CN112927166A - Infrared polarization angle image combined noise reduction method and device based on phase correction - Google Patents
Infrared polarization angle image combined noise reduction method and device based on phase correction Download PDFInfo
- Publication number
- CN112927166A CN112927166A CN202110307391.1A CN202110307391A CN112927166A CN 112927166 A CN112927166 A CN 112927166A CN 202110307391 A CN202110307391 A CN 202110307391A CN 112927166 A CN112927166 A CN 112927166A
- Authority
- CN
- China
- Prior art keywords
- image data
- light intensity
- intensity information
- polarization angle
- guide
- 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
Links
- 230000010287 polarization Effects 0.000 title claims abstract description 182
- 238000000034 method Methods 0.000 title claims abstract description 58
- 230000009467 reduction Effects 0.000 title claims abstract description 57
- 238000001914 filtration Methods 0.000 claims abstract description 86
- 238000004364 calculation method Methods 0.000 claims abstract description 37
- 230000000694 effects Effects 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 230000008569 process Effects 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 5
- 238000003384 imaging method Methods 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 4
- 238000003331 infrared imaging Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G06T5/70—
-
- 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
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Abstract
The invention discloses an infrared polarization angle image combined noise reduction method and a device based on phase correction, which are characterized in that polarization information extraction is carried out on a polarization angle image to be processed to obtain light intensity information image data S0 corresponding to total light intensity, light intensity information image data S1 corresponding to light intensity difference with the polarization direction between 0 DEG and 90 DEG and light intensity information image data S2 corresponding to light intensity difference with the polarization direction between 45 DEG and 135 DEG, then interframe noise reduction processing and Gaussian filtering processing are carried out on S0 by utilizing the mutual correlation among S0, S1 and S2 to reduce the use of logic resources, then S0 is taken as a reference to carry out guide filtering processing on S1, S1 is taken as a reference to carry out guide filtering processing on S2 to realize combined noise reduction processing, and finally a polarization angle image is obtained by a polarization angle calculation formula of phase correction, namely a polarization angle phase correction formula to improve the noise reduction effect, reduce image distortion and improve image quality.
Description
Technical Field
The invention relates to the technical field of infrared polarization angle images, in particular to an infrared polarization angle image combined noise reduction method and device based on phase correction.
Background
Compared with the traditional infrared imaging technology, the target detection is realized by detecting the difference of the infrared radiation intensity of the target and the background, the infrared polarization imaging technology can comprehensively acquire the multi-dimensional characteristic information such as the target intensity, the polarization degree, the polarization angle and the like, and the defect that the traditional infrared imaging technology cannot effectively identify and detect the target in the complex background and the camouflage radiation is overcome. The specific implementation principle of the infrared polarization imaging technology is that light intensity information of four polarization angles of 0 degrees, 45 degrees, 90 degrees and 135 degrees of an image is collected through an infrared polarization detector, then the polarization angle and the polarization degree of the image are calculated through a Stokes vector, but the polarization angle of the image is calculated by directly utilizing the Stokes vector, the obtained infrared polarization angle image has obvious noise and serious distortion, and the single filtering processing mode has limited noise reduction capability, so that an effective polarization angle noise filtering method is urgently needed.
The invention discloses a method and a system for removing infrared polarization angle image noise, which are applied for patent application No. 202010250993.3, namely invention patent application No. 2020, 4, 1, and the method and the system are characterized in that inter-frame noise reduction is carried out on image data S1 corresponding to light intensity difference between 0 degrees and 90 degrees in the polarization direction and image data S2 corresponding to light intensity difference between 45 degrees and 135 degrees in the polarization direction, then filtering is carried out on S1 and S2 after noise reduction, and finally a polarization angle image is obtained through an original polarization angle calculation formula. The method achieves the effect of denoising the polarization angle image to a certain extent, but the processing of the data of the two images S1 and S2 requires more logic resources in an actual infrared polarization imaging system, and the relation between the phases S1 and S2 is not considered in the processing process, so that the calculated polarization angle image has larger noise and poor image quality.
