Hikvision DS-2TP21B handheld camera - an interesting unit

[Fraser]

I have just added a Hikvision DS-2TP21B-6AVF/W Human temperature screening thermal camera to my inventory. Details of the camera may be seen here……

https://www.test4less.co.uk/hikvision-fever-screening-handheld-camera.html

As I have previously stated, I have been watching for heavily discounted COVID-19 countermeasure related thermal imaging cameras and have been very fortunate in finding some real bargains. The Hikvision DS-2TP21B-6AVF/W is the bigger brother to the DS-2TP31B that I previously highlighted as a bargain brand new camera at only £60 each ! At the time the DS-2TP21B-6AVF/W had not appeared on the disposals market but that recently changed. I negotiated the purchase of this brand new camera at a great price and it arrived today. Something to note is that this version of the camera currently sells for £3594, whereas the standard general use version from the same dealer sells for £958. The COVID-19 effect ?

I was interested to see how this 160 x 120 pixel thermal camera performs as I like the DS-2TP31B and it has excellent build quality. The DS-2TP21B is no different …. It feels very solid in the hand and when I switched it on I was presented with a very respectable image that appears better than it’s 160 x 120 pixels. The touch screen 3.5” LCD display is a very nice sharp 640 x 480 pixel assembly so the thermal image is upscaled to match it. There appears to be more than just upscaling going on though. I am wondering whether this camera is using Hikvision’s Super-Resolution feature but see no mention of it in sales literature. I shall do some more tests on the camera to ascertain it’s effective resolution in use. First impressions are good though.

The DS-2TP21B has a very nice set of options in its menus and it is very easy to navigate. Important features like manual mode for span and centre temperature are included along with a comprehensive set of measurement options. Sadly, as this is the B (health screening) model, the stated temperature measurement range is +30C to +45C with +/-0.5C tolerance. I have yet to establish the actual limits of measurement but the manually selectable span and centre temperature scale is -20C to +150C so the range is decent, even if measurements within that range is software limited. NETD is stated as better than 40mK and the lens is F1.1. The HFOV is 25 Degrees so in terms of image detail this 160 x 120 pixel camera will produce something very similar to a 320 x 240 pixel camera that is equipped with a 50 Degree lens. The choice of lens is very nice as it produces a reasonable IFOV of 2.74 mRad. The lens is manual focus which is welcomed as it is possible to achieve best possible target focus very easily. Oh and the camera has an image refresh rate of 25fps 

The kit that I received includes the. Camera, charger, USB cable, Desk charger and two battery packs. These are protected in a solid hard case.

For anyone hunting a bargain thermal Camera but with no need of measurements or monitoring of temperatures above 150C, it would be worth keeping your eyes open for one of these cameras. There is one on eBay, brand new in its case fir £300 at the moment. I paid a lot less than that for mine

I will write more about this camera once I have had time to use it for a while. For what it cost me I am very pleased with the kit 

Fraser

[Fraser]

My previous post on the DS-2TP31B, the little brother to the DS-2TP21B 

https://www.eevblog.com/forum/thermal-imaging/hikvision-ds-2tp31-series-thermal-camera-teardown-307840/msg3942805/#msg3942805

I managed to buy that camera for only £60, new in the box.

Fraser

[Fraser]

After a fellow forum member advised of a hidden menu on the DS-2TP31B camera I tried the same process to see whether one was present in the 21B model. It is  To access the hidden menus you press the “Back” button three times and then one of the navigation buttons. Each navigation button selects a different hidden menu list. One is in English whilst the other three are in Chinese. The English menu advises the temperature of various parts of the camera whilst the Chinese menus (translated with Google visual translator) offer some unusual options, the purpose of which is not currently known. Sadly there is no option to change the cameras measurement range. These cameras were calibrated to be very accurate over a narrow temperature range and this means they likely cannot be converted to the general use version. For me this is not a problem though.

