To project or unproject regularly-spaced raster data that is associated with a geographic
or map reference object, you must first create a coordinate grid that matches the size of the
raster. Use different grid creation functions depending on which way you are projecting. When
you project latitude-longitude coordinates to x-y
coordinates, create a grid using the meshgrat
function. When you unproject
x-y coordinates to latitude-longitude coordinates,
create a grid using the meshgrid
and intrinsicToWorld
functions.
After transforming the raster data, you can display it on a map using visualization
functions such as mapshow
and geoshow
. Use mapshow
for projected x-y coordinates
and geoshow
for unprojected latitude-longitude coordinates.
To project data that is associated with a geographic raster reference object, first create a grid of latitude-longitude coordinates for each point in the raster. Then, project the geographic coordinates to x-y map coordinates.
For example, import elevation raster data as an array and a geographic cells reference object. Get the latitude-longitude coordinates for each point in the raster by using the meshgrat
function.
[Z,R] = readgeoraster('n39_w106_3arc_v2.dt1');
[lat,lon] = meshgrat(Z,R);
Now that you have your grid, select a map projection to use when projecting the coordinates. For this example, create a projcrs
object for UTM zone 13 in the northern hemisphere. Then, project the latitude-longitude coordinates to x-y coordinates.
p = projcrs(32613); [x,y] = projfwd(p,lat,lon);
Display the projected raster as a surface by calling mapshow
and specifying the x-y coordinates and elevation array. Add axis labels and apply a colormap appropriate for elevation data.
figure mapshow(x,y,Z,'DisplayType','surface') xlabel('x (meters)') ylabel('y (meters)') demcmap(Z)
If the geographic CRS of the latitude-longitude coordinates does not match the geographic CRS of the projected CRS, then the projected coordinates may be inaccurate. You can find the geographic CRS of a projcrs
object or a geographic raster reference object by querying their GeographicCRS
properties.
p.GeographicCRS.Name
ans = "WGS 84"
R.GeographicCRS.Name
ans = "WGS 84"
The DTED file used in this example is courtesy of the US Geological Survey.
To unproject data that is associated with a map raster reference object, first create a grid of x-y coordinates for each point in the raster. Then, unproject the x-y map coordinates to geographic coordinates.
For example, import an image of Boston as an array and a map cells reference object. Get information about the map projection as a projcrs
object by querying the ProjectedCRS
property of the reference object.
[Z,R] = readgeoraster('boston.tif');
p = R.ProjectedCRS;
Create a coordinate grid by getting the x-y coordinates for each point in the raster. To do this, first create a grid of intrinsic coordinates that matches the size of the raster by using the meshgrid
function. The intrinsic y-coordinates increase from row to row and the intrinsic x-coordinates increase from column to column. Then, transform the intrinsic coordinates to x-y coordinates by using the intrinsicToWorld
function.
s = size(Z); [xIntrinsic,yIntrinsic] = meshgrid(1:s(2),1:s(1)); [x,y] = intrinsicToWorld(R,xIntrinsic,yIntrinsic);
Unproject the x-y coordinates to latitude-longitude coordinates by using the projinv
function and specifying the projcrs
object and coordinate grid.
[lat,lon] = projinv(p,x,y);
Display the unprojected image by calling geoshow
and specifying the latitude-longitude coordinates and image array. By default, geoshow
displays coordinates using a Plate Carre projection. Then, add axis labels.
figure geoshow(lat,lon,Z) xlabel('Longitude (degrees)') ylabel('Latitude (degrees)')
intrinsicToWorld
| meshgrat
| meshgrid
| projfwd
| projinv