I am trying to calibrate a monocular camera using the
camera_calibration package and the camera1394 drivers. I am able to
successfully calibrate and view a rectified image using a scale value
of 0, but setting the scaling to 1.0 in the cameracalibrator doesn't
seem to have any effect on the saved calibration data. Changing the
scale in cameracalibrator to 1.0 does change the output shown in the
calibration GUI to have black borders so that all pixels in the
original image show up in the rectified image. If I try to view the
rectified image output by image_proc for the corresponding scale of
1.0 gives me the same rectified output as a scale of 0.0.

I have put a few prints in the calibrator to ensure that the set_alpha
method is actually getting called and it is, but when I save
calibration data (or upload it to the camera), the saved calibration
data is the same for both scale values (and at least one of the
matrices should be different correct?). Any ideas why the rectified
image according to image_proc does not match the rectified image in
the calibration GUI?

I'm running latest on Ubuntu 9.10.

- Eric