ICI and DALI thermal camera software ?

[zev]

A little more info:

I got analog out set up on the camera now.  As per the documentation (D8X3CUserManual attached earlier in this thread) analog video out is pin 1 of the 50 pin header on the board.  A little continuity testing showed that that pin is connected to the following traces, picture 1 shows where it is on the camera module and picture 2 shows it on the unit main board.

I spliced in so that the analog video connector in the rear cable is connected direct to the camera board instead of the main board, circumventing the main board entirely.

This in interesting, because it appears the analog out directly from the camera is a much higher quality signal than what is output over the ethernet port.  I wonder if the digital capture being done by the ethernet board is not taking the full 14bit depth or some other bottle neck, it is less sharp and has less dynamic range and sensitivity than the analog channel.

One interesting next direction to go would be to directly control the camera module.  This can be done with pins 3-7, and should allow full control of the camera without using the ICI main board at all.  As is, if you try to run the module without connecting it to the main board it only NUCs once at startup, which quickly makes it not a great image.  Very interesting that the analog video looks so much better, definitely points to the fact that a custom data capture 'backpack' could squeeze much more performance out of this camera.

[Bobertsawesome]

To add to this thread on different variants of the ICI FM 640 P...

I snagged one off eBay for ~$300, still a good price IMO for what's inside. It was absolutely caked with dust on the inside of the housing and must've been used for a while or in an industrial setting. It has a variation of the visible camera I've never seen before in teardowns (although noted as being on the original model by radziorr). A HIKVision (DS-2CD7027FWD) camera equipped with electronic focus control.

The core is the DLD640 and the backpack/core combo is the Dali-Tech D843NT, which is Dali's own module. I'm hoping that this will provide better implementation possibilities if I can get Dali to provide core documentation, plus it's got a much smaller footprint than the later ICI "Plus" models which had a separate processing board.

The login password is different when connecting via IP 192.168.1.102 than the newer modules noted in this thread. Exact same interface otherwise.
Username: admin
Password: admin


One thing I did notice is this whole module gets rather warm, which I'm sure isn't great for image sensitivity. I used another thermal camera to image the Dali module (I recognize the irony) and it seems the Altera Cyclone is the main heat source. It would be interesting to see if I can add some heatsinking or cooling to remove this heat. After several minutes the whole core housing becomes heated. I do have a stirling crycooler at my disposal, the Superlink Hybrid 98, which may call for some experimentation.

[Fraser]

Bobertawsome,

Thank you for sharing details of your cameras internal modules. Very interesting 

Regarding localised heating of the thermal camera module, this is not uncommon and not normally a problem. The Microbolometer normally operates in thermal equilibrium at around +32 Degrees Celsius. The die temperature is monitored by the electronics package. Cryogenic cooling of a microbolometer is not normally needed or recommended. If a component in the electronics package is generating too much localised heat it is often dealt with by adding forced air cooling to the enclosure and a small finned heatsink to the offending component. Another option is to use a copper heat pipe or solid copper ‘finger’ attached to the offending component and a large heat sink or thermal reservoir. As I say, heat from the electronics package is not normally a problem in larger cores where air can circulate. It is the miniature imaging cores with hot processors in close proximity to the rear of the microbolometer that can cause thermal gradients and other undesirable effects on the image output.

Fraser

[Bobertsawesome]

Thanks for the additional info 

Looking at the product page I believe the camera core is TEC temp controlled, so it would make sense that the die is probably a lower temp and the additional heat is sinking off the hot side of the Peltier.

[Fraser]

If a Peltier module is being used that is good news as it greatly improves microbolometer stability. The down side of the Peltier temperature stabiliser is that the core will be quite power hungry, especially at first switch on. You will see the Peltier stabilising the microbolometer by monitoring the current draw when the core is switched on. The current will drop significantly once operating temperature is reached. I personally like cores that have temperature stabilisation so it is good that yours has this feature 

Fraser

[Fraser]

Several nice thermal cameras were rehomed in private sales. I have multiple examples of some cameras so sold some excess inventory

I still buy the odd thermal camera if it is interesting enough and the price is right (it has to be inexpensive!) I no longer buy cameras purely for teardown and analysis posts on this forum though as my funds are very limited. Times have changed from when I used to buy many cameras to learn about the designs and technology within. I have now accumulated enough technical knowledge to repair both old and new cameras 

I have also moved onto repairing specialist fully rugged laptops, as used by the military, law enforcement and hazardous environment teams. That new interest distracts me a little from the thermal imaging scene  That said, a nice high-end industrial thermal imaging camera connected to a mint Dell Latitude Rugged 7414 laptop is a thing of beauty to me 

[CRC]

I just got one of the 640 type 2 units in the mail today. Wow,  this is really a very neat camera. I am able to see the thermal side via VLC using the address previously posted but I'm not understanding what address carries the regular camera video.

