Need to account for binning and ROI in mosaic positions

[mickp]

I recently implemented support for binning and ROI for the Andor CCD cameras in MicronOxford/microscope. Image display required a minor tweak to the imageViewer to track texture size and check against new image size. The mosaic will be a bit more involved.

When images are added to the mosaic, they are centred on the current image position. This means that a camera ROI that is not symmetrically cropped (so the image centre corresponds to the detector centre) will be placed at the wrong position on the mosaic. We'll need to:

• check ROI and adjust the image placement, possibly accounting for any image transforms on the camera side, too;
• scale or resample binned images.

I will implement ROI and binning on the MicronOxford/microscope test camera so we can test this easily without the need for hardware.

[iandobbie]

For this to work we need to move pixel size from the objective to the camera, otherwise what happens when even identical cameras have different binning?

[iandobbie]

The mosaic basically works, Some strangeness but binning breaks the mosaic image positioning. I suggest we open a new ticket as the problem is really the pixel size should go with the camera image not the objective.

Also relates to saving snap images #391

[iandobbie]

The new interactive ROI code works very well but the image position in the mosaic doesn't reflect the position of the ROI on the chip, just the stage position. Not sure if binning has any affect but when we fix the position we should take account of binning as well.