More about this interesting camera to follow 

Fraser

[Fraser]

For those interested in the specifications of this camera……

Hikvision - Fever Screening Handheld Camera


Key Features:

160× 120 resolution (thermal), and 8 MP resolution (optical)
Thermographic accuracy up to ±0.5 °C
Display fusion of thermal view and optical view.
640 × 480 resolution 3.5” LCD touch display
Support live viewon PC, mobile device or external monitor
Support audio intercom
Support color highlight alarm and audio alarm


Thermal Module   Image sensor:   Uncooled focal plane arrays
Resolution   25Hz: 160 x 120
Pixel Pitch   17μm
Response waveband   8μm to 14μm
NETD   <40 mk (@ 25 °C,F#=1.0)
Field of view   25° × 18.7°
Focal length   6.2 mm (Manual focus)
IFOV   2.74 mrad
Aperture   F 1.1
Optical Module   Picture Resolution   Configurable: 2 MP, 5 MP, 8 MP
Video Resolution   640 × 480
Image Display Monitor   640 × 480 resolution 3.5’’ LCD touch display
Palettes   Black hot, White hot, Red hot, Rainbow, Ironbow, Fusion, Rain
Digital Zoom   Thermal：1×, 2×, 4×
Thermal / Optical   Thermal/Optical/Fusion image/Picture in picture
Picture in Picture   Support Thermal View in Optical View
Fusion   Support (Thermal View and Optical View Combined)
Record Video   Support on-board video recording
Thermography   Rules   Max. Temperature; Min. Temperature; Center Temperature; Custom Point/Line/Area Thermography
Range   30°C to 45°C
Accuracy   ±0.5°C
Battery   Battery Type   Dismountable and rechargeable Li-ion Battery
Battery Operating Time   5  hoursin  normal  temperature,  with  Wi-Fi,  Bluetooth, and supplement light off.4 hoursin normal temperature, with Wi-Fi, Bluetooth, and supplement light on.
Storage   Built-in memory module (16G)
Photo   Thermal/Optical/Fusion image
Video   Thermal/Optical/Fusion image
Wi-Fi   Support
Text Comment   Support
Voice Comment   Support
Language   English
General   Power   5VDC/2A
Power Consumption   Max.4.5W
Hardware Interface   Type-C Interface
Working Temperature   15°C to 35°C
Protection level   IP54
Drop Test Height   2 m (6.56 feet)


Worthy of note are the following…..

17um VOx pixels used for 160 x 120 pixel microbolometer. VOx normally out-performs A-Si in terms of S/N
25fps frame refresh rate
Stated <40mK NETD at 25C F1.0
6.2mm F1.1 manual focus lens provide FOV of 25 x 18.7 Degrees
3.5” 640 x 480 pixel touch screen LCD display
Removable Micro card storage so not limited to internal memory size
Video recording function in addition to still images
Easily removable battery packs
Wireless connectivity to other display devices such as a mobile phone.

The two limitations of the camera are it’s 160 x 120 pixel microbolometer resolution and its software restricted temperature measurement range. The camera MAY be using Hikvision Super resolution feature for image enhancement but this needs further investigation. The choice of lens FOV is excellent for a 160 x 120 pixel microbolometer. The constrained temperature measurement range is a function of the cameras intended use as a very accurate human fever detection system and is not easily overcome. The camera can still display its full thermal range but it cannot officially provide measurements outside the stated +30C to +45C. The actual limits of the measurement function will be tested soon.

Fraser

[Fraser]

The DS-2TP21 and DS-2TP21B cameras appear to be based upon the sister companies (Hik Micro tech) M10 industrial cameras…..

https://www.hikmicrotech.com/en/industrial-products/m-series-handheld-thermal-imager/

https://www.pass-thermal.co.uk/hikvision-ds-2tp21-6avf-w-handheld-thermal-imaging-camera

https://isswww.co.uk/hikmicro-m10-hand-held-thermal-imaging-camera

The M10 firmware update is available for download but is likely encrypted, as found with their CCTV cameras…..

https://www.hikmicrotech.com/en/support/download-center/firmware-download/

The DS-2TP21B firmware update is also available on the Hikvision site.

[Fraser]

Just a tiny update…

Tonight I quickly tested the camera against boiling water and our fridge contents 

The result was unexpected ! …… the boiling water read +102C and the fridge read +2.5C. We know water boils at around 100C and our fridge is set to +3C, so despite not optimising the Emissivity for the test targets (water at +100C and plastic at +3C) the camera provided decent measurements  With the boiling water I noted a black speckling that I also saw when a Infiray P2 Pro was viewing a thermal scene that was above its recommended maximum temperature in high sensitivity mode. This makes me think the ‘B’ version of this camera is optimised for temperatures below +100C. Temperatures above that of a healthy human are shown in red flashing digits on the measurement display. I will see whether there is an option to disable this ‘alarm’ function as there was for the associated audible alert tone. Not a big issue though as at least the measurement may still be read.

