How do I determine heat in a thermal image using image processing?

Hi!
I have a project. What happens is that I put a thermal image in image processor and then I want MATLAB to determine where in the image is the heat (the red-yellow) part of the image.
can you please help me?

2 Comments

Image Analyst can you teach me. now i want to change thermal image to temperature scale by matlab
I don't know what that means. The thermal image already has units of temperature, or some factor times temperature.

Sign in to comment.

 Accepted Answer

First of all you need to get the original data, not some pseudocolored RGB image. That means you'll have to have the raw image file from the camera and a reader for it. What format is the original image in? Don't tell me PNG, JPG, or similar because that's no good - that will be just a displayed image with no real temperature data inside it.

3 Comments

hi! thank you for your reply.
whats needed isnt actually complicated. i just dont know how to do it. the photo's in jpeg (thats what i got from my professor) and what i need to do is to simply just determine in the photo where is the heat. coz in the photo the background is blue and the heat is red. so basically i just need to determine the red part.
i dunno if you understood what i said tho but i can give you more info.
That's actually the complicated way - a lot more. At the very least see if you can get it in grayscale, not pseudocolor. If there is absolutely no possible way around working with a low resolution pseudocolored image (poor guy), then you'll have to obtain the colormap somehow and make a look up table to convert the color into a temperature. Your temperature precision may be horrible with this bad approach.
You can call FLIR and ask them which menu item gives you the actual temperature data rather than gray scale or RGB colors. I'm not sure which menu item would do it. But you want floating point values around 0-100 degrees C rather than integers in the range 0-255.

Sign in to comment.

More Answers (4)

shazia sk
shazia sk on 11 Apr 2017
Edited: shazia sk on 11 Apr 2017
If I have understood your question correctly, it simply means you need to identify the red pixels from the image, since they are indicative of heat. I assume the thermal image has been subjected to some colour palette representing heat map of the image. From image processing point of view, channel split functions can segregate the RGB channels, and you can simply use the Red channel image for your purpose.

3 Comments

No, you can't as you'll realize once you think about it. You have to use the full colormap. I attach a demo.
Error in the thermal_image_color_to_gray scale.m program.The error line is histogram(thermalImage,'Normalizaton', 'probability'): How to solve this problem?
I just ran it again and it ran fine. For some reason you accidentally deleted an i in Normalization.
Change 'Normalizaton' to 'Normalization' and it should work.
But I'm attaching the newest version which gives the histogram in degrees rather than gray levels.

Sign in to comment.

This is a little bit more complicated than "go to other color palete". Flir save at the moment radiometric data as a radiometric jpeg file which is an extension of standard image jpeg file. The problem is that in file are only energy value not a themperature value. I try to use Flir SDK for Matlab at the moment but I do not have any results yet. There is not exacly true about "RAW Data", if You have FLIR ResearchIR You can export themperature value as a for example csv file and read them in Matlab. Of cource there are not raw data but You can calculate heat flow.
Bro I used to have this question for long time till I reach solution, you can extract temperature by using mat file that its generate it by thermal camera after that you can convert it to image and deal with it as image but when you want obtain temperature you must use mat files, for FLIR camera they have platform which deal with matlab effectively http://www.flir.com.au/science/display/?id=74637 please have look and check their code , good luck :)

1 Comment

Thank you Mohammed. It looks like that will be very useful to people using FLIR cameras.

Sign in to comment.

Categories

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!