Dahua Thermal Temperature Monitoring System Tested

By Ethan Ace and John Honovich, Updated Nov 19, 2020, 08:25am EST

Dahua's thermal temperature monitoring system has gained mass attention as a fever detection solution for organizations around the world, including Amazon.

IPVM Image

As such, it certainly works, at some level, but how well does it work to detect fevers (or as they market it 'elevated skin temperature detection')? And what does Dahua do to estimate human temperatures?

We bought and tested the Dahua system. Inside this report, we explain:

  • Dahua's 'confidential' 'compensation algorithm' and the practical impact of this controversial technique
  • How Dahua delivers such 'accurate' readings even under bad conditions
  • How Dahua misses 'elevated skin temperature' detections
  • How software setup impacts performance and measurements
  • How Dahua's engineering recommendations conflict with Dahua's sales and marketing


New Firmware Reduces Rigging System

New firmware has significantly corrected Dahua's thermal measurement system delivering near-"normal" readings in poor conditions, raising low measurements to close to body temperature.

For example, in our original test, Dahua delivered a nearly-normal 97°F measurement of a subject zooming by on a skateboard. However, in new firmware, the system delivers extremely low measurements, and more sporadically than before.

IPVM Image

Similarly, in our original test, the Dahua camera produced a 96.8°F measurement of a subject whose forehead and inner canthus was covered by cardboard. However, new firmware performs differently to provide more realistic measurement in key ways:

Get Notified of Video Surveillance Breaking News
Get Notified of Video Surveillance Breaking News

  • First, the camera fails to detect faces if the eyes are not visible at all, so in most cases subjects' inner canthus is visible, resulting in a measurement similar to those taken without the cardboard in place.
  • Second, when switched into Dahua's new forehead-only mode, the camera produces drastically low readings nowhere near body temperature, e.g. 77.7°F seen below.

Both these factors are significantly different from Dahua's original performance, where the camera shifted to measure the subject's nose, shown below.

IPVM Image

New Firmware Versions

While Dahua refuses to explain, the first firmware we noticed these differences was dated 2020-07-06. A more recent release, dated 2020-10-15, performs similarly. We cannot confirm specifics of changes made in these versions, as Dahua does not publish release notes and did not respond to repeated requests for comment.

Firmware versions from this and older tests are as follows:

  • Original test firmware: V2.631.0000000.23.T, Build Date: 2020-04-03
  • July firmware: V2.631.0000000.31.T, Build Date: 2020-07-06
  • October firmware: V2.631.12JW000.0.T, Build Date: 2020-10-15

System Vs. Core Firmware

Dahua now uses two separate firmware for their thermal measurement cameras, referred to as "System" and "Core", which must be installed sequentially when updating cameras.

IPVM Image

Users may find their "Core" version in the camera's UI, below the software version:

IPVM Image

It is unknown exactly what each of these firmware components impacts, as Dahua does not document it publicly and has not responded to our requests for comment.



Our testing of Dahua’s system found that it produced remarkably “normal” human body temperature readings, even with hard to read conditions like occlusions (hats, masks, hands, hairs, while walking, and even partial headshots). This creates a feeling of comfort and safety that was impressive. Moreover, Dahua could not be fooled by hot water bottles on foreheads at 110°F, for example.

Missed Issues

On the other hand, our testing found Dahua regularly missed elevated skin temperatures of 100 - 102, raising a serious risk about the system missing many people with fevers, especially when subjects were moving, wearing hats, or moving next to others.

Compensating Algorithm

Moreover, our testing revealed that the remarkably normal temperature readings Dahua produces are artificially enhanced by what Dahua described to us as a ‘compensating algorithm’ that we found dynamically increases the offset to transform low / bad reads into more normal temperature readings.

Missed Elevated Temperatures

Our testing found that Dahua regularly missed subjects with elevated skin temperature in scenarios shown in their own marketing, simply walking and even more so when wearing hats and sunglasses, ski masks, multiple subjects, etc.

To simulate fever, our subjects elevated the temperature of the forehead using a hot water bottle in the low 100s° F. The subject was then measured by the Dahua camera, and finally by a handheld IR thermometer (Extech IR200).

IPVM Image

While health and safety measures prevented measuring actual sick or coronavirus infected people, this approach is similar to what manufacturers like Dahua use to prove their systems can alarm. Our approach differed by controlling the heated temperatures to realistic fever levels and verifying those.

Our methodology is conservative in that (1) we heated the entire forehead and (2) we had Dahua measure the subject before we use the IR gun and in the intervening few seconds between Dahua and the IR gun's measurement, the subject will, minimally, cool slightly.

In the video below, we show examples of subjects with fevers being missed, due to hats, ski masks, hair, and simply low readings by the Dahua camera.

In the first example, after heating his forehead with a hot water bottle, the Dahua camera measures the subject at 99.6°F, while he is measured via thermometer at 100.6°F, a full degree difference.

IPVM Image

Directly following two others through the scene, the subject is measured at 99.9°F by Dahua, but again 100.6°F using an IR thermometer, over the camera's alarm threshold.

IPVM Image

Wearing a hat and sunglasses, as shown in Dahua marketing, the subject is measured by Dahua at a max of 100°F (under the camera's alarm threshold), but a much higher 102°F using the IR thermometer.

IPVM Image

Similarly, wearing a knit hat/beanie, the subject was measured at 98.5°F, but 101.5°F when removed.

