Products > Thermal Imaging

I got my Seek Thermal SDK!

(1/4) > >>

Ben321:
I went to developer.seek.com and registered an account there, including an explanation of my intended application (that was a required field to fill out on the registration form). I explained that even though I didn't have one of the OEM cores, that I did have a Seek Compact Pro for Android with micro-USB connector, as well as a micro-USB to USB adapter to allow it to connect to a PC, and that I planned to use the SDK to develop Windows software for personal use to experiment with my Seek Compact Pro on my Windows PC. A few hours later I got back an email that said my developer account was activated! I'm surprised they didn't outright deny my request for access to the SDK, since the SDK is normally intended to normally be used to develop software for the OEM cores, and is also clearly aimed at businesses (with the registration form even going so far as to ask what company you work for that will be developing a thermal imaging product, and making that a required question that you need to answer, which I simply filled out with the answer "none").

And I'm not joking. To prove that they did actually give me access to the developer portal, here's a screenshot its main page (no confidential info is here, just the links to pages with such info and downloads for the SDK, so I think it's safe to post this screenshot).

Ben321:
Ok, I just tried the oldest version of their viewer (the one that comes with SDK 3.8) with my Seek Compact Pro connected to my laptop via a micro-USB to USB adapter. And it WORKS! By the way I had to first uninstall the WinUSB driver it was using (that was what I'd installed with Zadig before, in order to use 3rd party software with it), and then install their OFFICIAL Seek driver for it. The software locked up when using the WinUSB driver. Here's a screenshot of the display from this older viewer software.

Ben321:
I also tried the viewers that come with the newer SDKs (4.0 and 4.1) and these have terrible flat-field correction! The flat field correction that is performed by the viewer in SDK 3.8 is great, but all the later ones leave a HUGE gradient. No matter how long you leave the camera running, it NEVER corrects for this gradient. I'm not sure if this is an issue with the programming of the viewer itself, or an issue with the DLL that contains the API functions (I haven't looked myself to see if FFC is done in the viewer's source code, or if it calls a DLL function to do it). Here's a screenshot from the viewer in SDK 4.1 to show you just how bad the gradient is. This was captured while the Seek Compact Pro was looking at an object with a constant temperature across its surface.

Ben321:
Here's a screenshot of the same object, but this time using the viewer that comes with SDK 3.8. Yes, there's a lot of noise, but that's to be expected with an object that is of constant temperature, as the viewer is using AGC to stretch even small temperature differences (mostly from noise) of the flat temperature surface into a wide range of values for display in the viewer window.

Ben321:
I just realized that there was a difference in the AGC mode between these two viewers. The 4.1 and 4.0 versions I had set to use linear AGC, but the 3.8 version viewer only offered one mode (though it wasn't stated) which clearly histogram equalizing AGC. Even then though, that doesn't make up for the terrible FFC of the 4.1 and 4.0 versions of the viewer (and possibly an issue with the underlying FFC function in the DLL if it has one). I know this because I went back and set the viewer for SDK 4.1 to use histogram equalizing AGC, and still the gradient is very obvious.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version