Disclosure of Invention
The technical problem to be solved by the present invention is that the existing method for removing noise of infrared polarization angle images has poor denoising effect, so that the calculated polarization image noise is large and the image distortion is serious, therefore, the present invention provides a method and a device for reducing noise of infrared polarization angle images based on phase correction, which do not directly process the image data S1 corresponding to the light intensity difference with the polarization direction between 0 ° and 90 ° and the image data S2 corresponding to the light intensity difference with the polarization direction between 45 ° and 135 °, and utilize the correlation among S1, S2 and the image data S0 corresponding to the total light intensity to perform inter-frame noise reduction on the image data S0, reduce the use of logic resources, process the image data S1 with the image data S0 as the reference, process the image data S2 with the image data S1 as the reference, and finally calculate the image data S1 and the image data S2 by the improved polarization angle calculation formula, and a polarization angle image is obtained, the image noise is reduced, and the image quality is improved.
The invention is realized by the following technical scheme:
an infrared polarization angle image combined noise reduction method based on phase correction comprises the following steps:
step 1: acquiring a polarization angle image to be processed, and extracting polarization information of the polarization angle image to be processed to obtain three light intensity information image data, namely light intensity information image data S0, light intensity information image data S1 and light intensity information image data S2;
step 2: performing interframe noise reduction processing and Gaussian filtering processing on the light intensity information image data S0 to obtain reference guide image data S0_ P1;
and step 3: performing guiding filtering on the light intensity information image data S1 based on the reference guide image data S0_ P1 to obtain first guide image data S1_ P1;
and 4, step 4: performing guiding filtering on the light intensity information image data S2 based on the first guide image data S1_ P1 to obtain second guide image data S2_ P1;
and 5: and calling a polarization angle phase correction formula, and performing polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1 to obtain a polarization angle image.
Further, the extracting polarization information of the polarization angle image to be processed to obtain three types of light intensity information image data includes:
processing the polarization angle image to be processed through a polarizing film at the focal plane of the infrared detector to obtain linearly polarized light intensity data with polarization angles of 0 degree, 45 degrees, 90 degrees and 135 degrees;
and calculating the linearly polarized light intensity data through a Stokes vector to obtain light intensity information image data S0 corresponding to total light intensity, light intensity information image data S1 corresponding to light intensity difference with the polarization direction between 0 and 90 degrees and light intensity information image data S2 corresponding to light intensity difference with the polarization direction between 45 and 135 degrees.
Further, the inter-frame noise reduction processing and gaussian filtering processing on the light intensity information image data S0 to obtain reference guide image data S0_ P1 includes:
processing the light intensity information image data S0 through an interframe noise reduction processing algorithm to obtain a noise reduction processing image;
and performing Gaussian filtering processing on the noise reduction processing image through a Gaussian filtering processing algorithm to obtain reference guide image data S0_ P1.
Further, the inter-frame noise reduction processing algorithm specifically includes:
when k is 1, S0(i, j, k) is S0(i, j, k);
when k is more than or equal to 2,
in the formula: s0(i, j, k) is data of the ith row and the jth column corresponding to the kth frame image in the light intensity information image data S0, diff | S0(i, j, k) -S0(i, j, k-1) |, where diff is an absolute difference value of the kth frame data and the kth-1 frame data in the light intensity information image data S0 at the same pixel, S0(i, j, k) is an mxn matrix image, i is greater than or equal to 1 and less than or equal to M, j is greater than or equal to 1 and less than or equal to N, and thresholld is an absolute difference threshold.
Further, the gaussian filtering algorithm specifically includes:
wherein, S0(i, j, k)The data of the ith row and the jth column corresponding to the kth frame image in the light intensity information image data S0,
h (i, j, k) is a Gaussian kernel normalization parameter, G (p, q) is a Gaussian kernel coefficient of 5 x 5, and p is in the range of [1,5 ]]，q∈[1,5]。
Further, the pilot filtering the light intensity information image data S1 based on the reference pilot image data S0_ P1 to obtain first pilot image data S1_ P1 includes:
performing box filtering processing on the reference guide image data S0_ P1 and the light intensity information image data S1 to complete calculation, extraction and conversion of correlation coefficients and parameters to obtain a first guide filtering window linear variation parameter ar1And br1；
Based on the first guide filtering window linear variation parameter ar1And br1Calculating the light intensity information image data S1 by a first guide image calculation formula to obtain first guide image data S1_ P1; the first guide image calculation formula is specifically S1_ P1 ═ ar1S1+br1；
The process of performing the box filtering process on the reference guide image data S0_ P1 and the light intensity information image data S1 is specifically as follows:
wherein, ω isrTo filter the window, | ωrL is the number of pixels in the window, mur1、σr1Guide image data S0_ P1 at ωrThe mean and the variance in the filtering window,
image data S1 at omega for intensity informationrThe average value epsilon in the filtering window is a punishment value and is used for improving the filtering effect.
