Thermal Emission Evaluation Criteria

One of the worst conditions for evaluating thermal emission is during a hot summer’s day. The different surface areas and background play a major role in attenuation of the heat photon. I would like to use my short research version description, used as part of the education towards thermal camera users, as a guideline and then evaluate the image accordingly.

Thermal Image Evaluation: Attenuation Effects Across Ground, Tar, Sand and Grass Surfaces

1. Introduction

Thermal imaging systems detect infrared (IR) radiation emitted by objects as a function of their surface temperature and emissivity. In perimeter security applications, the performance of thermal cameras is strongly influenced by thermal attenuation effects caused by surface materials, background environments, and distance. This evaluation examines how different ground surfaces e.g. tar, sand, grass, and general terrain, affect thermal contrast and detection performance, and explains how heat radiation attenuates over distance, including the relevance of the inverse square law.

2. Thermal Behaviour of Ground and Surface Materials

Different surfaces exhibit distinct thermal characteristics based on their emissivity, heat capacity, and thermal conductivity.

Tar (Timeless Technologies) – asphalt – typically has high emissivity (≈0.95) and high thermal mass, allowing it to absorb and retain heat for extended periods. As a result, tar surfaces may remain warm after sunset, reducing thermal contrast between a human target and the background during early night conditions. This phenomenon can degrade detection performance due to thermal blending.

Sand, (TBA) – particularly dry desert sand, heats rapidly during the day and cools quickly after sunset due to low thermal mass. This rapid temperature transition can improve thermal contrast at night but may introduce thermal noise during late afternoon periods when sand emits strong residual heat signatures.

Grass (Helderberg Village) – covered terrain exhibits evaporative cooling and lower surface temperatures compared to hard ground. Grass generally enhances thermal contrast between human targets and the background, particularly at night, making it one of the most favourable surfaces for thermal detection.

Mixed (TBA) – or uneven terrain introduces spatial temperature variations caused by shadows, moisture differences, and material heterogeneity. These background inconsistencies may generate false alarms or reduce detection confidence if not mitigated by advanced image processing and analytics.

3. Thermal Attenuation Over Distance

Thermal radiation emitted by a target weakens as distance increases due to geometric spreading, atmospheric absorption, and scattering. Although thermal cameras detect emitted radiation rather than reflected energy, the signal reaching the sensor diminishes with range, reducing signal-to-noise ratio (SNR) and image clarity.

Atmospheric attenuation is influenced by humidity, dust, fog, and temperature gradients, all of which absorb or scatter long-wave infrared radiation (8–14 µm). In coastal or desert environments, airborne moisture and particulate matter can significantly reduce effective detection ranges.

4. Inverse Square Law and Its Relevance

The inverse square law states that radiative intensity decreases proportionally to the square of the distance from the source. In thermal imaging, this principle applies to the radiant power emitted by a target: as the distance between the target and the sensor doubles, the received radiant intensity is reduced by approximately a factor of four.

While modern thermal cameras employ high-sensitivity detectors and image enhancement algorithms to mitigate this effect, the inverse square law remains a fundamental physical limitation. Its impact is most evident in long-range detection scenarios, where small temperature differentials become increasingly difficult to resolve against background noise.

Importantly, the inverse square law acts in combination with surface-dependent background emissions. For example, a human target viewed over warm tar at long range will experience compounded attenuation, both reduced radiant intensity and diminished contrast, compared to the same target over cooler grass.

5. Conclusion

Thermal image evaluation must account for both surface-dependent thermal behaviour and distance-related attenuation effects. Tar and sand surfaces can reduce detection performance under certain thermal conditions, while grass and cooler natural terrain generally enhance contrast. Thermal attenuation over distance, governed in part by the inverse square law and exacerbated by atmospheric effects, imposes practical limits on detection range. Understanding these interactions is essential for accurate system design, camera placement, and performance modelling in perimeter security applications.

The above guideline is solid (surfaces + distance + atmosphere + inverse-square framing), but thermal image quality in the real world is usually “won or lost” by a few additional factors that sit around those physics. If you add the items below to your evaluation, you’ll catch most of the failure modes that only show up on site.

6) Scene physics not listed (but will bite in summer)

Thermal crossover / equalisation windowsOn hot days you often get periods (late afternoon, pre-dawn) where target ΔT vs background collapses. Performance drops even if the camera is “good”. Make sure you test specifically in these windows, not only at “nice” times.

Sky and horizon effects (cold background trap)A human against “cold sky” looks fantastic; the same human against sun-loaded rocks or heated fence fabric looks mediocre. Always score image quality in multiple backgrounds: sky/horizon, ground, vegetation, structures.

Viewing angle and emissivity shiftEmissivity isn’t a single number in practice. At shallow angles, and on some materials, you can get more reflected apparent temperature and less true emission, especially on smoother/denser surfaces and painted metals. If your line of sight is low and long, test that geometry.

Wind and microclimateWind can cool surfaces (and people’s clothing) fast, changing contrast minute-to-minute. Coastal wind + fog is a classic “same day, different result” trap. Record wind speed/direction during each test run.

