Camera Calibration Technical Report:
Image Correction

Image Correction

The calibration procedure results in a conversion table from measured value to a density. To determine the density values in a particular image requires a corresponding dark-field (zero light) and bright-field (no object) image. Since the dark-field image clearly does not depend on the microscope optics or illumination, the dark-field image captured during calibration may be used for subsequent measurements. However the bright-field image must be re-captured whenever the slide, objective, condenser or lamp brightness are adjusted. The bright-field image is obtained as follows:

1.

Set up the microscope for the slide of interest, using the required objective. Set the lamp brightness to give a good image on the video monitor. Close down the condensor field apperture so as to just enclose the required image frame and so minimise glare.

2.

Move to an area of the slide which contains mostly clear field, and focus the objective and condensor on this region. Digitise several frames (we usually find three or four to be adequate) at different slide positions.

3.

Compute the final bright-field image by selecting the brightest value across the captured frames at each pixel.

The bright-field may be used to correct subsequent images from the same slide as long as the microscope settings are not adjusted. To correct an image subtract the dark-field value, convert each pixel value to a density and subtract the bright-field density:

$$OD = \rho(g - g_d) - \rho(g_b - g_d),$$

where g is the measured value, g_d the corresponding dark-field value and g_b the corresponding bright field value.