DOI Heft:
Nerudová, Zdeňka; Neruda, Petr; Sadovský, Petr: Open software „HROT”. Digital 2D technology for the description of archaeological analysis
DOI Seite / Zitierlink: https://doi.org/10.11588/diglit.26514#0140
136
Zdeńka Nerudova, Petr Neruda, Petr Sadovsky
in all pictures, it is necessary to rotate it and adjust
the main axis to the yertical position. To rotate the
leaf point, the longest linę joining two contour po-
ints must be found. The rotation means to move the
whole image in a way the previously found longest
linę is parallel to the yertical picture edge. As soon as
the image is rotated, another segmentation is carried
out to crop picture edges caused by the rotation.
Measuring the Basic Parameters
The edge coordinates are conyerted from the co-
ordinates of the leaf point image to the coordinates
with zero in the lowest point of the artefact. Sub-
seąuently dimension standardization is done using
a constant obtained from recalculation of the scalę
size. The result represents a dot matrix which de-
scribes the artefact in Cartesian coordinates whose
yalues correspond to the real sizes of the leaf point.
The measurement itself is very easy. The only thing
to do is to find out maximum and minimum yalues
of the coordinates. The maximal value of the y-axis
indicates the leaf point height. The minimal and
maximal yalues of the x-axis indicate the leaf point
width or thickness. As the measurement is based on
pictures taken from front and side views, we obtain
two yalues for the leaf point height and their arith-
metic average makes the real size gained from the
image morę accurate.5
The biggest error appears in the value of thickness
because the shape is very irregular. While a slide gau-
ge measures only thickness between two opposite po-
ints, the measurement done on an image works with
the longest distance between two points which may
not be opposite in reality (see Fig. 1. A,B).6
Determination of Tip Shape
The goal of the next measuring phase is to obtain
data which would enable to characterize each leaf
point. For now, the distances of a point from the mi-
nimal and maximal yalues measured under a certain
angle were chosen as determinants of the shape. It
means there are lines drawn under angles of 75°, 80°
and 85° both in the first ąuadrant and symmetrical-
ly to the y-axis in the second ąuadrant of the distal
extremity. All those lines starts in zero of the refe-
rence frame. Other lines were likewise constructed
from the proximal extremity (Fig. l.C). Values of di-
stances are diyided by the leaf point height to enable
further comparison of all leaf points. The size and
the leaf point shape are in relative coordinates and
thus comparable. Values in the set of determinants
describing the leaf point belong to interval <0,1 >
and they define the shape of the proximal and distal
part of the leaf point.
Fig. l.D also shows reasons for error occurren-
ce. Measuring from the maximum value, there are
certain points lying close to the starting zero. Those
points are, considering sought distances, incorrect
even they meet a mentioned condition. The error
may be void by applying a condition of minimal di-
stance. It is also the reason why we decided to me-
asure from the maximum and minimum; smaller
angles (30° or 45° for example) generate a big num-
ber of such incorrect points and they influence the
result considerably. It can be clearly seen especially
for the angle of 15° in Fig. l.D.
Calculation of a Distal Tip Angle and a Proximal Tip
To describe the shape of a leaf point, angles of
the distal and proximal tips should be known. We
chose a test zonę in one fifth of the leaf point height
to check algorithms for parameter determination.
The measuring itself consists in examination of the
angles from the beginning (the maximum) to all po-
ints meeting the 1/5-height criterion. Those angles
are used for an arithmetic average which indicates
approximately the angle of the zonę in ąuestion.
Though, it is not an angle which would keep precise
track of edges (Fig. l.E,F). The ambition is to find a
suitable set of parameters with the minimal mutual
correlation which would describe leaf point shapes
in the best possible way.
Gained data are continuously or in batches
displayed in a separate application window as
common text in the ASCII format. No special
program was created for data processing because
it seems morę useful to export data in a format
suitable for a common commercial spreadsheet
(MS Excel for example). It also ensured better
compatibility with statistic programs (it was te-
5 A loop check of such calculated values and their comparison to classical measuring with a slide gauge shows only minimal difference in
tenths of milimetres to 1 mm (notę Z. N.).
