1
$\begingroup$

Say, you captured a high-res image of a scene and a low-res image of the same scene, using a lower-res camera, ensuring that the lower-res camera is using the same technologies, except for the image sensor's resolution.

If I wanted to drop my higher-res image to the same lower-res image via digitally processing the image, what would be the right approach?

I reckon interpolation or naive downsampling are not the right way-to-go. They seem kinda illogical. I was thinking of using the RAW images and, perhaps, deploying these downsampling therein, and, then, using the pre-processing algorithms for converting the image into the final output. Is there some standard approach available, though?

$\endgroup$
  • $\begingroup$ basically you need to low pass filter before you down sample to avoid aliasing. If you want to reduce your resolution by an integer, it is pretty straight forward. if you want to do something like 7/9 it is a bit more complicated. $\endgroup$ – Stanley Pawlukiewicz Dec 8 '17 at 19:45
0
$\begingroup$

If you are using a “single-shot” sensor you could simply just not debayer the raw channels. That would cut your sampling without adding the extra overhead that is already baked into the image acquisition. A benefit of this approach is that you could probably boost your frame rate at the same time.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.