IPVM Image

Dahua also claims to detect elevated temperatures when wearing ski masks/balaclavas. However, the subject below was measured by the Dahua camera at 98.4°F while covering his forehead, which increases to 101.5°F when removed.

IPVM Image

Finally, simply having hair covering the forehead prevented proper measurement of elevated temperatures. As the subject enters the scene below, Dahua measures his temperature at 99.2°F, but 102°F when his hair is pushed aside.

IPVM Image

Response From Dahua on Misses

Below is Dahua's response, in full, on the misses:

> Big temperature difference while object is moving

After days of test, we cannot seems to replicate this issue, need to identify what’s the cause, attached is two video taken in both R&D and field.

> Regular face with heat applied, not detecting elevated temperature

We really cannot replicate this issue, temperature difference might be exist, but the video show more than 1 degree difference, need to identify what’s the cause

> Baseball hat sun glass, beanie not detecting elevated temperature.

Understand and agree with your point, north America will update our marketing material to reflect on detection on conditions where the eye canthus is showing.

We are confident is the person with beanie is really on an elevated temperature, we can detect it since his eye canthus will show elevated temperature instead of just forehead.

"Compensating Algorithm" Investigated

Dahua acknowledged they used a "compensating algorithm" in response to our questions about why the temperature offset varied between the actual surface temperature of objects and the reported human body temperature when Dahua detected a face. Dahua explained:

We have shared a lot of information about how the technology is, how our solution works, how to make necessary setup to show best test result in the past couple of emails.

If you are asking specifically about the algorithm, this is a highly confidential intellectual property of our company, I hope you understand why we cannot share any further details.

You can think like this, a thermal sensor can work out of the box, anybody can buy a thermal sensor and get read outs from it, what makes that thermal sensor to a body temperature solution is the compensating algorithm. That’s why not every thermal company has human body temperature solution.

While the algorithm is confidential, we were able to determine how it acts under varying temperatures. The lower the temperature reading Dahua has of a face, the more it increases the offset to make the reading look closer to 'normal' or statistically average human temperature.

IPVM Image

We were able to uncover this because Dahua does not have liveness detection and thinks printouts of faces are actual human faces. Using that, we placed controlled heat sources behind the printout face (e.g., a blackbody, that controlled the heat level behind the printed face) to determine how Dahua would report specific surface temperatures to as 'human' temperatures, as we demonstrate in this video below:

Though this is certainly very clever in that it increases user confidence that the system is working well, this is deceptive because it transforms low / 'bad' reads into reads that look good. And because literally 98% of humans have temperatures between 96°F - 99° F, to a casual observer it looks like it is 'working'.

Examples Bad Reads Become More Normal Temperature

While thermography systems tend to generate low readings on bad angles or positioning (e.g., 91°F, 92°F, 94°F, etc.) Dahua never did so. This is consistent with our research showing that Dahua boosts bad reads into nearly normal measurements.

For example, even when the subject covered the top half of their face, obscuring the eyes and forehead (the warmest parts of his face), Dahua displayed a 96.8°F measurement.

Similarly, in an even less realistic scenario, the subject glides past on a skateboard, detected and measured by Dahua at 97°F.

Finally, when simply looking down or to the side, even at harsh angles, the camera continues to produce nearly normal measurements. In the example below, the subject is measured at 97.9°F when looking down, increasing to 98.8° when looking up, just a 1°F difference.

Subject Temperature Similarities

In our tests, Dahua tended to measure subjects in a very small temperature range, while IR gun thermometers varied much more. We tested this using 12 subjects of varying gender and age (ranging from infant to middle age). When walking, 8 of 12 subjects were measured within the range of 97.3-97.7°F. However, when measured using an IR thermometer, those same subjects varied by 1.2°, from 98°F to 99.2°F, a much greater difference.

IPVM Image

Moving Vs. Standing Measurement Differences

During testing, subjects were frequently measured significantly lower when moving than standing, with temperatures increasing slowly the longer subjects remained in the scene. For example, when walking normally, the subject below was measured at 97.6°F. However, when remaining still for 3-4 seconds, his temperature is measured at 98.2°F.

IPVM Image

These effects were seen on all subjects, with some warming faster than others when remaining in the scene.

Camera Configuration Impact On Accuracy

Several settings in the Dahua thermal camera are important for measurement accuracy, including blackbody setup, human temperature compensation, and ambient temperature. These settings should also be regularly reviewed to ensure accuracy as human temperature and ambient temperature change throughout the day and blackbodies may be bumped or knocked out of alignment.

We review these settings in this video:

Engineering Recommendations Conflict With Sales and Marketing Claims

Our testing also revealed a conflict between what Dahua engineering recommends and what Dahua sells and markets publicly, particularly as it comes to people moving and people's heads being covered, two of the most distinctive claims compared to conventional thermal providers.

While Dahua's marketing shows various examples of people's heads being covered, Dahua engineering objected to us using them. For example, from Dahua's website [UPDATE Dahua removed the ski mask, sunglasses and hat and helmet images after IPVM called it out]:

IPVM Image

Dahua's engineering team understandably responded, "since the heated area(forehead) is covered the camera physically couldn’t get a proper reading". We covered the forehead to the same level as Dahua's marketing images. So while we agree with Dahua engineering that these coverings negatively impact accuracy, Dahua marketing had prominently claimed they can deliver accurate readings with such coverings.

