You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In particular for captures over water, glare can be significant in the input images. Ideally we could (1) detect glare in the source images and (2) during texturing, de-prioritise selecting pixels classified as glare.
While currently texturing favours pixels around the image center for nidar captures, this change would mean it could opt for pixels towards the image edge if they are less affected by glare (as is the case for this example below).
For detecting glare, rather than trying to classify each image individually, I'm thinking it would work better for this to happen after the images are aligned, so we can look across the pixels from different source images at the same ground location.
Activity