CN111462011A - Method and system for removing infrared polarization angle image noise - Google Patents

Method and system for removing infrared polarization angle image noise Download PDF

Info

Publication number
CN111462011A
CN111462011A CN202010250993.3A CN202010250993A CN111462011A CN 111462011 A CN111462011 A CN 111462011A CN 202010250993 A CN202010250993 A CN 202010250993A CN 111462011 A CN111462011 A CN 111462011A
Authority
CN
China
Prior art keywords
equal
frame
polarization angle
image
threshold2
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010250993.3A
Other languages
Chinese (zh)
Other versions
CN111462011B (en
Inventor
赵嘉学
卢云龙
米冠宇
王五一
周云
吕坚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202010250993.3A priority Critical patent/CN111462011B/en
Publication of CN111462011A publication Critical patent/CN111462011A/en
Application granted granted Critical
Publication of CN111462011B publication Critical patent/CN111462011B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G06T5/70—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration by the use of local operators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10048—Infrared image

Abstract

The invention discloses a method for removing infrared polarization angle image noise, which comprises the following steps: acquiring infrared polarization information of a target; s1 is generated by performing interframe noise reduction processing1_p1(i, j, k); s2 is generated by performing interframe noise reduction processing2_p1(i, j, k); to S1_p1(i, j, k) conditional filtering to generate S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering to generate S2_p2(i, j, k); according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image. The invention also discloses a system for removing the infrared polarization angle image noise. According to the method and the system for removing the infrared polarization angle image noise, the S1 and S2 images are processed, so that the AOP image noise is obviously reduced, the imaging quality is greatly improved, more original detail information of the images is reserved, the AOP polarization angle image can show more target information, and the identification of the target is improvedThe performance of the infrared polarization detection system is improved due to the difference of the tracking capacity.

