<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Daniel,</p>
    <p>You rightly spotted <a class="moz-txt-link-freetext" href="https://github.com/OSGeo/gdal/pull/6069">https://github.com/OSGeo/gdal/pull/6069</a> as
      the enabler for that capability.<br>
    </p>
    <p>if your existing target dataset has a geolocation array attached
      to do it, this should just be a matter of doing:<br>
    </p>
    <p>target_ds = gdal.Open( filename, gdal.GA_Update )<br>
    </p>
    <p>gdal.Warp(target_ds, source_ds, ... other options here ...)</p>
    <p>If the target dataset doesn't have a geolocation array attached
      to it, you can point to an external one with the DST_GEOLOC_ARRAY
      tranformer option<br>
    </p>
    <p>gdal.Warp(target_ds, source_ds,
      transformerOptions=["DST_METHOD=GEOLOC_ARRAY",
      "DST_GEOLOC_ARRAY=/path/to/geoloc_dataset"], ... other options
      here ...)</p>
    <p>Even<br>
    </p>
    <div class="moz-cite-prefix">Le 16/05/2023 à 19:35, Daniel Scheffler
      a écrit :<br>
    </div>
    <blockquote type="cite"
      cite="mid:c3158573-ee48-2e4f-1516-ca80b4e6aa13@gfz-potsdam.de">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      Hi!<br>
      <br>
      Some time ago, I asked for an inversion of gdal.Warp based on
      GEOLOCATION arrays (longitude/latitude). Back then, my question
      was: How can I transform an image with projected coordinates back
      to cartesian/image coordinates, given that a geolocation array
      tells GDAL where to put which pixel in the output?<br>
      <br>
      In 11/2021, this was unfortunately not implemented yet. However,
      is seems like in the meantime someting like this has been added:<br>
          - <a class="moz-txt-link-freetext"
        href="https://github.com/OSGeo/gdal/pull/5520"
        moz-do-not-send="true">https://github.com/OSGeo/gdal/pull/5520</a><br>
          - <a class="moz-txt-link-freetext"
        href="https://github.com/OSGeo/gdal/pull/6069"
        moz-do-not-send="true">https://github.com/OSGeo/gdal/pull/6069</a><br>
          -
      <a class="moz-txt-link-freetext"
href="https://github.com/OSGeo/gdal/blob/c92b22d02c99eae0152f49595947fb3747ddc280/autotest/gcore/geoloc.py#L396"
        moz-do-not-send="true">https://github.com/OSGeo/gdal/blob/c92b22d02c99eae0152f49595947fb3747ddc280/autotest/gcore/geoloc.py#L396</a><br>
      <br>
      But I am not quite sure if that is what I want. If so, how would a
      Python implementation based on gdal.Warp look like? Is that
      documented somewhere?<br>
      <br>
      Best,<br>
      Daniel<br>
      <br>
      <br>
      <br>
      <div class="moz-cite-prefix">Am 30.11.2021 um 14:20 schrieb Daniel
        Scheffler:<br>
      </div>
      <blockquote type="cite"
        cite="mid:b1cf6b86-4922-ca9a-bdd4-ec4fff2883d7@gfz-potsdam.de">
        <meta http-equiv="Content-Type" content="text/html;
          charset=UTF-8">
        Ok thanks, too bad that this is not implemented. I think the
        inversion of this transformation would be a nice feature to be
        added in GDAL. It would be very useful to me (especially if it
        is accessible via the Python bindings) and would ease the
        implementation in a processing pipeline for the upcoming EnMAP
        hyperspectral satellite. Should I open a feature request in the
        GDAL issue tracker on GitHub?<br>
        <br>
        <br>
        <div class="moz-cite-prefix">Am 30.11.2021 um 14:02 schrieb Even
          Rouault:<br>
        </div>
        <blockquote type="cite"
          cite="mid:1791a553-3e0e-f5e6-659d-11ac6f708ab3@spatialys.com">
          <meta http-equiv="Content-Type" content="text/html;
            charset=UTF-8">
          <p><br>
          </p>
          <div class="moz-cite-prefix">Le 30/11/2021 à 12:52, Daniel
            Scheffler a écrit :<br>
          </div>
          <blockquote type="cite"
            cite="mid:c2cf3a49-200b-91a6-dd81-1b3059859558@gfz-potsdam.de">
            <meta http-equiv="content-type" content="text/html;
              charset=UTF-8">
            Thanks a lot for taking the time, Even, I got the
            transformation from cartesian to projected coordinates to
            work in memory with the GTiff driver. With MEM, NUMPY or VRT
            it does not work because these formats are either not
            readable from /vsimem/ or don´t have a regular file path
            which is needed to set the <font size="2">X_DATASET and </font><font
              size="2">Y_DATASET keys in the GEOLOCATION metadata</font>.<br>
          </blockquote>
          Ah indeed for X_DATASET/Y_DATASET, you can't use a MEM or
          NUMPY dataset. But a /vsimem/xxx dataset in another format
          should work.<br>
          <blockquote type="cite"
            cite="mid:c2cf3a49-200b-91a6-dd81-1b3059859558@gfz-potsdam.de">
            <div class="moz-forward-container"> <br>
              <font size="2">Regarding the inversion of this
                transformation, i.e., </font><font size="2">from
                projected coordinates to pixel/line:</font><br>
              <blockquote type="cite">The logic of <span class="pl-en">GDALCreateGenImgProjTransformer2()
                  around <a class="moz-txt-link-freetext"
href="https://github.com/OSGeo/gdal/blob/master/alg/gdaltransformer.cpp#L1825"
                    moz-do-not-send="true">https://github.com/OSGeo/gdal/blob/master/alg/gdaltransformer.cpp#L1825</a>
                  which is for the source dataset should be ported a few
                  lines after for the target dataset.  Probably with a