So despite this being a quick and very ugly test, I now know that this DS-2TP21B model can measure temperatures beyond its stated +30C to +45C coverage, unlike with the DS-2TP31B where measurement ceases outside the stated range for human screening. Such a pity that they did that to the 31B  :'(

I was impressed with the quality of the images produced by the camera on its very nice display.

A more scientific test with Blackbodies will follow in due course.

[Fraser]

I have been asked what I paid for my DS-2TP21B.

It was purchased for £220 from an eBay seller after some price negotiation.
The kit included : The camera, Desk charger, power supply, two batteries, usb C cable, lanyard and manual. New and unused.

Do I think it was good value at £220 ?

Yes. It offers acceptable resolution (160 x 120 pixels), good frame rate (25fps), manual focus and plenty of features including wireless connectivity. Whilst it would have been great to see 256 x192 pixels in this model, it remains a very capable thermal imaging system never the less.

Will I be keeping it ? That is as yet undecided as my finances are very limited and I have plenty of other cameras to use. My current focus and priority is a AGEMA Thermovision 550 cooled MWIR camera that cost me £200. I did not budget for buying both ! Time to do some more work for my old employer to earn some pennies me thinks 

Fraser

[Fraser]

I tested the temperature measurement capabilities of the camera today using a calibrated Blackbody. I was pleased to see that the camera measures from at least Zero Celsius (minimum tested due to the issues of frosting below 0C) up to +100 Celcius. At around +95 Celsius I noted the onset of black pixels in the image indicating that I had hit the end stop of the ADC or measurement firmware. At +100 Celsius the black pixels had grown in number so measurement beyond +100 Celsius was not tested.

The camera works well and the only disappointment is that the saved images appear less sharp than those of the LCD display.

The Laser and LED flashlight functions are not present/activated on the 'B' version of ths camera.

I am very pleased with the build quality and performance of this camera. This camera achieves a lot with its 160 x 120 pixels  The manual focus lens is great to have and I have decided to keep this unit. For what I paid for the DS-2TP21B camera, I got myself a very good deal 

Fraser

[c.camber]

Hi Fraser,

I followed your topic with interest and that's why I Also  bought the DS-2TP21B camera it's working great also below zero until about -25 Celcius.

I have 3 questions for you: did you take it apart like the 31B and if you did do you know if the laser and the LED are present in the unit, I can't see them from the outside. If they are in there can we make them work?

What software do you use for analysing the pictures? The IVD 4800 from Hikvision?

Thanks for your research.

[c.camber]

And do you know if the firmware of the M10 can be used in the Hikvision version?

Thanks in advance👍

[Nic Fiddler]

Hi, I'm a newbie both here and to thermal imaging, so after buying a DS-2TP31B on-spec I was gratified to come across this blog and find that it wasn't too bad a buy after all, though not exactly best suited to my particular needs.  (I'd like to both monitor peripheral blood circulation and scan roof insulation.)  The seller had notified me of a key mistake in the description, so I returned it and kept looking.

Then a DS-2TP21B came up for less than £300, so I bought that.  This seemed to do all I'd need it to and more, but despite making all manner of adjustments the thermal images are nowhere near as crisp as on the lesser model.  The visible-light camera works fine, but doesn't help with delineating fine thermographic detail.  The attached pic is of an upper corner in an unheated room.  Alas, I don't have a comparative shot from the 31B, but it would have shown very much finer detail.

I called Hikvision UK, but was told they couldn't help end-users, though they we're able to identify the original vendor as Fortus, so I phoned them.  A very attentive chap suggested that the device may have been dropped, shocking the thermal lens array slightly out of alignment.

As investigation and repair is likely to prove more costly than I can afford, before I return it too for a refund I thought I'd ask if Fraser or others can offer any further opinion or advice, please? 

[Bill W]

It is only a 160x120 imager, so will not look that good on its' own.

