THERMAL CAM IMAGER / CEM DT-9885 - experience

[bozidarms]

Hi to all,

have someone any experiences with this camera (CEM DT9885)?
http://cem-instruments.de/en/products/cem-dt-9885-thermal-imager-384x288-pixel/

Please share your thoughts.

[Fraser]

Well they are one of several Chinese thermal camera manufacturers who produce cameras that have the appearance of other manufacturers products. They are not clones however. They are totally different inside the case. This one looks like a sort of FLIR Exx TESTO 880 hybrid

The DT-980 Appears to be a rip off of the FLIR iX series

The cameras are likely to contain a thermal sensor that is bought in from outside China. A favoured source is ULIS in France.

The Chinese manufacturer can use ULIS (or other) development boards and software to create a pretty decent thermal camera. Provided they use a reasonable lens the hardware can be very capable. Expect Chinese budget build quality though !

Sadly where the Chinese thermal cameras can fall down is in their firmware and software. In a thermal camera the firmware performance is very important in that it carries out the image processing that makes a pretty nasty raw microbolometer image presentable and useful to the user. Poor quality firmware can turn capable hardware into a pig of a camera to use.

The PC software for analysing images captured by the camera is also important. Poor quality software can be frustrating to use. Chinese thermal camera manufacturers are not normally known for high quality firmware or software.

You really need to see one of these cameras working and test it before shelling out a lot of money for it. The PC software may be available for download from the CEM site. It may be worth taking a look at it to get an idea of the effort that has been invested (or not) in its production.

http://www.cem-instruments.com/en/Product/type2/id/683

If you google CEM thermal imager you will find You Tube videos showing their cameras working. It is worth taking a close look at these reviews. They may not be the exact model of camera that you are interested in but they give you an idea of how well built they are and how they tend to perform.

I will not say that a Chinese made thermal camera is not good, but my experiences with them has not been great to date. Do your research and spend your money wisely.

Fraser

[bozidarms]

Hi Fraser,

first, thank you very much for detailed reply.
Unfortunately, i haven't a possibility to test the camera (have only a reasonable offer for one new).
I have seen some videos and that all looks very promising,
but, i must say - that is completely new filed for me,
i haven't any experience with thermal cameras at all.
I have wondering if  someone have experiences with that (CEM) Chinese firm?

Best regards

[Fraser]

Just for info, this camera is sold under other brand names as well.
The price is normall just under $1000.

Fraser

[bozidarms]

Camera has arrived
Results are way out of expectations

[Fraser]

Excellent news  I am pleased that it is providing good images for you.

As I previously advised, it is likely to be a quality ULIS microbolometer inside the camera.
The noise level in that image looks pretty good. How does it perform at smaller temperature spans of less than 10C ?
It is normal for images with smaller spans to contain higher levels of noise, but providing such is not excessive it is not normally a problem.

How good is the PC software ? Sadly Chinese products often come with software that is a little on the basic side.

It may be possible to use other image analysis software with the cameras images, depending upon their format. Joe-C on this forum is expert in the analysis of thermal camera image formats.

Fraser

[bozidarms]

Hi Fraser,

at smaller temperature spans is some nose, but when you carefully make measurements results are quite acceptable.
I have noticed importance of  right focusing, manual adjustment of level and span, and also the best results are than  in monochrome modes (gray, sepia,...). Later, i put some pictures.
In case that in the field of view exist one object with bigger temperature difference (even a small one), the whole picture becomes dramatically better.
I don't know is that behavior normal with others cameras too?
Anyway, in greater temp. spans results are superb.
I have no experience with other camera software, so i cant tell much - in supplyed software, one can make quite a lot  settings, and i think that pictures appears better than on camera itself.

Regards

[bozidarms]

Hire are the pictures with small span:

[Fraser]

Those are excellent images for a small span and low thermal contrast targets.

The vertical lines in some images is perfectly normal with an Amorphous Silicon microbolometer and small spans.

Well done on this purchase, you appear to have got yourself a Chinese thermal camera that delivers really good images I suppose it was only a matter of tIme before the achinese OEM's upped their game and sorted out decent noise reduction in their cameras firmware.

Fraser

[bozidarms]

Yes, definitely, thank you.
I am totally happy with camera
But i need some time to learn how to proper use it.
Here, some more low range images (span only 4 degree Celsius):
I have also noticed a rather big difference between JPEG-s and PNG-s.
PNG-s are much more pleasant to see.

Regards

[Fraser]

You can download come decent thermal imaging guides for free from FLIR and from FLUKE. Most FLIR and FLUK user manuals also contain a section at the back detailing the basics of thermography. Try downloading the FLIR Ex ans Exx series user manuals. The FLUKE TiR10, TiR20 etc manuals may also contain similar.

I have several thermography reference books but they tend to be e pensive for the beginner and likely too detailed for basic usage. FLUKE did a dedicated fee book on thermography called 'Thermography A practical approach'. It was a downloadable PDF but I think a new version is now available for free download. I have the hardback book version that came with some of their cameras.

