Obtaining image mask
In an example, a method and apparatus for obtaining an image mask is provided. After a magnitude image and a phase image of a to-be-processed image is obtained, magnitude coherent data of each pixel point in the magnitude image and phase coherent data of each pixel point in the phase image may be calculated. Then, a binarization threshold processing may be performed on the magnitude coherent data of each pixel point in the magnitude image to obtain a magnitude image mask. A binarization threshold processing may be performed on the phase coherent data of each pixel point in the phase image to obtain a phase image mask. In this way, an image mask of the to-be-processed image may be obtained by using the magnitude image mask and the phase image mask.
1. A method for obtaining an image mask, the method comprising:
obtaining a magnitude image and a phase image of a to-be-processed image;
obtaining magnitude coherent data of each pixel point in the magnitude image by calculating coherent data of each pixel point in the magnitude image within a preset range from the pixel point, and
obtaining phase coherent data of each pixel point in the phase image by calculating coherent data of each pixel point in the phase image within the preset range from the pixel point;
obtaining a magnitude image mask by performing binarization threshold processing on the magnitude coherent data of each pixel point in the magnitude image, and
obtaining a phase image mask by performing binarization threshold processing on the phase coherent data of each pixel point in the phase image; and
obtaining an image mask for the to-be-processed image by using the magnitude image mask and the phase image mask.
2. The method of claim 1 , wherein before calculating the coherent data of each pixel point in the phase image within the preset range from the pixel point, the method further comprises:
performing homodyne high-pass filtering or phase unwrapping on the phase image.
3. The method of claim 1 , wherein obtaining the image mask for the to-be-processed image by using the magnitude image mask and the phase image mask comprises:
calculating a union of the magnitude image mask and the phase image mask, as the image mask of the to-be-processed image.
4. The method of claim 1 , wherein obtaining the image mask for the to-be-processed image by using the magnitude image mask and the phase image mask comprises:
calculating an intersection of the magnitude image mask and the phase image mask, as the image mask of the to-be-processed image.
5. The method of claim 1 , wherein calculating the coherent data of each pixel point in the magnitude image within the preset range from the pixel point comprises:
calculating the magnitude coherent data Q i,j,k AM of a pixel point (i, j, k) in the magnitude image according to a following formula:
Q
i
,
j
,
k
AM
=
1
α
AM
d
x
d
y
d
z
∑
l
=
i
-
d
x
-
1
2
l
=
i
+
d
x
-
1
2
∑
m
=
j
-
d
y
-
1
2
m
=
j
+
d
y
-
1
2
∑
n
=
k
-
d
z
-
1
2
n
=
k
+
d
z
-
1
2
p
^
l
,
m
,
n
;
wherein Q i,j,k AM denotes the magnitude coherent data of the pixel point (i, j, k) in the magnitude image,
α AM denotes a normalization constant,
d x ,d y ,d z respectively denote values in direction x, direction y and direction z within the preset range from the pixel point, wherein direction z is perpendicular to a plane in which direction x and direction y locate in, and
|{circumflex over (p)} l,m,n | denotes a magnitude of a pixel point (l, m, n) in the magnitude image,
l
∈
(
i
-
d
x
-
1
2
,
i
+
d
x
-
1
2
)
,
m
∈
(
j
-
d
y
-
1
2
,
j
+
d
y
-
1
2
)
,
n
∈
(
k
-
d
z
-
1
2
,
k
+
d
z
-
1
2
)
.
6. The method of claim 1 , wherein calculating the coherent data of each pixel point in the magnitude image within the preset range from the pixel point comprises:
calculating the magnitude coherent data Q AM (i,j,k) of each pixel point in the magnitude image according to a following formula:
Q
AM
(
i
,
j
,
k
)
=
1
α
AM
FFT
-
1
[
LFilter
(
a
,
b
,
c
)
·
FFT
(
p
^
(
i
,
j
,
k
)
)
]
;
wherein, Q AM (i,j,k) denotes the magnitude coherent data of each pixel point in the magnitude image,
α AM denotes a normalization constant,
FFT −1 denotes an inverse Fourier transform,
LFilter(a,b,c) denotes a filter, the filter is a frequency domain filter, and a cutoff frequency of the filter falls in a preset range,
FFT denotes a Fourier transform, and
|{circumflex over (p)}(i,j,k)| denotes a magnitude of each pixel point in the magnitude image.
