Orthophoto of river mouth

I am trying to put together an ortho of sand banks at a shallow river mouth so the local coast guard can look for a channel to navigate through to deeper water. I am acquiring the images with flights using a multicopter at 300 - 400 feet.

The water is clear enough and shallow enough to see the undulations in the sand under the water. However the images will not stitch together unless there is a land feature in the image. Is there a way around this issue?

Hi, unfortunately water is quite impossible for photometry.
If there is clear pattern underwater that do not move between two pictures it could work.
Which software do you use ? Try using different software (Pix4D and Photoscan for example have free demo version).

Thanks. I’m using Photoscan and will try Pix4D.
Is the same issue there with other featureless terrain like even grassed areas at low level or densely vegetated areas?

Hi, yes its the same problem. At least with trees. The problem is, orthophotos are made of more than one picture. The software (photoscan, etc.) looks for identical parts of the images to recognize ground features. But trees and waves and other things move due to wind between pictures taken by your multicopter. That means these features are not identical (in the eyes of photoscan), and therefor can’t be merged. Grass is not so bad. It depends on camera resolution and altitude - whether little details are visible on the picture or not.

The less features you have the more ground control points you have to put down is generally the rule.

For any real accurate photogrammetry we try to have 12 to 15 GCP’s minimum.

The software can then key off those points.

If you can’t align the photos manually then then the software won’t be able to either is the accepted view.
Have tried doing it manually in good 'ol photoshop?