Systems and methods for estimating depth and visibility from a reference viewpoint for pixels in a set of images captured from different viewpoints
View Patent ↗Systems in accordance with embodiments of the invention can perform parallax detection and correction in images captured using array cameras. Due to the different viewpoints of the cameras, parallax results in variations in the position of objects within the captured images of the scene. Methods in accordance with embodiments of the invention provide an accurate account of the pixel disparity due to parallax between the different cameras in the array, so that appropriate scene-dependent geometric shifts can be applied to the pixels of the captured images when performing super-resolution processing. In a number of embodiments, generating depth estimates considers the similarity of pixels in multiple spectral channels. In certain embodiments, generating depth estimates involves generating a confidence map indicating the reliability of depth estimates.
1. A camera array, comprising:
a plurality of cameras configured to capture images of a scene from different viewpoints, wherein at least one camera in the plurality of cameras is a monochrome camera and at least one camera in the plurality of cameras includes a Bayer color filter;
a processor; and
memory containing an image processing application;
wherein the image processing application stored in memory directs the processor to:
separately configure the imaging parameters for each of the plurality of cameras;
read out image data from the plurality of cameras including a set of images captured from different viewpoints;
store the image data in the memory;
select a reference viewpoint relative to the viewpoints of the set of images captured from different viewpoints; and
determine depth estimates for pixel locations in an image from the reference viewpoint, where a depth estimate for a given pixel location in the image from the reference viewpoint is determined by:
determining the similarity of pixels that correspond to the given pixel location in the image from the reference viewpoint based upon expected disparity at a plurality of depths, where similarity of pixels across different spectral channels is evaluated based upon similarity of neighborhoods of pixels containing the corresponding pixels; and
selecting the depth from the plurality of depths at which corresponding pixels have the highest degree of similarity as a depth estimate for the given pixel location in the image from the reference viewpoint.
2. The camera array of claim 1 , wherein similarity of pixels across different spectral channels is evaluated using a cost function that evaluates similarity for two neighborhoods of pixels.
3. The camera array of claim 2 , wherein the cost function evaluates similarity of neighborhoods of pixels using cross-correlation between the pixels in the neighborhoods.
4. The camera array of claim 2 , wherein the cost function evaluates similarity of neighborhoods of pixels using normalized cross-correlation between the pixels in the neighborhoods.
5. The camera array of claim 2 , wherein the cost function evaluates similarity of neighborhoods of pixels using a metric that measures the similarity of the relative values of two sets of pixels.
6. The camera array of claim 5 , wherein the measure of the similarity of the relative values of two sets of pixels is an entropy measure.
7. The camera array of claim 2 , further comprising adjusting a weighting applied to a color channel in the cost function based on a map of a signal-to-noise ratio (SNR) generated with respect to the reference image.
8. The camera array of claim 7 , wherein a strong SNR in the reference image is used to reduce the weighting applied to the other color channels.
9. The camera array of claim 8 , wherein generating the SNR comprises estimating SNR along epipolar lines which connects a pixel location (x,y) in an alternative viewpoint to the reference viewpoint.
10. The camera array of claim 9 , further comprising searching the epipolar lines for regions of high SNR.
11. The camera array of claim 2 , wherein the cost function is spatially filters as follows:
Filtered CV ( x,y,d )=Filter x n ,y n εN(x,y) {Cost( x n ,y n ,d )}
where the Filter is applied in a neighborhood N(x,y) surrounding pixel location (x,y).
12. The camera array of claim 2 , wherein the cost function is a weighted aggregated cost function as follows:
FilteredCV
(
x
,
y
,
d
)
=
1
Norm
∑
(
x
1
,
y
1
)
ε
N
(
x
,
y
)
CV
(
x
1
,
y
1
,
d
)
×
wd
(
x
,
y
,
x
1
,
y
1
)
×
wr
(
I
Ref
(
x
,
y
)
-
I
Ref
(
x
1
,
y
1
)
)
where N(x,y) is the immediate neighborhood of the pixel (x,y), which can be square, circular, rectangular, or any other shape appropriate to the requirements of a specific application,
Norm is a normalization term,
I Ref (x,y) is the image data from the reference camera,
wd is a weighting function based on pixel distance, and
wr is a weighting function based on intensity difference.
13. The camera array of claim 1 , further comprising calculating a gradient edge map.
14. The camera array of claim 13 , further comprising using the gradient edge map as a confidence metric within a confidence map.
15. The camera array of claim 1 , wherein the image processing application further directs the processor to select the viewpoint of the Bayer camera as the reference viewpoint.
16. The camera array of claim 1 , wherein the image processing application further directs the processor to generate a depth map using depth estimates for pixel locations in the image from the reference viewpoint.
17. A camera array, comprising:
a plurality of cameras configured to capture images of a scene from different viewpoints, wherein at least one camera in the plurality of cameras is a monochrome camera and at least one camera in the plurality of cameras includes a Bayer color filter;
a processor; and
memory containing an image processing application;
wherein the image processing application stored in memory directs the processor to:
separately configure the imaging parameters for each of the plurality of cameras;
read out image data from the plurality of cameras including a set of images captured from different viewpoints;
store the image data in the memory;
select a viewpoint of a camera including a Bayer color filter as a reference viewpoint;
determine depth estimates for pixel locations in an image from the reference viewpoint, where a depth estimate for a given pixel location in the image from the reference viewpoint is determined by:
determining the similarity of pixels that correspond to the given pixel location in the image from the reference viewpoint based upon expected disparity at a plurality of depths, where similarity of pixels across different spectral channels is evaluated using a cost function that evaluates similarity of neighborhoods of pixels containing the corresponding pixels; and
selecting the depth from the plurality of depths at which corresponding pixels have the highest degree of similarity as a depth estimate for the given pixel location in the image from the reference viewpoint; and
generate a depth map using depth estimates for pixel locations in the image from the reference viewpoint.
18. A method of generating a depth map from images captured by a plurality of cameras that includes a monochrome camera and a camera that includes a Bayer color filter, comprising:
selecting a viewpoint of the camera including a Bayer color filter as a reference viewpoint;
determining depth estimates for pixel locations in an image captured by the camera including the Bayer color filter using a processor configured by an image processing application, where a depth estimate for a given pixel location in the image from the reference viewpoint is determined by:
determining the similarity of pixels that correspond to the given pixel location in the image from the reference viewpoint based upon expected disparity at a plurality of depths, where similarity of pixels across different spectral channels is evaluated using a cost function that evaluates similarity of neighborhoods of pixels containing the corresponding pixels; and
selecting the depth from the plurality of depths at which corresponding pixels have the highest degree of similarity as a depth estimate for the given pixel location in the image from the reference viewpoint; and
generating a depth map from the depth estimates for pixel locations in the image captured by the camera including the Bayer color filter using the processor configured by the image processing application.