A visible light camera is in many ways a 'cheat' as it assumes any visible details should be there thermally and fuses the images to suit, so apparently sharpening the thermal image but actually adding extracted detail out of the visible image.

If the 'drop' theory is correct (doubtful, but the first excuse) then if you show it a well defined thermal image such as someones' outstretched hand & fingers at 2m or so then you may be able to refocus by turning the lens in the housing.

[IR_Geek]

Echo what Bill W said. 

I'm not super familiar with this unit but from a quick compare:  The 31 just has more bells and whistles, but has the exact same uncooled microbolometer.     I'd recommenced switching to grayscale if you can access the 'raw' array data.   color mapping has it's uses but to me it washes out fine detail.

In reality the lens is what really makes or breaks the usability of a thermal sensor.    With an FPA this small they try to give a large field of view (FOV) while sacrificing the instantaneous FOV (IFOV).    Of course then it wouldn't be a cheap system if they provided a quality optic.

[Nic Fiddler]

Thanks BIll and IR for your helpful replies.

Slight confusion, IR ... the 21B is the higher spec unit.  Both are 160 x 120 thermographically, though the 21B has 4 times the number of display pixels.

To try to better illustrate my problem, the attached pics show an unpowered mains extension socket 'rail' at 2m, captured in thermal, fusion and optical views with rainbow and white-hot palettes. 

Key settings:
Emissivity - 0.95
Distance - 2.0m
Reflection temp - 25C
Humidity - 50%
Parallax correction - 1.0m (notes nothing I've tried has successfuly changed this setting - it just reverts to 1.0m)
Picture Resolution - 8MP
Range - Auto

Prior to capturing these images, the Measurement tool Initialisation and SD Card Initialisation were each re-set and the device then also restored to Default Settings.  Manual focussing was carried out before each shot.

The Fusion shots clearly show a mis-match between the thermal and optical views.  Is this normal, or might it too be an indication of internal misalignment caused by being dropped?

eBay apart, current offerings online for this model, new, include around £1.9K (pass-thermal.co.uk) and £5.9K! (isswww.co.uk), so considering this indicates that it's a serious instrument and not merely similar in performance to cheap Chinese alternatives - along with with Fortus chap being clear that the thermal image should be very crisp and even better than than 31B - I don't think I'll be convinced until I can compare it with another similar example.  Might there be someone in the Bristol UK area who has one that I could drop-by to check with, please?

Being one who's not afraid of tinkering, I've considered dismantling it to see if there's anything obviously out of place that I might try to correct, but the rear casing doesn't readily come off and I don't want to risk breaking the plastic.  Does anyone have any tips on how to do this, or know where I can view or get a copy of a service manual, by any chance, please?

Thanks again for reading - any further observations gratelfully received.  If I'm to return it, I only have a few days left  .

Nic

[Bill W]

The optical looks like the one out of focus.  Obviously that won't help the fused ones.

The pure thermal images just look like too much sharpening used, breaking up the image of the cords.

The misalignment could be 'correct' as it is about 15mm.  Are the two lenses about the same separation as the shift ?
if so that is simply parallax with two parallel camera axes and no realignment.  You'd expect either a clever realign in software (by reading the focus position) or a convergence around 2 - 4m

The images on this page from my website will give you an idea what a pure 160x120 sensor should achive - and they are via analogue video.
http://www.fire-tics.co.uk/project1/index.htm

Bill

[Nic Fiddler]

Thank you.  The inter-lens distance is ~40mm.  Manual focussing doesn't get any sharper than shown - at any distance.   

Any ideas why I can't get the Parallax setting to change?  It offers 1.0, 2.0, 4.0 & 8.0 metre options, but is stuck on 1.0.

  

[mauada]

Just for the record: do not update your unit to 5.5.19, as I did on mine.

While the latest firmware will enable wireless connection to the Hikmicro app, it will limit the display temperature ranges to 25..50°C, similar to the 2TP31B.
I managed to find the original firmware but downgrade does not seem to be supported

[Fraser]

Thank you for the warning. That is very bad news that this firmware restricts the temperature range. When I bought my unit I was so pleased to discover that it provided a really useful, broad temperature measurement range. As the Health screening cameras normally cost so much more than the standard model it seems counter intuitive that they would make this change to effectively hobble it. It is not as if all of the surplus units will be updated to this latest firmware that restricts their usefulness. Thanks to your post we are now warned to not update the firmware. Regrettably I do not know how to reinstall the original firmware.