Likewise, Dahua has publicly repeatedly touted its ability to do readings with people walking in a straight line one after the other, most notably in their marketing video that has received 150,000+ views:

However, Dahua's installation instructions and recent 'whitepaper', confirmed by Dahua engineering to IPVM, contradicts the straight line, one after the other, approach. Instead, Dahua calls for

IPVM Image

Moreover, the instructions emphasize to use it only indoors and even indoors not aimed at entrances and to use a queuing area, as shown above and below:

IPVM Image

However, in Dahua's case study video, they show an outdoors setup, excerpted below:

IPVM Image

Consistent with Dahua's installation instructions and contrary to the case study, we generated numerous false alerts, as the example below shows:

IPVM Image

The Dahua system had read the temperature of the pavement directly behind the person's head which was ~100°F on a ~70°F day.


Dahua's "Thermal Temperature Monitoring Solution" sells online for ~$12,000-13,000 USD, total, similar to body temperature measurement systems from Hikvision and Sunell OEMs (~$12,000-15,000 USD).

Individual prices for components (online):

  • DH-TPC-BF5421-T camera: ~$8,300 USD
  • JQ-D70Z blackbody calibrator: ~$3,600
  • DHI-NVR5216-16P-I 16-channel NVR: ~$1,000

Physical Overview

The DH-TPC-BF5421-T camera is physically similar to other Dahua bullets, with the addition of the second thermal lens/imager. The camera contains a white light illuminator and speaker, which trigger on over-temperature alarms by default.

The JQ-D70Z blackbody is similar to others in our tests, such as Hikvision and Sunell, about 4" square by 7" long, and adjustable from 30°C to 50°C (86° - 122°F), though defaults to 35° (95°F).

We review the construction of the camera and blackbody in this video:

Live Monitoring Temperature Events

The simplest means to view temperature events from the Dahua camera is using the local interface of their NVR (DHI-NVR5216-16P-I), which displays events along with statistics as temperature measurements occur (demonstrated below). The camera/NVR web interface or Dahua's Smart PSS client do not display these events, with the face bounding box turning red the only indication of an over-temperature event. This makes monitoring more difficult as temperatures frequently only shifted to alarm levels for one or two frames in our tests, which could be easily missed if viewing only a live stream.

These events are integrated to Dahua's DSS Express VMS, but this requires an additional license, unlike their Smart PSS CMS client.

ONVIF Conformant

In addition to monitoring methods above, the DH-TPC-BF5421-T is ONVIF Profile S and T conformant and visible and thermal streams integrated to Avigilon, Exacq, Genetec, Milestone, and Network Optix VMSes in our tests with no issues. However, only Milestone integrated temperature events, via an add-on plugin.


Dahua's remarkably normal temperature readings do a great job of giving people peace of mind but our testing shows that it comes with a significant risk of missing real fevers. For those who want the convenience of allowing people to wear whatever they want and move as quick as they want, without prioritizing fever detection, Dahua excels.

22 reports cite this report:

Directory of 237 "Fever" Detector Suppliers on Jan 21, 2021
This directory provides a list of "Fever" scanning providers to help you see...
2021 Video Surveillance Cameras State of the Market on Dec 28, 2020
Each year, IPVM explains the main advances and changes for video surveillance...
FLIR Finds Fault In Temperature Screening Cameras on Dec 02, 2020
FLIR has found a hardware fault in some thermal screening cameras which may...
Dahua Rigs Fever Cameras, Covers Up on Nov 20, 2020
IPVM testing has verified Dahua rigged their fever cameras, with the company...
Six Flags' FDA Violating Outdoor Dahua Fever Cameras on Oct 26, 2020
As Six Flags scrambled to reopen parks amid plummeting revenues caused by the...
Bosch Announces "Human Skin Temperature Detection" Camera on Oct 07, 2020
Bosch announced its "Human Skin Temperature Detection" system at the 2020...
Avigilon Elevated Temperature Detection Camera Tested on Sep 17, 2020
Avigilon has entered the temperature screening market with the release of...
Beware Rigged China Fever Cameras on Sep 08, 2020
Many China fever camera manufacturers have rigged algorithms dynamically...
Thermology Expert: "95-99%" Doing Fever Screening Wrong, Unjustified Compensating Algorithms "Insane" on Aug 27, 2020
A thermology expert tells IPVM "95 to 99% of people" are doing fever...
US Surgeon General Unwittingly Showcases Sanctioned Dahua Temperature System on Jul 28, 2020
The US' top public health spokesperson, the Surgeon General, posted a photo...
Wrong Dahua Australia Medical Device Approved on Jul 20, 2020
Dahua's body temperature system is now in Australia's medical device...
Dahua Temperature Measurement Station Tested on Jun 17, 2020
Dahua hit the market hard with its original 'fever' camera (tested here)....
Mobotix Thermal Detection Camera Tested on Jun 09, 2020
For years Mobotix has struggled but now sales are surging driven by Mobotix's...
FLIR A Series Temperature Screening Cameras Tested on Jun 04, 2020
FLIR is one of the biggest names in thermal and one of the most conservative....
"Fever Camera" Online Show June 2020 - On-Demand Recordings on Jun 03, 2020
IPVM has successfully completed the world's first "Fever Camera" show....
Seek Scan Thermal Temperature Screening System Tested on May 28, 2020
Now that IPVM has tested Dahua, Hikvision, and Sunell, we are returning to...
ZKTeco Body Temperature and Mask Detection Reader Tested on May 26, 2020
While dedicated fever cameras emerged first, now tablet/kiosk fever detectors...
Hikvision Temperature Screening Tested on May 20, 2020
Hikvision has ramped up the promotion of its 'temperature screening' system,...
Kogniz Health Thermal Screening System on May 15, 2020
Kogniz, a Silicon Valley AI surveillance company is claiming to offer...
Sunell Panda Cam Body Temperature Measurement Camera Tested on May 14, 2020
Sunell is far less well known than its gargantuan domestic competitors Dahua...
Dahua Faked Coronavirus Camera Marketing on Apr 01, 2020
Dahua has conducted a coronavirus camera global marketing campaign centered...
Detecting Coronavirus Fevers With Thermal Cameras on Mar 15, 2020
MAY 2020 Update: This post was our early examination of these systems being...

