Methods and apparatuses for double sided dark reference pixel row-wise dark level non-uniformity compensation in image signals
View Patent ↗Methods and apparatuses for row-wise dark level non-uniformity compensation of imaging sensor pixel signals. A column dependent dark reference value is determined as one of a linear and parabolic function of signal values from two areas of dark reference pixels and a column location and then used for dark level non-uniformity compensation of signal values from imaging pixels.
1. A method of adjusting imaging pixel signals produced by pixels in a row of a pixel array, the method comprising:
acquiring first dark signal values from a first area of dark reference pixels in the row;
acquiring second dark signal values from a second area of dark reference pixels in the row;
acquiring pixel signals values from an area of imaging pixels in the row between the first and second areas, wherein each pixel signal has a respective column location;
determining a column dependent dark reference value for a target pixel having a respective acquired pixel signal value, wherein the column dependent dark reference value is based on the acquired first and second dark pixel values and changes in accordance with a column location of the target pixel; and
determining an adjusted pixel signal value for the target pixel as a function of the column dependent dark reference value and the acquired target pixel signal value, wherein the column dependent dark reference value changes in accordance with a linear function based on a column location of the target pixel and wherein the determined adjusted pixel signal value is determined by:
V
′
(
col
)
=
V
(
col
)
-
(
P
0
-
P
1
2
+
P
1
Width
·
col
)
where col is the column location of the target pixel, V′(col) is the determined adjusted pixel signal value for the target pixel, V(col) is the acquired pixel signal value for the target pixel, Width is the total number of columns of active pixels of the row, P 0 is
P
0
=
V
(
dark_left
)
+
V
(
dark_right
)
2
where V(dark_left) is the average of the acquired first dark signal values and V(dark_right) is the average of the acquired second dark signal values and P 1 is
P
1
=
V
(
dark_right
)
-
V
(
dark_left
)
2
.
2. The method of claim 1 , further comprising determining a column dependent dark reference value for a target pixel at each column location having a respective acquired pixel signal value and determining an adjusted pixel signal value for each target pixel.
3. A method of adjusting imaging pixel signals produced by pixels in a row of a pixel array, the method comprising:
acquiring first dark signal values from a first area of dark reference pixels in the row;
acquiring second dark signal values from a second area of dark reference pixels in the row;
acquiring pixel signals values from an area of imaging pixels in the row between the first and second areas, wherein each pixel signal has a respective column location;
determining a column dependent dark reference value for a target pixel having a respective acquired pixel signal value, wherein the column dependent dark reference value is based on the acquired first and second dark pixel values and changes in accordance with a column location of the target pixel; and
determining an adjusted pixel signal value for the target pixel as a function of the column dependent dark reference value and the acquired target pixel signal value, wherein the column dependent dark reference value changes in accordance with a parabolic function based on a column location of the target pixel.
4. The method of claim 3 , wherein the determined adjusted pixel signal value is determined by:
V
′
(
col
)
=
V
(
col
)
-
(
P
0
-
P
1
2
+
P
1
Width
·
col
+
P
1
·
(
P2
·
(
2
·
col
Width
-
1
)
2
-
P
2
)
)
where col is the column location of the target pixel, V′(col) is the determined adjusted pixel signal value for the target pixel, V(col) is the acquired pixel signal value for the target pixel, Width is the total number of columns of active pixels of the row, P 0 is
P
0
=
V
(
dark_left
)
+
V
(
dark_right
)
2
where V(dark_left) is the average of the acquired first dark signal values and V(dark_right) is the average of the acquired second dark signal values and P 1 is
P
1
=
V
(
dark_right
)
-
V
(
dark_left
)
2
and P 2 is a predetermined parabolic parameter that defines a curvature of a parabolic curve fitting for the acquired pixel signal values.
5. The method of claim 4 , wherein P 2 is within the range of approximately 0 to approximately 0.3.
6. An imaging device comprising:
an imaging sensor comprising an array of pixels having dark reference pixels and imaging pixels each having a respective column location, wherein a first area of dark reference pixels is located at one side of the imaging pixels and a second area of dark reference pixels is located at an other side of the imaging pixels; and
a signal processing circuit for adjusting imaging pixel signals using respective column dependent dark reference values derived from the dark reference pixels and the respective column locations of the imaging pixels, wherein the column dependent dark reference values are determined by a function which produces different column dependent dark reference values in accordance with a column location of an imaging pixel, wherein the function is a parabolic function.
7. The imaging device of claim 6 , wherein the dark reference pixels comprise at least one of optical black pixels and tied pixels.
8. The imaging device of claim 6 , wherein the determined respective column dependent dark reference values are determined from a parabolic dark reference value function using a parabolic parameter that defines a curvature of a parabolic curve fitting for the imaging pixel signal values.
9. The imaging device of claim 8 , wherein the parabolic parameter is set to approximately 0.2.