Thank you, but why so many control points? For better precision? Usually 4-5 well distributed (in different corners) points is enough. The part about saving the raster was useful though!
Hi! Thanks for the input. True 4 to 5 well distributed points works well with Affine transformation. My problem was the I had a poor distribution of points (and some colinear points) due to the fact that the vector data was considerable older than the image.
@@LuisCBerrocal Also if is an airphoto it could have additional distortions (relief displacement, crabbing) that could be adjusted for with a 2d order transformation and lots of control points. Especially if you don't have time to get a DEM and due true ortho photo creation.
@@curtvprice yes Curtis! Thank you for pointing that out. This tutorial is focused to my students running the intro to GIS course. I will make another video attending this generalization.
Great tutorial, but not convinced about the resampling type when saving the raster data. Your spatial references were buildings and other objects with well defined boundaries, thus discrete objects. Why using bilinear interpolation, then?
It still is an airphoto so you want the output to be smooth. Nearest neighbor would give the image a blocky appearance making it harder to interpret. So bilinear or even cubic is appropriate for the imagery.
@@LuisCBerrocal I think Parallels runs ArcMap great, but you need to make sure you have enough memory to allocate to Parallels (ArcMap really needs 8G RAM on the VM to run well). I run Parallels off a boot camp partition, this allows me to reboot and run 100% Windows if I need more resources to do something more intensive. Thank you for this tutorial!
Great tutorial. Thank you. I have to do this on so many TIFFs tomorrow at work!
Thank you, but why so many control points? For better precision? Usually 4-5 well distributed (in different corners) points is enough. The part about saving the raster was useful though!
Hi! Thanks for the input. True 4 to 5 well distributed points works well with Affine transformation. My problem was the I had a poor distribution of points (and some colinear points) due to the fact that the vector data was considerable older than the image.
@@LuisCBerrocal Also if is an airphoto it could have additional distortions (relief displacement, crabbing) that could be adjusted for with a 2d order transformation and lots of control points. Especially if you don't have time to get a DEM and due true ortho photo creation.
@@curtvprice yes Curtis! Thank you for pointing that out. This tutorial is focused to my students running the intro to GIS course. I will make another video attending this generalization.
Great tutorial, but not convinced about the resampling type when saving the raster data. Your spatial references were buildings and other objects with well defined boundaries, thus discrete objects. Why using bilinear interpolation, then?
The quality of the picture was very poor and the prescion of the building layer was unknown. Please advise on what interpolation you would recommend.
It still is an airphoto so you want the output to be smooth. Nearest neighbor would give the image a blocky appearance making it harder to interpret. So bilinear or even cubic is appropriate for the imagery.
ArcGIS Desktop is support on Apple already now
Unfortunately as of this date ArcGIS does not run on Mac OsX I'm using a Windows emulator called Parallels...
Wow... Is it good workly?
Not that I know of.
@@LuisCBerrocal I think Parallels runs ArcMap great, but you need to make sure you have enough memory to allocate to Parallels (ArcMap really needs 8G RAM on the VM to run well). I run Parallels off a boot camp partition, this allows me to reboot and run 100% Windows if I need more resources to do something more intensive. Thank you for this tutorial!
@@curtvprice Hi!! yes that is what I use Parallels to work with ArcGIS in my Mac
Thank you very much, great explanation.