Comments (78)

Only IPVM Members may comment. Login or Join.

I want to thank our team for working hard over the past 10 days to get this test published. I also want to recognize Dahua’s engineering team who repeatedly provided feedback and input. As we mention in the post, we found a clear divergence between what the engineering team understands and what the sales and marketing team is selling.

This is a polarizing topic as it is arguably the hottest growing industry product and also the most controversial ever. Regardless of one’s position on this, I ask you to be respectful in the comments.

For those who are selling Dahua’s systems, if you have criticisms of our test, that is welcome. In particular, if you believe the methodology is flawed or we missed other test scenarios, please post them in the comments and we are happy to review each one. We own this equipment and we also bought the Hikvision one (testing in progress) so we certainly plan to do more testing, comparative testing, more scenario testing (e.g., if it ever gets warm in PA we can test how summer heat impacts these systems, etc).

IPVM this is great work, and an analysis sorely and urgently needed. Who else is doing anything like this? Nobody....

Hello there

We may do the work from a statistical aspect very easily. But we cannot buy a camera and even if we could it would need to be tested in a hospital or a place where patients are coming in.

To do a medical study you need a lot of efforts and money beyond buying a camera. IPVM made a nice work given their potential, I agree but again testing accuracy is a special study and this has not to be done yet by any company.

Please read my doc if you wish to see how a study like this looks like.

Thermal Camera comparative tables - Fever Accuracy definitions

We've had great success selling this product and every customer has been extremely satisfied! We also have one setup on demo in our showroom for local customers and everyone that has tried it has loved it. No complaints here!

Brandon, thanks! How many fevers and or false alerts positives have they had?


Hi John,

One feedback I got from a dahua partner is that they placed the dahua system on test in a hospital which receives 30-50 covid patients on an average.

What the partner told is that ,when testing people with normal temperature the systems works perfect.

But if the system showed multiple with fever to be normal temperature and this was considered to be unacceptable.

He seemed confused how normal temp is perfect and higher temp is not detecting.

After reading this article ,now it makes sense.

Would the customers be "extremely satisfied" if they understood the product was falsifying the readings?

Meth dealers can say the same thing; "great selling success and every customer has been extremely satisfied"

So just to get this straight because the conclusion above seems too generous...

In summary, Dahua thermals use an algorithm which alters information captured when it detects a face, and is thus inherently inaccurate, read: unreliable?

In summary, Dahua thermals use an algorithm which alters information captured when it detects a face, and is thus inherently inaccurate, read: unreliable?

From what I have seen in the field all of these thermal camera based systems are pretty inaccurate, FDA 510K or not. What will be very controversial in this report is that Dahua appears to be skewing the results to make it appear more accurate in their software when a face is detected.

This article contains great testing / investigation by the IPVM team. IMO this is among the greatest reports of all time.

Dahua appears to be skewing the results to make it appear more accurate in their software when a face is detected.

Dahua's counterpoint was that IR guns do that as well (not detecting face, but varying the offset in 'human' mode). As we mentioned in the report, we did find that in the Extech IR200 but a much smaller range than Dahua did.

More importantly, historically, these thermal 'fever' cameras with FDA approval did straight thermography (i.e., reading the actual temperature measurements), not estimating based on the thermal input (i.e., the read is low but we think it's a human so let's boast it to the levels almost all humans are).


Dahua is accurate in that 99% of people don't have a fever, so not detecting most subjects as "normal" temperature will be mostly correct in the vast majority of cases.

However, shown in the multiple examples above, the camera was unreliable in our tests in detecting the small number of people who may have a fever, whether outright missed or obscured.

i’m sure this will all be fixed in the next firmware version, once Ethan’s and Derek’s facial profiles are permanently baked into all subsequent releases :)

In summary, Dahua thermals use an algorithm which alters information captured when it detects a face, and is thus inherently inaccurate, read: unreliable?

No, it's pure AI :)

Right. It is Actually Inaccurate.

because Dahua does not have liveness detection

Just curiously, has anyone done a good job at liveness detection?

Just curiously, has anyone done a good job at liveness detection?

Anyvision's tablet uses stereographic imaging (2 cameras) to generate a 3D model of the face to verify it is not a flat picture. We have yet to see the actual unit, just sales pitches. "Liveness" in this case seems like it could be defeated by a statue/bust of some kind.

seems like it could be defeated by a statue/bust of some kind.

ZKTeco SF1008+

Really useful report, as usual! Looking forward to seeing if anyone actually does a great job in this arena!

Was there a physical exertion test done, where a person exercised, built-up heat in a normal way in their body and then got measured in front of the thermal camera?

Was there any occasions where the camera actually reported the elevated skin temperature correctly?

We have done some testing with physical exertion, mostly running laps around the building, but it has never triggered a temperature alarm. One thing to keep in mind as people are physically exerting themselves is that they tend to perspire, which results in them being measured lower than they otherwise would be.

The camera detected elevated skin temperature on several occasions (100.3°F and up) while we were testing the scenarios in the report. It was more frequent when forehead temperature was >101°F, misses were more common closer to the threshold.

I will pay money to see videos of Ethan running laps around the IPVM building.

"Compensating Algorithm

Moreover, our testing revealed that the remarkably normal temperature readings Dahua produces are artificially enhanced by what Dahua described to us as a ‘compensating algorithm’ that we found dynamically increases the offset to transform low / bad reads into more normal temperature readings."


I guess there is none "Compensation algorithm" in sense as described in review. Just tester omit fact that paper on which they printed a face has too different emissivity than skin. So if you measure heated paper with emissivity about 0.7 and camera think it must have emissivity at least 0.95 than result is very different.

Edit: Originaly I guess you did contact mesurement to compare measurement with real surface, now I see you are comparing two contactless devices. It is much more problematic when measuring using two contactless devices which has different emissivity setup (it will not be beig in numbers, something in range 0.93-0.99) and mesuring almost unknown emissivity surface which we guess is around 0.7. All others I mention originaly is same:

If so, you did in fact what every contactless thermometric measurement need which is calibration and linearization. I put it into datasheet and calculated that Dahua simply using emissivity set to 0.95 (correct!) and using linearization multiplication ~0.3 + bias factor ~200. Really none artificial results synthesis, just simple sensor calibration and linearization.Do not forget than all curves are power 4 dependant! Stefan-Boltzman equation for radiated energy depends on emissivity (linearly) and on absolute temperature (hardly non-lineary). I suggest you to repeat this test but on heated area glue some material with emissivity similar to human skin. In my experiments very well working piece of jackets from pig skin, but should be ok any organic materials.

IPVM can you please provide me real test data of this "Compensating algorithm"? Then I can recalculate and provide data in .xls to exactly show you that there is probably nothing artificial on those measurements.

Outside of emissivity, the concern we raised is simply how aggressively Dahua attempts to "normalize" measurements when it sees a face. It could be a paper face or a human face, results are similar in that the camera attempts to present a body temperature close to average when it is not the case. We actually saw very similar results when using humans to measure results of their algorithm, as well. Someone standing outside on a cold day or in front of air condition may show surface temperature of ~89-90°F measured on their face, but be shown by the camera at 97-98°F, very similar to the paper Rob face. Estimating core body temperature from 90° skin temperature is subject to far too many variables to be much more than a guess.

Regarding emissivity: paper varies significantly. We used white bond paper, heavier weight inkjet paper, which various sources I've seen show as having emissivity between 0.9-9.95, vs. 0.95-0.99 for human skin, so not terribly different. Other papers vary from 0.6-0.9 from what I've seen, but this is close.

Finally, it's worth mentioning that we demonstrated this to Dahua engineering and they did not object to our test method.

Richard, to emphasize on Ethan's response, the concern is how the offset varies. If the offset was the same at every temperature, that would not be a concern. And even if the emissivity was lower, whether a little or a lot, the difference would be the same across temperatures measured, it would not diminish as it got closer to 'normal' human body temperature.

Just today we made own tests on this Dahua system with small "blackbody" which we build - it is temperature stabilized piece for +/-0,2degC with correct emissivity layer which is precious in range about 35-39degC. Engineer simply can wear it on their forehead and freely walk in detection area, it is battery powered.

And system properly alarming if we set this piece over trigger level 37,5C (99,5 F) and does not if we are bellow tolerance limit. This blackbody allow us to prove as near as possible correct system behavior. And because is small enough it completely mitigate AI filters from any unwelcomed "guessing". Visually thermal part of system cannot recognize if we have such piece on head, it is about 5x3cm flat and in thermal there is practically none edge. Can show you more, well it is still prototype, do not expect much high tech. We build it exactly for reason to prove for customers that the system is doing his job for hospitals.

What Ethan mention - yes, it all can be correct but without any artificial compensation, say without any statistical method (like f.e. Volkswagen developed to pass lab emission tests) . It results from calculation I tried to briefly make, because graphics about "compensation algorithm" looks alarming. However when I put there data which are really around "working point" which is about 36-38C (96,8 - 100,4F) the data starts to make sense. I still think there is nothing such "algorithm", even it looks so when we are a bit more far from working point.

John exactly offset plays important role but there is couple of point which need to be calibrated on every sensor and where little emissivity difference can play significant role. Dahua definitely has know-how how they calibrated equation, but this is about how they fit nonlinear sensor behavior to correct values. So I believe that if you repeat "static face test" you will see that what original looks like AI job is noise, liner shift, calibration job etc.

Ok, maybe AI add there some filter which f.e. dropping away some non-sense results and doing it on some geometrical rules. Yes, can be still possible (and it is welcomed in fact). However our test above did not noticed that this can have any negative effect on main functionality.

Really "compensation algorithm" scared me - nothing can be worse that safety sensor which is guessing results! However this need strong observations to prove it. So more detailed measurements to verify such are necessary.

