[pdal] tweaking parameters in filters.smrf

Fri Jul 12 05:18:41 PDT 2019

Hello Silvia,

Like you I am new to PDAL and have been trying to optimize the output and
filtering steps. As you suggested, I would also like to participate in a
collabration efforts to improve ground filtering efforts.

I have been looking at a similar approach as yours. I have a little
programming background, so am working on an automated process (custom or
using pipeline). From what I have seen, pmf & smrf tends to include the low
vegetation as ground points.
About clustering: I do not understand, how clustering would help with
classifying ground. With default settings, the whole file might end within
a single cluster. Best use I can think of is filtering the noise points.

I confess, I am not an expert and dont understand many of the Innards.

Can you explain what is normalZ and its relation to ground?

Let me know what you think.

Warm Regards,
Thejus

On Fri, Jul 12, 2019 at 2:17 AM Silvia Franceschi <
silvia.franceschi at gmail.com> wrote:

> Hi Karl,
> i am also trying to use the PDAL to extract a good DTM with my data on a
> mixed vegetated terrain.
> I am new to PDAL so I can't really help you, but we could share a bit the
> experience: commands, parameters and flowchart of the approach, what do you
> think?
>
> My own approach is the follow:
> 1. denoise the data: using filters.elm and filters.ouliers and after this
> 2. classify ground/no ground points: alternatively used filters.smrf,
> filters.pmf and filters.mongus (this unfortunately still do not work on my
> data, but I will try again to test it during the next week)
> 3. create the DTM: export the data with writers.gdal
>
> The DTM I obtained is not really nice and I am trying to fix it by working
> on the parameters and looking for further analysis to add. There are now
> two possibilities:
> A. try to clusterize as suggested by Howard: did you already tried? is
> this a good solution for a good final product? which threshold of the
> dimension of the cluster did you use to select the ground points to keep
> and ignore?
> B. try to follow the instructions given by Luigi and do the following
> steps:
>     - filters.pmf
>     - calculate normalZ and assign: normalZ ! [0:0.6] : Classification ->2
>     - calculate kDistance and assign: KD <1 : classification = 2
>     - apply pmf again on the new dataset
> The problem in the second case is to store the results and do the checks
> on each point. I am blocked here because even if I can store the results in
> an extra_dimension of the las, then I am not able to do the check on the
> values in the interval [0:0.6] using PDAL and I don't have any other
> possibilities for a separate software.
> C. find a better solution for the final step: interpolation to create the
> DTM, what do you use for this purpose?
>
> I'll go on with some tests these days, finger cross!!!
>
> Regards
>
> Silvia
>
>
>
> On Wed, Jun 12, 2019 at 8:41 PM Karl North <karln at surdex.com> wrote:
>
>> Moving along here, and have encountered a new question…
>>
>>
>>
>> I’m getting some promising results from filters.cluster but would
>> certainly like to be able to apply this filter ( and others ) multiple
>> times with different settings while using writers.las after each filter so
>> that I can review the progress of the overall pipeline processing after
>> each step completes.
>>
>>
>>
>> Is there a way to get my pipeline to write a LAS after each of the
>> steps/filters?  I’d like to save interim results for review, but cannot
>> find a good example in the docs or in the list archive.
>>
>>
>>
>>
>>
>>
>>
>> *From:* Howard Butler <howard at hobu.co>
>> *Sent:* Wednesday, June 12, 2019 9:41 AM
>> *To:* Karl North <karln at surdex.com>
>> *Cc:* Bradley Chambers <brad.chambers at gmail.com>; pdal at lists.osgeo.org
>> *Subject:* Re: [pdal] tweaking parameters in filters.smrf
>>
>>
>>
>>
>>
>>
>>
>> On Jun 12, 2019, at 9:36 AM, Karl North <karln at surdex.com> wrote:
>>
>>
>>
>> Brad:
>>
>>
>>
>> Thanks for the suggestions.
>>
>>
>>
>> I’ve done some reading and begun playing around with the cluster filter.
>> One problem is that I’m new enough to PDAL that I cannot find a way to
>> add/define a dimension and store the cluster ID into either LAS or LAZ
>> format on output.  I think it should be possible to just hijack an existing
>> field in the point data records for current testing purposes.  Does this
>> sound possible?  I’m floundering a bit trying to find an example of
>> something similar.  Can you point me in the right direction?
>>
>>
>>
>> Use the filters.ferry to copy ClusterID => PointSourceId downstream of
>> your filters.cluster invocation.
>>
>>
>>
>> https://pdal.io/stages/filters.ferry.html
>> <https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpdal.io%2Fstages%2Ffilters.ferry.html&data=02%7C01%7C%7C5dc2119c10ca44f84e9008d6ef43fd22%7Cbf4323beeac04c8eb466c7924493cbee%7C0%7C0%7C636959472628150563&sdata=hKaN0thXKKf7fE%2F5nqkAZcSq8YMCR26WM7xkkz5IINI%3D&reserved=0>
>>
>>
>>
>>
>>
>> I can get a BPF output, but currently have no way to review the result.
>> I’m now downloading 64-bit QT Reader, after finding a thread by HoBu saying
>> that this viewer would work for a quick review.
>>
>>
>>
>> I don't know if that is still viable.
>>
>>
>>
>> You can also write to LAS with extra bytes
>>
>>
>>
>> --writers.las.extra_dims=ClusterID
>>
>>
>>
>> but you'll need something that can view extra LAS dimensions.
>>
>>
>>
>>
>>
>> Howard
>>
>>
>> _______________________________________________
>> pdal mailing list
>> pdal at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/pdal
>
>
>
> --
> ing. Silvia Franceschi
> Via Roma, 64
> 38030 Castello di Fiemme (TN)
>
> tel: 0039 -3384501332
> _______________________________________________
> pdal mailing list
> pdal at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/pdal
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20190712/6f864c47/attachment-0001.html>