Mount the digital camera on microscope (make sure there are no mirrors in the optical path), load batteries of camera, empty PCMCIA card (storage medium of camera). Adjust/center microscope, perform Köhler illumination, insert analyzer/polarizers, have lambda and lambda/4-plates ready, mount tilt stage on Kreuztisch, place thin section in tilt stage. Connect camera to computer via SCSI cable.
Make sure that the Kreuztisch, the border of the tilt stage, the crosshair of the oculars and the orientation of the camera are well aligned.
There are two possible ways of collecting the data: (a) one where the thin section is rotated while the analyzer/polarizer/lambda plate assembly remains stationary; and (b) one where the analyzer/lambda/polarizer plate assembly (ALP-assembly) is rotated while the thin section remains immobile.
The advantage of (a) is that illumination is constant with respect to all images and does not have to be corrected, the advantage of (b) is that the sequence of images is already matched and only slight shifts will be necessary in the re-matching step. The disadvantage of (a) is that all images have to be rotated, i.e., transformed. This means that in the course of the necessary bilinear interpolation calculations, the input gray values of any given pixel are not the original ones anymore but recalculated, "spread out" as it were, over the neighbouring pixels. This represents some sort of data smoothing which may or may not be advantageous. A further disadvantage of method (a) is the fact that only the center portion of the captured image can be used: usually, 780*620 matrices can be retrieved from 1268*1012 scanned images. The disadvantage of method (b) is the relative rotation of the lighting with respect to the scanner: since the ALP-assembly polarizes the light, and since the illuminating light is usually reflected on a mirror in the base of the microscope causing it to be slightly polarized, the lighting intensity received in the scanner will change as a function of rotation of the ALP-assembly. Later, in the calculation, this background oszillation has to be measured and corrected.
The camera is set to manual, the exposure meter is set to "evaluative metering" (entire field of view). The camera MUST be on manual (ensuring constant exposure times for all takes, see below) but the meter setting is not critical as it is used only to determine the exposure time once, at the beginning of the series, and because the correct exposure time is not determined via the exposure meter but by evaluating the grey value histogram of the acquired images, as discussed below.