Let me know if you cannot find the Fluke and FLIR e-books and I will find them for you.

Fraser

[Fraser]

Ok some e-book and other useful links for you........

https://s3-eu-west-1.amazonaws.com/approvedproducts/manufacturer/FLUKEthermographyprinciples.pdf

http://www.akruralenergy.org/2016/2016_REC_Practical_Thermal_Imaging_Application-Mike_McGrath.pdf


http://www.flirmedia.com/MMC/THG/Brochures/T820325/T820325_EN.pdf

http://www.flirmedia.com/MMC/THG/Brochures/T559213/T559213_EN.pdf

http://www.flirmedia.com/MMC/THG/Brochures/T820264/T820264_EN.pdf

http://www.flirmedia.com/MMC/THG/Brochures/T559243/T559243_EN.pdf

That should do for starters. I will link to the E4 and PM series user manuals as well as these contain useful information too.

Fraser

[Fraser]

FLIR Ex and PM series user manuals. These contain useful information as well...

http://thinkbox.case.edu/sites/engineering.case.edu.thinkbox/files/images/manual_flir_e6.pdf

http://www.flir.com/legacy/view/?id=51532

Fraser

[bozidarms]

Oh boy ,
thanks a lot - regards,
Bozidar

[Fraser]

A 160 x 120 resolution version of this camera is also sold under the 'REED' brand

http://www.itm.com/files/Thermal-Imager-Comparison-Chart.pdf


http://www.reedinstruments.com/category/thermal-imagers


http://www.reedinstruments.com/product/reed_r2100_thermal_imaging_camera

Brochure:

http://www.reedinstruments.com/pdfs/cache/www.reedinstruments.com/r2100/datasheet/r2100-datasheet.pdf

Manual:

http://www.reedinstruments.com/product/reed_r2100_thermal_imaging_camera

Software:

http://www.reedinstruments.com/files/softwares/R2050-R2100-Software.zip


Fraser

[Ben321]

Do these Chinese CEM thermal cameras save radiometric data embedded in the JPEG? Or is it just a plain JPEG image?

[bozidarms]

Hi,
camera has radiometric data embedded in the JPEG.
You can change some parameters in original software (PCIRMeter),
look at this video (note that video describes data from DT980 - 80x80 pixel, and not from DT9885 - 384x288 pixel):

Probably, not the ultimate software, but could do the job.
I don't know if exist  alternative software?

Altogether, it became apparent  that this camera is
superb bit of kit that represents exceptional value for money.

[Fraser]

Excellent. I am pleased that you are pleased with the camera. Thank you for sharing so that others may consider it when deciding upon a purchase.

Fraser

[Ben321]

Just did some digging into the JPEGs from this camera, using a hex editor, and Gimp (using its raw image loader feature, which lets you specify offset, width, height, and raw pixel format), and have figured out a way to get to the raw data in the DT-9885 thermal imager.

Scan the file for the unicode string FPZdnx (yes, it is case sensitive). Because it is unicode, it's 2 bytes per letter. So this is a 12 byte string you are looking for, with the following hexidecimal values:
46 00 50 00 5A 00 64 00 6E 00 78 00

After finding this byte string, the raw 16-bits-per-pixel image data starts exactly 96 bytes past the last byte of that string. So after your byte string scan finds this byte string, jump to the end of the byte string (12 byte jump), and then skip 96 bytes (the raw image header is 96 bytes long) to get to the start of the actual raw image pixel data. Alternatively do a 108 byte jump from the start of the 12 byte string (this will eliminate one step in writing a program that is designed to load the raw data). Unlike FLIR's FFF format, which (when embedded in a JPEG) is split up into JPEG sized chunks (each with a header that states the size of that chunk), the format used by the CEM DT-9885 doesn't use JPEG format in any way for the raw data. It's simply appended to the JPEG section of the image file. To find where it starts, use the technique I mentioned above. Not sure where the width and height are stored though, as I've not reverse engineered the header yet. I will be sure to provide more info as I dig deeper into the header of the raw data section. I assume the same basic format is used for other thermal imagers made by CEM (including their 160x120, and 640x480 thermal imagers). The DT-9885 is a 384x288 thermal imager. The JPEGs it saves have 640x480 for the JPEG compressed image, because they use upscaling, but the raw image stored at the end of the JPEG file is not upscaled at all (unlike the cropped then upscaled raw thermal image saved by the FLIR One).

It was great that you attached pics from your DT-9885 in your posts here. If you, or anybody else here has any other CEM thermal imagers in various resolutions, please post pics from them here too. It would help me to reverse engineer their format.

[bozidarms]

Hi Ben321,

Thanks for your effort - unfortunately i am not a software expert at all
and hard understand what are you talking about (old-fashion RF analog electronics are my things).

I have try to open a photo in Gimp and that was on ordinary way, don't understand what would be benefit
of your way?
This is my first thermal camera so i have no experience with other software for thermal images,
also don't know what is all possible to do with them?

Best regards