Description

Method and system for removing infrared polarization angle image noise
Technical Field
The invention relates to an infrared polarization detection technology, in particular to a method and a system for removing infrared polarization angle image noise.
Background
The traditional infrared detector can only acquire the intensity information of the target, and the infrared polarization detection technology can calculate the intensity, the polarization degree and the polarization angle information of the target according to the acquired information, so that the infrared polarization detector can acquire more detailed information of the target, and the effect is better.
The infrared polarization information of the target can be obtained by adding a polarizing film in a light path of the infrared detection pixel for receiving the target radiation, and under a general condition, the infrared polarization imaging system needs to obtain light intensity information of 4 polarization angles including 0 degree, 45 degrees, 90 degrees and 135 degrees.
For polarized light, the polarization information of the target is generally represented by the stokes vector:
Figure BDA0002435481360000011
in the formula ,I0、I45、I90 and I135Respectively represent light intensities with polarization directions of 0 °, 45 °, 90 ° and 135 °; i isR and ILRespectively a right-handed polarized light component and a left-handed polarized light component; s0 represents the total light intensity; s1 represents the intensity difference between 0 ° and 90 ° in the polarization direction; s2 represents the intensity difference between 45 ° and 135 ° for the polarization direction; s3 represents the difference in intensity of the left-and right-hand circularly polarized components of light.
In practical cases, the circularly polarized light component is generally small, and therefore S3 is generally regarded as 0. So that the degree of polarization of the polarized light is obtained from the stokes vector as:
Figure BDA0002435481360000012
the polarization angle is:
Figure BDA0002435481360000013
by utilizing a calculation formula of the Stokes vector, the infrared polarization information of the target can be calculated according to the acquired information of the 4 polarization angles.
However, in practical applications, the calculated polarization angle AOP has poor image quality and serious noise, which affects the imaging quality. When the polarization angle image is directly filtered, the denoising effect is not good, and therefore an effective method for removing the polarization angle noise is urgently needed.
Disclosure of Invention
The invention aims to solve the technical problem that the denoising effect is not good when the polarization angle image is directly filtered in the prior art, and aims to provide a method and a system for removing infrared polarization angle image noise to solve the problem.
The invention is realized by the following technical scheme:
a method for removing infrared polarization angle image noise comprises the following steps: step 1: acquiring infrared polarization information of a target; the infrared polarization information comprises an intensity difference S1 of which the polarization direction is between 0 and 90 degrees and an intensity difference S2 of which the polarization direction is between 45 and 135 degrees; step 2: performing inter-frame noise reduction processing on S1 to generate S1 first processed data S1_p1(i, j, k); performing inter-frame noise reduction processing on S2 to generate S2 first processed data S2_p1(i, j, k); and step 3: to S1_p1(i, j, k) conditional filtering is performed to generate S1 second processed data S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering is performed to generate S2 second processed data S2_p2(i, j, k); and 4, step 4: according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image.
When the invention is applied, the inventor firstly analyzes the reason of the poor quality of the polarization angle image:
suppose S1 and S2All have errors of (S)1) and ε(S2) Then according to the error transfer formula, finallyThe absolute error of the polarization angle AOP is obtained as
Figure BDA0002435481360000021
Therefore, when S is1 and S2When all values of (A) are small, S1 and S2May result in large AOP polarization angle errors. The following examples illustrate: e.g. S1 and S2The exact value of (2) is 3, but due to error, the value may change, assuming that 3 cases occur:
(1)S1becomes 1, S2The value of (d) becomes 6.
At this time, AOP is 1/2 × arctan (6/1) is 40.27 °.
(2)S1Has a constant value of 3, S2The value of (d) is constant and is 3.
At this time, AOP is 1/2 × arctan (3/3) is 22.5 °.
(3)S1Becomes 6, S2The value of (b) becomes 1.
At this time, AOP is 1/2 × arctan (1/6) is 4.73 °.
Can see S1 and S2Since a small error in the polarization angle causes a large error in the polarization angle, it is necessary to deal with the polarization angle noise according to the cause of the polarization angle noise.
Based on the above theoretical basis, the following technical scheme is applied:
the noise of the AOP polarization angle image is subjected to S1 and S2Therefore, the effect of directly performing filtering and noise reduction processing on the AOP polarization angle image is not good. Invention pair S1 and S2Filtering to obtain better S1 and S2And (4) data are obtained, so that an AOP polarization angle image with better quality is obtained through calculation, and the purpose of noise reduction is achieved.
First, the present invention obtains infrared polarization information of a target, in which circularly polarized light components are generally small, and therefore S3 is generally regarded as 0, so that the processing object of the present invention is S1 and S2. Inter-frame noise reduction is first performed for S1 and S2, by which inter-frame noise reductionNoise, errors can be effectively reduced; and then, performing conditional filtering on the data subjected to the inter-frame noise reduction, and reserving details in the data, so that the integrity of the data during later-period synthesis is ensured, and finally, performing data synthesis. According to the invention, the AOP polarization angle images are not directly filtered, but processed by S1 and S2 images, the AOP image noise is obviously reduced, and the imaging quality is greatly improved. The denoising method provided by the invention reserves more original detail information of the image, so that the AOP polarization angle image can show more target information, the target identification and tracking capability is improved, and the performance of the infrared polarization detection system is improved.
