Vol. XXXVIII, Part 7B 
In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
21 
n. It is an information 
s through the mobile 
o make use of the 
ce (Shu Wang 2008). 
there are nine types of 
Physical surroundings, 
'Stem properties, Time 
Iture and Orientation 
t of the defined context 
targets many users and 
unication environment, 
ivided for both pull and 
ot rely on previous user 
rridor be threatened by 
ation messages will be 
e 7, technical based 
<rs (those could be any 
the corridor situation, 
application purpose), 
technology, and data 
nd service 
ing information and send 
isly, the provision of the 
proach would involve a 
tn, since the positions of 
uously transmitted to the 
to be send to the devices 
6). Supervisory and user 
through the internet with 
Dver the ROW, the actual 
ined from the positioning 
of occurrence is sent via 
d gateway. Consequently, 
mge messages among a 
met. At the same time it 
;nsors. A server reads the 
n analyzer. Moreover, the 
mshes it to the receiver. 
>n on whether the corridor 
end, results are sent to the 
;met gateway or mobile 
es and scanner, a pipeline 
the ground based remote 
sensing technique. Figure 8 shows top level schema of three 
simple steps from receiving signal and extracting process aim 
to detect moving object over corridor and to make warning 
about intrusion. 
Figure 8. Top Level Framework 
As it’s mentioned on earlier paragraph and depicted in figure 9, 
receiver input signal consist of data from sensor, reference data 
which calls pattern data and feedback from the previous 
records. The next step, of the extracting moving object has 
three functions of revealing, de-noising and detection. This step 
is usually programmed based on different remote sensing 
application. The last step consists in the function of making 
warn about intmsion to decision maker, estimation and event 
history. Finally event history makes feed back to first step. 
Receiver Input signal Extracting Moving Object Moving Object 
Figure 9. Classification Schema 
Image and data that are gathered recently will be compared to 
the reference. For instance, if hardware processes just a few 
points (no more than 5), because there is high flow rate of data, 
we need a lot of time and expensive hardware. For this reason 
and according to the Figure 10 the problem with scanners 
especially LADAR is considered in three components. 
Figure 10: Process of Object Detection schema 
The first component is Classification and Filtering Process. As 
far as this technique is supposed to helps us to concern certain 
point instead of all the point in the corridor. Figure 10 shows 
how Classification and Filtering can reduce the time to 
producing report and the size of its. To put it another way, 
system is looking for the similarities between recent data and 
reference. For this reason the elimination of the frequent data 
interaction will take into consideration. Secondly, to reduce the 
process time Defining Reference process is introduced. 
Additionally, Discrimination technique is of assistance 
software to match recent data with typical shape and suitable 
format for moving and static objects. For instance, red colour 
and green can assign to moving and static objects respectively. 
On the other hand, it is possible to identify whether it is human 
or device? What it looks like. Regarding to the power of code, 
it is also possible to recognize the dimension of objects. 
For the computation of optical flow of 2D image motion the 
following (Yokoyama and Poggio 2000) function is presented: 
/ (x, t) . I (x + dx,t + dt) (1) 
Where /Ci, t) = the spatiotemporal function of image 
intensity 
^ = intensity of the time 
t + dt = object is assumed to remain the same 
For instance; no illumination change are supposed to occur. 
Equation 1 can be expanded using Taylor order as follows: 
\/ T l.v+ lt= 0 (2) 
Where ^ ~ ^ x ‘ = the gradient 
it = temporal derivative of i (•*• O 
^ = = image velocity 
According to the (Barron, Fleet et al. 1994) one linear equation 
is insufficient to determine 2D velocity that is known as 
aperture problem. However it seems good enough to detect 
moving objects. For instance, moving object on geometric 
shape, (such as line circle, rectangular and....), indicates 
intrusion activities. 
3.2 Detecting Object 
It is defined the similarity between an object of previous frame 
Rprov C r 0 and an object of current frame ^c«r( ra ), 
using estimated position of lines by optical flow. 
{Pjlf A(n)}6 5 
Equation (Bernhard Gruber SW. Location Based Services using 
a Database Federation. Institute for Geoinformation) is 
Detected and stored line of the current frame. 
{5 4 1 S*(¿Vim )} 6 Rfprev (m) (4) 
Whereas equation (4) is defined all background line as a 
previous frame, then similarity is defined as follows: 
p(wi, m) 
y n ; , 
r£[s,l 
Where l“*/ I 
two objects 
u 
(5) 
then these 
is the number of element in S, and '’iJ 
Rprtvtmyand R eur (m ) are considered as 
the same object if the similarity of pOn.m ) 
value. 
is non-zero 
t il if Pi* R C ur{m) where P t € 5 ; 
lj to otherwise . (6) 
RpreArni = R cur (m ),if p(m,m r ) * 0.