Handle Missing Map Sections
Sometimes a map might be missing a section, even though the pictures for that section have been uploaded. This is a sign that something went wrong during the reconstruction process.
While it might be tempting to blame the software, 9 out of 10 times a missing section is caused by an underlying issue with the input images.
For starters, we recommend people read our guide on creating successful maps and our instructions on image capture, which include specific information for handling different scenes. Do not assume that the mission planning software suggested the correct overlap, or that the overlap requirements for one scene type automatically apply to another.
Below we report steps that attempt to mitigate the problem if recapture of a mission is not possible.
The mitigation steps sometimes work, but they should never be used as a replacement for best data capture practices.
Causes
1. Insufficient Overlap
If images don't have sufficient overlap, the software might be unable to process the entire dataset. See the image capture page for overlap requirements.
Forest areas are especially prone to having missing sections caused by insufficient overlap. Use at least 85% overlap and 85% sidelap for these areas.
2. Flying Too Low
Fly at the optimal altitude for your desired target resolution. Don't fly lower than you absolutely need to. Higher elevation enhances feature matching, reduces distortion near building edges, and allows for greater image overlap and coverage. It also reduces the data capture time.
This is especially true for forest areas. Trees are difficult to reconstruct at low altitude, as wind causes leaves to move, which complicates the image matching process
3. Motion Blur
Caused by an inadequate shutter speed or excessive flight speed.
4. Flying Manually
Use a mission planner to reliably capture a dataset. Flying manually is sometimes necessary, but it's easy to miss the overlap requirements.
Mitigations Steps
If you are unable to fly again, you can try to tweak some of the software's task options and reprocess the dataset.
If you need to reprocess the same dataset multiple times, consider purchasing a monthly plan to avoid consuming credits or use a smaller subset of images for testing purposes to reduce your cost.
- Increase min-num-features to
18000
. - Change feature-type to
akaze
. - Lower feature-quality to
medium
(yes, it's counter-intuitive, but it can help reconstruct certain difficult forest areas).
It can be frustrating to have the section of a map missing. By following best practices for data acquisition, you can significantly reduce the chances of this happening.