Further, step 1 comprises the following substeps:
by additionally arranging the polaroid in the light path of the infrared detection pixel for receiving the target radiation, the infrared polarization information of the target can be acquired.
Further, step 2 comprises the following substeps:
acquiring S1 first processing data S according to the following formula1_p1(i，j，k)：
When k is 1, S1_p1(i，j，k)＝S1(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000031
wherein Diff (i, j, k) ═ S1(i，j，k)-S1(i，j，k-1)|
in the formula ,S1_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s1(i, j, k) is data of ith row and jth column of the kth frame of the S1 image;
acquiring S2 first processing data S according to the following formula2_p1(i，j,k)：
When k is 1, S2_p1(i,j,k)＝S2(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000032
wherein Diff (i, j, k) ═ S2(i,j，k)-S2(i，j，k-1)|
in the formula ,S2_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s2(i, j, k) is data of the ith row and the jth column of the kth frame of the S2 image.
When the invention is applied, when the difference absolute value Diff (i, j, k) is greater than the threshold value, the target moves, so that the data of the current frame is directly used; when the absolute value of the difference is smaller than or equal to the threshold value, the target change is proved to be small, and the average processing is carried out with the previous frame image at the moment, so that the error can be effectively reduced.
Further, step 3 comprises the following substeps:
acquiring S1 second processed data S according to the following formula1_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S1_p2(i,j,k)＝S1_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000041
wherein ,
Figure BDA0002435481360000042
Figure BDA0002435481360000043
in the formula ,S1_p2(i, j, k) is the result after filtering; w (p, q) is weekWeights of 9 pixels in a region of 3 × 3 (p and q are integers, p is more than or equal to 1 and less than or equal to 3, q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold;
acquiring S2 second processed data S according to the following formula2_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S2_p2(i,j,k)＝S2_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000044
wherein ,
Figure BDA0002435481360000045
Figure BDA0002435481360000046
in the formula ,S2_p2And w (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, and q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold.
When the invention is applied, when | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold2, at which time S1_p1(i, j, k) and S2_p1Since the error of (i, j, k) does not greatly affect the calculation result of the polarization angle, S is directly used1_p1(i, j, k) and S2_p1(i, j, k), no processing; when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2, when S is present1_p1(i, j, k) and S2_p1The absolute values of (i, j, k) are all small, S1_p1(i, j, k) and S2_p1Small errors in (i, j, k) also have a large effect on the AOP polarization angle calculation, and thus on S at this time1_p1(i, j, k) and S2_p1(i, j, k) conditional mean filtering. With S1_p1(i, j, k) for example, for S1_p1Calculating weight values at 9 points in the 3 × 3 area around (i, j, k), and if a certain point and S1_p1If the difference value of (i, j, k) is greater than or equal to the threshold value threshold3, the point is considered as detail information of the image and does not participate in the S pair1_p1(i, j, k) is weighted, the weight is 0, otherwise, if the point is S1_p1If the difference (i, j, k) is less than the threshold3, the point is considered to belong to S1_p1(i, j, k) in the smoothing region, participate in the pair S1_p1And (i, j, k) weighting operation, wherein the weight is 1. By weighted average calculation, S can be obtained1_p2(i, j, k). Similarly, S can be calculated2_p2(i,j,k)。
Further, step 4 comprises the following substeps:
obtaining an AOP polarization angle image according to the following formula:
Figure BDA0002435481360000051
in the formula, AOP (i, j, k) is the calculated AOP polarization angle image.
When the invention is applied, because S1_p2(i, j, k) may be 0, in which case S1_p2(i, j, k) cannot be calculated as denominator, and is therefore directly based on S2_p2The positive or negative of (i, j, k) is determined. When S is2_p2When (i, j, k) is not negative, see S2_p2(i,j,k)/S1_p2(i, j, k) is infinite, and AOP (i, j, k) is 45 deg. In the same way, when S2_p2When (i, j, k) is negative, see S2_p2(i,j,k)/S1_p2(i, j, k) is minus infinity, and AOP (i, j, k) is-45 deg. When S is1_p2When (i, j, k) ≠ 0, it can be calculated according to the formula.
A system for removing infrared polarization angle image noise, comprising:
an acquisition unit: the infrared polarization information acquisition device is used for acquiring infrared polarization information of a target; the infrared polarization information comprises an intensity difference S1 of which the polarization direction is between 0 and 90 degrees and an intensity difference S2 of which the polarization direction is between 45 and 135 degrees;
inter-frame dropA noise unit: for inter-frame noise reduction processing of S1 to generate S1 first processed data S1_p1(i, j, k); performing inter-frame noise reduction processing on S2 to generate S2 first processed data S2_p1(i,j,k)；
A conditional filtering unit: for pair S1_p1(i, j, k) conditional filtering is performed to generate S1 second processed data S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering is performed to generate S2 second processed data S2_p2(i,j,k)；
A synthesis unit: for according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image.
