temperature measurement of gas turbine blade

i apply thermal paint to gas turbine blade i want to measure the temperature distribution on blade using image processing by considering the RGB value please help me this is my college project

Answer 1

[Image Analyst]

You've given us nothing to help on yet, but we can try. In the meantime, see my attached thermal example

In addition, see my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image to find blobs, and then measure things about the blobs, and extract certain blobs based on their areas or diameters.

[Tejas]

can you please help me i want image processing for thermal measurement

[Image Analyst]

I'll try. What can I do for you? Can you be more specific?

If you have any more questions, then attach your data and code to read it in with the paperclip icon after you read this:

TUTORIAL: How to ask a question (on Answers) and get a fast answer

[Tejas]

i apply thermal paint on gas turbine Nozzle guide. when thermal paint came in contact with temperature it undergoes permenent color change. we can measure the temperature of nozzle guide vane by analysing the color change. but i want to measure the temperature by image processing . whenever i move the pointer of mouse at any point the temperature should be shown . can you plz help me

i capture image of nozzle guide vane using 1+ camera

[Image Analyst]

You need to make up calibration images. First take a plate and heat it up to a known temperature. Then spray paint on it and capture an image of it. Do this over the whole range of temperatures you expect to encounter. Make sure the camera is in manual mode so that it does not do anything unwanted to the picture to try to "improve" it for you. Make a minimum of 3 images like this but the more the better, to account for non-linear color change with temperature. I'd recommend around 8 to 16 images. Now you can measure the color for each temperature and make up a formula to get temperature as a function of color. If you still can't figure it out, attach your calibration images when you've created them.

[Tejas]

i want to measure temperature at every pixel like this.

[Tejas]

i apply thermal paint on gas turbine Nozzle guide. when thermal paint came in contact with temperature it undergoes permenent color change. we can measure the temperature of nozzle guide vane by analysing the color change. but i want to measure the temperature by image processing . whenever i move the pointer of mouse at any point the temperature should be shown . can you plz help me

i capture image of nozzle guide vane using 1+ camera

i also provided temperature measurement strips. by using the strips we can measure the temperature manually.

but i want to measure by image processing, if possible provide me code also

[DGM]

You're going to have to do something about your illumination.

inpict = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/1367464/temperature%20strips.jpg');
inpict = imrotate(inpict,-90);

% these are spot averages taken from each test plate
% the illumination and coating thickness vary, so this is highly questionable information
CT0 = [125 49 35; 117 43 30; 101 39 24; 115 54 35; 112 58 32; 111 66 37; 103 74 40; ...
    120 85 45; 112 84 45; 108 92 59; 101 84 54; 102 92 57; 128 108 58; 139 115 53; ...
    139 115 53; 142 119 51; 146 133 54; 158 156 79; 152 160 98; 148 159 103; 141 158 103; ...
    142 159 105; 159 172 129; 136 151 114; 135 154 122; 164 166 91; 137 159 121; 126 153 118; ...
    113 146 117; 159 169 117; 117 151 124; 110 148 121; 100 133 109; 98 132 107; 137 159 123]/255;
T0 = [350 365 380 395 410 425 440 455 470 485 500 515 530 545 560 575 590 600 ...
    620 635 650 665 680 695 710 725 740 755 770 785 800 815 830 845 860];

% the trajectory in RGB
% ctpath(CT0,'rgb') % THIS IS MIMT-ONLY

it's hard to see in the image, but the trajectory between the samples is a tangled mess. In order to get anywhere with interpolation, this should be an orderly progresion through the (any) color space.

You could try to use it as-is, but even trying to test it on the same image, the results aren't accurate, since there are many cases where points are geometrically closest to points which are distant in the sequence.

% i'm not even going to try to interpolate the CT
% just use it as-is and see if samples map back to themselves
T = rgb2ind(inpict,CT0,'nodither');
T = T0(T+1); % this is a crude estimated map of temperature

% use the datatip tool
hi = imshow(T,[]);
dt = datatip(hi,591,134);

This particular sample does map back, but a lot don't.

[Tejas]

i want temperature measurement of NGV front which i attached in que. that temperature strip are color value

[DGM]

Yes, I know. What I'm demonstrating is the importance of having sufficient calibration information and good lighting control. If the information that's available is insufficient to accurately map the calibration image itself to the known temperatures, then there's no reason to expect other images to map accurately.

Think about it. How do you propose to compensate for the varied lighting in the calibration image and the working image? How do you propose to compensate for coating opacity that may vary from "opaque" to "none"? If you have a vane that had a uniform temperature, but one part is brightly lit and another part is in a shadow, and another part has only a very thin transparent layer of paint, how do you propose to map those all to the same temperature?

[Tejas]

we only paint on 7-8 blades.which is in red colour. i only want temperature measurement of that particular blades