Master Thesis - Fachbereich Informatik

4.4. TUBE LOCALIZATION 69
Global Threshold The classification into BG and TUBE is based on the general assumption
that the mean intensity of objects is darker than the background. In more detail,
taking the mean value of the smoothed profile Psmooth as a global reference and calculating
the local mean value for each segment s(i), the classification C can be expressed as:
�
TUBE , mean(s(i)) ≤ Psmooth
C1(s) =
(4.10)
BG , otherwise
In image segmentation the mean value is widely used as an initial guess of a threshold
separating two classes of data distinguishable via the gray level [48, 2]. There are many
more sophisticated approaches for threshold selection including histogram shape analysis
[57, 63, 26], entropy [54], fuzzy sets [20, 14] or cluster-based approaches [55, 46]. The different
techniques are summarized and compared in several surveys [59, 47, 60]. However,
in this application the threshold is used for classification and it is not intended for calculation
of a binary image that segments the tubes from the background. Since processing
time is strictly limited and critical in this application, it is essential to save computation
time if possible. As introduced before, the actual segmentation is based on strong vertical
edges in the profile, but does not include any semantic meaning of the segments. In the
classification step, the mean turned out to be a reliable and fast choice to distinguish between
foreground and background segments both for black and transparent tubes if there
is a uniform and sufficient contrast between tubes and the background over the whole image.
In this case there is no need for another threshold than the mean - saving additional
operations.
Insteadofcomparingtheglobalmeanwiththelocalmean,thelocalmediancouldbe
observed to result in a more distinct measure for discrimination:
�
TUBE , median(s(i)) ≤ Psmooth
C2(s) =
(4.11)
BG , otherwise
The better performance of measure C2 originates in the characteristic of the median
tobelesssensitivetooutlierscomparedtothemean[32]. Thisisimportantsincethe
input data can be very unsteady due to the background texture or printing visible on
transparent tubes (independent of the additional camera noise level). As mentioned before,
thesmoothingoftheprofileatthefirststepalsoblursthetubeedgescausingthesegment
boundaries not to be totally precise. In this case, the local mean tends to move closer
to the global mean, which does not have to implicate a misclassification. The median,
however, turned out to be more distinct in most cases. Figure 4.14 shows the smoothed
profile of (a) a transparent and (b) a black tube respectively. The examples represents the
states entering + centered and entering + centered + leaving. The segment boundaries,
which correspond to the locations of the strongest peaks in the first derivative of the
profile, are visualized as well as the global mean and the local median. Segments that
have a median above the global mean are classified as background.
Regional Threshold One drawback of the global threshold approach is that different
background segments are assumed to be almost equal in image brightness, i.e. the tubebackground
contrast is approximately uniform within one image. This assumption, however,
does not hold if there are larger variations in background brightness (for example
due to material properties or dirt on the belt). Such variations can occur between images,