Further, the performing a guiding filtering on the light intensity information image data S2 based on the first guiding image data S1_ P1 to obtain a second guiding image data S2_ P1 includes:
performing box filtering processing on the first guide image data S1_ P1 and the light intensity information image data S2 to complete calculation, extraction and conversion of correlation coefficient parameters to obtain a second guide filtering window linear variation parameter ar2And br2；
Based on the second guide filter window linear variation parameter ar2And br2Calculating the light intensity information image data S2 by a second guide image calculation formula to obtain second guide image data S2_ P1; the second guide image calculation formula is specifically S2_ P1 ═ ar2S2+br2；
The process of performing the box filtering process on the first guide image data S1_ P1 and the light intensity information image data S2 is specifically:
wherein, ω isrTo filter the window, | ωrL is the number of pixels in the window, mur2、σr2At ω for the first guide image data S1_ P1, respectivelyrThe mean and the variance in the filtering window,
image data S2 at omega for intensity informationrAnd E is a penalty value, and the average value in the filtering window is used for improving the filtering effect.
Further, before the invoking of the polarization angle and phase correction formula, the phase correction-based infrared polarization angle image joint noise reduction method further includes:
obtaining an original polarization angle calculation formula, wherein the original polarization angle calculation formula specifically comprises:
wherein AoP (i, j, k) is the polarization angle corresponding to the data of the ith row and the jth column corresponding to the kth frame image, S1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the light intensity information image data S1, and S2(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the light intensity information image data S2;
performing phase correction on the original polarization angle calculation formula through the phase difference between the light intensity information image data S1 and the light intensity information image data S2 to obtain a polarization angle phase correction formula;
the formula for correcting the polarization angle phase specifically comprises the following steps:
further, the invoking of the polarization angle phase correction formula performs polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1, and obtains a polarization angle image specifically as follows:
where AoP (i, j, k) is the polarization angle corresponding to the data of the ith row and the jth column corresponding to the kth frame image, S1_ P1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the first guide image data S1_ P1, and S2_ P1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the second guide image data S2_ P1.
An infrared polarization angle image combined noise reduction device based on phase correction comprises:
the polarization information extraction module is used for acquiring a polarization angle image to be processed and extracting polarization information of the polarization angle image to be processed to obtain three types of light intensity information image data, namely light intensity information image data S0, light intensity information image data S1 and light intensity information image data S2;
a light intensity information image data S0 processing module, configured to perform interframe noise reduction processing and Gaussian filtering processing on the light intensity information image data S0 to obtain reference guide image data S0_ P1;
a light intensity information image data S1 processing module for performing a guide filtering on the light intensity information image data S1 based on the reference guide image data S0_ P1 to obtain first guide image data S1_ P1;
a light intensity information image data S2 processing module for performing guiding filtering on the light intensity information image data S2 based on the first guide image data S1_ P1 to obtain second guide image data S2_ P1;
and the polarization angle image acquisition module is used for calling a polarization angle phase correction formula and carrying out polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1 to obtain a polarization angle image.
The invention provides an infrared polarization angle image combined noise reduction method and device based on phase correction, which are characterized in that polarization information extraction is carried out on a polarization angle image to be processed to obtain light intensity information image data S0 corresponding to total light intensity, light intensity information image data S1 corresponding to light intensity difference with the polarization direction between 0 DEG and 90 DEG and light intensity information image data S2 corresponding to light intensity difference with the polarization direction between 45 DEG and 135 DEG, then interframe noise reduction processing and Gaussian filtering processing are carried out on S0 by utilizing the mutual correlation among S0, S1 and S2 to reduce the use of logic resources, then S0 is taken as a reference to carry out guide filtering processing on S1, S1 is taken as a reference to carry out guide filtering processing on S2 to realize combined noise reduction processing, and finally a polarization angle image is obtained by a polarization angle calculation formula of phase correction, namely a polarization angle phase correction formula to improve the noise reduction effect, reduce image distortion and improve image quality.
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 step diagram of an infrared polarization angle image combined noise reduction method based on phase correction according to the present invention.
Fig. 2 is a raw AoP polarization angle image obtained without any processing in an embodiment of the present invention.
FIG. 3 is a phase corrected AoP polarization angle image with the original polarization angle formula in an embodiment of the present invention.
Fig. 4 is an AoP polarization angle image after being processed by original polarization angle formula phase correction and gaussian filtering according to an embodiment of the present invention.
Fig. 5 is an effect diagram of processing by using the infrared polarization angle image joint noise reduction method based on phase correction in the embodiment of the present invention.