Just today we made own tests on this Dahua system with small "blackbody" which we build - it is temperature stabilized piece for +/-0,2degC with correct emissivity layer which is precious in range about 35-39degC. Engineer simply can wear it on their forehead and freely walk in detection area, it is battery powered.

Richard, cool! Can you send a picture of what this looks like? My first thought is it looks something like this:

IPVM Image

As for:

you will see that what original looks like AI job is noise, liner shift, calibration job etc.

What we have seen on multiple systems is that errors or mistakes tend to be mean seeking, i.e., these systems tend to make reads lower than average (i.e. 37°C) get higher and ones higher than average go lower.

Yes :-) Just we need wires - one is PT100 sensor and second is heater/cooler. Regulator and powerbank holds engineer in hand.

IPVM Image

Reasons for "mean seeking" by Dahua can be at least couple. What I suspect as reason is: doubling measurement error made by using another contactless device as reference; not giving enough test objects in range above 37C.

Great test and article. We have a number of customers very strongly looking into this but we don't want to make them a guinea pig. From what we have found, the Viper Imaging system seems to be the most accurate in multiple conditions. We need high, accurate throughput and quick reading of the tear ducts with a blackbody is the best option so far. Of course this comes at a substantial cost.

Mark what kind of substantial cost are we talking about? How much better is Viper's accuracy?

A live demo as well as their marketing video showed about a 1-2 second read targeting the tear ducts. Not for crowds but we were impressed with accuracy and speed. Regarding cost I believe their pricing is on the website.

Very interesting test, I find the presence of a Compensation algorithm particularly interesting.

In your testing try walking straight up to the camera so your within 1m, we found this gave very inaccurate readings.

But if you set the camera at an angle so the subject never gets within 1m it works well.

We also found all the problems you have seen with the environmental setup.

But we did not see the level of missed high temperature readings you did.

We will carry out further testing as this is a little worrying.

Heathrow just announced thermal screening - does anyone know what company they are using?

The one thing I notice with all IPVM testing is that the blackbody parameter box isnt precise. I was told from Dahua (and some instructions that they gave me) that the box should be drawn within the heat source and no lines should be touching the outter edges. I feel like the blackbody is what give this system the .3 C accuracy. That being said, is it perfect? No. Can it be tricked? Yes. We have been selling it as honest as possible saying it is a first line defense screening tool, and that further protocols should be put in place if an alarm is triggered.

I feel like the blackbody is what give this system the .3 C accuracy.

At this point I think it is pretty clear that the unit is not accurate at all. You could argue whether it is the blackbody, or the compensation algorithm, that gives it the perceived accuracy. But the unit itself is not a reliable source of temperature readings, IMO.

Clearly alarms are not triggered consistently when it matters hence your customers are also in that world of a false sense of security. According to the test results this in my opinion is not a solid first line of defense at all.

The Blackbody reference square doesn't look right in some of the videos.See attached image of how I have been setting it up.

IPVM Image

Why is your Thermal Color scheme not the default IronBow2 ? and does the different Colorization change any of the results ?Also what is the version of the firmware used on the camera?

All of our clips were reviewed with Dahua's engineering team prior to publication and they did not question our blackbody setup. We were on a live webinar with them for over an hour, as well. Additionally, in our initial setup, we were specifically told by Dahua that the box did not have to be tight to the inside of the blackbody, because I questioned it.

As far as colorization, we changed it to make it easier to visualize hot areas. We did so while on a call with Dahua, actually. It does not change temperature measurement, just display.

Firmware is: V2.631.0000000.23.T, Build Date: 2020-04-03, which was the latest sent to us by Dahua.

Good to know.Thanks for the update Ethan.

This looks reliable...

That guy with the hat and mask coming in the from outside and being measured 3 seconds later is like 4 what-not-to-do things all in one.

Thanks for posting that. I'm assuming it's unintentional, but that system actually shows a high temperature alarm on the second subject. It's hard to see, but you can make out the red overlay which indicates over-temperature.

IPVM Image

It's too small to see where it was detecting, but in our tests against an open door, it detected hot pavement behind the subject as part of his "face" and triggered an alarm (you can make out the cursor to the right of his head below):

IPVM Image

As a reseller / Integrator, I will again state my disagreement / disappointment with you quoting any pricing other than MSRP or MAP. Since you allow anyone to subscribe, And now you're providing pricing in your Facebook comments? I for one feel that you should only be quoting what the manufacture established. Online reseller pricing is not reality. These players do not support you or your customers and all they want to do is sell at the cheapest pricing possible. Integrators work hard to provide a value added service beyond what a customer gets on the web. You have clearly indicated why these products are not in the DIY category any professional knows where to buy these and what they cost. Your end users should be directed to contact a trusted integrator to help them understand and provide the correct process to install / maintain.

Whether you like a product or not has no bearing on this. If you think the product is not working up to your expectations, that's fair. Continue to do your investigative reporting. I appreciate your opinions and observations. Sometimes I disagree with your findings but that's what you are paid to do.

You clearly value your publication's value, what about the value of your Real Subscribers?

IPVM Image

You quoted the following


Dahua's "Thermal Temperature Monitoring Solution" sells online for ~$12,000-13,000 USD, total, similar to body temperature measurement systems from Hikvision and Sunell OEMs (~$12,000-15,000 USD).

