MATLAB: How do you use ‘polynomial’, ‘piecewise linear’ or ‘lwm’ with the ‘fitgeotrans’ function

Image Processing Toolbox

Do you know how I can build a transform with ‘polynomial’, ‘piecewise linear’ or ‘lwm’ in the ‘fitgeotrans’ function?

Best Answer

To get the 'polynomial', 'piecewiselinear' or 'lwn' points, you need to have at least as many specified control points for the transformations. Assuming that there are 6 points in 'movingPoints' ans 'fixedPoints', then the calls look like:

pl = fitgeotrans(movingPoints,fixedPoints,'polynomial',2)
pwl = fitgeotrans(movingPoints,fixedPoints,'pwl')
lwm = fitgeotrans(movingPoints,fixedPoints,'lwm', 6)

You can find control points in many ways, including using the 'cpselect' tool.

Here is an example with some dummy points:

im1 = checkerboard;
im2 = imrotate(im1,30);
fixedPoints = ...
[11.0000 11.0000;
41.0000 71.0000;
70.6261 10.3739;
50.8784 30.4099;
20.7523 40.3559;
65.4369 65.5811];
movingPoints = ...
[14.0000 44.0000;
70.0000 81.0000;
66.5991 14.8694;
59.2658 42.0225;
38.6532 64.6171;
89.9865 64.6171];
pl = fitgeotrans(movingPoints,fixedPoints,'polynomial',2)
pwl = fitgeotrans(movingPoints,fixedPoints,'pwl')
lwm = fitgeotrans(movingPoints,fixedPoints,'lwm', 6)

Please visit the following link for additional information:

https://www.mathworks.com/help/images/ref/fitgeotrans.html#bvonasz

Related Solutions

MATLAB: How to use functions “fitgeotrans” and “transformPointsForward” instead of “cp2tform” and “tformfwd”

It is possible to use the function "cp2tform" to get a 'tform' structure and use it as an input to the function "tformfwd". However, the documentation for the function "cp2tform" recommends using the function "fitgeotrans" instead. The function "fitgeotrans" returns a 'tform' object, which is not a valid input to the function "tformfwd". This issue can be resolved by using the function "transformPointsForward" instead of "tformfwd".

The following is an example of how one can use the function "fitgeotrans" followed by "transformPointsForward" instead of "cp2tform" followed by "tformfwd". The only additional step required in the new workflow is to create a 'projective2d' object from the 'tform' object returned by the function "fitgeotrans". This 'projective2d' object is a valid input to the function "transformPointsForward".

Example:

Please refer to the following example that shows how to use the function "cp2tform":

% Transform an image, use the cp2tform function to return the transformation, and 
% compare the angle and scale of the TFORM to the angle and scale of the original 
% transformation:
I = checkerboard;
J = imrotate(I,30);
fixedPoints = [11 11; 41 71];
movingPoints = [14 44; 70 81];
cpselect(J,I,movingPoints,fixedPoints);
t = cp2tform(movingPoints,fixedPoints,'nonreflective similarity');
% Recover angle and scale by checking how a unit vector 
% parallel to the x-axis is rotated and stretched. 
u = [0 1]; 
v = [0 0]; 
[x, y] = tformfwd(t, u, v); 
dx = x(2) - x(1); 
dy = y(2) - y(1); 
angle = (180/pi) * atan2(dy, dx);
scale = 1 / sqrt(dx^2 + dy^2);

In this example, an image is transformed and the function "cp2tform" is used to return the transformation as a 'tform structure'. Subsequently, the angle and scale of the 'tform' is obtained using the function "tformfwd", and compared to the angle and scale of the original transformation.

The MATLAB script "usingFITGEOTRANS.m" demonstrates how the same example can be rewritten using functions "fitgeotrans" and "transformPointsForward".

Please note the following key differences in these two examples:

(1) Using the function "fitgeotrans" instead of "cp2tform":

% Using "cp2tform"
t = cp2tform(movingPoints,fixedPoints,'nonreflective similarity');
% Using "fitgeotrans"
t = fitgeotrans(movingPoints,fixedPoints,'nonreflectivesimilarity');

(2) Creating a 'projective2d' object from the 'tform' object, and using it with the function "transformPointsForward" instead of "tformfwd":

% Using "tformfwd"
[x, y] = tformfwd(t, u, v); 
% Using "transformPointsForward"
projectiveObj = projective2d(t.T);
[x, y] = transformPointsForward(projectiveObj, u, v);

MATLAB: How to transform given XY coordinates using ‘lwm’ tform

A forward version of the lwm transform is not provided because it is not guaranteed to be a 1-1 transformation, but perhaps it would be useful to you simply to do the inverse fit

tform_inv = fitgeotrans(img2_coordinates, img1_coordinates, 'lwm', 10);

and then apply transformPointsInverse.

Best Answer

Related Solutions

MATLAB: How to use functions “fitgeotrans” and “transformPointsForward” instead of “cp2tform” and “tformfwd”

MATLAB: How to transform given XY coordinates using ‘lwm’ tform

Related Question