Page 213 - MATLAB Recipes for Earth Sciences

208 8 Image Processing

The four corners of the image correspond to the pixels in the four corners of
the image that we store in a variable named basepoints.

basepoints(1,:) = [1,4200];
basepoints(2,:)= [1,1];
basepoints(3,:)= [4100,4200];
basepoints(4,:)= [4100,1];

The function cp2tform now takes the pairs of control points input-
points and basepoints and uses them to infer a spatial transformation
matrix tform.

tform = cp2tform(inputpoints,basepoints,'affine');
This transformation can be applied to the original RGB composite naiva-
sha_rgb in order to obtain a georeferenced version of the satellite image
newnaivasha_rgb.

[newnaivasha_rgb,x,y]=imtransform(naivasha_rgb,tform);

Subsequently, an appropriate grid for the image may be computed. The grid
is typically defined by the minimum and maximum values for the longitude

and the latitude. The vector increments are then obtained by dividing the
longitude and latitude range by the array dimension and by subtracting one
from the result.

X = 36.096003 : (36.770406-36.096003)/8569 : 36.770406;
Y = 0.319922 : (0.958743-0.319922)/8400: 0.958743;

Hence, both images can be displayed for comparison (Fig. 8.4 and 8.5).
iptsetpref('ImshowAxesVisibl','On')
imshow(naivasha_rgb), title('Original ASTER Image')
figure
imshow(newnaivasha_rgb,'XData',X,'YData',Y);
xlabel('Longitude'), ylabel('Latitude')
title('Georeferenced ASTER Image')
grid on
The command iptsetpref makes the axis of the image visible. Exporting
the results is possible in many ways, such as

print -djpeg70 -r600 naivasha_georef.jpg
as JPEG fi le naivasha_georef.jpg compressed at 70% and at a resolution
of 600 dpi.