Fig. 6 is a schematic diagram of an infrared polarization angle image processing apparatus based on phase correction 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. The infrared polarization angle image combined noise reduction method based on phase correction is specifically applied to an application scene schematic diagram shown in fig. 1.
Example 1
As shown in fig. 1, the present invention provides a phase correction-based infrared polarization angle image joint noise reduction method, which includes:
step 1: and acquiring a to-be-processed polarization angle image, and extracting polarization information of the to-be-processed polarization angle image to obtain three types of light intensity information image data, namely light intensity information image data S0, light intensity information image data S1 and light intensity information image data S2.
The polarization angle image to be processed refers to an infrared polarization angle image to be processed.
Specifically, a polarization angle image to be processed is processed through a polarizing film at a focal plane of the infrared detector, so that linearly polarized light intensity data with polarization angles of 0 degrees, 45 degrees, 90 degrees and 135 degrees are obtained; and then calculating the linear polarization light intensity data through the Stokes vector to obtain light intensity information image data in different polarization directions.
The Stokes vector for polarized light is expressed as:
wherein, I0、I45、I90And I135Respectively representing light intensities with polarization directions of 0 °, 45 °, 90 ° and 135 °, IRAnd ILRespectively are right-handed circularly polarized light component and left-handed circularly polarized light component; s0 is light intensity information image data corresponding to total light intensity, S1 is light intensity information image data corresponding to a light intensity difference between 0 ° and 90 ° in the polarization direction, and S2 is light intensity information image data corresponding to a light intensity difference between 45 ° and 135 ° in the polarization direction; s3 is light intensity information image data corresponding to the light intensity difference between the left-handed and right-handed circularly polarized components of light, and in practical applications, the circularly polarized light component is extremely low, and S3 is set to 0 as default.
Step 2: the inter-frame noise reduction processing and the gaussian filter processing are performed on the light intensity information image data S0, resulting in reference guide image data S0_ P1.
Specifically, after obtaining the light intensity information image data S0, the light intensity information image data S0 is processed by an inter-frame noise reduction algorithm to obtain a noise reduction processed image; then, gaussian filtering processing is performed on the noise reduction processed image by a gaussian filtering processing algorithm, and reference guide image data S0_ P1 is obtained.
The interframe noise reduction processing algorithm specifically comprises the following steps:
when k is 1, S0(i, j, k) is S0(i, j, k).
When k is more than or equal to 2,
in the formula: s0(i, j, k) is data of the ith row and the jth column corresponding to the kth frame image of the intensity information image data S0, diff is | S0(i, j, k) -S0(i, j, k-1) |, where diff is an absolute difference value of the kth frame data and the kth-1 frame data in the intensity information image data S0 at the same pixel, S0(i, j, k) is an mxn matrix image, i is greater than or equal to 1 and less than or equal to M, j is greater than or equal to 1 and less than or equal to N, and thresholded is a preset absolute difference threshold. The absolute difference threshold in this embodiment may be determined according to the filtering effect. When the difference between the absolute difference values of the kth frame data and the kth-1 frame data in the light intensity information image data S0 at the same pixel is larger, namely larger than a preset absolute difference threshold, the currently output frame data is the kth frame data in the light intensity information image data S0; when the difference between the absolute difference values of the kth frame data and the kth-1 frame data in the light intensity information image data S0 is smaller, i.e., smaller than or equal to the predetermined absolute difference threshold, the currently output frame data is the average value of the sum of the kth frame data and the kth-1 frame data in the light intensity information image data S0.
The Gaussian filter processing algorithm is specifically as follows:
wherein S0(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image of the light intensity information image data S0,
h (i, j, k) is a Gaussian kernel normalization parameter, G (p, q) is a Gaussian kernel coefficient of 5 x 5, and p is in the range of [1,5 ]]，q∈[1,5]。
And step 3: the light intensity information image data S1 is subjected to guide filtering based on the reference guide image data S0_ P1, resulting in first guide image data S1_ P1.
