System and method for fusing images
View Patent ↗A system and method for generating a fused image is provided. The system comprises processing circuitry configured to receive a plurality of images, filter each received image using an edge preserving filter, compute a weight for each received image based on the corresponding filtered image and the received image and generate a fused image based on the weights of each received image. The system further comprises a memory device configured to store the fused image.
1. A method for generating a fused image, the method comprising:
receiving a plurality of images;
bi-laterally filtering, including spatial domain filtering and range filtering, each of the plurality of received images;
computing a weight of each pixel in the plurality of received images based on a corresponding bi-laterally filtered image and the plurality of received images by subtracting each pixel in the corresponding bi-lateral filtered image from corresponding pixel in the plurality of received images; and
generating a fused image based on the weight of each pixel in each of the plurality of received images.
2. The method of claim 1 , wherein the weight of each of the plurality of received images is computed on a pixel-by-pixel basis by comparison of pixel values of each of the plurality of received images with a respective filtered value.
3. The method of claim 1 , wherein a sum of weights for the plurality of images is equal to unity.
4. The method of claim 1 , wherein the plurality of images is received at a corresponding plurality of wavelengths.
5. The method of claim 1 , wherein the weight of each pixel in an image I(x, y, λ i ) in a set of ‘M’ images is computed based upon th relationship:
w
i
(
x
,
y
)
=
I
(
x
,
y
,
λ
i
)
-
I
BF
(
x
,
y
,
λ
i
)
+
K
∑
I
=
1
M
(
I
(
x
,
y
,
λ
i
)
-
I
BF
(
x
,
y
,
λ
i
)
+
K
)
wherein the image I BF (x, y, λ i ) is the bi-laterally filtered image, K is a positive real number, and λ i represents an index number of each of the plurality of received images.
6. The method of claim 1 , wherein the plurality of received images represent information of a geographical region.
7. A memory device comprising computer executable instructions that, in response to being executed, perform operations comprising:
receiving a plurality of images;
bi-laterally filtering, including spatial domain filtering and range filtering, each of the plurality of received images;
computing a weight of each pixel in the plurality of received images based on a corresponding bi-laterally filtered image and the plurality of received images by subtracting each pixel value in the corresponding bi-laterally filtered image from a corresponding pixel value in the plurality of received images; and
generating a fused image based on the weight of each pixel in the plurality of received images.
8. The memory device of claim 7 , wherein a sum of weights of a pixel of the plurality of images is equal to unity.
9. The memory device of claim 7 , wherein the plurality of images includes hyperspectral images.
10. The memory device of claim 7 , wherein the weight of an image I(x, y, λ i ) ina set of ‘M’ images is computed based upton the relationship:
w
i
(
x
,
y
)
=
I
(
x
,
y
,
λ
i
)
-
I
BF
(
x
,
y
,
λ
i
)
+
K
∑
I
=
1
M
(
I
(
x
,
y
,
λ
i
)
-
I
BF
(
x
,
y
,
λ
i
)
+
K
)
wherein the image I BF (x, y, λ i ) is the filtered image, K is a positive real number, and λ i represents each received image.
11. A system for generating a fused image, the system comprising:
processing circuitry configured to:
receive a plurality of images;
bi-laterally filter, including spatial domain filtering and range filtering, each received image;
compute a weight for each pixel in the plurality of received images based on a corresponding bi-laterally filtered image and the plurality of received images by subtracting each pixel in the corresponding bi-lateral filtered image from a corresponding pixel in the plurality of received images
generate a fused image based on the weights of each pixel in each of the plurality of received images; and
a memory device configured to store the fused image.
12. The system of claim 11 , wherein a sum of weights for the plurality of received images is equal to unity.
13. The system of claim 11 , wherein the plurality of images represent information of a geographical area.
14. The system of claim 11 , further comprising hyperspectral sensors configured to generate the plurality of images.
15. The system of claim 11 , whererin the processor is configured to compute the weight of an image I(x, y, λ i ) in a set of ‘M’ images is computed based upon the relationship:
w
i
(
x
,
y
)
=
I
(
x
,
y
,
λ
i
)
-
I
BF
(
x
,
y
,
λ
i
)
+
K
∑
I
=
1
M
(
I
(
x
,
y
,
λ
i
)
-
I
BF
(
x
,
y
,
λ
i
)
+
K
)
wherein the image I BF (x, y, λ i ) is a filtered image, K is a positive real number, and λ i represents each received image.
16. The method of claim 1 , wherein the bi-lateral filtering includes measuring spatial closeness of a neighborhood pixel and a surrounding point.
17. The memory device of claim 7 , wherein the bi-lateral filtering includes measuring spatial closeness of a neighborhood pixel and a surrounding point.
18. The system of claim 11 , wherein the processing circuitry is configured to bi-laterally filter each of the plurality of received images by measuring spatial closeness of a neighborhood pixel and a surrounding point.