Agisoft Metashape
Agisoft Metashape => General => Topic started by: Yoann Courtois on August 28, 2017, 06:41:24 PM
-
Hello !
I would like to get smooth contours (+- 5cm) on a model which has high resolution data (Area = 17.5ha with GSD = 1.6cm)...
There is indeed no tools to smooth those contours on a hand, and no possibility to get low resolution DEM on the other one...
Any help would be appreciate again !
Regards
-
Hi Yoann,
Use QGIS (or ArcGIS) to generate and smooth contour lines ... and add SMOOTH CONTOURS to the feature request section of this forum.
Regards
SAV
-
Hi SAV !
Thanks for this advice !
I wasn't sure it was already possible to do that in PhotoScan so I posted here in General section.
I'm going to move it to Feature request section.
Regards
-
Hi Yoann,
As far as I know there is only a 'simplify contours' option in PhotoScan (see attached image). But 1) I am not sure what it really does and 2) you cannot choose the level of 'simplification/smoothing'.
Regards
SAV
-
Hi SAV !
Guess you still work on 1.2 version of PhotoScan because your screenshot come from its panel.
In 1.3, "Simply contours" became "Prevent intersections".
No idea what's the aim of one or the other but still..
Regards
-
Yep, version 1.2.6. Too many Python scripts that I don't want to adapt/change so they run in 1.3.
Prevent intersections? Interesting. Contour lines, by definition, cannot intersect because that would mean there are two different elevation values for the same pixel/point.
Not sure how PS extracts contour lines, but theoretically there should be no need for the 'prevent intersection' option.
Regards
SAV
-
Ahah of course... I understand you...
Since, you really miss a lot of improvements of 1.3 !
Just regarding the basic calculations which are really faster and give us better results !
With 1.3 we indeed had a completely new API vocabulary. But I don't thing it gonna change again in 1.4.
Maybe you should wait a while and update your scripts with next 1.4 !
Yes you right, I though the same the moment I saw it.
I don't know either how contours are extracted, but it should have a bit of smoothing already. And if you ask for really small gap between contours (one each cm), the smoothing on random altitude area may give you intersected contours.
Regards
-
I have read in the change log file that PS is now using CPU + GPU for image alignment (instead of CPU alone), which should speed up the whole process. Do you have an estimate of how much?
What do you mean with better results? Have they also improved some other parts of the code?
Yes, version 1.4 should be coming out soon. I'll probably 'bit the bullet' then and upgrade PS and my scripts ;-)
Regards,
SAV
-
I never tried the test because I would need to put 1.2 back on my computer.
I would also say it depend on your set up.
On my computer as an exemple the depth maps calculation (for dense cloud) is made at 5% by CPU and 95% by the GPU.
For the alignment, detecting and matching points are calculated on both CPU and GPU, whereas the camera location estimation is calculated on CPU only. The time-saving is huge !
For other better result, I would say the ground extraction tool, which had several problem in 1.2, is working perfectly on 1.3 (I have extracted ground from flowers by using 5 cm max distance...)
Many new tools are pretty useful, such has shapes on pictures, or ruler which can measure point coordinates directly on pictures, model or orthomosaic...
You really miss something ;)
Regards
-
Thanks for sharing your experience.
I have never really had good/consistent results using the ground extraction tool. Good to know that this has changed since version 1.3
Time savings in processing are always welcome ;-) I might upgrade sooner than I initially anticipated.
Regards
SAV
I never tried the test because I would need to put 1.2 back on my computer.
I would also say it depend on your set up.
On my computer as an exemple the depth maps calculation (for dense cloud) is made at 5% by CPU and 95% by the GPU.
For the alignment, detecting and matching points are calculated on both CPU and GPU, whereas the camera location estimation is calculated on CPU only. The time-saving is huge !
For other better result, I would say the ground extraction tool, which had several problem in 1.2, is working perfectly on 1.3 (I have extracted ground from flowers by using 5 cm max distance...)
Many new tools are pretty useful, such has shapes on pictures, or ruler which can measure point coordinates directly on pictures, model or orthomosaic...
You really miss something ;)
Regards