This is by far the most helpful video I could find for georeferencing, even though it's from 2 years ago. Thank you very much This saved me a lot of time
wow alex! you so helped me. your video is so concise and well spoken, what a pleasure to find you. i'm going to watch all your videos, i just wanted you to know how well spoken you are...and no background music - yay!
Thanks - this is very well done. The only thing I’d add is that you need a few known checkpoints that are not used in calculating the model to serve as a reference for accuracy.
Thanks! I agree. For real-world jobs we always have check points to assess the work. This video is intended for students who are just starting out and getting the basics of the workflow. I'm glad you liked the video :)
Is the mesh important for the process or is it just for us to see better what we're doing? What are projections and what should we do if we don't have many, or any on some GCP's? I just located 4 GCP's in about 10 photos each but I only have 9. 6. 4 and 0 projections
Hi, can you please tell us what are your best practices for photogrammetry. what height do you fly? what overlap? what camera angle? ... Thanks in advance!
Oftentimes, it depends on what you are trying to photocapture/map, but in general I fly somewhere between 25 to 60m (depending on the ground sample distance I want). I set the overlap between 70-80% and run a double grid. Camera angle about 70 degree, slightly off nadir. But again, depends on the terrain.
@@alexelvisbadillo Thank you. The problem i experience with non-nadir pictures is that as soon as you have vertical structures (any kind of building really), there will always be a face that will not be captured well, resulting in blurry models and orthomosaics. What about double grid with nadir pics?
@@arnaudlrc If I am working in a landscape that has buildings, and I want to capture the building, I combine the nadir (top-down) imagery with oblique imagery (usually 45 degree angle). This way, I can ensure that I am capturing the sides of the buildings.
You may have left the cameras checked in the reference pane. You will want to uncheck all cameras so that when you hit the update transform button, the transformation is based on the markers(GCPs) rather than the cameras and markers.
Mine does not seem to be automatically doing projections in the tab. It moves the points though. Does it need to say projected in order for it to take it as true? is there a button accept all auto projections? thanks
After placing a few markers on three or more GCPs, I hit the update transform button at the top of the reference pane. Then I continue adding the other GCPs, and at the end I hit the update transform button a final time.
@@alexelvisbadillo Circling back, when you clicked one manually, the blue flags came up automatically. What exactly does this mean and is there a setting to turn on for this? Mine just stay grey. Thanks
@@Nardzz019 this is a mystery to me too. What your describing is guided marker placement. In theory, after running a low resolution mesh (200,000), placement of the markers will be aided through this feature. I have noticed that on my laptop this works, but on my other lab computers, it doesn't seems to work. Please let me know if you find out why this is. Check out p. 77 of the Agisoft Metashape manual where they describe this.
The GCP altitude measurements should be more accurate than the altitude measurement from the geotagged photos. That's why we use the GCPs as reference data.
