thermal images

I want to do research using the thermal images to detect temperature for fever person. I want to ask, given the thermal images,and I have to input the image in matlab, how to verify the temperature in face to detect fever, is it based on the color in the image? Next, I have to use colorbar as indicator for temperature in the human face, how to set the color in colorbar parallel with the color in thermal image?

I really need your ideas regarding this matter since I am new for this toolbox. Thanks

Answer 1

[Image Analyst]

You need to analyze the original grayscale image, not the rgb pseudocolored image. To verify you should look at pixel values and compare the intensity at that location with the temp measured with another device, like a thermometer, and use on individuals known to have, or not have, fever.

To apply pseudocolors, use the colormap() function. There are a variety of built-in colormaps you can choose from, in addition to building your own custom colormap.

[nur izzati]

So, the temperature indicator of the thermal image is not based on color of thermal image, but based on intensity value of the image?

I try using colormap,

I=imread('url.jpg');

colormap(I)

and show this error, Error using ==> colormap at 89

Colormap must have 3 columns: [R,G,B].

Name Size Bytes Class Attributes

I 1135x1200x3 4086000 uint8

I dont understand why this happen since this is RGB image.

For the colorbar, how to make the color in colorbar same with the color in image?

Thanks

[Image Analyst]

The image is not a colormap - it's an image. You need to pass it an M by 3 array or normalized RGB colors. Try

colormap(gray(256));

Make sure you're supplied with a grayscale image, not the saved pseudocolored image which won't do you any good. What does

size(I)

say?

[nur izzati]

Dear Analyst,

Thanks for your answer, I've got colormap already.

For fever detection using thermal image, do you mean that I have to compare pixel value in grayscale image for both fever and non fever person?

The comparison with other device like thermometer only may be done after I may detect the temperature in thermal image, is it?

[Image Analyst]

Yes, grayscale on both normal and fevered individuals. In addition to thermal cameras you could use handheld spot thermometers, like those that project a laser spot - you've probably seen them.

[nur izzati]

I have compare the pixel value for both fever and normal person and I take the reading in the area of eye, the pixel value for fever person is nearly 255, and the normal person is about 125++, so is that mean the pixel value for fever person is nearly 255, and we just may conclude that for fever person , the pixel value is ~255?

Thank you for your advice

[Image Analyst]

Well you said it is - I have no way of knowing. Your camera must have some relationship between gray level and actual temperature, like 125 = 98 degrees F, and 255 = 106 degrees F and values in between are temperatures in between. Fever is a continuous thing you know, it's not binary. It's not like you either have it or you don't, it's a continuum from normal on up. You can set the "normal" range to be whatever you want. Maybe fever starts at 100 degrees, maybe it starts at 101 or 99. Pick whatever you want.

[nur izzati]

ow..ok..Thank you for your information.

[Lucaci Diana]

Did you manage to assign those value into temperatures? like 125 = 98 degrees F, and 255 = 106 degrees F, this section, because I'm doing something related to thermal images and I can't figure how to do this to my image and it would be a great help. Thanks.

[Image Analyst]

You might be able to undo the colormap if there is a colomap embedded in the image. See attached demo.

Answer 2

[prasad vaidya]

if we want yo perform thermal image processing on matlab the which type of camera we can use?? is it a thermal imaging cam or normal one?????

[Image Analyst]

You would not be able to "detect temperature for fever person" with a regular visible light camera. I got scanned for fever as I walked past an stand at an airport last summer - can't remember whether it was Osaka or Beijing - but they were thermally imaging all the incoming passengers.

Answer 3

[Md Sohrab Ali]

how to detect the hotspot in thrmal image