Fingerprint sensor and correction method thereof
A fingerprint sensor and a correction method thereof are provided. The fingerprint sensor includes a sensing part to be touched by a finger, including sensing units, a processing unit, and a correction unit; the sensing unit is configured to sense, when correction starts, a fingerprint signal and output first data; the processing unit is configured to calculate, according to a mathematical model, a new correction coefficient of the corresponding sensing unit by the first data; and the correction unit is configured to correct, according to the new correction coefficient, second data output by the corresponding sensing unit. The fingerprint sensor can automatically modify a correction coefficient when a physical characteristic of the fingerprint sensor changes, thereby reducing the product return rate and prolonging the product service life, omitting a step of correction before delivery, and reducing the manufacturing difficulty and the manufacturing costs.
1. A fingerprint sensor, comprising a sensing part to be touched by a finger, the sensing part comprising a plurality of sensing units, wherein the fingerprint sensor further comprises a processing unit and a correction unit, wherein:
the sensing unit is configured to sense, when correction starts, a fingerprint signal and output first data;
the processing unit is configured to calculate, according to a mathematical model, a new correction coefficient of the corresponding sensing unit by using the first data; and
the correction unit is configured to correct, according to the new correction coefficient, second data output by the corresponding sensing unit;
wherein the processing unit comprises a first calculation subunit, configured to calculate, according to the mathematical model, a new correction coefficient of the corresponding sensing unit by using the first data and the old correction coefficient of the corresponding sensing unit,
k
new
=
(
1
-
σ
)
k
old
+
σ
y
F
(
Y
)
b
new
=
(
1
-
σ
)
b
old
+
σ
[
y
-
Avg
(
Y
)
]
wherein k new and b new represent new correction coefficients of the sensing unit; k old and b old represent old correction coefficients of the sensing unit; y represent the first data output by the sensing unit; Y represents an image constituted by first data output by all the sensing units; F(Y) represents data of the corresponding sensing units after performing median filtering on Y; Avg(Y) represents data of the corresponding sensing units after calculating an average of Y; σ represents a learning factor, wherein a range of σ is [0,1] and
σ
=
x
old
-
F
(
Y
)
x
old
+
F
(
Y
)
;
and x old represents data of the sensing unit obtained after correction process by means of the old correction coefficients.
2. The fingerprint sensor according to claim 1 , further comprising a storage unit, configured to store a correction coefficient, wherein if an old correction coefficient is stored in the fingerprint sensor, the old correction coefficient is replaced with the new correction coefficient, or the old correction coefficient is replaced with a weighted sum of the new correction coefficient and the old correction coefficient.
3. The fingerprint sensor according to claim 2 , further comprising a determining unit, configured to determine whether a correction condition is satisfied, wherein the correction condition comprises one or more of the following: the fingerprint sensor is enabled for the first time, a preset correction time is reached, a correction instruction is received, or an effect of correction by means of the old correction coefficient does not reach a designated requirement.
4. The fingerprint sensor according to claim 3 , wherein the effect of correction by means of the old correction coefficient is evaluated by using an image smoothness or a root-mean-square error of data that is output by all the sensing units and corrected by means of the old correction coefficient, and if the an image smoothness or the root-mean-square error is greater than a preset threshold, the designated requirement is not reached.
5. The fingerprint sensor according to claim 1 , wherein the correction unit is specifically configured to:
calculate
x
new
=
y
′
-
b
new
k
new
,
wherein k new and b new represent new correction coefficients of the sensing unit, y′ represents the second data output by the sensing unit, and x new represents data of the sensing unit after correction by means of the new correction coefficients.
6. A correction method for a fingerprint sensor, the fingerprint sensor comprising a sensing part to be touched by a finger, the sensing part comprising a plurality of sensing units, wherein the method comprises steps of:
S 10 : when correction starts, sensing a fingerprint signal and obtaining first data output by each sensing unit;
S 20 : calculating, according to a mathematical model, a new correction coefficient of the corresponding sensing unit by using the first data; and
S 30 : correcting, according to the new correction coefficient, second data output by the corresponding sensing unit;
wherein step S 20 specifically comprises: calculating, according to the mathematical model, a new correction coefficient of the corresponding sensing unit by using the first data and the old correction coefficient of the corresponding sensing unit,
k
new
=
(
1
-
σ
)
k
old
+
σ
y
F
(
Y
)
b
new
=
(
1
-
σ
)
b
old
+
σ
[
y
-
Avg
(
Y
)
]
wherein k new and b new represent new correction coefficients of the sensing unit; k old and b old represent old correction coefficients of the sensing unit; y represents the first data output by the sensing unit; Y represents an image constituted by first data output by all the sensing units; F(Y) represents data of the corresponding sensing units after performing median filtering on Y; Avg(Y) represents data of the corresponding sensing units after calculating an average of Y; σ represents a learning factor, wherein a range of σ is [0,1] and
σ
=
x
old
-
F
(
Y
)
x
old
+
F
(
Y
)
;
and x old represents data of the sensing unit obtained after correction process by means of the old correction coefficients.
7. The correction method according to claim 6 , wherein after step S 20 but before step S 30 , the method further comprises:
if an old correction coefficient is stored in the fingerprint sensor, replacing the old correction coefficient with the new correction coefficient, or with a weighted sum of the new correction coefficient and the old correction coefficient.
8. The correction method according to claim 7 , wherein before step S 10 , the method further comprises:
determining whether a correction condition is satisfied, and executing S 10 if the correction condition is satisfied, wherein the correction condition comprises: the fingerprint sensor is enabled for the first time, a preset correction time is reached, a correction instruction is received, or an effect of correction by means of the old correction coefficient does not reach a designated requirement.
9. The correction method according to claim 8 , wherein the effect of correction by means of the old correction coefficient is evaluated by using an image smoothness or a root-mean-square error of data that is output by all the sensing units and corrected by means of the old correction coefficient, and if the an image smoothness or the root-mean-square error is greater than a preset threshold, the designated requirement is not reached.
10. The correction method according to claim 6 , wherein step S 30 specifically comprises:
correcting the second data according to a formula
x
new
=
y
′
-
b
new
k
new
,
wherein k new and b new represent new correction coefficients of the sensing unit, y′ represent the second data output by the sensing unit, and x new represents data of the sensing unit after correction by means of the new correction coefficients.