Furthermore, the acquiring unit can acquire the infrared polarization information of the target by additionally arranging a polarizing film in a light path of the infrared detection pixel for receiving the target radiation.
Further, the inter-frame noise reduction unit acquires S1 the first processed data S according to the following equation1_p1(i,j,k)：
When k is 1, S1_p1(i,j,k)＝S1(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000052
wherein Diff (i, j, k) ═ S1(i,j,k)-S1(i,j,k-1)|
in the formula ,S1_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s1(i, j, k) is data of ith row and jth column of the kth frame of the S1 image;
the inter-frame noise reduction unit acquires S2 first processing data S according to the following formula2_p1(i,j,k)：
When k is 1, S2_p1(i,j,k)＝S2(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000061
wherein Diff (i, j, k) ═ S2(i,j,k)-S2(i,j,k-1)|
in the formula ,S2_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s2(i, j, k) is data of the ith row and the jth column of the kth frame of the S2 image.
Further, the conditional filtering unit acquires S1 second processed data S according to the following equation1_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S1_p2(i,j,k)＝S1_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000062
wherein ,
Figure BDA0002435481360000063
Figure BDA0002435481360000064
in the formula ,S1_p2W (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1, and q is less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold;
the conditional filtering unit acquires S2 second processed data S according to the following equation2_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S2_p2(i,j,k)＝S2_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000071
wherein ,
Figure BDA0002435481360000072
Figure BDA0002435481360000073
in the formula ,S2_p2And w (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, and q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold.
Further, the synthesizing unit acquires an AOP polarization angle image according to the following formula:
Figure BDA0002435481360000074
in the formula, AOP (i, j, k) is the calculated AOP polarization angle image.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the method and the system for removing the infrared polarization angle image noise, the S1 and S2 images are processed, so that the AOP image noise is obviously reduced, the imaging quality is greatly improved, more original detail information of the images is reserved, the AOP polarization angle image can show more target information, the target identification and tracking capacity is improved, and the performance of an infrared polarization detection system is improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
FIG. 1 is a schematic diagram of the process steps of the present invention;
FIG. 2 is an unfiltered raw AOP polarization angle image according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating the filtering effect of directly filtering an AOP polarization angle image according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating the processing effect of the method according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention.
Examples
As shown in fig. 1, the method for removing infrared polarization angle image noise of the present invention includes the following steps: step 1: acquiring infrared polarization information of a target; the infrared polarization information comprises an intensity difference S1 of which the polarization direction is between 0 and 90 degrees and an intensity difference S2 of which the polarization direction is between 45 and 135 degrees; step 2: performing inter-frame noise reduction processing on S1 to generate S1 first processed data S1_p1(i, j, k); performing inter-frame noise reduction processing on S2 to generate S2 first processed data S2_p1(i, j, k); and step 3: to S1_p1(i, j, k) conditional filtering is performed to generate S1 second processed data S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering is performed to generate S2 second processed data S2_p2(i, j, k); and 4, step 4: according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image.
In the implementation of the present embodiment, the inventors first analyzed the reason for the poor polarization angle image quality:
suppose S1 and S2All have errors of (S)1) and ε(S2) Then, according to the error transfer formula, the absolute error of the polarization angle AOP is finally obtained as
Figure BDA0002435481360000081
Therefore, when S is1 and S2When all values of (A) are small, S1 and S2May result in large AOP polarization angle errors. The following examples illustrate: e.g. S1 and S2The exact value of (2) is 3, but due to error, the value may change, assuming that 3 cases occur:
(1)S1becomes 1, S2The value of (d) becomes 6.
At this time, AOP is 1/2 × arctan (6/1) is 40.27 °.
(2)S1Has a constant value of 3, S2The value of (d) is constant and is 3.
At this time, AOP is 1/2 × arctan (3/3) is 22.5 °.
(3)S1Becomes 6, S2The value of (b) becomes 1.
At this time, AOP is 1/2 × arctan (1/6) is 4.73 °.
Can see S1 and S2Since a small error in the polarization angle causes a large error in the polarization angle, it is necessary to deal with the polarization angle noise according to the cause of the polarization angle noise.
Based on the above theoretical basis, the following technical scheme is applied:
the noise of the AOP polarization angle image is subjected to S1 and S2Therefore, the effect of directly performing filtering and noise reduction processing on the AOP polarization angle image is not good. Invention pair S1 and S2Filtering to obtain better S1 and S2And (4) data are obtained, so that an AOP polarization angle image with better quality is obtained through calculation, and the purpose of noise reduction is achieved.
