How to Tell a Hot Surface from a Reflection in a UVH2 Thermal Inspection
UNITED UAV Official
Commercial disclosure:
UNITED UAV sells the UVH2. The observations below are hypothetical examples for planning a buyer's acceptance demonstration.
Official UVH2 product image. The aircraft is shown from the front.
A bright patch on a thermal image of a rooftop equipment housing can trigger an unnecessary maintenance visit. The patch may be real surface heating, sunlight warming one side of the housing, or infrared radiation from a nearby surface reflected in a shiny panel. The operator has to distinguish those possibilities while the original scene and operating conditions are still available. A screenshot of a bright region and a pin on a map leave most of that decision to the person receiving the report.
The UVH2 is listed as a fixed wing VTOL aircraft with visible and thermal observation. That combination is useful for pairing a thermal indication with a recognizable feature such as a seam, fastener, or inspection label. The pairing is the start of an investigation. For this acceptance exercise, the important outcome is whether a second reviewer can follow the evidence and decide which finding deserves another look. Keep the exercise within the agreed inspection area and have facility personnel manage access to the equipment.
Start with the decision the image must support
Define what counts as a useful result before the demonstration. A facilities lead may only need to decide which enclosure should receive a follow-up inspection. A different contract may require a defensible temperature comparison at a specific surface. Those deliverables have different evidence needs. The first can use a well documented persistent indication and assign a maintenance priority. The second requires measurement conditions, an appropriate instrument and recording chain, and a way to evaluate uncertainty against the agreed decision threshold.
Ask the site team for a small asset inventory before collecting images. Record which panels are painted, which fasteners or trim pieces are bare metal, and which comparable assets are operating under similar loads. Add likely sources of reflected radiation to the scene sketch. These can include a sunlit parapet, a nearby exhaust duct, a warm roof membrane, or a large region of open sky. A map of those features gives the reviewer hypotheses to test. It also helps the crew select reference areas that share a material and exposure with the suspected area.
Give each observation a precise location in the language of the asset. “Upper right corner of the image” changes when the view changes. “Painted panel above hinge H2 on enclosure M17” is a reference that two people can check against the visible image. Include enough of the enclosure and the adjacent surroundings in the first pair of images to identify the panel without relying on the operator's memory. Then collect detail views only from observation positions that the qualified crew and facility have approved.
Understand what a reflective surface can show
A thermal camera receives radiation from the surface, from the surroundings reflected at that surface, and from the intervening atmosphere. The relative importance of those contributions depends on the material, finish, view and conditions. For a similar physical temperature, a matte painted area and a clean metal edge can present very different signals. A reflective edge can look warm when it reflects a warm object. It can look cool when the scene reflected toward the camera is cooler. A bright stripe along an edge therefore deserves a check of the material boundary and scene before a defect is entered in the report.
Emissivity describes how effectively a surface emits thermal radiation compared with a reference ideal emitter. Reflected apparent temperature describes the radiation arriving from the surrounding scene that the target reflects toward the camera. Air temperature is a separate observation. Someone writing a quantitative report should record which values were measured, estimated, or assumed and which regions those values apply to. A single value assigned to an entire enclosure is a poor substitute for knowing that the comparison crosses a painted panel, a metal latch, and a glass indicator window.
Display scaling is another variable. A thermal viewer may assign a new color range when the composition changes or another warm object enters the field. A region may change color while its underlying signal remains similar. Save the display range and palette with the evidence and compare views on a consistent scale when the available tools allow it. If the equipment only provides rendered thermal images, treat colors as observations within the documented display settings. For a quantitative deliverable, ask separately which underlying values can be retrieved and what metadata survive export.
Separate a change of direction from a change of condition
A useful starting sequence is to record a baseline view, a second approved view from a different direction, and then a return to the baseline view. The return makes the sequence a test instead of just a pair of pictures. If the first and third views agree while the middle view changes, geometry is a plausible explanation for the change. If the baseline itself has changed, a cloud, a load change, a wetting event or a drifting camera display could be involved. Record the capture times and ask the site representative to log any known change in equipment operation during the sequence.
Keep the feature large enough to identify in the thermal image and pair each thermal frame with a visible frame. Document the change in direction using asset landmarks or the available position record. Try to keep viewing distance and target coverage comparable where the inspection plan permits. When the second direction forces a poorer view, record the difference and classify that comparison accordingly. The absence of a stripe in a coarse or partly blocked image can be explained by reduced target coverage, so it has little value as evidence about reflection.
Look at where the indication sits relative to a physical reference. Does the bright patch move along a bare metal strip when the viewing direction changes? Does it remain at a particular painted seam? Does a reflection in neighboring bare trim move in the same way? Save a marked copy that points to those features alongside the unmarked originals. The marks help a reviewer find the areas quickly. The originals allow the reviewer to check whether the indicated boundary was visible in the first place.
Use controls that answer different questions
Include a reference region on a similar finish under similar conditions. A comparison with the next housing can be useful when its finish, orientation, sun exposure and operating state are similar and documented. If those conditions differ, use it as scene context and look for a closer match. A neighboring panel that has been in shade for hours is weak evidence about a panel that has been in direct sun. Matching on material and exposure often changes the interpretation more than adding another frame of the original target.
For a demonstration, a facility team can prepare an accessible, de-energized comparison sample with adjacent matte and reflective regions. Have the team document the surface materials and select a safe position. Inspect the sample from more than one approved viewing direction while its physical condition remains stable. An apparent change concentrated on the reflective region is useful evidence of view sensitivity. If the buyer needs to show a repeatable heating indication too, ask for a separate safe reference with an independently established temperature or contrast, under conditions controlled by the demonstration team.