Specifically, after the reference guide image data S0_ P1 is acquired, the reference guide image data S0_ P1 and the light intensity information image data S1 are first subjected to box filtering processing, and the calculation, extraction and conversion of the correlation coefficient parameters are completed to obtain the first guide filter window linear variation parameter ar1And br1(ii) a Then based on the first guide filter window linear variation parameter ar1And br1The light intensity information image data S1 is calculated by the first guide image calculation formula, resulting in first guide image data S1_ P1. The first guide image calculation formula is specifically S1_ P1 ═ ar1S1+br1。
The process of performing the box filtering process on the reference guide image data S0_ P1 and the light intensity information image data S1 specifically includes:
wherein, ω isrTo filter the window, | ωrL is the number of pixels in the window, mur1、σr1Guide image data S0_ P1 at ωrThe mean and the variance in the filtering window,
image data S1 at omega for intensity informationrThe mean value, epsilon, in the filtering window is a penalty value.
And 4, step 4: the light intensity information image data S2 is subjected to guide filtering based on the first guide image data S1_ P1, resulting in second guide image data S2_ P1.
Specifically, after the first guide image data S1_ P1 is obtained, the first guide image data S1_ P1 and the light intensity information image data S2 are first subjected to box filtering processing, and the calculation, extraction and conversion of the correlation coefficient parameters are completed to obtain the second guide filter window linear variation parameter ar2And br2(ii) a Then based on a second guide filter window linear variation parameter ar2And br2The light intensity information image data S2 is calculated by the second guide image calculation formula, resulting in second guide image data S2_ P1. The second guide image calculation formula is specifically S2_ P1 ═ S
ar2S2+br2。
The process of performing the box filtering process on the first guide image data S1_ P1 and the light intensity information image data S2 is specifically as follows:
wherein, ω isrTo filter the window, | ωrL is the number of pixels in the window, mur2、σr2Respectively the first guide image numberAt ω according to S1_ P1rThe mean and the variance in the filtering window,
image data S2 at omega for intensity informationrThe mean value, epsilon, in the filtering window is a penalty value.
And 5: the polarization angle phase correction formula is called, and the polarization calculation is performed on the first guide image data S1_ P1 and the second guide image data S2_ P1, so that a polarization angle image is obtained.
Specifically, as can be seen from the above Stokes vector expressions, S1, S2 are related to S0, and the formula is calculated from the original polarization angle
It can be seen that the polarization angle image quality is not only related to the image quality corresponding to S1 and S2, but also related to the phase information between the two. Therefore, in the present embodiment, the original polarization angle calculation formula is modified by the phase information of the light intensity information image data S1 and the light intensity information image data S2 to obtain a polarization angle phase modification formula, so as to reduce image distortion and improve image quality. Wherein AoP (i, j, k) is the polarization angle corresponding to the ith row and jth column data corresponding to the kth frame image, S1(i, j, k) is the ith row and jth column data corresponding to the kth frame image in the light intensity information image data S1, and S2(i, j, k) is the ith row and jth column data corresponding to the kth frame image in the light intensity information image data S2.
The corrected polarization angle phase correction formula is specifically as follows:
after the first guide image data S1_ P1 and the second guide image data S2_ P1 are obtained, a polarization angle phase correction formula is called, and polarization calculation is performed on the first guide image data S1_ P1 and the second guide image data S2_ P1, so as to obtain a polarization angle image, specifically:
where AoP (i, j, k) is the polarization angle corresponding to the data of the ith row and the jth column corresponding to the kth frame image, S1_ P1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the first guide image data S1_ P1, and S2_ P1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the second guide image data S2_ P1.
According to the infrared polarization angle image joint noise reduction method based on phase correction, the inherent relevance of S1, S2 and S0 is considered, on the basis that inter-frame noise reduction processing and Gaussian filtering processing are not directly performed on S1, S2 and polarization angle images, but on the basis that inter-frame noise reduction processing and Gaussian filtering processing are performed on S0, then guiding filtering processing is sequentially performed on S1 and S2, joint noise reduction is achieved, finally obtained polarization angle (AoP) image noise is well suppressed, meanwhile, the phase relation between S1 and S2 is considered, distortion of the polarization angle (AoP) image is reduced, and the quality of the polarization angle (AoP) image is obviously improved.
To further illustrate the working process of the infrared polarization angle image combined noise reduction method based on phase correction in the present embodiment, as can be seen from fig. 2 to fig. 5, fig. 2 is an original polarization angle AOP image obtained without any processing, the image quality is very poor, and the noise level is very strong; FIG. 3 is a polarization angle AOP image obtained by only performing the original polarization angle formula phase correction, which has significantly improved image quality but also has low image noise level; fig. 4 shows the result of performing gaussian filtering processing on the polarization angle AOP image obtained by phase correction of the original polarization angle formula, and due to filtering, although the image noise is further reduced, the image contrast is also greatly reduced, and the imaging level is not good. Fig. 5 is a processing effect diagram of the method of the present invention, and it can be seen that, by the processing of the method of the present invention, the noise of the polarization angle AOP image is greatly reduced, the image contrast is also improved to a certain extent, and a good imaging quality is obtained.
