Hello,
I have a set of 12 images with the camera traversing an object. How can OpenCV align (stabilization) the images? I want to create an animated GIF of this images.
Images: https://photos.app.goo.gl/gL4SCvuvZ9u...
Bad GIF :( https://photos.google.com/share/AF1Qi...
Thank you so much!
I cannot give you some code but maybe an idea: Its just an idea to discuss!
Well for face identification - the face is aligned before extracting the feature vector. This is done by using facial landmarks of person (eyes, nose, mouth, etc). Heres how its done
I think you could do the same for buildings. Extract landmark (windows, door, roof, etc) and use an algorithm to measure the angle and align the building.
How to extract building landmarks? Try this
Hope this helps, Good luck.
Asked: 2018-10-08 17:57:53 -0600
Seen: 283 times
Last updated: Oct 09 '18
Hmm berak do i understand you right - that - instead of post processing and correcting the misalignments - you just make sure you the error does not occur and use video stabilization instead?
This would make sense for me.
no, misunderstanding. it's not possible to "make the error not occur", right ? (that would be nice, though !)
video stabilization is an attempt at "post-processing and correcting the misalignments". have a peek at the videostab code. stuff like optical flow, inpainting, deblurring, and tons of least-square solving are used.
got it - thank you for clarification. I will take a look into it once i hit this problem but i know now a solution exists. That's great!
If you really want the error not occur - i think you can use special equipment like this https://en.wikipedia.org/wiki/Camera_...
hardware solution ;)
exactly :-)