                  new transformer option to be able to point to an
                  auxiliary dataset, such as a VRT one of your example,
                  to extract the geolocation metadata items from it,
                  that would be different from the target dataset
                  itself, because, except perhaps for the netCDF case,
                  most GDAL datasets that expose a GEOLOCATION metadata
                  domain must be read-only.</span></blockquote>
              <br>
              I don´t completely get what you mean here. To me, this
              sounds like there might be a way to do the inverted
              transformation using the C-API of GDAL. </div>
          </blockquote>
          No, I meant there's some missing code to do that.<br>
          <blockquote type="cite"
            cite="mid:c2cf3a49-200b-91a6-dd81-1b3059859558@gfz-potsdam.de">
            <div class="moz-forward-container">However, I am a Python
              developer and my C skills are a bit poor. Is there any way
              to use the Python bindings here?<br>
              <br>
              Kind regards,<br>
              Daniel<br>
              <br>
              <br>
                        <br>
              <br>
              <div class="moz-cite-prefix">Am 26.11.2021 um 12:38
                schrieb Even Rouault:<br>
              </div>
              <blockquote type="cite"
                cite="mid:2c9d72a2-3950-f846-b3e4-ee80500179a1@spatialys.com">
                <meta http-equiv="Content-Type" content="text/html;
                  charset=UTF-8">
                Daniel,<br>
                <blockquote type="cite"
                  cite="mid:8e78eb30-6a89-104b-d2c4-f9ea8231aced@gfz-potsdam.de">
                  <br>
                  I am trying to convert image data from cartesian/image
                  coordinates to projected coordinates AND vice versa
                  using geolocation arrays in GDAL. I have two
                  questions:<br>
                  <ol>
                    <li>Since this transformation is part of a
                      processing chain implemented in Python, I try to
                      transform the data directly in-memory, i.e,
                      without any disk access. This saves IO time and
                      avoids permission errors when trying to write
                      temporary data on Windows. How can this be done? I
                      got correct results with the code below, however,
                      only when I temporarily write the data to disk. I
                      tried to write the data to /vsimem/ using the MEM,
                      GTiff and NUMPY drivers. However, gdal.Warp can´t
                      find the data there (FileNotFoundError). I think,
                      also the gdal.Transformer class might be useful
                      and I found an interesting thread on that <a
href="https://lists.osgeo.org/pipermail/gdal-dev/2012-January/031502.html"
                        moz-do-not-send="true">here</a> and a related
                      test in the GDAL autotest suite (<a
href="https://github.com/OSGeo/gdal/blob/master/autotest/alg/transformgeoloc.py"
                        moz-do-not-send="true">here</a>). However, I
                      can´t get it to work for my specific case.</li>
                  </ol>
                </blockquote>
                <p>There's no reason it won't work with a in-memory
                  dataset.</p>
                <p>If you use a MEM dataset, then you need to provide
                  the dataset object itself as the input dataset of
                  gdal.Warp() (the name of a MEM dataset is completely
                  ignored. a MEM dataset can't be opened, just created).</p>
                <p>Similarly with a NUMPY dataset.   With GTiff and
                  /vsimem/, they behave as a regular file. If you pass
                  it by name as input of gdal.Warp(), you need to make
                  sure to close (ds = None typically) the dataset
                  before, so it is properly flushed and can be opened.
                  But you can also pass it as an object without that
                  constraint.</p>
                <p>You can also create purely in-memory VRT files by
                  assigning them an empty name. Then of course you need
                  to provide them as objects to gdal.Warp()<br>
                </p>
                <p><br>
                </p>
                <blockquote type="cite"
                  cite="mid:8e78eb30-6a89-104b-d2c4-f9ea8231aced@gfz-potsdam.de">
                  <ol>
                    <li>My second question is how I can invert the
                      transformation, i.e., how can I transform an image
                      with projected coordinates back to cartesian/image
                      coordinates, given that a geolocation array tells
                      GDAL where to put which pixel in the output?
                      Background is a processing pipeline for satellite
                      data where some processing steps are running in
                      sensor geometry (image data as acquired by the
                      sensor, without any geocoding and projection) and
                      I need to provide corresponding AUX data which
                      originally come with projected coordinates.<br>
                    </li>
                  </ol>
                </blockquote>
                The logic of <span class="pl-en">GDALCreateGenImgProjTransformer2()
                  around <a class="moz-txt-link-freetext"