In addition, as can be seen from comparison between fig. 2 and fig. 3, the polarization angle image obtained by directly performing the phase correction on the original polarization angle formula has a greatly suppressed noise level, but the noise in the image is still not low, and as can be seen from comparison between fig. 3 and fig. 4, when the polarization angle image is directly subjected to the gaussian filtering denoising, although fig. 4 is compared with fig. 3, the noise right in front of the vehicle is suppressed to a certain extent, the overall contrast of the vehicle body becomes much lower, where fig. 4 adopts a gaussian filter of order 5, and the gaussian parameter is a filtering result when 0.8 is taken as the gaussian parameter. Fig. 5 is an image processed by the method of the present invention, as compared to all previous methods: the method not only greatly eliminates the polarization angle image noise, but also improves the contrast of the image.
It should be noted that the infrared polarization angle image joint noise reduction method based on phase correction protected in the embodiment of the present invention is based on the internal relationship among S0, S1, and S2, and performs inter-frame noise reduction processing and gaussian filtering processing on S0 in advance, performs guided filtering processing on S1 with reference to S0, and performs guided filtering processing on S2 with reference to S1, so as to finally achieve the purpose of joint noise reduction of polarization angle images (AoP), thereby reducing distortion of polarization angle images. The method is not limited to S0 Gaussian filtering and S1 and S2 guided filtering combined noise reduction, and other filtering combined noise reduction methods based on the idea belong to the protection content of the invention.
The method for correcting the phase of the original polarization angle calculation formula is provided based on polarized light with different angles obtained by detecting polarization pixels of a polarization detector, and is not limited to 0 degree and 45 degrees. The invention provides phase correction at 90 degrees and 135 degrees, and other methods for correspondingly performing phase correction based on different angles, which belong to the protection content of the invention.
Example 2
As shown in fig. 6, this embodiment provides a phase correction-based infrared polarization angle image joint noise reduction apparatus corresponding to the phase correction-based infrared polarization angle image joint noise reduction method of embodiment 1 one to one, including:
the polarization information extraction module 10 is configured to obtain a polarization angle image to be processed, and extract polarization information of the polarization angle image to be processed to obtain three types of light intensity information image data, which are light intensity information image data S0, light intensity information image data S1, and light intensity information image data S2.
And the light intensity information image data S0 processing module 20 is used for performing interframe noise reduction processing and Gaussian filtering processing on the light intensity information image data S0 to obtain reference guiding image data S0_ P1.
The light intensity information image data S1 processing module 30 is configured to perform guiding filtering on the light intensity information image data S1 based on the reference guide image data S0_ P1 to obtain first guide image data S1_ P1.
The light intensity information image data S2 processing module 40 is configured to perform guiding filtering on the light intensity information image data S2 based on the first guide image data S1_ P1 to obtain second guide image data S2_ P1.
The polarization angle image obtaining module 50 is configured to invoke a polarization angle phase correction formula, and perform polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1 to obtain a polarization angle image.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
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. An infrared polarization angle image combined noise reduction method based on phase correction is characterized by comprising the following steps:
step 1: acquiring a polarization angle image to be processed, and extracting polarization information of the polarization angle image to be processed to obtain three light intensity information image data, namely light intensity information image data S0, light intensity information image data S1 and light intensity information image data S2;
step 2: performing interframe noise reduction processing and Gaussian filtering processing on the light intensity information image data S0 to obtain reference guide image data S0_ P1;
and step 3: performing guiding filtering on the light intensity information image data S1 based on the reference guide image data S0_ P1 to obtain first guide image data S1_ P1;
and 4, step 4: performing guiding filtering on the light intensity information image data S2 based on the first guide image data S1_ P1 to obtain second guide image data S2_ P1;
and 5: and calling a polarization angle phase correction formula, and performing polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1 to obtain a polarization angle image.
2. The method of claim 1, wherein the extracting polarization information from the polarization angle image to be processed to obtain three types of light intensity information image data comprises:
processing the polarization angle image to be processed through a polarizing film at the focal plane of the infrared detector to obtain linearly polarized light intensity data with polarization angles of 0 degree, 45 degrees, 90 degrees and 135 degrees;
and calculating the linearly polarized light intensity data through a Stokes vector to obtain light intensity information image data S0 corresponding to total light intensity, light intensity information image data S1 corresponding to light intensity difference with the polarization direction between 0 and 90 degrees and light intensity information image data S2 corresponding to light intensity difference with the polarization direction between 45 and 135 degrees.
