Rotate a Matrix

[Optiker]

I have a matrix of values (180 X 220) representing surface heights from a CCD. The data is rotated ( anywhere from -3 to + 15 degrees) and I need to rotate it back so the surface heights line up with the ccd axes. The rotation is strictly around the z axis and is data-centric ( rotates around the center of the data). Any ideas?
Bill

 

[GlennUK]

Can you tell me what kind of thing you are expecting for those cells near the origin?

for example, let's suppose that the numbers 5's and 8's are the centre 4 cells of the matrix. What would expect a rotation of 15 degrees to result in?
[1][2][3][2]
[4][5][5][4]
[7][8][8][7]
[6][9][9][6]


Cheers, Glenn.

Did you hear about the literalist show-jumper? He broke his nose jumping against the clock.

 

[Optiker]

I would expect there would need to be an expansion of the matrix, and interpolation between the phase values. For the case of 45 degrees:

0 degrees

[1][2][3][2]
[4][5][5][4]
[7][8][8][7]
[6][9][9][6]

45 degrees

[2]
[3] [4]
[2] [5] [7]
[1] [5] [8] [6]
[4] [8] [9]
[7] [9]
[6]
(sorry, looked OK with notepad -- should be symmetrical diamond shaped matrix)

I would expect the relationship of the values to each other to remain the same.

 

[GlennUK]

Yes, Optiker, I see what 45 degrees anticlockwise would look like, and I'd already worked that out. But what happens for amounts less than a multiple of 45 degrees. You could work out co-ordinates for those parts of the matrix well away from the origin, where the change might result in a shift that matches a good fit for cell location, but what about those cells near the origin?



Cheers, Glenn.

Did you hear about the literalist show-jumper? He broke his nose jumping against the clock.