with high numbers of outliers well. These capabilities make
the
SLIM
be a very reliable as well accurate and practical,
especially in terms of computational time.
The advantages and specifications of the proposed
MGL
s and
MGP
s along with the innovative
SLIM
are categorized below:
• Linear features are extracted from different methods
in image and object spaces. In addition, a line may be
fragmented to some segments or its position may be
changed due to noise, occlusion or applied extraction
methods. Therefore, the end-points of matched line-
segments are not conjugate. However, the end-points of
matched
MGL
s in both spaces are conjugate.
• In the case of true end-points of line-segments, it is not
feasible to use some geometrical-relations such as the
ratio of lines’ length or the coordinates of end-points in
the matching process. However, by introducing the
MGL
s
and
MGP
s, the geometrical-relations of
MGL
s are possible
for use in matching procedure. Moreover, using both
geometrical relations between lines and the potential
of
MGP
s as the control points simultaneously, make the
procedure faster and also more accurate and reliable.
• Because the process of automatically extracting control
points is a very tedious work, extracting linear features,
and then using the
MGP
s as the control points can sim-
plify the process.
Plate 1. Final matched-lines in image (left figures) and object spaces (right figures) for the first to third datasets (a to c). The lines with
similar colors are indicated as matched-lines. Additionally, the black lines are outliers.
374
May 2016
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