First, the present invention obtains infrared polarization information of a target, in which circularly polarized light components are generally small, and therefore S3 is generally regarded as 0, so that the processing object of the present invention is S1 and S2. Inter-frame noise reduction is firstly carried out on S1 and S2, and errors can be effectively reduced through the inter-frame noise reduction; and then, performing conditional filtering on the data subjected to the inter-frame noise reduction, and reserving details in the data so as to ensure the data during later synthesisAnd (4) integrity, and finally data synthesis. According to the invention, the AOP polarization angle images are not directly filtered, but processed by S1 and S2 images, the AOP image noise is obviously reduced, and the imaging quality is greatly improved. The denoising method provided by the invention reserves more original detail information of the image, so that the AOP polarization angle image can show more target information, the target identification and tracking capability is improved, and the performance of the infrared polarization detection system is improved.
To further illustrate the working process of the present embodiment, step 1 includes the following sub-steps:
by additionally arranging the polaroid in the light path of the infrared detection pixel for receiving the target radiation, the infrared polarization information of the target can be acquired.
To further illustrate the working process of the present embodiment, step 2 includes the following sub-steps:
acquiring S1 first processing data S according to the following formula1_p1(i,j,k)：
When k is 1, S1_p1(i,j,k)＝S1(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000091
wherein Diff (i, j, k) ═ S1(i,j,k)-S1(i,j,k-1)|
in the formula ,S1_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s1(i, j, k) is data of ith row and jth column of the kth frame of the S1 image;
acquiring S2 first processing data S according to the following formula2_p1(i,j,k)：
When k is 1, S2_p1(i,j,k)＝S2(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000092
wherein Diff (i, j, k) ═ S2(i,j,k)-S2(i,j,k-1)|
in the formula ,S2_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s2(i, j, k) is data of the ith row and the jth column of the kth frame of the S2 image.
In the implementation of this embodiment, when the absolute difference value Diff (i, j, k) is greater than the threshold, it indicates that the object has moved, so the data of the current frame is directly used; when the absolute value of the difference is smaller than or equal to the threshold value, the target change is proved to be small, and the average processing is carried out with the previous frame image at the moment, so that the error can be effectively reduced.
To further illustrate the working process of the present embodiment, step 3 includes the following sub-steps:
acquiring S1 second processed data S according to the following formula1_p2(i，j，k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S1_p2(i,j,k)＝S1_p1(i，j，k)
when | S1_p1(i，j,k)|<threshold2 and | S2_p1(i，j，k)|<threshold2 time:
Figure BDA0002435481360000101
wherein ,
Figure BDA0002435481360000102
Figure BDA0002435481360000103
in the formula ,S1_p2(i, j, k) is the result after filtering, w (p, q) is the weight (p sum) of 9 pixels in the surrounding 3 × 3 areaq is an integer, p is more than or equal to 1 and less than or equal to 3, q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold;
acquiring S2 second processed data S according to the following formula2_p2(i，j，k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S2_p2(i,j，k)＝S2_p1(i，j，k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000104
wherein ,
Figure BDA0002435481360000105
Figure BDA0002435481360000106
in the formula ,S2_p2And w (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, and q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold.
When this embodiment is implemented, when | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold2, at which time S1_p1(i, j, k) and S2_p1Since the error of (i, j, k) does not greatly affect the calculation result of the polarization angle, S is directly used1_p1(i, j, k) and S2_p1(i, j, k), no processing; when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2, when S is present1_p1(i, j, k) and S2_p1The absolute values of (i, j, k) are all small, S1_p1(i, j, k) and S2_p1Small errors in (i, j, k) also have a large effect on the AOP polarization angle calculation, and thus on S at this time1_p1(iJ, k) and S2_p1(i, j, k) conditional mean filtering. With S1_p1(i, j, k) for example, for S1_p1Calculating weight values at 9 points in the 3 × 3 area around (i, j, k), and if a certain point and S1_p1If the difference value of (i, j, k) is greater than or equal to the threshold value threshold3, the point is considered as detail information of the image and does not participate in the S pair1_p1(i, j, k) is weighted, the weight is 0, otherwise, if the point is S1_p1If the difference (i, j, k) is less than the threshold3, the point is considered to belong to S1_p1(i, j, k) in the smoothing region, participate in the pair S1_p1And (i, j, k) weighting operation, wherein the weight is 1. By weighted average calculation, S can be obtained1_p2(i, j, k). Similarly, S can be calculated2_p2(i,j,k)。
To further illustrate the working process of the present embodiment, step 4 includes the following sub-steps:
obtaining an AOP polarization angle image according to the following formula:
Figure BDA0002435481360000111
in the formula, AOP (i, j, k) is the calculated AOP polarization angle image.
This example is implemented because S1_p2(i, j, k) may be 0, in which case S1_p2(i, j, k) cannot be calculated as denominator, and is therefore directly based on S2_p2The positive or negative of (i, j, k) is determined. When S is2_p2When (i, j, k) is not negative, see S2_p2(i,j,k)/S1_p2(i, j, k) is infinite, and AOP (i, j, k) is 45 deg. In the same way, when S2_p2When (i, j, k) is negative, see S2_p2(i,j,k)/S1_p2(i, j, k) is minus infinity, and AOP (i, j, k) is-45 deg. When S is1_p2When (i, j, k) ≠ 0, it can be calculated according to the formula.
A system for removing infrared polarization angle image noise, comprising:
an acquisition unit: the infrared polarization information acquisition device is used for acquiring infrared polarization information of a target; the infrared polarization information comprises an intensity difference S1 of which the polarization direction is between 0 and 90 degrees and an intensity difference S2 of which the polarization direction is between 45 and 135 degrees;
an inter-frame noise reduction unit: for inter-frame noise reduction processing of S1 to generate S1 first processed data S1_p1(i, j, k); performing inter-frame noise reduction processing on S2 to generate S2 first processed data S2_p1(i,j,k)；
