6.4 Complex Methodology for Reverse Engineering …
555
P x X =
1
k
k−1
i=0
P x X i ; P y X =
1
k
k−1
i=0
P y X i
P x Y =
1
k
k−1
i=0
P x Y i ; P y Y =
1
k
k−1
i=0
P y Y i
The offset value will be calculated as the average from all possible values. The
use of the average offset from the entire data array is due to the fact that the density
of the offset distribution function is close to the normal probability law.
After calculations for the selected minimum of the morphological correlation
function are completed, the relevant program analyzes the result based on two criteria:
– Definition of the minimum R m (K);
– Position of the minimum inside the definition domain.
K =
R m cp. − R m min
R m cp.xs
.
Analysis of the definition of the minimum is performed using the calculated
coefficient
After determining the numeric value of the K coefficient, it is compared to the
e value set by the user. If K > e, the minimum is considered clearly defined. If K
< e, the definition domain for kτ and kτ ort is expanded, and all calculations are
performed once again. Going beyond the definition domain indicates failure of the
alignment process; in this case, PX = PX. If the above condition is met, the frame
coordinates are adjusted accordingly, and the frame takes the corresponding place.
If the alignment result is unsatisfactory, it is possible to increase the number of
the acquisition, number of iterations, and s. However, it is necessary to remember
that an increase in the acquisition and number of iterations leads to an increase in
the calculation time; an increase in the number s in turn reduces the precision of
alignment.
Algorithm for calculation of offsets of PX and PX overlaps during alignment of
the /st and the I + 1st frame along the X-axis is shown in Fig. 6.33.
6.4.4 Description of the Process of Aligning a Group
of Image Frames
Calculation of the average function value for each pair of frames requires a great
amount of hardware resources, since the number of aligned frames usually exceeds a
hundred [11]. In order to minimize alignment costs, another alignment method was
suggested in the basis of the original material.
The essence of this method is that after obtaining the entire array of frames, it
is necessary to determine pairs of frame offset values during movement along the
555
P x X =
1
k
k−1
i=0
P x X i ; P y X =
1
k
k−1
i=0
P y X i
P x Y =
1
k
k−1
i=0
P x Y i ; P y Y =
1
k
k−1
i=0
P y Y i
The offset value will be calculated as the average from all possible values. The
use of the average offset from the entire data array is due to the fact that the density
of the offset distribution function is close to the normal probability law.
After calculations for the selected minimum of the morphological correlation
function are completed, the relevant program analyzes the result based on two criteria:
– Definition of the minimum R m (K);
– Position of the minimum inside the definition domain.
K =
R m cp. − R m min
R m cp.xs
.
Analysis of the definition of the minimum is performed using the calculated
coefficient
After determining the numeric value of the K coefficient, it is compared to the
e value set by the user. If K > e, the minimum is considered clearly defined. If K
< e, the definition domain for kτ and kτ ort is expanded, and all calculations are
performed once again. Going beyond the definition domain indicates failure of the
alignment process; in this case, PX = PX. If the above condition is met, the frame
coordinates are adjusted accordingly, and the frame takes the corresponding place.
If the alignment result is unsatisfactory, it is possible to increase the number of
the acquisition, number of iterations, and s. However, it is necessary to remember
that an increase in the acquisition and number of iterations leads to an increase in
the calculation time; an increase in the number s in turn reduces the precision of
alignment.
Algorithm for calculation of offsets of PX and PX overlaps during alignment of
the /st and the I + 1st frame along the X-axis is shown in Fig. 6.33.
6.4.4 Description of the Process of Aligning a Group
of Image Frames
Calculation of the average function value for each pair of frames requires a great
amount of hardware resources, since the number of aligned frames usually exceeds a
hundred [11]. In order to minimize alignment costs, another alignment method was
suggested in the basis of the original material.
The essence of this method is that after obtaining the entire array of frames, it
is necessary to determine pairs of frame offset values during movement along the