If I have a multiple exposure image, is there a way to deconvolve/separate the multiple layers of the image into separate pictures?
My camera can't do high speed but it can do multiple exposures so I was wondering if there's a way to post-process the multi exposure images and recover each image individually. The image is quite simple, it just shows a quasi-spherical flame front propagating over an initially black background.
Thank you!
It might be difficult to recover the actual original image contents. Note that the multiple-exposure image may not be a simple linear combination of the individual exposures. There may be non-linear operations used by the camera in producing the combined image. E.g. if there is an area of motion or significant change between the exposures, the fusion algorithm may have selected the content from one of the images over others, to avoid motion blur or artifacts in the fused/combined image. In other areas, it may be a linear combination of multiple images.
It would help if you knew exactly the algorithm that is being used in the camera to generate the combined image. If you have a more accurate "forward" image formation model, you have a better chance at inverting it.
However, I wondered what your end goal is. If the final goal is to do a measurement of the flame front, e.g. detecting the outline, or measuring the area, or something like that, then you don't have to actually recover the image content. Instead, you could try to do the measurements directly on the multi-exposure image, using all the prior knowledge you have.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing