How can I filtered the color of the combustion image?

Hallo everyone,

During my spray combustion experiment, I faced the problem like below:

• I accidentally forgot to attach the short pass filter (600nm) when capturing the combustion images using a high-speed camera.
• The filter supposedly to block the wavelength above 600 nm.
• This resulted in overexposed images, with the maximum luminosity reaching 65520 au, instead of the expected range of 0-255 au.

My initial idea was to filter out the color that represents wavelengths above 600nm (which corresponds to colors like orange, around #FFBE00) in original images. This supposedly would help me to remove contributions from emissions outside the OH* chemiluminescence region, which supposed to isolate with the short pass filter.

Instead of processing all wavelengths, the idea is to remove the specific color which represent wavelength range (above 600nm) using a color filter in MATLAB. This probably would allow me to approximate the effect of the missing physical filter by digitally excluding those higher wavelengths from the images.

I also attached the image and the code, but it seems it doesnt change anything from the original image.

Could anyone suggest another approach regarding this issue?

Best Regards,

Thank you very much

Answer 1

[Image Analyst]

Since the blue and green channels in your camera are sensitive to wavelengths in the 400 - (roughly) 600 nm light wavelengths, you can simply ignore wavelengths longer than 600 by just setting the red channel to 0.

rgbImage(:,:,1) = 0;

[I Komang Gede Tryas Agameru Putra]

Dear image analyst,

Thank you for answering the question.

But would it be altering so much the image by looking like below image if the red channel is removed?

However, this below image is the expected results if physical short pass filter is applied on high speed camera.

Without filter, the image is over exposure and appear brighter so it dimished the some of the flame detail.

Is it possible to resolve this issue in MATLAB?

Thank you for your support.

[Image Analyst]

If it's saturated, the data is clipped at 255 and lost forever. You would not know if the value should have been 257 or 459 or 1035 or any value above 255. So there is no repairing or recovering that image.

That said, all may not be lost depending on what you want to do. For example if you were going to take a particular color channel and threshold it at, say, 160, and get the area of the flame, then you can still do that, though the specific value of the threshold would be different than with an optically filtered, non-saturated image.

[I Komang Gede Tryas Agameru Putra]

Dear Image Analyst ,

Thank you for your comment regarding this issue.

I got your points now.

Best Regards.