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
Hi, great plugin, thanks.
It doesn't look like there isn't any active development, but I think it is worth filing these issues in case anyone decides to work on it in the future.
I've been abusing the plugin, using it to deform a DEM.
I imagine this means I am deforming it a lot further than people usually would when rubbersheeting, which may make it easier to notice any problems - I guess it would be worth testing extreme deformations even with a normal image.
Here are the problems:
With large inward deformations you will notice that part of the original image still shows up at the edge, duplicating part of what you correctly see in the deformed image. When I test with an aerial that I haven't clipped, this problem seems to be fixed or worked around by using a buffer, but that doesn't solve it with the DEM (which has null values around the edge), or with an aerial that I have clipped so that it has null values around the edge. I believe in these cases the plugin should erase the original values in that area, replacing them with null values.
When working with a DEM the plugin creates some pixels around the outside with extremely low values. I'm not sure if this is cause by the null values around the outside, or by some sort of extrapolation. It doesn't seem that anything like this occurs when working with an aerial; maybe it only works properly with RGB images or something.
The plugin also creates some pixels along the triangulation lines with extremely low values. Again, it isn't obvious if there is a similar problem with an aerial.
It looks like the triangulation algorithm could be improved; it creates some edges crossing over each other. Try using the attached pairs layer to see.
Hi, great plugin, thanks.
It doesn't look like there isn't any active development, but I think it is worth filing these issues in case anyone decides to work on it in the future.
I've been abusing the plugin, using it to deform a DEM.
I imagine this means I am deforming it a lot further than people usually would when rubbersheeting, which may make it easier to notice any problems - I guess it would be worth testing extreme deformations even with a normal image.
Here are the problems:
With large inward deformations you will notice that part of the original image still shows up at the edge, duplicating part of what you correctly see in the deformed image. When I test with an aerial that I haven't clipped, this problem seems to be fixed or worked around by using a buffer, but that doesn't solve it with the DEM (which has null values around the edge), or with an aerial that I have clipped so that it has null values around the edge. I believe in these cases the plugin should erase the original values in that area, replacing them with null values.
When working with a DEM the plugin creates some pixels around the outside with extremely low values. I'm not sure if this is cause by the null values around the outside, or by some sort of extrapolation. It doesn't seem that anything like this occurs when working with an aerial; maybe it only works properly with RGB images or something.
The plugin also creates some pixels along the triangulation lines with extremely low values. Again, it isn't obvious if there is a similar problem with an aerial.
It looks like the triangulation algorithm could be improved; it creates some edges crossing over each other. Try using the attached pairs layer to see.
example files.zip
The text was updated successfully, but these errors were encountered: