1. A method for generating an enhanced output image, comprising the steps of:
acquiring an ambient image of a scene illuminated by ambient light;
acquiring a flash image of the scene illuminated with a flash unit;
determining a gradient flash image from the flash image;
determining a gradient ambient image from the ambient image;
combining the gradient flash image and the gradient ambient image to produce a combined gradient image;
constructing an enhanced output image from the combined gradient image; in which the flash image is due to a flash radiance \u03a6 scaled by a flash power P, and the ambient image is due to an ambient radiance \u03b1 scaled by an exposure time E, and wherein flash gradients in the gradient flash image are \u2207\u03a6 and ambient gradients in the gradient ambient image are \u2207\u03b1;and
in which a coherence map for the gradient flash image and the gradient ambient image is
M
=
\uf603
\u2207
\u03a6
\xb7
\u2207
\u03b1
\uf604
\uf603
\u2207
\u03a6
\u2062
\uf605
\u2207
\u03b1
\uf604
.
2. The method of claim 1, in which a gradient coherence model is \u2207\u03a6=k\u2207\u03b1, where k is a scalar.
3. The method of claim 1, in which a symbol \u2192 denotes projecting, and the combining of the gradient flash image and the gradient ambient image is
\u2207{tilde over (\u03a6)}=(\u2207\u03a6\u2032\u2192\u2207\u03b1)=\u2207\u03b1(\u2207\u03a6\u2032\xb7\u2207\u03b1)\u2225\u2207\u03b1\u22252,
where the vector \u2207{tilde over (\u03a6)} is along the vector \u2207\u03a6 and k is a scalar value.
4. The method of claim 1, in which the combined gradient image is
\u2207\u03a6*=wS\u2207\u03b1+1(1\u2212wS)(M\u2207\u03a6\u2032+(1\u2212M)\u2207\u03b1,
wherein wS are weights.
5. The method of claim 4, in which the weights wS, are a function of normalized pixel intensities.
6. The method of claim 1, in which a symbol \u2192 denotes projecting, and the combined gradient image is
\u2207\u03b1*=wS\u2207\u03b1\u2032+(1\u2212wS)(\u2207\u03b1\u2032\u2192\u2207\u03a6),
wherein wS are weights.
7. The method of claim 1, in which the constructing further comprises:
integrating gradients in the combined gradient image.
8. The method of claim 1, further comprising:
generating a residual image according to \u2207\u03b7A=(1\u2212wS)\u2207\u03b1\u2032\u2212(\u2207\u03b1\u2032\u2212\u2207\u03a6),
where wS are weights, and \u2207\u03b7A represents noise in the ambient image.
9. The method of claim 1, in which a symbol \u2192 denotes projecting, and further comprising:
adding the gradient flash image and the gradient ambient image to produce an additive gradient image \u2207H, and the combined gradient image is
\u2207\u03a6*=wue\u2207H+(1\u2212wue)(\u2207H\u2192\u2207\u03b1),
where wS are weights.
10. The method of claim 1, in which the combined gradient field is
\u2207\u03a6\u2032=M \u2207\u03a6\u03b2+(1\u2212M)\u2207\u03b1,
where \u03b2 is a linear scaling term.
11. The method of claim 1, in which the flash image is acquired while the scene is also illuminated by the ambient light, and further comprising:
subtracting the ambient image from flash image before determining the gradient flash image.
12. A system for generating an enhanced output image, comprising:
a camera configured to acquire a first image of a scene illuminated by a first illumination condition and a second image of the scene illuminated by a second illumination condition;
means for determining a first gradient image from the first image;
means for determining a second gradient image from the second image;
means for comparing the first gradient image and the second gradient image to produce a combined gradient image;
means for constructing an enhanced output image from the combined gradient image;
wherein the first illumination condition is produced by a flash unit, and the second illumination condition is ambient lighting;
in which the first image is due to a flash radiance \u03a6 scaled by a flash power P, and the second image is due to an ambient radiance \u03b1 scaled by an exposure time E, and wherein flash gradients in the first gradient image are \u2207\u03a6 and ambient gradients in the second gradient image are \u2207\u03b1;and
in which a coherence map for the first gradient image and the second gradient image is
M
=
\uf603
\u2207
\u03a6
\xb7
\u2207
\u03b1
\uf604
\uf603
\u2207
\u03a6
\u2062
\uf605
\u2207
\u03b1
\uf604
.
13. A method for generating an enhanced output image, comprising the steps of:
acquiring a first image of a scene illuminated by a first illumination condition;
acquiring a second image of the scene illuminated by a second illumination condition;
determining a first gradient image from the first image;
determining a second gradient image from the second image;
comparing an orientation of gradients in the first gradient image and the second gradient image to produce a combined gradient image;
constructing an enhanced output image from the combined gradient image;
in which the first illumination condition is produced by a flash unit and the second illumination condition is produced by ambient lighting;
in which the first image is due to a flash radiance \u03a6 scaled by a flash power P, and the second image is due to an ambient radiance \u03b1 scaled by an exposure time E, and wherein flash gradients in the first gradient image are \u2207\u03a6 and ambient gradients in the second gradient image are \u2207\u03b1;and
wherein a coherence map for the first gradient image and the second gradient image is
M
=
\uf603
\u2207
\u03a6
\xb7
\u2207
\u03b1
\uf604
\uf603
\u2207
\u03a6
\u2062
\uf605
\u2207
\u03b1
\uf604
.
14. The method of claim 13, in which a gradient vector for each pixel in the combined gradient image has a magnitude of a gradient vector of a pixel at a corresponding location in the first gradient image and an orientation of a gradient vector of a pixel at a corresponding location in the second gradient image.
15. The method of claim 13, in which the combining further comprises:
projecting, for each pixel in the first gradient image, a gradient vector onto a gradient vector of a corresponding pixel in the second gradient image to obtain a combined gradient vector for a corresponding pixel in the combined gradient image.
The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.
1. A bipolar plate for PEM fuel cells comprising a plastic structure having a surface zone including gas-transport channels, the plastic structure comprising a polymer blend which is filled with conductivity-enhancing carbon fillers and which includes at least two mutually nonmiscible blend polymers,
wherein the polymer blend includes at least one polyamide and at least one polyether ketone or polyether sulfone as blend polymers,
wherein the at least two blend polymers form a co-continuous structure and the carbon fillers are at a higher concentration in one of the blend polymers or in the phase between the blend polymers, or wherein a blend polymer in which the carbon fillers are at a higher concentration forms a continuously conductive matrix in which the at least one further blend polymer is intercalated,
wherein the polymer blend comprises from 45 to 85 wt % of blend polymers and from 15 to 65 wt % of carbon fillers, and
wherein the polymer blend contains as carbon fillers
from 1 to 30 wt % of conductive black,
from 5 to 60 wt % of carbon fibers, and
from 1 to 25 wt % of carbon nanotubes,
in each case based on the total weight of the polymer blend, and
the weight ratio, in the polymer blend, of polyamide to polyether ketonepolyether sulfone is from 1:1.6 to 4:1.
2. The bipolar plate as claimed in claim 1, wherein the carbon fillers are selected from conductive black, graphite, carbon fibers, carbon nanotubes and mixtures thereof.
3. A method of fabricating bipolar plates as claimed in claim 1 by preparing and shaping the polymer blend filled with conductivity-enhancing carbon fillers.
4. A PEM fuel cell comprising bipolar plates as claimed in claim 1.