href="https://github.com/OSGeo/gdal/blob/master/alg/gdaltransformer.cpp#L1825"
                    moz-do-not-send="true">https://github.com/OSGeo/gdal/blob/master/alg/gdaltransformer.cpp#L1825</a>
                  which is for the source dataset should be ported a few
                  lines after for the target dataset.  Probably with a
                  new transformer option to be able to point to an
                  auxiliary dataset, such as a VRT one of your example,
                  to extract the geolocation metadata items from it,
                  that would be different from the target dataset
                  itself, because, except perhaps for the netCDF case,
                  most GDAL datasets that expose a GEOLOCATION metadata
                  domain must be read-only.<br>
                </span>
                <blockquote type="cite"
                  cite="mid:8e78eb30-6a89-104b-d2c4-f9ea8231aced@gfz-potsdam.de">
                  <ol>
                    <li> <br>
                    </li>
                  </ol>
                  <p>Here is the code I already have to convert a sample
                    image from cartesian to projected coordinates:</p>
                  <blockquote>
                    <p><font size="2">import os<br>
                        from tempfile import TemporaryDirectory<br>
                        from osgeo import gdal, osr<br>
                        import numpy as np<br>
                        from matplotlib import pyplot as plt<br>
                        <br>
                        <br>
                        # get some test data<br>
                        swath_data = np.random.randint(1, 100, (500,
                        400))<br>
                        lons, lats = np.meshgrid(np.linspace(3, 5, 500),<br>
                                                 np.linspace(40, 42,
                        400))<br>
                        <br>
                        with TemporaryDirectory() as td:<br>
                            p_lons_tmp = os.path.join(td, 'lons.tif')<br>
                            p_lats_tmp = os.path.join(td, 'lats.tif')<br>
                            p_data_tmp = os.path.join(td, 'data.tif')<br>
                            p_data_vrt = os.path.join(td, 'data.vrt')<br>
                            p_data_mapgeo_vrt = os.path.join(td,
                        'data_mapgeo.vrt')<br>
                        <br>
                            # save numpy arrays to temporary tif files<br>
                            for arr, path in zip((swath_data, lons,
                        lats), (p_data_tmp, p_lons_tmp, p_lats_tmp)):<br>
                                rows, cols = arr.shape<br>
                                drv = gdal.GetDriverByName('GTiff')<br>
                                ds = drv.Create(path, cols, rows, 1,
                        gdal.GDT_Float64)<br>
                                ds.GetRasterBand(1).WriteArray(arr)<br>
                                del ds<br>
                        <br>
                            # add geolocation information to input data<br>
                            wgs84_wkt = osr.GetUserInputAsWKT('WGS84')<br>
                            utm_wkt =
                        osr.GetUserInputAsWKT('EPSG:32632')<br>
                            ds = gdal.Translate(p_data_vrt, p_data_tmp,
                        format='VRT')<br>
                            ds.SetMetadata(<br>
                            <br>
                                dict(<br>
                                    SRS=wgs84_wkt,<br>
                                    X_DATASET=p_lons_tmp,<br>
                                    Y_DATASET=p_lats_tmp,<br>
                                    X_BAND='1',<br>
                                    Y_BAND='1',<br>
                                    PIXEL_OFFSET='0',<br>
                                    LINE_OFFSET='0',<br>
                                    PIXEL_STEP='1',<br>
                                    LINE_STEP='1'<br>
                                ),<br>
                                'GEOLOCATION'<br>
                            )del ds<br>
                        <br>
                            # warp from geolocation arrays and read the
                        result<br>
                            gdal.Warp(p_data_mapgeo_vrt, p_data_vrt,
                        format='VRT', geoloc=True,<br>
                                      srcSRS=wgs84_wkt, dstSRS=utm_wkt)<br>
                            data_mapgeo =
                        gdal.Open(p_data_mapgeo_vrt).ReadAsArray()<br>
                            <br>
                        # visualize input and output data<br>
                        fig, axes = plt.subplots(1, 4)<br>
                        for i, (arr, title) in
                        enumerate(zip((swath_data, lons, lats,
                        data_mapgeo),<br>
                                                          ('swath data',
                        'lons', 'lats', 'projected data'))):<br>
                            axes[i].imshow(arr, cmap='gray')<br>
                            axes[i].set_title(title)<br>
                        plt.tight_layout()<br>
                        plt.show()</font></p>
                    <p><font size="2"><br>
                      </font></p>
                  </blockquote>
                  <p>Any help would be highly appreciated!</p>
                  <p>Best,</p>
                  <p>Daniel Scheffler<br>
                  </p>
                  <br>
                  <pre class="moz-signature" cols="72">-- 