A conditional filtering unit: for pair S1_p1(i, j, k) conditional filtering is performed to generate S1 second processed data S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering is performed to generate S2 second processed data S2_p2(i,j,k)；
A synthesis unit: for according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image.
To further illustrate the operation of this embodiment, the obtaining unit may obtain the infrared polarization information of the target by adding a polarizer to a light path of the infrared detection pixel receiving the target radiation.
To further explain the operation of the present embodiment, the inter-frame noise reduction unit acquires S1 the first processed data S according to the following equation1_p1(i,j,k)：
When k is 1, S1_p1(i,j,k)＝S1(i,j,k)；
When k is more than or equal to 2,
Figure BDA0002435481360000112
wherein Diff (i, j, k) ═ S1(i,j,k)-S1(i,j,k-1)|
in the formula ,S1_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s1(i, j, k) is data of ith row and jth column of the kth frame of the S1 image;
the inter-frame noise reduction unit acquires S2 first processing data S according to the following formula2_p1(i,j,k)：
When k is 1, S2_p1(i,j,k)＝S2(i，j，k)；
When k is more than or equal to 2,
Figure BDA0002435481360000121
wherein Diff (i, j, k) ═ S2(i,j，k)-S2(i，j，k-1)|
in the formula ,S2_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s2(i, j, k) is data of the ith row and the jth column of the kth frame of the S2 image.
To further explain the operation of the present embodiment, the conditional filtering unit acquires S1 the second processed data S according to the following equation1_p2(i，j，k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S1_p2(i，j，k)＝S1_p1(i，j，k)
when | S1_p1(i，j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000122
wherein ,
Figure BDA0002435481360000123
Figure BDA0002435481360000124
in the formula ,S1_p2W (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold;
the conditional filtering unit obtains S2 a second processed number according to the following equationAccording to S2_p2(i，j，k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S2_p2(i,j,k)＝S2_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure BDA0002435481360000131
wherein ,
Figure BDA0002435481360000132
Figure BDA0002435481360000133
in the formula ,S2_p2And w (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, and q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold.
To further illustrate the operation of this embodiment, the synthesis unit acquires an AOP polarization angle image according to the following formula:
Figure BDA0002435481360000134
in the formula, AOP (i, j, k) is the calculated AOP polarization angle image.
In order to further explain the working process of this embodiment, according to the method for removing AOP polarization angle image noise proposed by the present invention, a method for performing denoising processing on the 3 rd frame AOP polarization angle image is specifically provided. The following were used:
(1) first, S of a first frame is obtained1 and S2Since an image has only 1 frame image with k equal to 1, the inter-frame noise reduction result is S1_p1(i,j,1)＝S1(i,j,1)，S2_p1(i,j,1)＝S2(i, j, 1). At this time, it is aligned with S1_p1(i, j,1) and S2_p1(i, j,1) conditional filtering the image, and calculating a filtered result S according to a formula1_p2(i, j,1) and S2_p2And (i, j,1), and then calculating the corresponding AOP polarization angle image to obtain the AOP (i, j, 1). The denoised first frame AOP polarization angle image can be output.
(2) Reacquiring S of a second frame1 and S2And in the image, k is 2, and k-1 is 1. Therefore, the method for reducing noise between frames is as follows: according to S1(i, j,2) and S1(i, j,1) calculating S1_p1(i, j,2) according to S2(i, j,2) and S2(i, j,1) calculating S2_p1(i, j, 2). At this time, it is aligned with S1_p1(i, j,2) and S2_p1(i, j,2) conditional filtering the image, and calculating a filtered result S according to a formula1_p2(i, j,2) and S2_p2And (i, j,2), and then calculating the corresponding AOP polarization angle image to obtain the AOP (i, j, 2). And outputting the denoised second frame AOP polarization angle image.
(3) S of third frame is acquired again1 and S2And in the image, k is 3, and k-1 is 2. Therefore, the method for reducing noise between frames is as follows: according to S1(i, j,3) and S1(i, j,2) calculating S1_p1(i, j,3) according to S2(i, j,3) and S2(i, j,2) calculating S2_p1(i, j, 3). At this time, it is aligned with S1_p1(i, j,3) and S2_p1(i, j,3) conditional filtering the image, and calculating a filtered result S according to a formula1_p2(i, j,3) and S2_p2And (i, j,3) and calculating the corresponding AOP polarization angle image to obtain the AOP (i, j, 3). And outputting the denoised third frame AOP polarization angle image.
For further explanation of the working process of this embodiment, it can be seen from fig. 2 to fig. 4 that the original AOP polarization angle image in fig. 2 is very noisy and has poor image quality. Fig. 3 shows the result of directly performing filtering processing on the AOP polarization angle image, and due to the filtering, the image loses much detail information, but the image still has much noise and the denoising effect is not good. Fig. 4 is a processing effect diagram of the method of the present invention, and it can be seen that the noise of the AOP polarization angle image is greatly reduced and more detailed information is retained by the processing of the method of the present invention. Compared with the prior art, the method has better denoising effect.
As can be seen from the comparison between fig. 3 and fig. 4, fig. 3 shows a large amount of noise at the lamp position of the leftmost vehicle and loses a lot of contour details on the background building contour, while fig. 4 not only eliminates a large amount of noise but also retains complete details, wherein fig. 3 shows the filtering result when gaussian filtering is adopted and the gaussian parameter is 0.8; fig. 4 illustrates the filtering in the manner of the present embodiment.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (10)