6 The optics also plays its role because images taken from a short distance may be distorted especially in case of an object with big depth of
focus. The location of the scalę is also important.
Zdeńka Nerudova, Petr Neruda, Petr Sadovsky
in all pictures, it is necessary to rotate it and adjust
the main axis to the yertical position. To rotate the
leaf point, the longest linę joining two contour po-
ints must be found. The rotation means to move the
whole image in a way the previously found longest
linę is parallel to the yertical picture edge. As soon as
the image is rotated, another segmentation is carried
out to crop picture edges caused by the rotation.
Measuring the Basic Parameters
The edge coordinates are conyerted from the co-
ordinates of the leaf point image to the coordinates
with zero in the lowest point of the artefact. Sub-
seąuently dimension standardization is done using
a constant obtained from recalculation of the scalę
size. The result represents a dot matrix which de-
scribes the artefact in Cartesian coordinates whose
yalues correspond to the real sizes of the leaf point.
The measurement itself is very easy. The only thing
to do is to find out maximum and minimum yalues
of the coordinates. The maximal value of the y-axis
indicates the leaf point height. The minimal and
maximal yalues of the x-axis indicate the leaf point
width or thickness. As the measurement is based on
pictures taken from front and side views, we obtain
two yalues for the leaf point height and their arith-
metic average makes the real size gained from the
image morę accurate.5
The biggest error appears in the value of thickness
because the shape is very irregular. While a slide gau-
ge measures only thickness between two opposite po-
ints, the measurement done on an image works with
the longest distance between two points which may
not be opposite in reality (see Fig. 1. A,B).6
Determination of Tip Shape
The goal of the next measuring phase is to obtain
data which would enable to characterize each leaf
point. For now, the distances of a point from the mi-
nimal and maximal yalues measured under a certain
angle were chosen as determinants of the shape. It
means there are lines drawn under angles of 75°, 80°
and 85° both in the first ąuadrant and symmetrical-
ly to the y-axis in the second ąuadrant of the distal
extremity. All those lines starts in zero of the refe-
rence frame. Other lines were likewise constructed
from the proximal extremity (Fig. l.C). Values of di-
stances are diyided by the leaf point height to enable
further comparison of all leaf points. The size and
the leaf point shape are in relative coordinates and
thus comparable. Values in the set of determinants
describing the leaf point belong to interval <0,1 >
and they define the shape of the proximal and distal
part of the leaf point.
Fig. l.D also shows reasons for error occurren-
ce. Measuring from the maximum value, there are
certain points lying close to the starting zero. Those
points are, considering sought distances, incorrect
even they meet a mentioned condition. The error
may be void by applying a condition of minimal di-
stance. It is also the reason why we decided to me-
asure from the maximum and minimum; smaller
angles (30° or 45° for example) generate a big num-
ber of such incorrect points and they influence the
result considerably. It can be clearly seen especially
for the angle of 15° in Fig. l.D.
Calculation of a Distal Tip Angle and a Proximal Tip
To describe the shape of a leaf point, angles of
the distal and proximal tips should be known. We
chose a test zonę in one fifth of the leaf point height
to check algorithms for parameter determination.
The measuring itself consists in examination of the
angles from the beginning (the maximum) to all po-
ints meeting the 1/5-height criterion. Those angles
are used for an arithmetic average which indicates
approximately the angle of the zonę in ąuestion.
Though, it is not an angle which would keep precise
track of edges (Fig. l.E,F). The ambition is to find a
suitable set of parameters with the minimal mutual
correlation which would describe leaf point shapes
in the best possible way.
Gained data are continuously or in batches
displayed in a separate application window as
common text in the ASCII format. No special
program was created for data processing because
it seems morę useful to export data in a format
suitable for a common commercial spreadsheet
(MS Excel for example). It also ensured better
compatibility with statistic programs (it was te-
5 A loop check of such calculated values and their comparison to classical measuring with a slide gauge shows only minimal difference in
tenths of milimetres to 1 mm (notę Z. N.).
6 The optics also plays its role because images taken from a short distance may be distorted especially in case of an object with big depth of
focus. The location of the scalę is also important.