[Proj] Natural Earth projection
Bernhard Jenny
jennyb at geo.oregonstate.edu
Thu Mar 8 18:46:20 EST 2012
In Reply to:
Oscar van Vlijmen ovv at hetnet.nl
Mar 4, 2012, at 8:33 AM
> Could be my error, but as closer I get to a pole, the more inaccurate the
> inverse gets.
> The error amounts to several degrees. Useless!
>
> Examples:
>
> R = 6371008.7714 m
…
> lat, lon = 89.0, 120.0
> -> x,y = 6604995.340672151, 9042363.459174621
> inverse, back to lat, lon: 86.95668, 113.64953
I'm unable to reproduce the problem. Below is the code I use (JavaScript from http://www.shadedrelief.com/NE_proj/code.html embedded in a HTML file). I apologize for posting JavaScript and HTML on this list.
Bernie
<html>
<head>
<script type="text/javascript">
var MAX_Y = 0.8707 * 0.52 * Math.PI;
function forward(lon, lat, xy) {
var lat2 = lat * lat;
var lat4 = lat2 * lat2;
xy[0] = lon * (0.8707 - 0.131979 * lat2 + lat4 * (-0.013791 + lat4 * (0.003971 * lat2 - 0.001529 * lat4)));
xy[1] = lat * (1.007226 + lat2 * (0.015085 + lat4 * (-0.044475 + 0.028874 * lat2 - 0.005916 * lat4)));
};
function inverse(x, y, lonlat) {
var yc, tol, y2, y4, f, fder;
if(y > MAX_Y) {
y = MAX_Y;
} else if(y < -MAX_Y) {
y = -MAX_Y;
}
// Newton's method
yc = y;
for(; ; ) {
y2 = yc * yc;
y4 = y2 * y2;
f = (yc * (1.007226 + y2 * (0.015085 + y4 * (-0.044475 + 0.028874 * y2 - 0.005916 * y4)))) - y;
fder = 1.007226 + y2 * (0.015085 * 3 + y4 * (-0.044475 * 7 + 0.028874 * 9 * y2 - 0.005916 * 11 * y4));
yc -= tol = f / fder;
if(Math.abs(tol) < 0.0000000001) {
break;
}
}
lonlat[1] = yc;
y2 = yc * yc;
lonlat[0] = x / (0.8707 + y2 * (-0.131979 + y2 * (-0.013791 + y2 * y2 * y2 * (0.003971 - 0.001529 * y2))));
};
</script>
</head>
<body>
<script type="text/javascript">
var R = 6371008.7714;
var lat = 89 / 180 * Math.PI, lon = 120 / 180 * Math.PI;
var xy = [];
forward (lon, lat, xy);
document.write((xy[0] * R).toFixed(10) + " / " + (xy[1] * R).toFixed(10) + "<br>");
inverse (xy[0], xy[1], xy);
document.write((xy[0] / Math.PI * 180).toFixed(10) + " / " + (xy[1] / Math.PI * 180).toFixed(10));
</script>
</body>
More information about the Proj
mailing list