1. A method for removing infrared polarization angle image noise is characterized by comprising the following steps:
step 1: acquiring infrared polarization information of a target; the infrared polarization information comprises an intensity difference S1 of which the polarization direction is between 0 and 90 degrees and an intensity difference S2 of which the polarization direction is between 45 and 135 degrees;
step 2: performing inter-frame noise reduction processing on S1 to generate S1 first processed data S1_p1(i, j, k); performing inter-frame noise reduction processing on S2 to generate S2 first processed data S2_p1(i，j，k)；
And step 3: to S1_p1(i, j, k) conditional filtering is performed to generate S1 second processed data S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering is performed to generate S2 second processed data S2_p2(i，j，k)；
And 4, step 4: according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image.
2. The method for removing the noise of the infrared polarization angle image according to claim 1, wherein the step 1 comprises the following sub-steps:
by additionally arranging the polaroid in the light path of the infrared detection pixel for receiving the target radiation, the infrared polarization information of the target can be acquired.
3. The method for removing noise from an infrared polarization angle image according to claim 1, wherein the step 2 comprises the following sub-steps:
acquiring S1 first processing data S according to the following formula1_p1(i，j，k)：
When k is 1, S1_p1(i,j,k)＝S1(i,j,k)；
When k is more than or equal to 2,
Figure FDA0002435481350000011
wherein Diff (i, j, k) ═ S1(i,j,k)-S1(i,j,k-1)|
in the formula ,S1_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s1(i, j, k) is data of ith row and jth column of the kth frame of the S1 image;
acquiring S2 first processing data S according to the following formula2_p1(i,j,k)：
When k is 1, S2_p1(i,j,k)＝S2(i,j,k)；
When k is more than or equal to 2,
Figure FDA0002435481350000012
wherein Diff (i, j, k) ═ S2(i,j,k)-S2(i,j,k-1)|
in the formula ,S2_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the k-th frameThe absolute value of the difference value of the pixel data corresponding to the (k-1) th frame; s2(i, j, k) is data of the ith row and the jth column of the kth frame of the S2 image.
4. The method for removing noise from an infrared polarization angle image according to claim 1, wherein the step 3 comprises the following sub-steps:
acquiring S1 second processed data S according to the following formula1_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S1_p2(i,j,k)＝S1_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure FDA0002435481350000021
wherein ,
Figure FDA0002435481350000022
Figure FDA0002435481350000023
in the formula ,S1_p2W (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold;
acquiring S2 second processed data S according to the following formula2_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S2_p2(i,j,k)＝S2_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure FDA0002435481350000024
wherein ,
Figure FDA0002435481350000025
Figure FDA0002435481350000026
in the formula ,S2_p2And w (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, and q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold.
5. The method for removing noise from an infrared polarization angle image according to claim 1, wherein the step 4 comprises the following sub-steps:
obtaining an AOP polarization angle image according to the following formula:
Figure FDA0002435481350000031
in the formula, AOP (i, j, k) is the calculated AOP polarization angle image.
6. A system for removing infrared polarization angle image noise, comprising:
an acquisition unit: the infrared polarization information acquisition device is used for acquiring infrared polarization information of a target; the infrared polarization information comprises an intensity difference S1 of which the polarization direction is between 0 and 90 degrees and an intensity difference S2 of which the polarization direction is between 45 and 135 degrees;
an inter-frame noise reduction unit: for inter-frame noise reduction processing of S1 to generate S1 first processed data S1_p1(i, j, k); performing inter-frame noise reduction processing on S2 to generate S2 first processed data S2_p1(i,j,k)；
A conditional filtering unit: for pair S1_p1(iJ, k) are subjected to conditional filtering to generate S1 second processed data S1_p2(i, j, k); to S2_p1(i, j, k) conditional filtering is performed to generate S2 second processed data S2_p2(i，j，k)；
A synthesis unit: for according to S1_p2(i, j, k) and S2_p2(i, j, k) generate an AOP polarization angle image.
7. The system for removing noise from an infrared polarization angle image of claim 6, wherein the obtaining unit is configured to obtain the infrared polarization information of the target by installing a polarizer in a light path of the infrared detection pixel receiving the target radiation.
8. The system for removing noise from infrared polarization angle images of claim 6, wherein the inter-frame noise reduction unit obtains S1 first processed data S according to the following formula1_p1(i，j,k)：
When k is 1, S1_p1(i，j，k)＝S1(i，j，k)；
When k is more than or equal to 2,
Figure FDA0002435481350000032
wherein Diff (i, j, k) ═ S1(i，j，k)-S1(i,j,k-1)|
in the formula ,S1_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s1(i, j, k) is data of ith row and jth column of the kth frame of the S1 image;
the inter-frame noise reduction unit acquires S2 first processing data S according to the following formula2_p1(i,j,k)：
When k is 1, S2_p1(i，j，k)＝S2(i，j，k)；
When k is more than or equal to 2,
Figure FDA0002435481350000041
wherein Diff (i, j, k) ═ S2(i,j,k)-S2(i,j,k-1)|
in the formula ,S2_p1(i, j, k) is the result obtained by inter-frame noise reduction; i and j are integers, i is more than or equal to 1 and less than or equal to M, and j is more than or equal to 1 and less than or equal to N; m is the number of lines of the image; n is the number of columns of the image; threshold is a first absolute value Threshold; diff (i, j, k) is the absolute value of the difference value of the pixel data corresponding to the kth frame and the kth-1 frame; s2(i, j, k) is data of the ith row and the jth column of the kth frame of the S2 image.
9. The system for removing noise from infrared polarization angle image of claim 6, wherein the conditional filtering unit obtains S1 second processed data S according to the following formula1_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S1_p2(i,j,k)＝S1_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure FDA0002435481350000042
wherein ,
Figure FDA0002435481350000043
Figure FDA0002435481350000044
in the formula ,S1_p2W (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold;
the conditional filtering unit is according toAcquiring S2 second processed data S2_p2(i,j,k)：
When | S1_p1(i, j, k) | ≧ threshold2 or | S2_p1(i, j, k) | ≧ threshold 2:
S2_p2(i,j,k)＝S2_p1(i,j,k)
when | S1_p1(i,j,k)|<threshold2 and | S2_p1(i,j,k)|<threshold2 time:
Figure FDA0002435481350000045
wherein ,
Figure FDA0002435481350000046
Figure FDA0002435481350000047
in the formula ,S2_p2And w (p, q) is the weight of 9 pixels in the surrounding 3 × 3 area (p and q are integers, p is more than or equal to 1 and less than or equal to 3, and q is more than or equal to 1 and less than or equal to 3), G (i, j, k) is a normalization coefficient, threshold2 is a second absolute value threshold, and threshold3 is a weight threshold.
10. The system for denoising an infrared polarization angle image according to claim 6, wherein the synthesizing unit obtains the AOP polarization angle image according to the following formula:
Figure FDA0002435481350000051
in the formula, AOP (i, j, k) is the calculated AOP polarization angle image.
CN202010250993.3A 2020-04-01 2020-04-01 Method and system for removing image noise of infrared polarization angle Active CN111462011B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010250993.3A CN111462011B (en) 2020-04-01 2020-04-01 Method and system for removing image noise of infrared polarization angle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010250993.3A CN111462011B (en) 2020-04-01 2020-04-01 Method and system for removing image noise of infrared polarization angle

