[pdal] Target Aliged Pixels (tap) for the GDAL driver

Håkon Åmdal hakon at aamdal.com
Fri May 8 04:05:53 PDT 2020


Thanks guys, that worked!

That said, such "-tap" functionality could be neat to have for the
rasterization writers.

Have a great weekend!
Håkon

On Thu, May 7, 2020 at 6:08 PM Håkon Åmdal <hakon at aamdal.com> wrote:

> Let me try that tomorrow. You'll hear from me. Thanks for quick feedback!
>
> On Thu, 7 May 2020 at 17:24, Andrew Bell <andrew.bell.ia at gmail.com> wrote:
>
>> Hi Håkon,
>>
>> I know nothing about gdalwarp, but I would think you could get perfect
>> alignment by using the bounds option in `writers.gdal`.  Are you having a
>> problem despite using this option?
>>
>> On Thu, May 7, 2020 at 5:16 PM Håkon Åmdal <hakon at aamdal.com> wrote:
>>
>>> Hi there,
>>>
>>> I'm currently working on a LIDAR data project where we classify "high
>>> vegetation" and buildings using PDAL. We convert them to rasters before we
>>> send them downstream.
>>>
>>> Now, we are looking to combine the rasters generated with
>>> satellite imagery. For this project, we believe it's important to have an
>>> 1:1 mapping between the pixels in the different sources.
>>>
>>> To align our satellite images with the underling projection, we use the
>>> "-tap" option in gdalwarp. However, no such feature exists for the
>>> GDAL-writer in PDAL. Which means the pixels are offset by up to half a
>>> meter (for UTM and 1.0 scale).
>>>
>>> I think we can overcome the above issue by using PDAL to generate a much
>>> higher resolution TIF file (say 0.1) and then resample with gdalwarp and
>>> the "-tap" option. Or even just keep the resolution, just use gdalwarp to
>>> resample.
>>>
>>> I was wondering if a similar "-tap" option belongs in PDAL, and how that
>>> would look like? I initially thought it was argument passed to the driver,
>>> but AFAIK the pixel alignment happens before it is passed to the driver. If
>>> I understand it correctly, it would then be an extra option to be passed to
>>> the "writers.gdal" component.
>>>
>>> I'd love to contribute, but I'm no C++ developer.
>>>
>>> Håkon
>>> _______________________________________________
>>> pdal mailing list
>>> pdal at lists.osgeo.org
>>> https://lists.osgeo.org/mailman/listinfo/pdal
>>
>>
>>
>> --
>> Andrew Bell
>> andrew.bell.ia at gmail.com
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20200508/e0fc9da2/attachment-0001.html>


More information about the pdal mailing list