Hello,
I have a series of photos from a drone that are of a building. Those images are 1920x1080 at 96dpi. They render great except I am losing detail on the sides due to the angle and height of the drone.
So, I also have a series of images taken from a camera of the front, corners and sides of the building. Those images are much higher resolution, they are ~4400x~3372 at 350 dpi. When I set control points on the camera images and matching points on my drone images, nothing happens when I perform a re-align.
- Could this be because of the resolution difference?
- I have ~4 CP’s on 3 or 4 different images, is that enough?
- Does the position of the CP’s matter? I mostly tried to make the points on the corner of windows, edge of the roof or edges of steps, etc.
- Can I set a CP on one image in component 2 (camera images) and then keep that image open and make the same CP on 3 or 5 other images from Component 1 (drone images)? Component 2, 3 and 4 are my camera images, Component 1 is the 100+ drone images that render ok.
- Periodically I also saw a red triangle on a couple of the images, but did now know why it was there, or what I did to get it to appear. Why does this appear?
- I read somewhere that the yellow triangle was ok. I believe this show up because I have no exif data on the photo, is that correct?
I have watched a couple tutorial videos, but I still cannot get any points to match so I must be missing something. I did try to lower the resolution on the camera images to match the drone and one of my tests did remove a component set, but I could not get any other images to align.
Sorry for the lengthy question but this has been frustrating me for a few days now so i thought it best to ask some people that may know!!!
Thanks for any help on the issue.
Best
Hi Tony Scamurra
Easy way to solve this 
Just add more images to the CP, that means have at least 3 images per CP and per COMPONENT. So for one CP you need at least 6 projections-images + spread them over the image like in my tutorial on this: https://www.youtube.com/watch?v=d8naLEtLqDY
Tony Scamurra wrote:
When I set control points on the camera images and matching points on my drone images, nothing happens when I perform a re-align.
You mean this does not result in registration of more cameras in new component(s) at all?
Tony Scamurra wrote:
- Could this be because of the resolution difference?
This should not be a problem.
Tony Scamurra wrote:
- I have ~4 CP’s on 3 or 4 different images, is that enough?
It is recommended to define at least 2 CPs in 2-5 images.
Tony Scamurra wrote:
- Does the position of the CP’s matter? I mostly tried to make the points on the corner of windows, edge of the roof or edges of steps, etc.
Of course, it matters. They need to be placed as precisely and correctly as possible.
Tony Scamurra wrote:
- Can I set a CP on one image in component 2 (camera images) and then keep that image open and make the same CP on 3 or 5 other images from Component 1 (drone images)? Component 2, 3 and 4 are my camera images, Component 1 is the 100+ drone images that render ok.
Yes, you can place the same CP on any project images. That is the point of doing it.
Tony Scamurra wrote:
- Periodically I also saw a red triangle on a couple of the images, but did now know why it was there, or what I did to get it to appear. Why does this appear?
The red triangle appears when you exceed the so-called “max feature reprojection error” set in the alignment settings.
Tony Scamurra wrote:
- I read somewhere that the yellow triangle was ok. I believe this show up because I have no exif data on the photo, is that correct?
See: Yellow triangle with an exclamation mark. More in the application Help coming out in the upcoming days.
Tony Scamurra wrote:
I have watched a couple tutorial videos, but I still cannot get any points to match so I must be missing something. I did try to lower the resolution on the camera images to match the drone and one of my tests did remove a component set, but I could not get any other images to align.
To merge components you have 2 options - add control points or more images - the latter is usually the faster way to do it.