Using a PIR sensor to trigger your surveillance? Probably best you don't read this then.

A researcher sets a camera at the edge of a pond, 2.5 metres from the water's edge, to record otters as they emerge. Over dozens of passes, the camera trap catches barely one in three. On the video setting, well under two-thirds of those triggers even register the animal in frame. Under the worst conditions, only 14% of passes end up as usable footage of an animal that was, unmistakably, there (Findlay et al., 2020).

That gap, between what happened and what got recorded, is the story of passive infrared (PIR). It's the sensor behind most "motion-triggered" cameras in use today, from wildlife monitoring to perimeter security. And decades of peer-reviewed testing show it has some very specific, very measurable blind spots.

A sensor built on 250-year-old physics

The effect PIR relies on is older than photography and was first documented in 1707. But it took until the early 1970s for that effect to become a security product, when the first infrared intrusion detector patents described a device that could flag intruders by sensing changes in ambient infrared radiation rather than emitting any signal of its own. That basic concept underpins essentially every PIR sensor built since, from burglar alarms and streetlights to the wildlife camera traps discussed throughout this piece: all inheriting the same basic architecture more than fifty years on.

What it actually measures

A pyroelectric crystal sits behind the sensor's lens. As a warmer or cooler object moves across the monitored zone, the changing infrared exposure induces a small, measurable electric charge on the crystal's surface. The sensor never forms an image. It has no concept of shape, colour or identity. It's watching a single value: whether the infrared hitting the crystal is changing fast enough, and by enough, to resemble something moving through the zone (Welbourne et al., 2016).

That's the entire input the sensor has to work with. If a subject isn't a few degrees warmer or cooler than its background, there's nothing to react to, no matter how large, close, or clearly "there" it is. Most studies put the practical threshold at around 2.7°C of contrast (Meek et al., 2012, cited in Findlay et al., 2020), comfortably met by a dry mammal on a cool day, and not remotely guaranteed for anything else.

Even in the best case, it misses one in five

Findlay, Briers and White (2020) ran CCTV alongside camera traps to measure exactly how often a genuine pass failed to produce a genuine trigger. Their finding: at least 20% of passes were missed across every species and scenario tested, including at close range, where detection is assumed to be near-certain.

Speed compounds the problem twice over. Faster subjects are both less likely to trigger the sensor and less likely to still be in frame once the camera activates. Trigger latency on video mode runs to a couple of seconds: enough time for anything moving briskly to be gone.

Small, slow or cool subjects are close to invisible

The temperature-contrast requirement means PIR is structurally weak against anything that doesn't run hot. Hobbs and Brehme (2017) compared a purpose-built alternative trigger against standard PIR units for amphibians, reptiles and small mammals, and measured an overall detection probability of 1.0 for their system against just 0.26 for PIR.

And it fails in the other direction too

The same sensitivity that misses real targets also fires on the wrong ones. Wind-driven vegetation, shifting shadows and solar heating of background surfaces are well-documented sources of false triggers. In one field trial of an Australian motion-based camera trap system, 75% of captured segments were false positives caused by wind moving shadows, leaves or grass through frame (Corva et al., 2022). A 2025 study quantified this precisely for conventional PIR units: at air temperatures of 30°C or higher, the probability of a false-positive trigger ranged from 0.10 right up to 1.00: essentially anywhere between "occasional nuisance" and "certainty of noise" (Kaltenbach et al., 2025).

The same physics, tested against people

Everything above comes from wildlife ecology, but the underlying physics doesn't change when the subject is a person rather than an animal, and the security and defence literature has tested it directly.

The standard reference in that field, Garcia's Vulnerability Assessment of Physical Protection Systems, frames PIR detection around a single variable: the "minimum resolvable temperature" difference between an intruder and the background. An intruder at the same temperature as the background is, in Garcia's words, invisible to the sensor. Slow, deliberate movement fails to produce enough rate-of-change across the detection zone to register at all. One industry technical reference calls a target walking straight toward the sensor "the most difficult test for a PIR motion sensor" (Katz, 2008).

This is not theoretical. A 2018 U.S. Department of Defense technical inquiry cites a field trial run by the U.S. Army's Cold Regions Research and Engineering Laboratory in which Special Forces soldiers attempted to defeat operational PIR intrusion-detection systems using nothing more than cold-soaked burlap and improvised poncho shelters. Under certain site conditions, the report notes, the countermeasures were "very effective," letting soldiers pass both the PIR systems and observers watching thermal imagery, undetected (Peck & Lacombe, 1998, cited in Garmon, 2018).

Detection range has the same ceiling

The same inverse-square falloff in emitted infrared that limits close-range triggering also caps how far a PIR sensor can be trusted at all. Commercial intrusion-grade units surveyed in the DSIAC review top out under 25 metres. Honeywell's IS3050, one of the more widely deployed models, is rated for a coverage area of roughly 16 by 22 metres (Garmon, 2018). Outside that rated zone, or off to the side of it, there is categorically no detection: the sensor has no way to resolve shape or distance, only a threshold crossing within a fixed lens geometry.

This is exactly why formal testing standards exist for higher-stakes deployments. Exterior intrusion sensors protecting U.S. national security facilities are evaluated against two standard figures (Probability of Detection (Pd) and Nuisance Alarm Rate (NAR)) against policy thresholds set by the Department of Energy and Department of Defense (Garcia, 2007; sensor testing standards maintained at Sandia National Laboratories). PIR sensors rarely clear that bar unassisted, which is precisely why government- and defence-grade perimeter security has moved to sensor fusion (PIR paired with microwave, radar, or video analytics) rather than relying on PIR's native range and geometry alone.

