Method for bayer mosaic image conversion
First Claim
1. A method for up-interpolating a Bayer mosaic image from input space to output space, said Bayer mosaic image comprising a plurality of four-pixel blocks, each pixel in said blocks being one of three different colors with two of said pixels being a dominant color and two of said pixels being nondominant colors, said method including the steps of:
- reading a two dimensional color plane of said Bayer image for each said color;
mapping said pixels of said dominant color from said input space to said output space by;
multiplying each ordinate of said input space by 1/{square root}2; and
scaling coordinates of said input space to a normalized coefficient kernel by multiplying said ordinates by 1/{square root}2;
mapping said pixels of said non-dominant colors by multiplying each ordinate of said input space by ½
;
for each color, convolving said input space pixels with a coefficient kernel for each color, wherein said coefficient kernel is the same for said two non-dominant colors but different for said dominant color; and
writing said mapped pixels to a storage location.
5 Assignments
0 Petitions
Accused Products
Abstract
A method of, and an apparatus for, image conversion. The three color channels of a Bayer mosaic image are up-interpolated from input space to output space. The pixels sampled from each two-dimensional color plane of the Bayer image are convolved with a coefficient kernel for each color. To facilitate reconstruction and resampling, the color space for a dominant pixel color of the Bayer image is effectively rotated.
The generation of each color value for a given pixel coordinate is carried out in parallel using an identical convolve unit and a dedicated coefficient kernel for each color.
26 Citations
7 Claims
-
1. A method for up-interpolating a Bayer mosaic image from input space to output space, said Bayer mosaic image comprising a plurality of four-pixel blocks, each pixel in said blocks being one of three different colors with two of said pixels being a dominant color and two of said pixels being nondominant colors, said method including the steps of:
-
reading a two dimensional color plane of said Bayer image for each said color;
mapping said pixels of said dominant color from said input space to said output space by;
multiplying each ordinate of said input space by 1/{square root}2; and
scaling coordinates of said input space to a normalized coefficient kernel by multiplying said ordinates by 1/{square root}2;
mapping said pixels of said non-dominant colors by multiplying each ordinate of said input space by ½
;
for each color, convolving said input space pixels with a coefficient kernel for each color, wherein said coefficient kernel is the same for said two non-dominant colors but different for said dominant color; and
writing said mapped pixels to a storage location. - View Dependent Claims (2, 3, 4)
-
-
5. A method of interpolating a Bayer image of red, green and blue pixels from an input space to an output space, said Bayer image having one of the colors being 50% of the pixels, while the other two colors each are 25% of the pixels the method including the steps of:
-
receiving the Bayer image; and
mapping each of the colors to the output space in accordance with the following equations;
x′
=(x/ops)+k1 y′
=(y/ops)+k2where x,y is a coordinate in the output space, x′
y′
is the coordinate in the input space, ops is the number of pixels in the output space per input space sample, and k1 &
k2 are either 0 or −
0.5 depending on the color and the desired relative rotational orientation of the image. - View Dependent Claims (6, 7)
-
Specification