7. The method of claim 1 , wherein calculating the coherent data of each pixel point in the phase image within the preset range from the pixel point comprises:
calculating the phase coherent data Q e,f,g VPH of a pixel point (e, f, g) in the phase image according to a following formula:
Q
e
,
f
,
g
VPH
=
1
α
VPH
(
1
-
1
d
x
d
y
d
z
∑
r
=
e
-
d
x
-
1
2
r
=
e
+
d
x
-
1
2
∑
s
=
f
-
d
y
-
1
2
s
=
f
+
d
y
-
1
2
∑
t
=
g
-
d
z
-
1
2
t
=
g
+
d
z
-
1
2
p
^
r
,
s
,
t
p
^
r
,
s
,
t
)
+
1
;
wherein, Q e,f,g VPH denotes the phase coherent data of the pixel point (e, f, g) in the phase image,
α VPH denotes a normalization constant,
d x ,d y ,d z respectively denote values in direction x, direction y and direction z within the preset range, wherein direction z is perpendicular to a plane in which direction x and direction y locate, and
p
^
r
,
s
,
t
p
^
r
,
s
,
t
denotes a phase of a pixel point (r, s, t) in the phase image,
r
∈
(
e
-
d
x
-
1
2
,
e
+
d
x
-
1
2
)
,
s
∈
(
f
-
d
y
-
1
2
,
f
+
d
y
-
1
2
)
,
t
∈
(
g
-
d
z
-
1
2
,
g
+
d
z
-
1
2
)
.
8. The method of claim 1 , wherein calculating the coherent data of each pixel point in the phase image within the preset range from the pixel point comprises:
calculating the phase coherent data Q VPH (e,f,g) of each pixel point in the phase image according to a following formula:
Q
VPH
(
e
,
f
,
g
,
)
=
1
α
VPH
[
1
-
FFT
-
1
(
LFilter
(
u
,
v
,
w
)
·
FFT
(
p
^
(
e
,
f
,
g
)
p
^
(
e
,
f
,
g
)
)
)
]
+
1
;
wherein, Q VPH (e,f,g) denotes the phase coherent data of each pixel point in the phase image,
α VPH denotes a normalization constant,
FFT −1 denotes an inverse Fourier transform,
LFilter(u,v,w) denotes a filter, the filter is a frequency domain filter, and a cutoff frequency of the filter falls in a preset range,
FFT denotes a Fourier transform, and
p
^
(
e
,
f
,
g
)
p
^
(
e
,
f
,
g
)
denotes a phase of each pixel point in the phase image.
9. An apparatus for obtaining an image mask, comprising: a processor and a non-transitory storage medium for storing machine-executable instructions corresponding to control logic for obtaining an image mask, and when executed, the machine-executable instructions causing the processor to:
obtain a magnitude image and a phase image of a to-be-processed image;
obtain magnitude coherent data of each pixel point in the magnitude image by calculating coherent data of each pixel point in the magnitude image within a preset range from the pixel point, and
obtain phase coherent data of each pixel point in the phase image by calculating coherent data of each pixel point in the phase image within the preset range from the pixel point;
obtain a magnitude image mask by performing binarization threshold processing on the magnitude coherent data of each pixel point in the magnitude image, and
obtain a phase image mask by performing binarization threshold processing on the phase coherent data of each pixel point in the phase image; and
obtain an image mask for the to-be-processed image by using the magnitude image mask and the phase image mask.
10. The apparatus of claim 9 , wherein the machine-executable instructions further cause the processor to:
perform homodyne high-pass filtering or phase unwrapping on the phase image.
11. The apparatus of claim 9 , wherein when obtaining the image mask of the to-be-processed image by using the magnitude image mask and the phase image mask, the machine-executable instructions further cause the processor to:
calculate a union or intersection of the magnitude image mask and the phase image mask, as the image mask for the to-be-processed image.
