I have been using Photoscan for a while now to combine images from a UAV into mosaic images.
Most of my projects can be done in a single chunk, but one appears to be too large because if I break the images into 3 sections, each section aligns fine (Attachments 1, 2, 3), but if I try them all together there are gaps in the camera positions that align.
The project is an agricultural field that doesn't vary in altitude by more than 2 feet over the entire area, and the images all contain altitude values that don't vary by more than a few inches.
When I align and create a mesh for each individual chunk, everything appears normal.
When I try to align the chunks using "point based" and include all three chunks, I don't get a viewable result after several hours of processing.
When I try to align the chunks using "camera based" the chunks appear to be aligned at one end of the field, but at the other there are significant altitude differences, as shown in the attached "merged chunks" screenshot.
I have also tried extending the range of each chuck so that there are some overlapping images across chunk boundaries.
Is there any way to get the alignment process to use the image altitudes properly?
Suggestions appreciated
Topic: Chunk Alignment - Photogrammetric Images (Read 3398 times)