Fraser

[mauada]

@c.camber, yes it is actually using the binary identical firmware as is provided for the Hikmicro M10.
I think Fraser mentioned this somewhere, it is some form of configuration or hardware ID that creates the specialized behavior, not the firmware image itself.

I contacted hikvision support, maybe there is still a chance to get back to the original firmware.

[mauada]

Update: No reaction from hikvision support.
I did ask hikmicro the same question for a possible downgrade of a M10. Their support was more helpful: they confirmed user downgrade is not possible but offered to have it send in to a service center - Hats off to hikmicro support!

[MetalHead]

Hello guys!

Fraser, thank you for the review and I also agree that this one is a nice, fast and good assembled camera!

Last week I got one also (second hand directly from china) and it's H10 model which can measure up to 350C and as far as I know it's being used in China along with custom macro lenses to spot problems on circuit boards. With it you can even see the temperature on different parts of singe IC, what is in fact amazing... they also use H11 version which has bigger screen with their H10Pro and H11Pro versions.
 
I bought this IR camera for less than 50$ but it has a problem. The problem is that it displays everything in yellow (from white to dark yellow). The menu, buttons - everything is working just fine!  I can enter menu, change options, no problem with that. Even if I point it on hot object I can recognize barely visible outline of what is hot (photo #2)

I'm attaching some photos to better understand what I mean.

So I'm wondering what is causing this problem - IR sensor and its circuitry, FPGA, or maybe it's TFT screen related and has to do with screen driver chip? Ideas

One more thing. I tried to make photo and save it to SD card. Last image is what I got.

[Fraser]

 I am trying to interpret your sample images but I am unclear as to what the scene being viewed actually is.

From what I can see, my first thoughts were that the LCD panel is faulty. This is supported by the correct colour rendition of the temperature bar in the saved image. Then there is the matter of the many “spots” and lines on the image. Sadly this reminds me of a microbolometer that has many dead pixels that are not being hidden by the dead pixel map software routine. From the images alone and without having inspected the camera, I would suggest that this camera has been for a swim ! Water ingress can cause havoc in a thermal camera. Is there any evidence of fluid having entered the camera…. Corrosion, water marks, contamination of the display layers or microbolometer lens etc ? I may be wrong, but I believe that I can see water marks in the LCD display panel layers.

Regards

Fraser

[Fraser]

If the camera has suffered water damage, you will need to completely strip it down for inspection, cleaning and repair of corroded areas on the PCB. The LCD display will be scrap and a new part will need to be sourced. The microbolometer itself is a sealed package, but the associated IC’s and FFC flag mechanism are vulnerable to corrosion. Can you hear the FFC flag operating ? It should click every time the flag is used to “calibrate” the camera. If it is not operating, the camera will not produce a decent image as it needs the flag to create the FFC offsets and this will be messed up if the camera does an FFC using a thermal scene instead. I personally would not undertake the restoration of this model of camera if it has been for a swim. The cost in terms of time and parts will exceed the cost to buy another unit in good working order.

Kind Regards

Fraser

[MetalHead]

PCB itself is clean without corrosion and any! signs of water, but I noticed the screws were corroded. So if there was water it evaporated eventually. I disassembled the sensor module. All FPC connectors are good and clean. Nothing suspicious on PCB itself.

Here is the image of a sensor itself, no water marks on it, but who knows it might have damaged something...

[Fraser]

Take a close look at the LCD display panel. Use a magnifying glass and torch if necessary. Are there signs of fluid ingress in the panel layers or is it contamination behind the vinyl screen protector ?

Fraser

[MetalHead]

Thanks Fraser!

I dismantled everything and reconnected without the housing of thermal imager. Same yellow screen with "dots and lines"(( All connectors are like new.  I haven't seen any signs of water neither on LCD and underneath it. Nothing is overheating, tried with finger. Also I used hot gun to check how ICs react to the increase of temperature and nothing...