[angelortega]

i have that same unit. go to the Hikvision USA  website and download the software if i remember its the sadp tool
it will scan for any Hikvision camera that you have on your network
i used it and it showed that addresses for my color camera had bin changed
good luck

[CRC]

i have that same unit. go to the Hikvision USA  website and download the software if i remember its the sadp tool
it will scan for any Hikvision camera that you have on your network
i used it and it showed that addresses for my color camera had bin changed
good luck

Thanks, I tried it but it didn't locate it. Advanced Network Scanner found the camera's base address but that's it. I assume it must be a different port.

[angelortega]

All these cameras are using the same address and ports . I think that for security reasons , people are  changing the normal settings to keep other's from logging on.
I used sadp tool to change the address on the color camera the hikvision ds-2cd7027fwd-a, and saved it And I was able logon and it worked . Good luck

[angelortega]

I just check my files and found ,that port addresses you can try are port 8000 or 8080 or just 80. 
good luck

[CRC]

I just received another camera and it's the very large case style with the large visual hikvision camera and the thermal with the ethernet integrated to the back like pictured above by bobertsawesome. I was able to remove the thermal from the housing with two screws, apply 12 volts, plug in an ethernet and was immediately met with a very nice image via VLC. Loosening a torx ring clamp allows for focus.

Now, to find where the analog output is.

I also like the Wago/Euroblock style connectors they used. They are very easy to plug and play.

So far,  I'm extremely happy. The possibilities are nearly endless with these things.

[angelortega]

On my hikvision camera it's the green and white wire, and for thermal camera it's two pin on the upper left as pictured

[CRC]

Oh, the top row of the empty green connector. That worked great using one of the 2 pin plugs included in the kit. Thanks!

The image looks great.

[angelortega]

As for the thermal camera the first pin plus and second is negative

[angelortega]

You can get great looking  video out of gray colored wire connector I made to the upper left of the camera header's. But there's a problem, changing the color plate doesn't effect the output it will stay the same, If you check the others posting on this site I was discovered that these Cameras have a direct video output before it's sent to the feature boards for network output witch is slower frame rates. good news is you now have two outputs one direct fast composite video and one network video you can record from both.  enjoy yourself ok

[ArsenioDev]

I need to get around to documenting all the hacks I've got up on these.
Got some stuff that I'm whangjangling before SUPERCON though, as i'll be there.

[CRC]

Thanks for the info! I wonder if a guy could poke around and find post feature board color pallete processing?

[zrq]

VIS feed: rtsp://192.168.1.102/ONVIFMediaS

Thanks for sharing. I think you have a typo for the VIS feed.

There is a third RTSP feed: rtsp://192.168.1.102/rawdata used by the ICI body temperature software which transports the raw radiometric data in uncompressed format (but only 8 bits dynamic range?). However nothing can handle these unencoded data neither VLC nor ffmpeg, I'd like to try to hack something together using this library once I get the time: https://github.com/marss/aiortsp/tree/master

Also the firmware update package ICI_20200826_restart .upg in the Dropbox link posted by others can be easily unpacked. binwalk should find a tar.gz in it and then you can explore the filesystem. It seems there is not much to hack around though.

[CRC]

VIS feed: rtsp://192.168.1.102/ONVIFMediaS

Thanks for sharing. I think you have a typo for the VIS feed.

There is a third RTSP feed: rtsp://192.168.1.102/rawdata used by the ICI body temperature software which transports the raw radiometric data in uncompressed format (but only 8 bits dynamic range?). However nothing can handle these unencoded data neither VLC nor ffmpeg, I'd like to try to hack something together using this library once I get the time: https://github.com/marss/aiortsp/tree/master

Also the firmware update package ICI_20200826_restart .upg in the Dropbox link posted by others can be easily unpacked. binwalk should find a tar.gz in it and then you can explore the filesystem. It seems there is not much to hack around though.

Thank you! I was trying to obtain the visual feed and couldn't figure it out.