There's a second, quieter range problem: PIR performance doesn't hold steady even within its rated zone over time. A 2022 study run under the U.S. Department of Energy's Solid-State Lighting program found that after accelerated aging, tested PIR sensors showed field-of-view shrinkage of up to 21.5% (U.S. DOE SSL Program, 2022). Range isn't just a spec-sheet number carved in stone: it degrades with the hardware.

What changes when the system can classify, not just detect

That same 2025 study is one of the first to put an edge-AI-enabled camera head-to-head against conventional PIR units in the field. The result was telling: at air temperatures of 30°C or above, the AI-enabled unit's false-positive probability dropped to nearly zero, while the conventional PIR units ranged across that full 0.10–1.00 band (Kaltenbach et al., 2025).

The clearest evidence for what's possible once a system moves from "did something change" to "what is that" comes from Norouzzadeh et al. (2018), published in PNAS. Applied to 3.2 million camera-trap images from the Snapshot Serengeti project, deep learning classification reached 93.8% accuracy: matching the 96.6% accuracy of crowdsourced human review teams, while automating 99.3% of the identification workload. That's not a marginal improvement on triggering. It's a different category of capability.

The takeaway

None of this is a knock on the engineers who designed PIR triggers. For a single, cheap, battery-sipping sensor solving a narrow problem, it's a genuinely clever piece of physics, and it will keep a legitimate place in low-power deployments. But the research is consistent and now spans wildlife ecology, conservation engineering and security literature: a sensor that can only ask "did the temperature change here" will always trade missed detections against false alarms, and it will always struggle with anything small, slow, cool, distant, or moving through wind-blown vegetation.

For government operations, conservation, security, pest control and primary production alike, that trade-off shows up as the same problem: a system that reacts to noise instead of understanding its environment. For anyone relying on these systems: on a fence line, in a reserve, on a work site, the question worth asking isn't which sensor to buy. It's whether the system understands what it's looking at.

References

• Berman, H. L. (1972). Infrared Intrusion Detector System. US Patent 3,703,718. https://patents.google.com/patent/US3703718A/en

• Lang, S. B. (2005). Pyroelectricity: From ancient curiosity to modern imaging tool. Physics Today, 58(8), 31–36. https://physicstoday.aip.org/features/pyroelectricity-from-ancient-curiosity-to-modern-imaging-tool

• Findlay, M. A., Briers, R. A., & White, P. J. C. (2020). Component processes of detection probability in camera-trap studies: understanding the occurrence of false-negatives. Mammal Research, 65, 167–180. https://doi.org/10.1007/s13364-020-00478-y

• Welbourne, D. J., Claridge, A. W., Paull, D. J., & Lambert, A. (2016). How do passive infrared triggered camera traps operate and why does it matter? Breaking down common misconceptions. Remote Sensing in Ecology and Conservation, 2(2), 77–83. https://doi.org/10.1002/rse2.20

• Randler, C., & Kalb, N. (2018). Distance and size matters: a comparison of six wildlife camera traps and their usefulness for wild birds. Ecology and Evolution, 8(14), 7151–7163. https://doi.org/10.1002/ece3.4240

• Hobbs, M. T., & Brehme, C. S. (2017). An improved camera trap for amphibians, reptiles, small mammals, and large invertebrates. PLOS ONE, 12(10), e0185026. https://doi.org/10.1371/journal.pone.0185026

• Corva, D. M., Semianiw, N. I., Eichholtzer, A. C., Adams, S. D., Mahmud, M. A. P., Gaur, K., Pestell, A. J. L., Driscoll, D. A., & Kouzani, A. Z. (2022). A Smart Camera Trap for Detection of Endotherms and Ectotherms. Sensors, 22(11), 4094. https://doi.org/10.3390/s22114094

• Kaltenbach, T. L., Mosley, J. C., McNew, L. B., & Beaver, J. T. (2025). Can edge AI mitigate environmental effects on camera trap performance? Wildlife Society Bulletin, 49(3). https://doi.org/10.1002/wsb.1598

• Norouzzadeh, M. S., Nguyen, A., Kosmala, M., Swanson, A., Palmer, M. S., Packer, C., & Clune, J. (2018). Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning. Proceedings of the National Academy of Sciences, 115(25), E5716–E5725. https://doi.org/10.1073/pnas.1719367115

• Garcia, M. L. (2007). Vulnerability Assessment of Physical Protection Systems. Butterworth-Heinemann/Elsevier.

• Garmon, P. (2018). Defeating Passive Infrared Motion Sensors. Defense Systems Information Analysis Center (DSIAC) Technical Inquiry, prepared with subject matter experts from the Georgia Tech Research Institute. https://dsiac.dtic.mil/technical-inquiries/notable/defeating-passive-infrared-motion-sensors/

• Peck, L., & Lacombe, J. (1998). Environmentally Dependent Countermeasures to Passive Infrared Detection. U.S. Army Cold Regions Research and Engineering Laboratory (cited in Garmon, 2018).

• Katz, F. (2008). Select the Perfect PIR Motion Sensor. Security Dealer & Integrator, 30(5), 90, 92–93 (cited in Garmon, 2018).

• U.S. Department of Energy, Solid-State Lighting Program (2022). Evaluation of the Reliability of Passive Infrared (PIR) Occupancy Sensors. https://www.energy.gov/sites/default/files/2023-07/ssl-rti_pir-sensors_mar2022.pdf