1221
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
Take the outdated road vector as initial position, along the
vertical direction of road segment, match every vector segment
to RS image using multi-scale template. Compute the value of
correlation coefficient [2] , the formula is as follows:
X X S,J -g'i+r.J+c-
p(c,r)-
| XXs,.,)(XX
' < = 1 j = 1 i = 1 j= 1
h<tl ¡,.,>‘11 t i fW. ~(tt )')
1 f
E snaked) = - J J£ int + E image + E co Jds
LSB-Snake [4] is an efficient model to extract liner-like features,
it describes Snake curve using B-spline with parameters, and
iterative to minimum energy by using the algorithm of least
square estimation, allocate the place of node points by the
complexity of B-spline.
In the formula, m and n represent the row and column of image
block respectively, r and c represent the searching scope, g and
g’ represent the grayscale of template and image respectively,
the maximal value of correlation coefficient must corresponds
the real place of the road.
After the multi-scale template matching, choose the max
correlation coefficient as the unique matching result of a vector
segment. If the unique result is larger than a given threshold,
it’s an effective result, each effective result must has an
corresponding template, compute the sum length (called L same )
of the vector segments who are same in template width, choose
the max L same named L sameMax , and the length of the road vector
(called L total ), if the ratio of L sameMax to L total is larger than a
given threshold, this road hasn’t disappeared or partial changed,
vice versa.
The template width corresponded with the maximum length
(LsameMax) is considered as an efficient template, and its width
of the road is considered as the width of road in image, the
corresponding dark or bright attribute of the template is
considered as the dark or bright attribute of the road in image.
After processing above, compute the sum length of successfully
matched road vector segments, and the ratio of it to the total
length of the road, if the ratio is less than a given threshold, the
road has disappeared or partially changed, or it’s unchanged.
Figure 6a is the change detection result of our method,
Figure 6b is the actual changed road. There are 87 road feature
in total, the actual disappeared of partial changed road is 8, our
method detected 41 disappeared of partial changed road, among
them 8 are right. The check-out-ratio is 100.00 % , the
correct-ratio is 62.07%.
a b
Figure 6 Change Detection Result of IKONOS and 1:2000 Map
Road Feature
2.1.2 Newly added road detection
Snake [3] model is a spline curve of lease energy, it has three
elements: inner force, outer force and image force. The inner
force restricts it’s shape, the outer force lead it’s action, and the
image force push it to notable image character. The energy
function of Snake E smte is defined as follows:
Before the extraction of road by LSB-Snake model, it need
manual input the width and dark or bright character of the road
to be extracted, Manual input may not be accuracy and hold
down the extract efficiency. Besides this, the LSB-Snake model
is not robust while the initial seed points are not dense enough.
In our method, we obtain each road’s width and dark or bright
attribute by self-adapt template matching and take it as initial
value of LSB-Snake, this value is accurate and trusty, and the
manual input is avoided. And, comparing with LSB-Snake
model in extraction of road, the method this paper put forward
can use not only the initial seed points, but also the new added
seed points created by self-adapt template matching method,
this makes the extracting result of roads more robust.
Figure 7a is the detection result to new added road by
LSB-Snake model, and Figure 7b is the result by our method.
The round points are inputted manually, the rectangle points are
added by multi-scale template matching method. The detection
to newly added road by our method is more robust than
LSB-Snake model, and the detect efficiency is high.
a b
Figure 7 Comparing of Newly Added Road Detection Results
2.2 Map revision
Take the result of change detection—partial changed or
diminished road as reference of revision, overlap them with the
outdated map road feature and remote sensing image, delete the
diminished road, and revise the partial changed road manually,
the result is the updating result of partial changed or diminished
road.
Take the newly added road produced by change detection
semi-automatically as the updating result of newly added road.
3. EXPERIMENTS
A program is developed to realize the change detection method
above using Visual C++ platform [5] , the revision is carried out
by AutoCAD. We take 5 map sheets with the scale of 1:2000
and corresponding IKONOS image of Beijing, using “revision
based on change detection” method and the current method
