Products > Thermal Imaging
Freeware Software for Thermal Analysis: Thermovision_JoeC
flotux:
Hi joe-c,
I'm a quite a new user of the UNI-T UTi260B thermal camera. So I really appreciate that you now support this model in your great software!
I just sent you an email with example pictures. If you need more examples please let me know. I will support you as good as I can.
I find it extraordinary how this great software gets even better from version to version. For me it is now the overall standard software for thermography.
Thanks for your great effort.
Best regards
Florian
flotux:
Hi joe-c,
I checked the debug version. It looks quite good for the first shot. A BMP can be opened and looks exactly like the one with the palette create by the camera.
Also the range (T min. and T max.) looks good.
But the automatic spots for min. and max. temperature are not working correctly. As you can see in the attached screenshot the data in the meassurement table is wrong.
Therefore the spots for T Min. and T Max. are not positioned correctly.
Keep the good work!
Florian
joe-c:
Hi All,
Finally... Release of 1.11.0.0
Google drive link: https://drive.google.com/file/d/1veBPcaGeGsFQ1YwvGq9vd2KxtRQ-2tvA/view?usp=sharing
Website (german): http://joe-c.de/pages/posts/version_1.11.0.0_219.php
Change notes:
-Mainform->Vision Toolbar: generic visual stream from Webcam A/B
-Tablet mode: improve scale handling of touch events
(removed toogle on long Press at scale)
-list only selected devices in imagebrowser
-add imagebrowser support for Uni-T, Bosch GTC 400, Variocam
-imagebowser add select and search images to mouse menu
-hide device is now a dynamic table, instead of single checkboxes
-Serial sensor support Optris-CT temperature
-rotation (if not disabled in setup) on each ImportThermalFrame
-remove vision rotation and add generic rotation (0°,180°,+90°,-90°)
-radioFile 2.2 (frame 2=planck raw,meas 2= with ranged box)
-save sequence (radiometric video) Disabled, need restructure...
(its planed to support the old files from Version 1.10.0.0 so far)
-autoload "seek" if start manually
-read Windows version at startup (extra dll call for Win10 for better view)
-tempswitch are now collected in function panel
-tempswitch setup one file for all items, with autoload from app.config
-bugfix webcam tool (dont start before)
-add interpolation x2 for raw frame
-hide devices: collection of check boxes -> dynamic table
-min/max auto 3x3 kernel disabled
-bugfix map & Planck cal
-autodisable Mapcal if not match
-resturcture ThermoVision Core
-PanckCalWindow: add planck Curve (display raw 0-0xffff conversion)
-now 2 generic Planck cals (Base for fixed TV and Global in Cal, overwritten by flir images)
-add Planck reverse (Temp to Raw)
-TempSwitch: add Alert to txt
-Image Processing: remove death pixel function (if diff to left pixel is over treshold -> replace with left pixel)
-FlirCameraCommander enum FlirCameraType { Normal_QtGui, Legacy_ThermaCam }
-add second 2p-cal (fever screening extension)
-add BoxRanged (define 5 max ranges in Area)
-buxfix while import
-bugfix 3x3 kernel
Histogram:
Now there are to ways to generate a Histogram
3x Fixed ranges
generate a list from min to max value with selected step range an count values for each Pixel. This type of histogram has a fixed stepsize and may have empty holes(slot for a temperature, what not exist in the image). Image quality markers are:
"Nr of Zeros" -> it shows how many entries has no matching pixel, lower is better
"Values" -> shows how many different pixel values are existing, higher is better
4x Dynamic ranges
this round each pixel value to selected range and add or count the pixel in a new list. This type of histogram can not have empty holes, but it can have different stepsizes. Image quality markers are:
"JustOne" -> number of pixelvalues which only exist once
"DiffValueCount" -> shows how many different pixel values are existing, higher is better
"StepDiff xxx" -> shows the minimal, mean and maximal difference between the temperature steps in the list, lower is better
Note: the Histogram can not differentiate between a sensitive camera and a noisy camera. It just shows a analysis of the existing Temperatures.
Changes for Cameras:
-improved: startup for DIY Thermocam
-improved: seek thermal raw range higher (before 6k-17k, now 1k-65k and adjustable)
-add camera support: Seek *.HIR read thermal frame (and *.pre)
-add camera support: Uni-T *.bmp files
-add camera support: Variocam *.irb files
-add camera support: Flir/Agema *.img files (experimental)
-add camera support: Bosch GTC 400 *.JPG (use linear Maping of raw values, maybe not accurate)
I attach a comparison of 320x240 Cameras with the new Histogram function.
best regards
santasemilla:
Dear Joe-c,
As usual, thanks for all your work and your kindness sharing it.
I just tried the 1.11 version and I'm amazed with the features and the UI polishing. Nevertheless, I have some questions about it. I'm still playing with version 1.09 as I see is the most stable for me. I'm using a Seek Compact camera and a Seek Compact PRO, and I'm testing all versions in a Windows tablet (up to date Win10 x64) with a CoreM processor and 4GB of RAM.
When trying my thermal cameras on version 1.11 all the problems are focused in the refresh-rate of the thermal cameras and the live image of the visual camera. For example, on version 1.11 if I load the Seek Compact on a live image, it works at ~12 /15 fps (that's fine!), but when adding a live spot measure, it drops to 1 or 2 fps. If I connect the Seek Compact PRO via the WinUSB driver, it lags as bad as 1fps or maybe less.
On the visual camera side, the image tab can barely update the image (regardless of which thermal camera is plugged, Compact or Compact PRO). Resolution of the visual camera could be 640*480, HD or fullHD, the effect is the same. Also, if I do a blend with thermal and visual, the update only occurs with the shuttering of the thermal camera. The behavior was quite the same on the previous 1.10 version for me, it lagged a lot, but now on 1.11 it is stronger.
I couldn't try editing saved images in 1.11, but in 1.10 images it used to freeze and update only if color scale was changed, same if I tried to load the visual image, one weird thing was that it kept the visual from the previous loaded thermal image, and it updated only with a color scheme change or some specific editing actions. I can update later if in 1.11 happens the same. On version 1.09, none of this happens, all is smooth and fluid.
I know my computer is not as powerful as a core i3/i5 and I also know you completely rewrote the program on version 1.10. Maybe I'm missing something, a driver a library (as all my issues seem related to image handling), but I just wanted to let you know as the program is excellent and super feature rich, but seems I need to stick to 1.09 by know.
Again, thanks for your hard work and please let me know if I can further test the software and give you more specific feedback.
Cheers!
joe-c:
Hi santasemilla,
thank you for the feedback.
It's very interesting, you could see a significant frame rate drop after enable the Spot measure. This should not happen.
I try to reproduce this on my tabled and try to figure out the reason for this.
I can tell you why the change of the resolution in the visual was not shown. This Visual image was taken from the MainIR overlay, therefore you can enable the Visuel relief stuff here. But the Overlay size is identical to the Ir size.
I need to add a second image stream for the visual, instead of just take the main IR and adjust the overlay parameter.
On of the biggest problems for performance is the Planck calculation for each pixel in each frame.
i should try out to use the temperature calculations only for enabled measurements.It's anyway not a really performant way to have each time both types of image available...if the camera delivers a uint16 pixel its converted to a temperature, if it delivers a temperature the reverse formula is used to give a uint16 (raw) value. I want to combine both on "import thermal frame" to be sure all calculations, histogram, measure and plot stuff works for all cameras similar.
i even use the .net framework 4.0 since it's the last version that supports winxp.
It's maybe time to go to a higher version including their performance improvements.
Thank you for sharing your observations.
best regards
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version