Image Processing Reference
In-Depth Information
c. Filter to detect horizontal line
2
4
3
5
1
222
1
1
h
(
n, m
) ¼
(
2
:
87
)
1
1
1
d. Filter to detect vertical line
2
4
3
5
12
1
h
(
n, m
) ¼
12
1
(
2
:
88
)
12
1
e. Filter to detect lines oriented at 45
2
4
3
5
1
12
h
(
n, m
) ¼
12
1
(
2
:
89
)
2
1
1
45
f. Filter to detect lines oriented at
2
4
3
5
2
1
1
h
(
n, m
) ¼
12
1
(
2
:
90
)
1
12
g. Filter for detecting horizontal edges
2
4
3
5
1
000
121
1
2
h
h
(
n, m
) ¼
(
2
:
91
)
h. Filter for detecting vertical edges
2
4
3
5
101
h
v
(
n, m
) ¼
202
(
2
:
92
)
101
As an example of FIR
filtering, consider edge detection based on the two
filters
speci
ed in g and h above. These
filters are known as Sobel operators. To create an
edge map of image f ,we
first compute the gradient of f in the x- and y-directions by
filtering the image using
filters h
h
(
n, m
)
and h
v
(
n, m
)
, respectively. That is,
G
x
(
n, m
) ¼
f
(
n, m
)
*
h
h
(
n, m
)
(
2
:
93
)
G
y
(
n, m
) ¼
f
(
n, m
)
*
h
v
(
n, m
)
(
2
:
94
)
To
find the edge map of the image, quantize the magnitude of the gradient image
estimated by
Search WWH ::
Custom Search