If the reference check is performed on the ground, document what it proves. It exercises the viewing and reporting method under controlled conditions. A successful bench or ground result leaves the team to demonstrate that the same type of evidence can be captured from the actual approved observation positions at the site. Image coverage, a short view of the target, occlusion and changes in ambient conditions can all affect that transfer. Procurement should treat ground verification and the representative field example as two distinct entries in the acceptance record.
Work through a reflective latch and a painted seam
Consider an inspection of two neighboring rooftop housings. On the first pass, the left housing shows a bright oval on its bare latch. The visible frame places a sunlit duct across the scene. On the second view, the oval shifts toward the latch edge, while an area of painted material above the latch remains uniform. On the return to the first view, the bright oval returns to its earlier position. The operating log records no known load change during the sequence, and the frame settings and target coverage are comparable.
The repeatable shift relative to the latch is evidence consistent with reflection. The reviewer should be able to locate the bright oval in each frame and see the painted comparison area. A useful report entry would say that the indication was view dependent on bare metal and that the source of the reflection was plausible but unconfirmed. It would preserve the observed position changes and avoid turning the apparent color into a component temperature. The site team can decide whether the indication warrants a closer observation by its usual inspection method.
Now suppose a patch above the hinge on the second housing remains in the same place on a matte painted surface in all three usable views. A comparable part of the same panel remains less prominent, and the return frame shows no meaningful change in target coverage. That supports a persistent surface indication under the observed conditions. It leaves several causes open. Solar exposure, a change in thermal contact, airflow across the housing, and internal heating could all produce a persistent result. The location and operating record are the handoff to the maintenance specialist who will distinguish among those explanations.
In a third case, the second direction puts the feature behind a railing and the baseline view has changed after cloud cover arrives. Record the result as inconclusive and keep the obstructed frame. It tells the next crew which comparison failed and why. A documented need for another observation is an acceptable outcome when the available evidence cannot support either of the other classifications. It is also a measurable requirement for the reporting software and for the crew's training.
Agree on the classification and the handover
| Evidence pattern | Suggested classification | What to retain |
|---|---|---|
| Feature shifts with view on a reflective finish and returns with the baseline view | Consistent with reflection | All three original observations, finish, orientation and matched reference |
| Feature stays at the same identified surface region in comparable usable views | Persistent surface indication | Repeated location, scene context and conditions for follow-up |
| One view obscures the region, resolves it poorly, or follows a change in conditions | Inconclusive | Unusable frame, reason and what the next visit should control |
Ask two people to classify the demonstration independently. Give each of them the same evidence set and the asset register, while keeping the operator's interpretation separate until after their first decision. Review disagreements at the level of individual evidence items. One person may have compared two different fasteners. Another may have assumed a change in color came from a change in signal when the scale changed. Require the team to say which observation would resolve each disagreement. This puts the focus on the quality of the inspection record instead of forcing agreement with the first report.
A handover should include the asset identifier, surface finish, acquisition times, usable view labels, the visible and thermal originals, any exported display settings, and the source of the operating and environmental observations. Add the marked comparison sheet, the reference region, the selected classification, and the reason it was chosen. If the field of view changes enough to affect the comparison, mention that in the same entry. Establish a naming convention that keeps each evidence pair together when the customer's reviewer receives the files.
Agree how the recipient will handle a quantitative request before the test. A buyer who needs a temperature value should ask how a target is resolved in the source thermal frame, how emissivity and reflected apparent temperature are handled, and how the measurement procedure is checked against a suitable reference. If a threshold drives an operational decision, ask what happens when the uncertainty overlaps it. Keep that requirement separate from the simpler question of whether a persistent indication can be reported and located for a follow-up inspection. Mixing the two tends to produce confident language unsupported by the records.
Put the evidence package in the quote request
Ask for a demonstration with the quoted aircraft and the recording and review workflow the vendor intends to supply. Include one representative reflective feature, one persistent reference indication, and at least one deliberately incomplete observation. A good deliverable shows how the reviewer notices the incomplete evidence and keeps it separate from a completed finding. Write the acceptance rule around observable outputs. The buyer should receive a comparison set in which each relevant region can be identified, the view change can be reconstructed, and each classification is tied to a specific observation.
Request the demonstration files in the formats the customer will actually receive. Ask which thermal and visible items can be exported, how their timestamps are represented, and what the reviewer can inspect without access to the operator's controller. If the supplier sends only a composite image or a clipped video, ask for a fresh sample showing the material boundaries and the comparison views. Include the roles responsible for collecting the evidence, classifying it, and assigning the follow-up. The person who owns the maintenance action should receive enough information to decide whether to schedule a visit.
See the UVH2 Dual-Sensor Fixed Wing VTOL Drone for Drone Mapping, Land Surveying and Thermal Inspection for the listed platform. When requesting a quote, include your sample asset, which finishes are present, the observation conditions you can reproduce safely, and whether you need screening evidence or a documented measurement. Ask the supplier to return an example of the complete evidence package before settling the acceptance language. It is easier to clarify what the buyer will receive while there is still time to specify the demonstration.
Technical references. Teledyne FLIR, “Achieving a good thermal image” (T810664), “The measurement formula” (T505006), and the Ex Series User’s Manual section “Thermographic measurement techniques.” These references explain general thermography and do not specify the UVH2 payload.