M.Sc. Geogr. Daniel Scheffler

Helmholtz Centre Potsdam
GFZ German Research Centre For Geosciences
Department 1 - Geodesy and Remote Sensing
Section 1.4  - Remote Sensing
Telegrafenberg, 14473 Potsdam, Germany

Phone:  +49 (0)331/288-1198
e-mail: <a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:daniel.scheffler@gfz-potsdam.de" moz-do-not-send="true">daniel.scheffler@gfz-potsdam.de</a></pre>
                  <br>
                  <fieldset class="mimeAttachmentHeader"></fieldset>
                  <pre class="moz-quote-pre" wrap="">_______________________________________________
gdal-dev mailing list
<a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:gdal-dev@lists.osgeo.org" moz-do-not-send="true">gdal-dev@lists.osgeo.org</a>
<a class="moz-txt-link-freetext" href="https://lists.osgeo.org/mailman/listinfo/gdal-dev" moz-do-not-send="true">https://lists.osgeo.org/mailman/listinfo/gdal-dev</a>
</pre>
                </blockquote>
                <pre class="moz-signature" cols="72">-- 
<a class="moz-txt-link-freetext" href="http://www.spatialys.com" moz-do-not-send="true">http://www.spatialys.com</a>
My software is free, but my time generally not.</pre>
              </blockquote>
              <br>
              <pre class="moz-signature" cols="72">-- 

M.Sc. Geogr. Daniel Scheffler

Helmholtz Centre Potsdam
GFZ German Research Centre For Geosciences
Department 1 - Geodesy and Remote Sensing
Section 1.4  - Remote Sensing
Telegrafenberg, 14473 Potsdam, Germany