Individual prices for components (online):

  • DH-TPC-BF5421-T camera: ~$8,300 USD
  • JQ-D70Z blackbody calibrator: ~$3,600
  • DHI-NVR5216-16P-I 16-channel NVR: ~$1,000

i'm not sure where your indignation comes from here...

if some product is available online (whether within the channel or not), this information should not be referenced by IPVM?

not sure I understand. please clarify your position.

Evan, you promoted this Dahua system to Reuters, saying: "They don't have to slow down, they don't have to stop, they don't have to look at the camera," embedded below:

We found a whole series of issues with the system you are promoting and your objection is that we mentioned the price we paid?

Whether you like a product or not has no bearing on this.

Then compare: End User Buying Axis At Prices Better Than Axis Gold. Maybe you will argue we don't "like" Axis either.

We have not made a big deal of Dahua pricing right now but it's certainly germane to the issue. For example, the Washington Post brought up that Dahua relabeller RedSpeed is selling these units at $30,000 apiece:

Amid the rush, the county had paid a heavy premium: RedSpeed’s setup, at roughly $30,000 a scanner, cost far more than similar systems sold by established competitors — including the industry leader, FLIR Systems, whose scanners range from $5,000 to $15,000.

A setup of components nearly identical to what RedSpeed sold Gwinnett can be bought online for roughly half of RedSpeed’s $30,000-a-system total bill. But the RedSpeed official defended its system as comparable to rival hardware and sold at a fair cost, saying the company had moved quickly to deliver and install the systems in a time of high demand.

Your end users should be directed to contact a trusted integrator to help them understand and provide the correct process to install / maintain.

you have mistaken IPVM for a CRM.

Agreed. Providing wholesale pricing to the public is not in IPVM's best interest.

wholesale pricing is PUBLICLY available.

what's the beef?

Being one, I get the angst of integrators. However, anyone can obtain pricing online.

The key point I took from all of this is that a system with questionable results has a high markup. That is relevant information for me, my customers, and anyone else reading the review. Trust begins with telling the truth.

Here is an example of online pricing at this etailer:

IPVM Image

It forced me to do a 'quick login' using a social media account and then immediately showed the ~$12k offer. If you all want to sell this for $30k or whatever, fine, but it's pretty easy to find this at this price range online.

Custom Video Security is an authorized Dahua dealer so we are not allowed to advertise products below MAP. Unlike other online e-tailers who claim to be authorized dealers, we strictly adhere to all manufacturers' minimum advertised price rules. The "quick login for better price" is an easy way for shoppers to view our competitive price points without violating MAP.

While I have to say that Ethan raised a good point about the public advertisement of street price, I do agree with John in the sense that if pricing is already public, then that pricing should be in the report. After all, most people expect to buy at the public price on the Internet; not pay full MSRP or anything close to it.

It should be the dealer's and manufacturer's responsibility to state the importance of buying from an authorized dealer: receive the full warranty and tech support from the manufacturer, etc.

From what I understand there shouldn’t be complaints about the online price,unless you are planning on making more than 100% margin as a reseller/integrator.

Do not forget that thermal cameras ARE NOT medical devices, and thus should not be treated like one.

We should try to use them as rough estimators of high temp in very densely populated areas and thus since 38 is considered high fever we should consider accuracy vs total solution cost, based on the fact that 37 is normal humane temperature with a variation of +/-0.5 degrees Celcius and considered temp as fever at 38 degrees. So the accuracy to be tested is 0.5 degrees and bellow you mays ee how to do so.

In my detailed paper, I describe how statistically accuracy can be estimated based on statistical metrology. Until we see such a study we can only check the camera's performance as this test did.

G&N Electronic and Medical Solutions posted on LinkedIn

IPVM team! Excellent and balanced report. We all want a quick and easy means to assess COVID safety in schools, office buildings, and workplaces. Thermal systems for fever monitoring might help if they are configured correctly and marketed with integrity.

Below are some thoughts related to schools. These considerations really would apply for most any setting.

  1. The FDA has released a statement to IPVM that cameras used for fever detection should have FDA approval. The process is not extremely difficult. The FDA is not currently enforcing the standard but reserves the right to revisit after the crisis is over.
    1. If a school invests $25,000 in test units, is this money wasted if the units don't meet FDA requirements?
  2. If a student is noted with a higher temperature, a secluded setup is necessary for secondary screening.
  3. At what point does HIPAA compliance come into play?
    1. Patients have a right to privacy, as do students. Is HIPAA an issue during the camera phase or at the secondary screening point?
    2. The only sustainable way to handle HIPAA is at the secondary screening, but what about a kid who is pulled out of line due to a false positive? Students need to be taught that someone being pulled out of the lined does not mean that they are COVID-19 positive. Bullying can become an issue here.
    3. If a camera identifies a student from a bus as having a high temp, does the whole bus wait until the student completes the secondary screening? If that student tests positive for the virus, does the entire bus go on quarantine?
  4. With 50% of virus carriers being asymptomatic, is fever screening productive? Would a school be better served to choose random buses to screen with a handheld unit? Would the math work (would it help alleviate the problem) if we set a goal of screening 20% of the student population every day?

A false sense of security, is not security.

False accept errors ("passing" subjects who should have been "failed") are the most egregious type of error in any security situation because they provide a sense of security where none actually exists.

If this system (or any system) green-lights people who clearly should have been flagged for secondary screening, under laboratory conditions & based on the system's own settings, then it is no better than an alarm system which reports doors secured when in fact they are frequently open.

