# US 2013/0258114 A1 · Digital processing for compensating signals emitted by photosites of a colour sensor

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2013/0258114 A1
- Document ID: US20130258114A1
- Status: Published application
- Kind: A1
- Application number: 13/641,809
- Filed: 2011-04-18
- Published: 2013-10-03
- Earliest claimed: 2010-04-20
- Assignees: Centre National de la Recherche Scientifique (CNRS)
- Inventors: David Alleysson
- Web page: https://patentgrep.com/patent/US20130258114A1

## CPC

- H04N 17/02: for colour television signals
- H04N 25/134: {based on three different wavelength filter elements}
- H04N 2209/047: using multispectral pick-up elements
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

The present invention relates to the acquisition of colour images by a sensor including a plurality of photosites comprising spectral sensitivities that can to be different between photosites. The invention provides for a digital processing of signals emitted by the photosites, which comprises applying a spatial-chromatic transfer function to said signals, taking into account a variety of spectral sensitivities of the photosites. The sensor can then be devoid of a physical matrix of colour filters. Spectral calibration of the sensor is also provided for this purpose, thus comprising the following steps: obtaining spectral sensitivity data of each photosite (S91), building a spatial-chromatic transfer function (S94) taking into account a variety of said spectral sensitivities of the photosites, and applying same to the signals emitted by the photosites in order to compensate for the aforementioned variety of sensitivities.

## Claims

### Claim 1

Method for spectral calibration of a color image sensor, said sensor comprising a plurality of photosites having spectral sensitivities likely to differ between photosites, said method comprising the steps of:
obtaining spectral sensitivity data for each photosite,
constructing a spatial-chromatic transfer function that takes into account a diversity in said spectral sensitivities of the photosites and is to be applied to the signals from the photosites to compensate for said diversity in sensitivities.

### Claim 2 (depends on 1)

Method according to claim 1, for spectral calibration of a color image sensor for images of H×W pixels and comprising H×W photosites, wherein it comprises the steps of:
a) for each sensor photosite of row i and column j, obtaining a spectral calibration function fi,j(λ) giving an intensity for the signal from the photosite i,j as a function of the light wavelength for the photosite,
b) from at least one standard image of dimensions H×W and of known light intensity levels at each pixel for at least K colors,
constructing a first matrix system X comprising:
for each row, n·m matrix elements respectively corresponding to intensities of pixels in the neighborhood of a central pixel of the standard image, said intensities each being modulated by a corresponding calibration function fi,j(λ) for row i and column j of a pixel of the standard image,
for W·H rows,
c) for said standard image, constructing a second matrix system Y comprising:
in each row, the K actual signal intensity levels for said K colors, for each pixel of the standard image,
for W·H rows,
d) and establishing a calibration matrix system H for the transformation between the first matrix system X and the second matrix system Y, such that H·X=Y,
the calibration matrix system H representing said spatial-chromatic transfer function,
in order to obtain, from a first current matrix system X′ constructed by applying step b) to a current image captured by the sensor, an actual distribution of color intensity levels according to a second current matrix system Y′ for each pixel of this current image, and such that H·X′=Y′.

### Claim 3 (depends on 2)

Method according to claim 2, wherein the first and second matrix systems comprise a third dimension specific to an index for a standard image among a plurality of standard images in a database, in order to establish a first matrix X and a second matrix Y averaged over the set of images in the database.

### Claim 4 (depends on 2)

Method according to claim 2, wherein the number of pixels in the neighborhood of a central pixel in step b) is greater than or equal to the number of colors K.

### Claim 5 (depends on 2)

Method according to claim 2, wherein in step a), for each photosite, an intensity value for the signal from the photosite for P successive wavelengths within a chosen spectral range is obtained, a set of P obtained signal intensities forming the spectral calibration function fi,j(λ),
and wherein the second matrix system Y is reduced for a current image of K colors, by averaging together about P/K columns to obtain K columns forming the second matrix system Y and each one corresponding to a color.

### Claim 6 (depends on 2)

Method according to claim 2, wherein the K colors are red, green, blue, with K=3.

### Claim 7 (depends on 2)

Method according to claim 2, wherein the K colors are black, cyan, yellow, and magenta, with K=4.

### Claim 8 (depends on 2)

Method according to claim 2, wherein the dimensions of the neighborhood n·m are greater than or equal to 7×7.

### Claim 9 (depends on 2)

Method according to claim 2, wherein the modulation by the calibration function fi,j(λ) in step b) amounts to calculating, for each matrix element Xi,j in the first matrix system X, an expression of the type:
X
i
,
j
=
∑
λ
=
1
U


R

(
λ
)

I
i
,
j

(
λ
)
,
where:
Ii,j(λ) is an intensity measurement for a signal from the photosite of row i and column j, at wavelength λ,
R(λ) is a spectral sensitivity function,
U is a number of wavelengths chosen within the visible spectral range, at a predetermined interval.

### Claim 10 (depends on 9)

Method according to claim 9, wherein the spectral sensitivity function is expressed by a Gaussian of the type:
R

(
λ
)
=
k


exp

[
-
(
λ
-
μ
σ
)
2
]

### Claim 11 (depends on 9)

Method according to claim 9, wherein the predetermined interval is about p=10 nm, yielding a total number U of about thirty wavelengths.

### Claim 12 (depends on 2)

Computer program, comprising instructions for implementing the method according to claim 2 when it is executed by a processor.

### Claim 13 (depends on 2)

Method for the acquisition of color images by a sensor comprising a plurality of photosites having spectral sensitivities likely to differ between photosites,
wherein it makes use of means for digitally processing signals from the photosites, said digital processing comprising the application to said signals of a spatial-chromatic transfer function taking into account a diversity in the spectral sensitivities of the photosites,
wherein said transfer function is obtained by making use of the calibration method according to claim 2,
and wherein said sensor is without a physical color filter array.

### Claim 14 (depends on 2, 13)

Acquisition method according to claim 13, said sensor being calibrated using the method according to claim 2 and the calibration matrix system H representing said spatial-chromatic transfer function, wherein the acquisition method includes a compensation for diversity in the spectral sensitivities of photosites, comprising the steps of:
a′) from at least one current image captured by the sensor and of dimensions W×H, constructing a first current matrix system X′ comprising:
for each row, n·m matrix elements respectively corresponding to the intensities of pixels within the neighborhood of a central pixel of the current image,
for W×H rows,
b′) applying the calibration matrix system H to the first matrix system X′ in order to obtain a second matrix system Y′, such that H·X′=Y′, the second matrix system Y′ comprising:
in each row, an actual distribution of color intensity levels for each pixel of the current image,
for W×H rows, each corresponding to one of the W×H pixels of the current image.

### Claim 15 (depends on 13)

Computer program, comprising instructions for implementing the method according to claim 13, when it is executed by a processor.

### Claim 16 (depends on 1)

Module for processing digital images from a color image sensor comprising a plurality of photosites, wherein it comprises means for implementing the method according to claim 1.

### Claim 17 (depends on 16)

System comprising a module according to claim 16, and a sensor connected to said module, wherein said sensor is without a physical color filter array.