Phone:  +49 (0)331/288-1198
e-mail: <a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:daniel.scheffler@gfz-potsdam.de" moz-do-not-send="true">daniel.scheffler@gfz-potsdam.de</a></pre>
            </div>
            <br>
            <fieldset class="mimeAttachmentHeader"></fieldset>
            <pre class="moz-quote-pre" wrap="">_______________________________________________
gdal-dev mailing list
<a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:gdal-dev@lists.osgeo.org" moz-do-not-send="true">gdal-dev@lists.osgeo.org</a>
<a class="moz-txt-link-freetext" href="https://lists.osgeo.org/mailman/listinfo/gdal-dev" moz-do-not-send="true">https://lists.osgeo.org/mailman/listinfo/gdal-dev</a>
</pre>
          </blockquote>
          <pre class="moz-signature" cols="72">-- 
<a class="moz-txt-link-freetext" href="http://www.spatialys.com" moz-do-not-send="true">http://www.spatialys.com</a>
My software is free, but my time generally not.</pre>
          <br>
          <fieldset class="mimeAttachmentHeader"></fieldset>
          <pre class="moz-quote-pre" wrap="">_______________________________________________
gdal-dev mailing list
<a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:gdal-dev@lists.osgeo.org" moz-do-not-send="true">gdal-dev@lists.osgeo.org</a>
<a class="moz-txt-link-freetext" href="https://lists.osgeo.org/mailman/listinfo/gdal-dev" moz-do-not-send="true">https://lists.osgeo.org/mailman/listinfo/gdal-dev</a>
</pre>
        </blockquote>
        <br>
        <pre class="moz-signature" cols="72">-- 

M.Sc. Geogr. Daniel Scheffler

Helmholtz Centre Potsdam
GFZ German Research Centre For Geosciences
Department 1 - Geodesy and Remote Sensing
Section 1.4  - Remote Sensing
Telegrafenberg, 14473 Potsdam, Germany

Phone:  +49 (0)331/288-1198
e-mail: <a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:daniel.scheffler@gfz-potsdam.de" moz-do-not-send="true">daniel.scheffler@gfz-potsdam.de</a></pre>
        <br>
        <fieldset class="moz-mime-attachment-header"></fieldset>
        <pre class="moz-quote-pre" wrap="">_______________________________________________
gdal-dev mailing list
<a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:gdal-dev@lists.osgeo.org" moz-do-not-send="true">gdal-dev@lists.osgeo.org</a>
<a class="moz-txt-link-freetext" href="https://lists.osgeo.org/mailman/listinfo/gdal-dev" moz-do-not-send="true">https://lists.osgeo.org/mailman/listinfo/gdal-dev</a>
</pre>
      </blockquote>
      <br>
      <pre class="moz-signature" cols="72">-- 

Dr. Daniel Scheffler

Helmholtz Centre Potsdam
GFZ German Research Centre For Geosciences
Department 1 - Geodesy and Remote Sensing
Section 1.4  - Remote Sensing and Geoinformatics
Telegrafenberg, 14473 Potsdam, Germany

Phone:  +49 (0)331/288-1198
e-mail: <a class="moz-txt-link-abbreviated moz-txt-link-freetext" href="mailto:daniel.scheffler@gfz-potsdam.de" moz-do-not-send="true">daniel.scheffler@gfz-potsdam.de</a></pre>
      <br>
      <fieldset class="moz-mime-attachment-header"></fieldset>
      <pre class="moz-quote-pre" wrap="">_______________________________________________
gdal-dev mailing list
<a class="moz-txt-link-abbreviated" href="mailto:gdal-dev@lists.osgeo.org">gdal-dev@lists.osgeo.org</a>
<a class="moz-txt-link-freetext" href="https://lists.osgeo.org/mailman/listinfo/gdal-dev">https://lists.osgeo.org/mailman/listinfo/gdal-dev</a>
</pre>
    </blockquote>
    <pre class="moz-signature" cols="72">-- 
<a class="moz-txt-link-freetext" href="http://www.spatialys.com">http://www.spatialys.com</a>
My software is free, but my time generally not.</pre>
  </body>
</html>