12. The apparatus of claim 9 , wherein when calculating the coherent data of each pixel point in the magnitude image within the preset range from the pixel point, the machine-executable instructions further cause the processor to:
calculate the magnitude coherent data Q i,j,k AM of a pixel point (i, j, k) in the magnitude image according to a following formula:
Q
i
,
j
,
k
AM
=
1
α
AM
d
x
d
y
d
z
∑
l
=
i
-
d
x
-
1
2
l
=
i
+
d
x
-
1
2
∑
m
=
j
-
d
y
-
1
2
m
=
j
+
d
y
-
1
2
∑
n
=
k
-
d
z
-
1
2
n
=
k
+
d
z
-
1
2
p
^
l
,
m
,
n
;
wherein, Q i,j,k AM denotes the magnitude coherent data of the pixel point (i, j, k) in the magnitude image,
α AM denotes a normalization constant,
d x ,d y ,d z respectively denote values in direction x, direction y and direction z within the preset range, wherein direction z is perpendicular to a plane in which direction x and direction y locate in, and
|{circumflex over (p)} l,m,n | denotes a magnitude of a pixel point (l, m, n) in the magnitude image,
l
∈
(
i
-
d
x
-
1
2
,
i
+
d
x
-
1
2
)
,
m
∈
(
j
-
d
y
-
1
2
,
j
+
d
y
-
1
2
)
,
n
∈
(
k
-
d
z
-
1
2
,
k
+
d
z
-
1
2
)
.
13. The apparatus of claim 9 , wherein when calculating the coherent data of each pixel point in the magnitude image within the preset range from the pixel point, the machine-executable instructions further cause the processor to:
calculate the magnitude coherent data Q AM (i,j,k) of each pixel point in the magnitude image according to a following formula:
Q
AM
(
i
,
j
,
k
)
=
1
α
AM
FFT
-
1
[
LFilter
(
a
,
b
,
c
)
·
FFT
(
p
^
(
i
,
j
,
k
)
)
]
;
wherein, Q AM (i,j,k) denotes the magnitude coherent data of each pixel point in the magnitude image,
α AM denotes a normalization constant,
FFT −1 denotes an inverse Fourier transform,
LFilter(a,b,c) denotes a filter, the filter is a frequency domain filter, and a cutoff frequency of the filter falls in a preset range,
FFT denotes a Fourier transform, and
|{circumflex over (p)}(i,j,k)| denotes a magnitude of each pixel point in the magnitude image.
14. The apparatus of claim 9 , wherein when calculating the coherent data of each pixel point in the phase image within the preset range from the pixel point, the machine-executable instructions further cause the processor to:
calculate the phase coherent data Q e,f,g VPH of a pixel point (e, f, g) in the phase image according to a following formula:
Q
e
,
f
,
g
VPH
=
1
α
VPH
(
1
-
1
d
x
d
y
d
z
∑
r
=
e
-
d
x
-
1
2
r
=
e
+
d
x
-
1
2
∑
s
=
f
-
d
y
-
1
2
s
=
f
+
d
y
-
1
2
∑
t
=
g
-
d
z
-
1
2
t
=
g
+
d
z
-
1
2
p
^
r
,
s
,
t
p
^
r
,
s
,
t
)
+
1
;
wherein, Q e,f,g VPH denotes the phase coherent data of the pixel point (e, f, g) in the phase image,
α VPH denotes a normalization constant,
d x ,d y ,d z respectively denote values in direction x, direction y and direction z within the preset range, wherein direction z is perpendicular to a plane in which direction x and direction y locate, and
p
^
r
,
s
,
t
p
^
r
,
s
,
t
denotes a phase of a pixel point (r, s, t) in the phase image, wherein
r
∈
(
e
-
d
x
-
1
2
,
e
+
d
x
-
1
2
)
,
s
∈
(
f
-
d
y
-
1
2
,
f
+
d
y
-
1
2
)
,
t
∈
(
g
-
d
z
-
1
2
,
g
+
d
z
-
1
2
)
.
15. The apparatus of claim 9 , wherein when calculating coherent data of each pixel point in the phase image within the preset range, the machine-executable instructions further cause the processor to:
calculate the phase coherent data Q VPH (e,f,g) of each pixel point in the phase image according to a following formula:
Q
VPH
(
e
,
f
,
g
,
)
=
1
α
VPH
[
1
-
FFT
-
1
(
LFilter
(
u
,
v
,
w
)
·
FFT
(
p
^
(
e
,
f
,
g
)
p
^
(
e
,
f
,
g
)
)
)
]
+
1
;
wherein, Q VPH (e,f,g) denotes the phase coherent data of each pixel point in the phase image,
α VPH denotes a normalization constant,
FFT −1 denotes an inverse Fourier transform,
LFilter(u,v,w) denotes a filter, the filter is a frequency domain filter, and a cutoff frequency of the filter falls in a preset range,
FFT denotes a Fourier transform, and
p
^
(
e
,
f
,
g
)
p
^
(
e
,
f
,
g
)
denotes a phase of each pixel point in the phase image.