Publications (2)

Publication Number Publication Date
CN111462011A true CN111462011A (en) 2020-07-28
CN111462011B CN111462011B (en) 2023-08-01

Family

ID=71685766

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010250993.3A Active CN111462011B (en) 2020-04-01 2020-04-01 Method and system for removing image noise of infrared polarization angle

Country Status (1)

Country Link
CN (1) CN111462011B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112927166A (en) * 2021-03-23 2021-06-08 电子科技大学 Infrared polarization angle image combined noise reduction method and device based on phase correction

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6717675B1 (en) * 1999-03-26 2004-04-06 Voith Sulzer Papiertechnik Patent Gmbh System and method for determining fiber orientation in fibrous material webs
US20080159643A1 (en) * 2006-12-27 2008-07-03 Altek Corporation Method of reducing image noise
US7737399B1 (en) * 2008-10-23 2010-06-15 The United States Of America As Represented By The Secretary Of The Army Infrared Mueller matrix acquisition and preprocessing system and method
CN104657965A (en) * 2015-03-12 2015-05-27 长春理工大学 Polarized image fusion method based on discrete continuous curvelet
CN108492274A (en) * 2018-04-03 2018-09-04 中国人民解放军国防科技大学 Long-wave infrared polarization feature extraction and fusion image enhancement method
CN109636766A (en) * 2018-11-28 2019-04-16 南京理工大学 Polarization differential and intensity image Multiscale Fusion method based on marginal information enhancing

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6717675B1 (en) * 1999-03-26 2004-04-06 Voith Sulzer Papiertechnik Patent Gmbh System and method for determining fiber orientation in fibrous material webs
US20080159643A1 (en) * 2006-12-27 2008-07-03 Altek Corporation Method of reducing image noise
US7737399B1 (en) * 2008-10-23 2010-06-15 The United States Of America As Represented By The Secretary Of The Army Infrared Mueller matrix acquisition and preprocessing system and method
CN104657965A (en) * 2015-03-12 2015-05-27 长春理工大学 Polarized image fusion method based on discrete continuous curvelet
CN108492274A (en) * 2018-04-03 2018-09-04 中国人民解放军国防科技大学 Long-wave infrared polarization feature extraction and fusion image enhancement method
CN109636766A (en) * 2018-11-28 2019-04-16 南京理工大学 Polarization differential and intensity image Multiscale Fusion method based on marginal information enhancing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
刘征 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112927166A (en) * 2021-03-23 2021-06-08 电子科技大学 Infrared polarization angle image combined noise reduction method and device based on phase correction
CN112927166B (en) * 2021-03-23 2023-02-03 电子科技大学 Infrared polarization angle image combined noise reduction method and device based on phase correction

Also Published As

Publication number Publication date
CN111462011B (en) 2023-08-01

Similar Documents

Publication Publication Date Title
Ji et al. Robust video denoising using low rank matrix completion
Zhu et al. Deconvolving PSFs for a better motion deblurring using multiple images
KR101524548B1 (en) Apparatus and method for alignment of images
WO2021114868A1 (en) Denoising method, terminal, and storage medium
CN111882504B (en) Method and system for processing color noise in image, electronic device and storage medium
EP3067863B1 (en) Image noise reduction
Song et al. An adaptive l 1–l 2 hybrid error model to super-resolution
Matsui et al. Image enhancement of low-light scenes with near-infrared flash images
GB2536429A (en) Image noise reduction
CN111553841A (en) Real-time video stitching algorithm based on optimal suture line updating
Sun et al. Deep maximum a posterior estimator for video denoising
TWI496115B (en) Video frame stabilization method for the moving camera
CN111462011A (en) Method and system for removing infrared polarization angle image noise
CN112927166B (en) Infrared polarization angle image combined noise reduction method and device based on phase correction
WO2021223127A1 (en) Global motion estimation-based time-domain filtering method and device, and storage medium
Thaipanich et al. Adaptive nonlocal means algorithm for image denoising
Rawat et al. Adaptive motion smoothening for video stabilization
CN113628123A (en) Training method and device of image recovery model, electronic equipment and readable medium
CN116883235A (en) Distributed photoelectric oriented image stitching method and device
WO2023056730A1 (en) Video image augmentation method, network training method, electronic device and storage medium
Wong et al. Turbo denoising for mobile photographic applications
CN115937021A (en) Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light
CN114519676A (en) Bayer format-based raw image denoising device and method
CN108648143B (en) Image resolution enhancement method using sequence image
Chen et al. Missing recovery: Single image reflection removal based on auxiliary prior learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant