.

What goes wrong:

• Measuring a bolt head, thin wire, or small connection that does not fill the spot mixes target radiance with background, usually pulling the reading toward ambient.
• Digital zoom or “spot meters” that look precise still obey optics—zoom is not free resolution.
• Standing too far away from small electrical targets is a classic Level I error that Level II must catch in QA.

Mitigations:

• Calculate or look up IFOV; ensure the target is several times larger than the IFOV spot for quantitative work (follow camera vendor and training guidance for measurement spot sizing).
• Move closer when safe, or use a lens with a smaller IFOV (telephoto IR lens).
• For pyrometers, verify D:S so the spot is smaller than the target.
• Report qualitative-only findings when the target is optically undersized.

Error Source 4: Atmosphere

Air between camera and target is not perfectly transparent. Distance, relative humidity, and atmospheric temperature affect transmission and the path’s own emission. Cameras that accept these parameters apply a simplified atmospheric model.

What goes wrong:

• Long outdoor paths, steam plumes, dust, or very humid air attenuate and add radiance.
• Leaving distance at “1 m” while standing 15 m from a target on a humid day injects systematic bias.
• Ignoring IR window transmittance (viewing ports) is a specialized but severe atmospheric-like error—windows need their own transmittance entry.

Mitigations:

• Enter realistic distance, RH, and air temperature for quantitative absolute temperatures.
• Prefer shorter paths when precision matters.
• Avoid sighting through exhaust steam, heavy dust, or precipitation.
• For IR windows, use measured or manufacturer transmittance and correct explicitly (Level II electrical work).

Error Source 5: Focus

An out-of-focus image spreads energy across pixels and softens edges. Temperatures of small hot targets are typically understated, and patterns needed for diagnosis blur.

Mitigations:

• Focus on the plane of the target of interest—not the fence in front of the substation.
• Use manual focus when autofocus hunts on low-contrast scenes.
• Refocus when distance changes; do not reuse a previous focus setting across bays.
• If the image is soft, treat spot temperatures as suspect until sharpness is restored.

Error Source 6: Viewing Angle

Emissivity of many surfaces is angle-dependent. As you move far from normal (perpendicular) viewing, effective emissivity often drops and reflectivity rises—especially beyond about 45° from normal for many materials. Specular metals complicate this further.

Mitigations:

• Image as close to normal incidence as safety and access allow.
• When forced to steep angles, reduce confidence in absolute temperatures and favor comparative same-angle measurements.
• Be consistent: baseline and follow-up images should repeat geometry when trending.

Integrated Error Picture

Errors compound. A low-ε bus viewed at a steep angle from too far away, with default RAT and default atmosphere, can produce a number that is not merely imprecise—it is misleading. Level II reports should not launder that number into a false NETA priority without correction or qualification.

Level II Mitigation Checklist (Use on Every Quantitative Survey)

Before imaging

1. Confirm camera calibration status and lens ID.
2. Set display units (°C/°F) to match the reporting standard.
3. Plan safe distances and target access; note whether contact verification will be possible.
4. For electrical work, confirm adequate load conditions when the procedure requires them.

At the camera

1. Focus critically on the measurement plane.
2. Enter emissivity appropriate to the actual surface (or measure it).
3. Measure and enter RAT.
4. Enter atmospheric temperature, RH, and distance (and window transmittance if applicable).
5. Verify target size versus IFOV / measurement tool; change position or lens if needed.
6. Keep viewing angle near normal; document forced exceptions.

While interpreting

1. Distinguish thermal patterns from reflections by changing angle or load when safe.
2. Compare similar components under similar conditions (comparative thermography).
3. Cross-check critical absolutes with contact sensors or process instruments when safe.
4. If uncertainty remains high, report qualitative findings and state limitations.

After the shot

1. Store radiometric images with parameters intact—not only screen JPEGs that strip metadata.
2. Annotate units, load, ε, RAT, distance, and environmental notes in the work product.
3. Peer-review Level I parameter coverage as part of QA/QC duties.

Professional Mindset

Measurement error is not a moral failing; it is physics meeting field constraints. The Level II differentiator is a repeatable error-control process: identify which terms dominate, attack them in order (usually ε and reflections first on metals; geometry first on tiny targets), and refuse to over-claim precision. Clients hire certified thermographers for trustworthy decisions—not for colorful palettes.

Summary for Level II

Control emissivity and reflections, obey size-of-source rules, correct atmosphere and windows, focus sharply, and view near normal. Use the mitigation checklist as standard operating procedure, and verify with contact or reference emitters whenever safety and access allow. That is how quantitative infrared temperatures become defensible.