False rejects (flagging people who are not actually over-temp, or over-temp but not ill) would be annoying but acceptable in a security screening, especially in a layered-screening approach where this is the "first layer." Security screening systems tend to be biased to over-reject at the first layer to provide a fail-safe environment at the risk of having to do more secondary screenings.

Numerous false accepts in the first screening layer renders the entire process unreliable and essentially a waste of time, because people passing the first screening don't get subjected to the others.

Thank you for this deep and articulate test. Ultimately, I've learned I can take aspirin and walk right in.

We spent this morning testing blackbody configuration after some questions from commenters and others.

Based on our tests, drawing the blackbody region box in several different ways results in extremely similar measurements. The subject below was measured at 98.1-98.3°F walking through the scene (using the serpentine queue setup Dahua recommends, then turning out of the scene at the blackbody) across several different attempts.

Blackbody region drawn tight on the center:

IPVM Image

Blackbody region drawn around the outside:

IPVM Image

Blackbody region drawn offset:

IPVM Image

Blackbody knocked out of alignment, nearly out of the region:

IPVM Image

If anything, I would have expected measurements to be skewed in that last example, but they remained steady in the same range as other setups.

We were also able to use a hot water bottle to create elevated forehead temperature, similar to our past tests, with the subject maxing at 99.8° on the camera, and at 100.6°F on an IR thermometer:

We noticed that it will affect readings after longer time and also when people move closer to the camera. But maybe it depends on firmware version?

I wonder what the results would be with no blackbody ?

I wonder how Amazon will feel about this report? They just spent $10 Million for the Dahua Thermal solution.....

Exclusive: Amazon turns to Chinese firm on U.S. blacklist to meet thermal camera needs - Reuters

Knowledge is Power!

In lieu with the current covid19 pandemic across the world, a lot of stakeholders and endusers around the world have been rushed to commit some kind of first defense against the spread within their working places.

Almost every single China based CCTV manufacturer has 1 type of bispectral thermal camera in their product line today! I am ware almost 99% used non-FLIR cheap low-res thermal sensors.

Many traditional CCTV suppliers have switched over and are preaching loudly about deploying such thermal monitoring. Some professionally while some just trying to make extra bucks (good money actually) at this moment of time.

This extremely informative publication based on the works here should be make AVAILABLE to the world. Especially to those who already committed. They have a right to know what they bought is actually within their expectation in term of the risk of not picking up a genuine case of person passing thru with higher forehead temp.

Thank you for this information. We've been vigorously researching solutions to offer our customer base and this is very helpful. I am curious if you have any thoughts on how elevated body cameras will roll under the NDAA guidelines? Additionally, with the rush of elevated body cameras being produced is your team able to create a OEM/ ODM chart. I know AV/Costar has an OEM partner and i'm sure others due as well.

Given our customer base NDAA compliance is very important to us, weather its a stand alone application or not.

Thanks again,

I am curious if you have any thoughts on how elevated body cameras will roll under the NDAA guidelines?

These (Dahua, Hikvision) are clearly network cameras and therefore fall under the same the ban.

What firmware version is the DHI-NVR5216-16P-I NVR running?

Thank you for taking the time to do this report. I have been spending a considerable amount of time dealing with distributors tripping over themselves to sell thermal fever detection cameras. It is obvious there is money to be made selling thermal cameras in these troubling times; however to do so with the promise they will accurately screen people with elevated body temperatures is dishonest. I appreciate the effort that went into researching the effectiveness of this system, which seems to be the one that is being marketed most the most aggressively.

At the price point, and performance accuracy, I believe it is a foolish investment for most end users when used to attempt to screen for covid-19 infected people.

"The subject was then measured by the Dahua camera, and finally by a handheld IR thermometer (Extech IR200)."

Do you check the calibration of IR200?

If Yes, how often?

and what do you use to calibrate?

During testing, we've been checking our IR200s regularly. I can't tell you how often, exactly. It varies, but it's been multiple times per day. We've been doing this for two weeks straight now, so we know what we all tend to measure, so we are stopping and checking when things seem at all odd.

Also we don't have just one IR200, we have four. We also have four other IR thermometers, three tympanic membrane thermometers, a temporal artery scanner, and an oral thermometer, with more on order. So we have a pretty good idea when things are drifting.

Do you check the calibration of IR200? If Yes, how often? and what do you use to calibrate?

...Que custodiet ipsos custodes...

Hi @MOD, will you test de Mobotix one?

Check this .

I have a lot of kits in our office and when every aspect its set correctly the messurement its very precise even if you work directly to the camera

i have an question for IPVM team some manufacturers say they dont need a black body for temp measurements and some advertise black body for accurate results DOES investing in black body help? and why was black body introduced when similar specs camera function without black body and competitor says black body is necessary

As it relates to Dahua and Hikvision's ongoing dispute about blackbodies (Dahua pro blackbody, Hikvision anti, or at least marketing it being 'optional'), our testing finds that Dahua is correct on this point. See the problems Hikvision had without using a blackbody in our test results.

dear team

thank for an reply , is there a possibility of you giving an apple to apple comparison of dahua and hik camera i mean side by side comparison in the same screen

Here you go, from the side - the Hikvision is a couple inches longer overall, with more on the cable whip:

IPVM Image

And from the front - main difference is the white light LEDs on the Dahua which can be flashed on alarm:

IPVM Image

Loading Related Reports