I attach some images. To me it looks like the problem is with microbolometer and its circuitry (like you said above). I will buy another good h10 and will try to replace IR sensor assembly. Will see if that will do the trick! this mess is either LCD or sensor or maybe both

[ser4]

Hi, I have 57 units of DS-2TP21B-6AVF/W - HIKVISION - $300. Located in Kazakhstan. Condition - new. Complete set. +77475587919 whats app/telegram, ser4_chet@mail.ru

[Fraser]

Metalhead,

Your camera certainly has an unusual fault.From what I can see, it has more than one fault, which is unusual unless something very bad happened to the camera. You have an LCD display panel that appears to have lost its @bility to display anything but the yellow spectrum and a microbolometer thermal image that contains many ‘spots’ that cannot be explained. The angled lines across the displayed image are also a mystery. In addition to these obvious issues, there are the corroded screws inside the camera, yet no evidence of water ingress. Another mystery. Those screws are very corroded so there has been a very damp atmosphere inside that camera.

Doing some “blue sky thinking” I am trying to come up with scenarios that could cause these symptoms but I can find no common denominator except water ingress but we know that is not supported by the physical state of the battery, PCB’s and alloy chassis part condition. I have considered the possibility of a significant drop event causing damage to the fragile LCD display panel and microbolometer but this is still not a “good fit” to your cameras fault symptoms. I have also considered whether someone has used your camera as a parts donor and put the faulty parts in the cameras casing to sell it. This might explain multiple non-related faults in a single camera. I have seen this done before and it is a pretty mean act by the seller. Is it the case here ? No idea.

Sadly this sort of fault can become a really deep rabbit hole when it comes to diagnostics. For this reason, such a camera is often uneconomic to repair in terms of time spent on it and the cost/availability of replacement parts. If you had another camera from which to borrow parts, you could quickly determine whether the LCD panel and Microbolometer modules are faulty and this would save much time probing signals with an oscilloscope. If the LCD panel is faulty, a replacement may be available if it is a generic model. If the microbolometer module is faulty, the camera becomes good for parts only as the module is not available and, even if it was, the camera would require a full calibration to use it.

Kind Regards

Fraser

[Fraser]

A good test to determine what is happening in a thermal camera is to image a hot coffee mug (fill with boiling water). Please do this and post the image captured. This simple test provides the camera with a high thermal contrast scene to image and the measurement system should show a sensible reading. I am considering weather your microbolometer has lost its tiny, but essential, vacuum as that destroys its performance.

Fraser

[MetalHead]

It doesn't seem that anybody was inside. Everything was glued together nicely and evenly and I used hot air gun to disassemble it.

Well, I started to look for spare parts, or some cheep body-type Hikvision model to take parts from it. I think the only thing that is different in H10 is FPGA and firmware on uC, screen and IR sensor are the same.

Regarding the cup of coffee, I tried but I can't see anything... too much noise on screen

When I will get parts or another H10 I will update you with what I found out! This may take some time however..

Thanks Fraser for all your help! 

[Fraser]

Metalhead,

The symptom whilst viewing a hot coffee mug makes me suspect loss of microbolometer vacuum, with potential pixel damage associated with the event. If it were not for the strange anomalies present I would suspect microbolometer power supplies and bias voltages but those faults would not cause the anomalies we are seeing.

I had already thought about a Human fever screening DS-2TP31B camera as a possible spares donor for the LCD display as those cameras can be quite cheap on eBay. Sadly you cannot use the microbolometer from such a camera without carrying out a full calibration of the camera as the NUC table, dead pixel map are specific to each unique microbolometer. Using another cameras microbolometer in your faulty unit would likely result in a marginal image with dead pixels visible and poor flat field. You cannot use the donor cameras calibration as it is set up for vHuman Fever screening and I not know how to transfer that data between these cameras in any case. Please think carefully before spending more money on your faulty H10. You may be better served to save your money and buy another camera to serve your needs. Losing $50 is hopefully not too serious a loss for you, but ‘wasting’ more money in an attempt to repair it would be rubbing salt into the wound. Maybe selling the M10 as “faulty, for parts” on eBay would recover some of your investment ?

Kind Regards

Fraser

[MetalHead]

Yes, maybe you are right! I will wait for a donor if it won't came up, I guess I probably sell it on local second hand auction as broken machine. Thanks! 😊