|
Size: 1449
Comment:
|
← Revision 26 as of 2019-11-08 16:47:21 ⇥
Size: 2319
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 4: | Line 4: |
| = STILL BEING EDITED! = |
|
| Line 8: | Line 6: |
| This block batch registers images. | This block batch registers images. Each time you run a batch run of run_script.b (typically would have a different make_scriptR.seed), then run [[blockFinish]] to log the results. |
| Line 10: | Line 8: |
| == Assumptions == * You have invoked a Bourne-Again Shell (bash) * You have completed the processing steps in Block 1 - Setup Directory * You have populated the working directory with new images following the steps in: * [[Block 2 - Ingest v2]] |
|
| Line 18: | Line 11: |
| === 2. Set-up === | === Set-up === For script, set the ID |
| Line 20: | Line 15: |
| Ensure make_scriptR.seed is linked to the script you want to run. | ID=312 |
| Line 23: | Line 18: |
| === 3. Set-up batch register files === | Ensure make_scriptR.seed is linked to the script you want to run. For starters, you can use one of the default ones. If needed, you can craft something specific and place it in lsupport. |
| Line 25: | Line 20: |
| # set-up batch register files ln -s lsupport/$imgListName make_script.in ln -s scripts/make_scriptR.SEED make_scriptR.seed make_scriptR ls -l |
relink.sh support/register-basic01.seed make_scriptR.seed relink.sh script/make_scriptR_02.seed make_scriptR.seed ls -l make_scriptR.seed | tee -a notes }}} === Set-up batch register files === Replace [[make_script.in]] with a list of images that you want to use. It is important that the first character is blank and the file is terminated with an "END" {{{ # Create a list of images that need to be processed. echo working with dataset active.$ID | tee -a notes echo y | cp lsupport/active.$ID make_script.in echo END >> make_script.in wc -l make_script.in | tee -a notes |
| Line 33: | Line 39: |
| === 4. Batch register images === | === Batch register images === Build the scripts. We use nohup to ensure that if the login window is lost, the process still runs. |
| Line 36: | Line 43: |
| # Batch register images register -v nohup sh run_script.b & |
make_scriptR nohup sh run_script.b |
| Line 41: | Line 47: |
| === 5. Monitor progress === | === Monitor progress === |
| Line 43: | Line 49: |
| # Monitor progress ls -l *.INN *.OOT |
ls *.INN | wc ls *.OOT | wc |
| Line 47: | Line 53: |
| === 6. Check results === | === Check results === This gives stats to help you see which things to work. |
| Line 49: | Line 56: |
| # Check results ls *OOT > tmpList sh support/after.sh 0.5 tmpList cat goodList cat badList sh support/evalReg.sh badList |
registerEval.sh wc eval/evalGood.txt eval/evalBad.txt | tee -a notes |
| Line 57: | Line 60: |
| === 7. Fix unregistered images === | If you so desire, especially early on, make R/B images to see the quality of the fit {{{ evalReg.sh make_script.in }}} === Fix unregistered images === |
| Line 63: | Line 71: |
| === 8. Clean directory === {{{ # Clean directory sh rem_script.b rm TESTFILES/* rm nohup.out }}} |
=== Rerun === |
| Line 71: | Line 73: |
| === 9. Complete procedure === | If evalBad.txt is large, then you can tweak [[make_scriptR.seed]] and rerun on the bad list |
| Line 74: | Line 76: |
| # Block Complete exit # Save block log mv typescript step-reg |
cut -c 1-12 evalBad.txt | sed -e "s/^/ /" > make_script.in echo "END" >> make_script.in relink ________.seed make_scriptR.seed echo "Rerunning with evalBad.txt" >> notes ls -l make_scriptR.seed >> notes make_scriptR nohup sh run_script.b |
| Line 81: | Line 88: |
=== Finish the Block === {{{ echo 2 | blockFinish ps$ID }}} |
|
| Line 83: | Line 95: |
| ''(Compiled by DL)'' CategoryHowTos |
''(Compiled by EEP)'' |
Block 3 - Register V2
Description
This block batch registers images. Each time you run a batch run of run_script.b (typically would have a different make_scriptR.seed), then run blockFinish to log the results.
Procedure
Set-up
For script, set the ID
ID=312
Ensure make_scriptR.seed is linked to the script you want to run. For starters, you can use one of the default ones. If needed, you can craft something specific and place it in lsupport.
relink.sh support/register-basic01.seed make_scriptR.seed relink.sh script/make_scriptR_02.seed make_scriptR.seed ls -l make_scriptR.seed | tee -a notes
Set-up batch register files
Replace make_script.in with a list of images that you want to use. It is important that the first character is blank and the file is terminated with an "END"
# Create a list of images that need to be processed. echo working with dataset active.$ID | tee -a notes echo y | cp lsupport/active.$ID make_script.in echo END >> make_script.in wc -l make_script.in | tee -a notes
Batch register images
Build the scripts. We use nohup to ensure that if the login window is lost, the process still runs.
make_scriptR nohup sh run_script.b
Monitor progress
ls *.INN | wc ls *.OOT | wc
Check results
This gives stats to help you see which things to work.
registerEval.sh wc eval/evalGood.txt eval/evalBad.txt | tee -a notes
If you so desire, especially early on, make R/B images to see the quality of the fit
evalReg.sh make_script.in
Fix unregistered images
# Fix unregistered images .......
Rerun
If evalBad.txt is large, then you can tweak make_scriptR.seed and rerun on the bad list
cut -c 1-12 evalBad.txt | sed -e "s/^/ /" > make_script.in echo "END" >> make_script.in relink ________.seed make_scriptR.seed echo "Rerunning with evalBad.txt" >> notes ls -l make_scriptR.seed >> notes make_scriptR nohup sh run_script.b
Finish the Block
echo 2 | blockFinish ps$ID
(Compiled by EEP)