3. The method of claim 1, wherein the inter-frame denoising and gaussian filtering are performed on the intensity information image data S0 to obtain reference guide image data S0_ P1, and the method comprises:
processing the light intensity information image data S0 through an interframe noise reduction processing algorithm to obtain a noise reduction processing image;
and performing Gaussian filtering processing on the noise reduction processing image through a Gaussian filtering processing algorithm to obtain reference guide image data S0_ P1.
4. The infrared polarization angle image combined noise reduction method based on phase correction according to claim 3, wherein the inter-frame noise reduction processing algorithm specifically comprises:
when k is 1, S0(i, j, k) is S0(i, j, k);
when k is more than or equal to 2,
in the formula: s0(i, j, k) is data of the ith row and the jth column corresponding to the kth frame image of S0 in the light intensity information image data, diff is | S0(i, j, k) -S0(i, j, k-1) |, where diff is an absolute difference value of the kth frame data and the kth-1 frame data in the light intensity information image data S0 at the same pixel, S0(i, j, k) is an mxn matrix image, i is greater than or equal to 1 and less than or equal to M, j is greater than or equal to 1 and less than or equal to N, and threshold is a preset absolute difference threshold.
5. The infrared polarization angle image combined noise reduction method based on phase correction according to claim 3, wherein the Gaussian filter processing algorithm specifically comprises:
wherein S0(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the light intensity information image data S0,
h (i, j, k) is a Gaussian kernel normalization parameter, G (p, q) is a Gaussian kernel coefficient of 5 x 5, and p is in the range of [1,5 ]]，q∈[1,5]。
6. The method of claim 1, wherein the performing a pilot filtering on the intensity information image data S1 based on the reference pilot image data S0_ P1 to obtain a first pilot image data S1_ P1 comprises:
performing box filtering processing on the reference guide image data S0_ P1 and the light intensity information image data S1 to obtain a first guide filter window linear variation parameter ar1And br1；
Based on the first guide filtering window linear variation parameter ar1And br1Calculating the light intensity information image data S1 by a first guide image calculation formula to obtain first guide image data S1_ P1; the first guide image calculation formula is specifically S1_ P1 ═ ar1S1+br1；
The process of performing the box filtering process on the reference guide image data S0_ P1 and the light intensity information image data S1 is specifically as follows:
wherein, ω isrTo filter the window, | ωrL is the number of pixels in the window, mur1、σr1Guide image data S0_ P1 at ωrThe mean and the variance in the filtering window,
image data S1 at omega for intensity informationrThe average value epsilon in the filtering window is a punishment value and is used for improving the filtering effect.
7. The method of claim 1, wherein the performing guided filtering on the intensity information image data S2 based on the first guide image data S1_ P1 to obtain second guide image data S2_ P1 comprises:
performing box filtering processing on the first guide image data S1_ P1 and the light intensity information image data S2 to obtain a second guide filter window linear variation parameter ar2And br2；
Based on the second guide filter window linear variation parameter ar2And br2Calculating the light intensity information image data S2 by a second guide image calculation formula to obtain second guide image data S2_ P1; the second guide image calculation formula is specifically S2_ P1 ═ ar2S2+br2；
The process of performing the box filtering process on the first guide image data S1_ P1 and the light intensity information image data S2 is specifically:
wherein, ω isrTo filter the window, | ωrL is the number of pixels in the window, mur2、σr2At ω for the first guide image data S1_ P1, respectivelyrThe mean and the variance in the filtering window,
image data S2 at omega for intensity informationrThe average value epsilon in the filtering window is a punishment value and is used for improving the filtering effect.
8. The method according to claim 1, wherein before the invoking of the polarization angle and phase correction formula, the method further comprises:
obtaining an original polarization angle calculation formula, wherein the original polarization angle calculation formula specifically comprises:
wherein AoP (i, j, k) is the polarization angle corresponding to the ith row and jth column data corresponding to the kth frame image, S1(i, j, k) is the ith row and jth column data corresponding to the kth frame image in the light intensity information image data S1, and S2(i, j, k) is the ith row and jth column data corresponding to the kth frame image in the light intensity information image data S2;
carrying out phase correction on the original polarization angle calculation formula through the phase information of the light intensity information image data S1 and the light intensity information image data S2 to obtain a polarization angle phase correction formula;
the formula for correcting the polarization angle phase specifically comprises the following steps:
9. the method of claim 1, wherein the polarization angle phase correction formula is invoked to perform polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1, and obtain a polarization angle image specifically:
where AoP (i, j, k) is the polarization angle corresponding to the data of the ith row and the jth column corresponding to the kth frame image, S1_ P1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the first guide image data S1_ P1, and S2_ P1(i, j, k) is the data of the ith row and the jth column corresponding to the kth frame image in the second guide image data S2_ P1.
