<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi,<br>
<br>
My issue is related to gdal, libgeotiff and proj.4, but the problem
appeared with gdalwarp so I'm posting here. I'm seeing different
output from the various tools that I can't really make sense of and
I was hoping someone here could clarify the meaning of the various
output values.<br>
<br>
I received a geotiff file from one of my clients that essentially
has all it's referencing metadata set to 'User-Defined'. Listgeo
-proj4 prints out:<br>
<pre>+proj=tmerc +lat_0=0.000000000 +lon_0=15.000000000 +k=0.999900 +x_0=500000.000 +y_0=-5000000.000 +a=6377397.155 +b=6356078.963 +units=m
</pre>
I've double checked the geotiff data manually and this is indeed the
information that is in the file. Transverse mercator projection with
respect to the Bessel 1841 ellipsoid in other words. My client
reprojected the geotiff to WGS84 using the following gdalwarp input<br>
<pre>./gdalwarp.exe -t_srs '+proj=tmerc +lat_0=0 +lon_0=15 +datum=WGS84'</pre>
<pre><span style="color: rgb(0, 0, 0); font-family: Verdana,sans-serif; font-size: 13px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; display: inline ! important; float: none;"></span></pre>
<span style="color: rgb(0, 0, 0); font-family: Verdana, sans-serif;
font-size: 13px; font-style: normal; font-variant: normal;
font-weight: normal; letter-spacing: normal; line-height: normal;
orphans: 2; text-align: -webkit-auto; text-indent: 0px;
text-transform: none; white-space: normal; widows: 2;
word-spacing: 0px; -webkit-text-size-adjust: auto;
-webkit-text-stroke-width: 0px; display: inline !important; float:
none; "><span class="Apple-converted-space"> </span></span>He
then loaded both geotiffs in our GIS tool (not based on osgeo
libraries) and noticed a substantial horizontal shift between the
two variants of the file. I'm now trying to figure out what is going
wrong. I listed what I've been able to find out so far below.<br>
<br>
Listgeo prints out the following information for the upper left
corner point:<br>
<pre>Upper Left ( 439702.628, 71450.647) ( 14d13'27.94"E, 45d47' 5.72"N)
</pre>
If I then take the same projected coordinate and input those into
cs2cs I get different lon lat values.<br>
<pre>cs2cs +proj=tmerc +lat_0=0.000000000 +lon_0=15.000000000 +k=0.999900 +x_0=500000.000 +y_0=-5000000.000 +a=6377397.155 +b=6356078.963 +units=m
14d13'59.5"E 45d8'31.057"N 0.000, 71450.647
</pre>
Shouldn't I be getting the same values here? If not, then what are
the two tools actually showing me; it's not clear to me with respect
to which datum the lon/lat coordinates are defined?<br>
<br>
I then tried to get cs2cs to produce WGS84 lat/lon coordinates, but
oddly enough that produced the same output as the previous command.<br>
<pre>cs2cs +proj=tmerc +lat_0=0.000000000 +lon_0=15.000000000 +k=0.999900 +x_0=500000.000 +y_0=-5000000.000 +a=6377397.155 +b=6356078.963 +units=m +to +proj=latlon +datum=WGS84 coords.txt
14d13'59.5"E 45d8'31.057"N 0.000, 71450.647
</pre>
I was able to get different output by adding an explicit +towgs84
parameter for the input CRS<br>
<pre>cs2cs +proj=tmerc +lat_0=0.000000000 +lon_0=15.000000000 +k=0.999900 +x_0=500000.000 +y_0=-5000000.000 +a=6377397.155 +b=6356078.963 +units=m +towgs84=0,0,0 +to +proj=latlon +datum=WGS84 coords.txt
14d13'59.5"E 45d8'33.211"N -706.488, 71450.647
</pre>
This leads me to conclude that without an +datum reference or a
+towgs84 parameter value proj (and possibly gdalwarp) no longer
performs any datum transformation whatsoever. In other words it is
not converting from one ellipsoid to the other via geocentric. Is
that correct? If so, what is the rationale behind this choice? Our
GIS toolkit takes a different approach and defaults to
+towgs84=0,0,0 instead. My hunch is that this difference in default
interpretation is the cause of the horizontal shift I'm seeing.<br>
<br>
Thanks in advance for any help,<br>
<br>
Pepijn Van Eeckhoudt<br>
</body>
</html>