So, a few minor changes to the original code. First, add 3 new variables:

int iCentreX, iCentreY;
int iDestCentreX, iDestCentreY;
int iWidth, iHeight;
int iDiagonal;

iDestCentreX, iDestCentreY and iDiagonal.

Next assign them as follows:

iWidth = bm.Width;
iHeight = bm.Height;
iDiagonal = (int)(Math.Ceiling(Math.Sqrt((double)(bm.Width * bm.Width + bm.Height * bm.Height)))) + 20;

iCentreX = iWidth / 2;
iCentreY = iHeight / 2;
iDestCentreX = iDestCentreY = iDiagonal / 2;

iDiagonal basically is the diagonal size of the source image. I added 20 pixels for buffer.

Then create the destination image as follows:

Bitmap bmBilinearInterpolation = new Bitmap(iDiagonal, iDiagonal);

Then initialise the destination image:

for (i = 0; i < iDiagonal; ++i)
{
for (j = 0; j < iDiagonal; ++j)
{
bmBilinearInterpolation.SetPixel(j, i, Color.Black);
}
}

Then with the assignment nested for loops, use this (note the end condition of the for loop):

for (i = 0; i < iDiagonal; ++i)
{
for (j = 0; j < iDiagonal; ++j)
{
// convert raster to Cartesian
x = j – iDestCentreX;
y = iDestCentreY – i;

…

The last change is when it's the centre of the source image. Then use this:

// centre of image, no rotation needed
bmBilinearInterpolation.SetPixel(j, i, bm.GetPixel(iCentreX, iCentreY));
continue;

That's it. You should get a destination image with no missing pixels. The edges will be a little jagged, because it's interpolated with black pixels. I think I'll write up a post with the full code. It's hard to put code in comments.

The more important point, is that I want to determine when (x,y) = (0,0), the centre of the image. Since I had an if clause to check for x=0, I might as well do some form of guarding.

Thank you for pointing that out.

Atan(y/x) doesn’t, granted, but Atan has a whole other host of problems. To wit, Atan(1/1) = Atan(-1/-1).

Excellent work.

William Milberry

I was in university then, and I knew only C and MATLAB college work. So yeah, obvious choice of language…