1
The details depend on the comparison factors you want to use when comparing the images.
For example, do you want to compare the images based on specific objects in the image?
Do you want to compare them based on the colors count, intensity and colors distribution (how many times each intensity level occurs in an image)?
In all cases, you might be able doing this by using one of the following approaches:
1. Read the images data pixel-by-pixel and compare the pixel's color with your color. This is going to be slow, especially of the last pixel is the color you are looking for.
2. You can use LEADTOOLS Imaging SDK, which provides image processing classes like CorrelationCommand class, HistogramCommand class, etc. that can help you to compare your images based on the comparison conditions you want to use. For more information, see the following links:
https://www.leadtools.com/help/leadtools/v19/dh/po/leadtools.imageprocessing.core~leadtools.imageprocessing.core.correlationcommand.html
https://www.leadtools.com/help/leadtools/v19/dh/pc/leadtools.imageprocessing.color~leadtools.imageprocessing.color.histogramcommand.html
Disclaimer: I am an employee of this library.
3. If all images are paletized (contains palette), then you might be able create a custom solution to search the image's palettes without loading them.

1
try this,..
https://www.codeproject.com/articles/374386/simple-image-comparison-in-net