OpenGL 3d stereo viewing for aerial stereo image (left, right images)

Question

Hi all! Please help me!

My purpose: - display overlapping area with whole left right satellite or aerial image

        - rotate image using roll, pitch yaw angles
        -image format is Tiff image

Now i'm doing OpengGL 3d stereo viewing for aerial stereo image (left, right stereo pair). I displayed my left right image using following code:

glDrawBuffer (GL_BACK_LEFT);
LeftBuf = new unsigned char [LeftCol * LeftRow *3];

for (int i = 0; i < LeftRow; i++){
 for (int j = 0; j< LeftCol; j++){

  LeftBuf [(i*LeftCol + j)*3] = tiff.LnBuf [i*LeftCol + LeftCol-(j-1)];
  LeftBuf [(i*LeftCol + j)*3+1] = tiff.LnBuf [i*LeftCol + LeftCol-(j-1)];
  LeftBuf [(i*LeftCol + j)*3+2] = tiff.LnBuf [i*LeftCol + LeftCol-(j-1)];

 }
}

glRasterPos2f(-1,-1);
glDrawPixels(LeftCol, LeftRow, GL_RGB, GL_UNSIGNED_BYTE, LeftBuf);

glDrawBuffer (GL_BACK_RIGHT);
RightBuf = new unsigned char [RightCol *RightRow *3];

for (int r = 0; r < RightRow; r++){
 for (int c = 0; c< RightCol; c++){

  RightBuf [(r*RightCol + c)*3] = tiff.RnBuf [r*RightCol + RightCol-(c-1)];
  RightBuf [(r*RightCol + c)*3+1] = tiff.RnBuf [r*RightCol + RightCol-(c-1)];
  RightBuf [(r*RightCol + c)*3+2] = tiff.RnBuf [r*RightCol + RightCol-(c-1)];

 }
}

glRasterPos2f(-1,-0.45);
glDrawPixels(RightCol, RightRow, GL_RGB, GL_UNSIGNED_BYTE, RightBuf);

now i need to 3D rotate my right image using Opengl following function: glRotatef(rightphoto.omega*180/PI,1.0f,0.0f,0.0f); glRotatef(rightphoto.phi*180/PI,0.0f,1.0f,0.0f); glRotatef(rightphoto.kappa*180/PI,0.0f,0.0f,1.0f); but i can't rotate my image.

Please help me how to rotate my image!!!!! Also i wanna to display left overlapped whole images of same part of left right images with 3d rotating. I don't know how to calculate overlapping regions coordinate of whole left right images. sorry for my poor english.

Answer 1

Pixels read or written using glDrawPixels() are not transformed by the model view matrix, which is what the glRotatef() changes. They are pixels, a more low-level primitive than actual models.

To solve this, consider drawing your content using e.g. a texture-mapped primitive such as a quad.

Answer 2

thank you very much