Ticket Change Details
Not logged in
Overview

Artifact ID: 94db0d879672f38572849e6baca4ed1fd2ec57b2
Ticket: 6e95dbcac98c8c5c2b1777a71d40dedd454a8453
RL2_WriteGeoTiff: bank images with 1-BIT MONOCHROME, FAX4
User & Date: mj10777 2017-08-28 08:31:02
Changes

  1. foundin changed to: "dev"
  2. icomment:
    In preparation for a QGIS QgsRasterLite2Provider, I am building a small Database with 4 RasterCoverages of the same 2000x2000 Meter area.
    <ul>
    1656 : 1-BIT MONOCHROME, FAX4<br>
    1811 : UINT8 RGB, JPEG<br>
    1910 : UINT8 RGB, JPEG<br>
    2010 : 1-BIT MONOCHROME, FAX4
    </ul>
    <hr>
    From existing RL2 Databases the attempt was mad to extract this are from the 4 with RL2_WriteGeoTiff.<br>
    For the UINT8 RGB, JPEG: no problem.<br>
    For the 1-BIT MONOCHROME, FAX4: the images are blank.<br>
    <hr>
    All 4 are exported from 1 script<br>
    - here the portion from the last image:
    <pre>
    -- -- ---------------------------------- --
    SELECT DateTime('now','localtime'),'ATTACH DATABASE ''../2010.berlin_k5.400_dpi.db'' AS db_import';
    ATTACH DATABASE '../2010.berlin_k5.400_dpi.db' AS db_import;
    -- -- ---------------------------------- --
    SELECT DateTime('now'),'2010.alt_berlin_coelln_friedrichsweder: pixel_size_x['||round((26000.0-24000.0)/0.3175065154982885)||'] pixel_size_y['||round((22250.0-20250.0)/0.3174918146641513)||']';
    -- -- ---------------------------------- --
    -- gdal_translate 2010.Berlin.5000.3068.FAX4_MINISWHITE.400.tif -projwin 24000.0 22250.000 26000.0 20250.000 2010.Alt_Berlin_Coelln_Friedrichsweder.5000.3068.tif
    -- gdalinfo 2010.Alt_Berlin_Coelln_Friedrichsweder.5000.3068.tif > 2010.Alt_Berlin_Coelln_Friedrichsweder.5000.3068.tif.info.txt
    -- -- ---------------------------------- --
    SELECT RL2_WriteGeoTiff
    (
     -- The dbPrefix argument is intended to specify the ATTACHED-DB where the Coverage is expected to be found
     'db_import',
     -- chosen name of raster_coverage
     lower('2010.berlin_k5'),
      -- directory where the files to be loaded are stored
     '2010.alt_berlin_coelln_friedrichsweder.5000.3068.tif',
     -- image-size width,height (calculate)
     -- 2143.730401266
     CAST(round((26000.0-24000.0)/0.3175065154982885) AS INTEGER),
     -- 2144.134100369
     CAST(round((22250.0-20250.0)/0.3174918146641513)AS INTEGER),
     -- area to extract,
     BuildMbr
     (
      24000.0,22250.000, 26000.0,20250.000,
      3068
     ),
     -- resolution horz/x
     0.3175065154982885,
    -- resolution vert/y
     0.3174918146641513,
     -- world-file 0=no ; 1=yes
     1,
     -- compression
     'FAX4',
    -- tile-size  (of tif)
     512
    );
    -- -- ---------------------------------- --
    SELECT DateTime('now','localtime'),'DETACH DATABASE ''../2010.berlin_k5.400_dpi.db'' AS db_import';
    DETACH DATABASE db_import;
    -- -- ---------------------------------- --
    </pre>
    The shown gdal command for the same task (from the original tiff) returns correct results.
    <hr>
    Note on the side:<br>
    When the extent is given and either
    <ul>
    image size <b>or</b>
    resolution
    </ul>
    are given - the missing value should be calculated.<br>
    In most cases, people will give the <b>extent</b> and <b>size</b><br>
    <ul>they should not have to bother about the resolution</ul>
    If no resolution or size is given (i.e. values are 0):<br>
    <ul>The original resolution should be taken to calculate the size<br>
    as I have done with this script</ul>
    This was a part of my patch which was never looked into.
    
  3. login: "mj10777"
  4. mimetype: "text/html"
  5. private_contact changed to: "a3ab3a740c17d1a955665bbfeda76d4f8b6739bf"
  6. severity changed to: "Critical"
  7. status changed to: "Open"
  8. title changed to:
    RL2_WriteGeoTiff: bank images with 1-BIT MONOCHROME, FAX4
    
  9. type changed to: "Code_Defect"