Up: Image Preprocessing
Figure 1: After subtracting background without care
Using image plates as detectors, several sample patterns and the apparative blind scattering have been measured. After subtraction of the background scattering the patterns show artificial edges which vary by strength and position. A typical case is shown in Fig. 1. The reason is that the accuracy of plate positioning in the two holders (detector holder and holder in the image plate scanner) is not good enough. What can we do?
Figure 2: The measured raw pattern
Although the measured raw pattern from Fig. 2 looks quite nice, we should first subtract the fog from the image
WAVE> sf_subfog,a,20
By doing so a background of 10 counts is subtracted from this image. The same operator applied on the background image (b) yields a fog level of 8 counts (Here both images show little fog).
Figure 3: Mask of the sample scattering at a level of 500.
After this we try to find the irradiated area by searching for the shade of the vacuum tube.
WAVE> m = a
WAVE> m.map = a.map GT 500
WAVE> sf_closing,m
WAVE> sf_show,m
yields the image shown in Fig. 3. We want to use the centre of the circular region for image positioning. Thus the shade of the primary beam stop and its holder must be filled. This can be achieved by calling
WAVE> sf_sharpen,m
Figure 4: The mask of the vacuum tube can be used for image plate
positioning
Now the mask appears somewhat eroded and the beamstop area has been filled by some special algorithm, which only works with a horizontal beam stop holder. We can now position the image in a way that the centre of this mask becomes the centre of the image.
WAVE> a = sf_isocenter(a,m)
% Compiled module: SF_ISOCENTER.
% Compiled module: SF_GRAVI.
% Compiled module: SF_IMGMOVE.
Old center 246 223
New center 228 224
WAVE>
If we do the same with the background pattern in the structure ``b'', we have two images which are at almost the same position. Thus now the subtraction
WAVE> ac = a
WAVE> ac.map = a.map - abso*b.map
is carried out with two images which are almost perfectly aligned (``abso'' is the absorption factor). The result is shown in fig. 5.
Figure 5: An almost perfect background subtraction