Menu
▾
▴
1 Attachments
#13 more passes needed
closed
nobody
done
program
major
1.0b2
enhancement
2010-05-01
2010-04-16
No
We have been comparing MCC-derived DEMs to those delivered by a lidar vendor. Looking at the output, and the large number of ground points remaining relative to the vendor DEM, I think there might be a problem using the max=20 passes constraint at each scale domain of MCC. We may need to remove or relax this constraint, especially on the final scale domain. What this means is MCC will be more aggressive in filtering out non-ground returns. The canopy returns from the latest lidar systems are so dense that MCC may just need more passes to burn the forest down to the ground. Beta-tester Ed Uebler also found that MCC was maxing the number of passes (n=60), suggesting that the filtering is being stopped prematurely, perhaps on all 3 scale domains.
Having learned more and thought more about this, it would be better to change the pass limit to 100 and change the convergence thresholds to 1% for scale domain 1, 0.1% for scale domain 2, and 0.01% for scale domain 3.
On third though, let's just change the pass limit to 100 and not alter the convergence thresholds at this time. We want to be able to diagnose the effect of changing one parameter before introducing an interaction effect by changing another.
Implemented in revision r109.
Related
Commit: [r109]