BBA - AAT tips

Technical questions and comments on EnsoMOSAIC

BBA - AAT tips

Postby Janne » Tue Jun 28, 2011 4:07 pm

Hello,

here are a couple questions from a user - the answers may be useful to others as well:

1. I have a project that is fairly large (16 flight lines) with some cloud cover in a few lines. Initially, when I ran the initial AAT 5-5 it finished with ¾ of the frames unattached. Therefore, I redid the project with more frames and the initial AAT 5-5 came out better (all tiles green). However, when I run the BBA it always ends in an error. I have added tie points (compute/update links before adding points), tried to run the BBA in small sections, and adjusted my weights. Do you have any suggestions on how to fix this issue?


Since after initial 5-5 all tiles are green you can proceed to BBA, as you have done. The steps you have taken are correct. The actions to take on the initial level 5 to be able to continue to Intermediate 4-0:
a. in Block Adjustment, deactivate “Use results of previous run for initialization”. If active, it will use the results of earlier runs which may be totally wrong and cause any new BBA to fail. Run BBA for a partial block.
b. if it did not help, go to the folder where you have TRP and GPA files, and delete file “block_name.BBA”. This way you will get rid of any possibly crazy results of any previous tests and badly ended runs.
c. Run block adjustment, as you did before, starting by a subset of the images and by growing the block after each successful run. In between the runs delete the points with the highest residuals. Remember to reactivate “Use results of previous run for initialization”
d. When you have all the images successfully solved and the highest residuals eliminated you can continue to Intermediate 4-x. Remember, that you CANNOT continue to Intermediate before you have a good result on Initial level, as Intermediate calculations cannot autocorrect large errors.




2. I have a project that has two images (in different lines) that have a different orientation than the adjacent frames. The mosaicked image is severely distorted in these areas. I have added multiple GCP and tie points, deleted bad autopoints, and then run the BBA again. However, it still comes out bad. Even though I have successfully completed the AAT Final 0-0 with these bad images attached, can I turn off these bad images? Would that help or cause an issue because I have automatic and manual tiepoints, in addition to GC points connecting it to the adjacent images?

If I add many tiepoints and GCPs should I only run the BBA or should I rerun the Intermediate AAT 1-1 through the Final ATT 0-0?


If you wanted to join those badly oriented images into the mosaic you would have to measure those well manually, or you should have joined those already well in the initial level. Now, if you don’t need those image in the mosaic, you can mark those OFF at the final level. The software will automatically deactivate the points that were measured on the deactivated images.

When you add manual points or control points you don’t have to do Intermediate AAT again. But you have to do BBA every time you edit the points in order to make the editions effective.
Janne
 
Posts: 67
Joined: Thu Jan 21, 2010 8:57 pm

Return to EnsoMOSAIC in English

Who is online

Users browsing this forum: No registered users and 0 guests

cron