SEARCH
0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Prev | Current Page 258 | Next

Yingshu Li, My T. Thai, and Weili Wu

"Wireless Sensor Networks and Applications"

??x and ??y for sensor S0 can be calculated as:
??x = X 8s2PAs0
Wx(xs, ys)Gx(xs, ys)Vs, (8)
??y = X 8s2PAs0
Wy(xs, ys)Gy(xs, ys)Vs. (9)
By observing equation (6) and (7), we know the basic idea behind the
Prewitt filters is to find the gradient in the x and y axes by deducting the
sum of grey levels of left pixels from the right ones and subtracting the sum
of grey levels of upper pixels from the below ones respectively. To simulate
this behavior, the new pair of ???convolution mask??? can be defined as:
Gx(xs, ys) = ?? 1 if xs ?? x0,
??’1 if xs < x0,
(10)
Gy(xs, ys) = ?? 1 if ys ?? y0,
??’1 if ys < y0.
(11)
Thus, the sum of Gx(xs, ys)Vs and Gy(xs, ys)Vs in equation (8) and (9)
represent the gradient of the values of event predicates for sensor S0 in the x
and y directions.
Fig. 3. The four quadrants around sensor S0, where n1+ = 4, n1??’ = 1, n2+ =
3, n2??’ = 2, n3+ = 1, n3??’ = 4, n4+ = 2, n4??’ = 3.
154
Chapter 6 Boundary Detection for Sensor Networks
Now, let??™s discuss the weighting function G(x, y). To take spatial information
into consideration, we can divide the region around sensor S0 into
four quadrant areas as shown in Figure 3. Let (ni+, ni??’)i=4
i=1 be the number
of sensors with 1 and 0 values of event predicates in each quadrant. A simple
weighting function is as follows:
Wx(x, y) = ( 1
n1++n1??’+n4++n4??’
if x < xs0 ,
1
n2++n2??’+n3++n3??’
if x > xs0 ,
(12)
Wy(x, y) = ( 1
n1++n1??’+n2++n2??’
if y > ys0 ,
1
n3++n3??’+n4++n4??’
if y < ys0 .


Pages:
246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270