Moisture on surfaces (dew / wet tar / wet sand / wet grass)Wetness changes emissivity and thermal conduction dramatically. Wet tar can “flatten” contrast; wet grass can become more uniform; wet sand can retain heat longer than dry sand. If you can, do a set of runs after sprinkler, rain, or morning dew.

7) Camera & optics performance factors (often mistaken as “physics attenuation”)

Focus stability and depth of field (critical)Many thermal evaluations are accidentally evaluating focus drift, not sensor quality. Check:

manual vs autofocus behaviour

focus repeatability after temperature swings

focus at near, mid, far (and whether the camera “hunts”)

Lens quality, f-number, and vignettingTwo cameras with the same resolution can look very different because of lens transmission, f/#, internal reflections, and edge performance. Evaluate centre vs edge sharpness and contrast.

Detector sensitivity & stability

NETD / temporal noise (how “grainy” it gets in low contrast)

NUC behaviour (does it freeze, smear, or “jump” during correction?)

Bad pixel handling (clusters become false alarms at long range)

Dynamic range and AGC (Automatic Gain Control) tuningAGC can make an image look “pretty” while crushing the details you need for

detection/ID. Test:

multiple AGC modes (linear / histogram / plateau / “smart scene”)

whether hot backgrounds cause target washout

whether cold backgrounds cause black clipping

Digital enhancement settingsEdge enhancement / DDE / local contrast can improve perceived sharpness but can also:

create “halos” that confuse analytics

amplify compression artifactsYou want a “baseline” test with minimal enhancement, then a “field” test with your preferred tuning.

8) The “system” problem: streaming, compression, VMS, and the operator’s screen

Compression can destroy thermal detail (and looks like atmospheric attenuation).Evaluate the same scene under:

H.264 vs H.265 (if available)

different bitrates and GOP sizes

CBR vs VBR

the exact VMS pipeline you will use (Bosch/Cathexis/etc.)

Display & scaling effectsIf operators watch a 640×512 stream scaled badly to 1080p/4K, you can get softness and aliasing.

Confirm:

client decoding quality

scaling mode (bilinear vs better scaling)

monitor brightness/contrast and viewing distance

Latency and frame rateFor moving targets, low frame rate and high latency reduce track confidence and “visual detectability” even if the raw image is fine.

9) Measurement discipline: how to evaluate consistently (so results are defendable)

Use a repeatable target setAt minimum:

human walking, stopping, crouching, rolling

human partially obscured (bush/fence)

human at multiple orientations (front/side/back)

a “false alarm” set (moving vegetation, hot rocks, small animals if relevant)

Score by task, not only by “looks good”Separate scoring into:

Detection (something is there)

Recognition (it’s a human, not a dog/bush)

Identification (who/what exactly)

If you want a practical rule of thumb, log pixels-on-target at each range and compare against your operational requirement (detection vs recognition vs ID), because perceived contrast alone is misleading at long range.

Log conditions for every run

For each clip/run, record:

ambient temperature, wind, humidity, visibility/fog, sun state

surface state (dry/wet), recent solar loading

range, target speed, background type

camera settings (AGC, palette, enhancement, bitrate)

This turns your evaluation into something you can defend to an ExCo / client when someone says: “but it looked better last week.”

10) Analytics and false alarms (if the camera will do perimeter work)

If you’re evaluating for perimeter detection (not just viewing), include:

VA sensitivity vs nuisance alarms in heat shimmer / moving grass

stability during NUC events

scene learning time and drift over day/night cycles

performance with “problem backgrounds” (warm tar, sun-heated sand, fence mesh)

11) A compact add-on checklist you can bolt onto your document

Add these headings to your evaluation form:

Environmental: thermal crossover windows; wind; wetness/dew; sky/horizon background; viewing angle/geometry

Optics & imaging: focus stability; edge/centre sharpness; NETD/noise; NUC behaviour; AGC mode impact; enhancement impact

Streaming/VMS: codec; bitrate/GOP; client scaling; frame rate/latency; recording settings

Operational: detection/recognition/ID scoring; pixels-on-target at ranges; false alarm testing; obstruction testing

My Summary References:

·         FLIR Systems (2019) Thermal Imaging Guidebook for Building and Perimeter Security. Wilsonville, OR: FLIR Systems.

·         Holst, G.C. (2000) Electro-Optical Imaging System Performance. 3rd edn. Bellingham, WA: SPIE Press.

·         Rogalski, A. (2011) Infrared Detectors. 2nd edn. Boca Raton, FL: CRC Press.

·         Vollmer, M. and Möllmann, K.-P. (2017) Infrared Thermal Imaging: Fundamentals, Research and Applications. 2nd edn. Weinheim: Wiley-VCH.

·         Holst, G.C. (2006) Electro-Optical Imaging System Performance. 4th edn. Winter Park, FL: JCD Publishing.

·         Rogalski, A. (2011) Infrared Detectors. 2nd edn. Boca Raton: CRC Press.

·         NATO (2008) STANAG 4347: Measurement of Minimum Resolvable Temperature Difference (MRTD) of Thermal Cameras (title wording varies by edition). Brussels: NATO Standardization Office.

·         ISO (various years) Standards on atmospheric attenuation/visibility measurement and optical testing methods (use the specific ISO you apply in your test method section).

By Tinus Diedericks

CEO of Timeless Technologies