10. An infrared polarization angle image combined noise reduction device based on phase correction is characterized by comprising:
the polarization information extraction module is used for acquiring a polarization angle image to be processed and extracting polarization information of the polarization angle image to be processed to obtain three types of light intensity information image data, namely light intensity information image data S0, light intensity information image data S1 and light intensity information image data S2;
a light intensity information image data S0 processing module, configured to perform interframe noise reduction processing and Gaussian filtering processing on the light intensity information image data S0 to obtain reference guide image data S0_ P1;
a light intensity information image data S1 processing module for performing a guide filtering on the light intensity information image data S1 based on the reference guide image data S0_ P1 to obtain first guide image data S1_ P1;
a light intensity information image data S2 processing module for performing guiding filtering on the light intensity information image data S2 based on the first guide image data S1_ P1 to obtain second guide image data S2_ P1;
and the polarization angle image acquisition module is used for calling a polarization angle phase correction formula and carrying out polarization calculation on the first guide image data S1_ P1 and the second guide image data S2_ P1 to obtain a polarization angle image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110307391.1A CN112927166B (en) | 2021-03-23 | 2021-03-23 | Infrared polarization angle image combined noise reduction method and device based on phase correction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110307391.1A CN112927166B (en) | 2021-03-23 | 2021-03-23 | Infrared polarization angle image combined noise reduction method and device based on phase correction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112927166A true CN112927166A (en) | 2021-06-08 |
| CN112927166B CN112927166B (en) | 2023-02-03 |
Family
ID=76175517
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110307391.1A Active CN112927166B (en) | 2021-03-23 | 2021-03-23 | Infrared polarization angle image combined noise reduction method and device based on phase correction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112927166B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114066759A (en) * | 2021-11-18 | 2022-02-18 | 电子科技大学 | FPGA-based infrared image real-time distortion correction method and system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030227859A1 (en) * | 2002-02-27 | 2003-12-11 | Hideaki Hirai | Optical pickup and optical information processing apparatus |
| US20100147674A1 (en) * | 2008-12-11 | 2010-06-17 | Anteos, Inc. | Method for fabrication of aligned nanowire structures in semiconductor materials for electronic, optoelectronic, photonic and plasmonic devices |
| CN103530853A (en) * | 2013-10-17 | 2014-01-22 | 中北大学 | Infrared intensity image and infrared polarization image enhancement and fusion method |
| US20150362698A1 (en) * | 2014-06-11 | 2015-12-17 | Olympus Corporation | Image Sensor for Depth Estimation |
| CN105279747A (en) * | 2015-11-25 | 2016-01-27 | 中北大学 | Infrared polarization and light intensity image fusing method guided by multi-feature objective function |
| CN106846289A (en) * | 2017-01-17 | 2017-06-13 | 中北大学 | A kind of infrared light intensity and polarization image fusion method based on conspicuousness migration with details classification |
| 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 |
| CN111462011A (en) * | 2020-04-01 | 2020-07-28 | 电子科技大学 | Method and system for removing infrared polarization angle image noise |
-
2021
- 2021-03-23 CN CN202110307391.1A patent/CN112927166B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030227859A1 (en) * | 2002-02-27 | 2003-12-11 | Hideaki Hirai | Optical pickup and optical information processing apparatus |
| US20100147674A1 (en) * | 2008-12-11 | 2010-06-17 | Anteos, Inc. | Method for fabrication of aligned nanowire structures in semiconductor materials for electronic, optoelectronic, photonic and plasmonic devices |
| CN103530853A (en) * | 2013-10-17 | 2014-01-22 | 中北大学 | Infrared intensity image and infrared polarization image enhancement and fusion method |
| US20150362698A1 (en) * | 2014-06-11 | 2015-12-17 | Olympus Corporation | Image Sensor for Depth Estimation |
| CN105279747A (en) * | 2015-11-25 | 2016-01-27 | 中北大学 | Infrared polarization and light intensity image fusing method guided by multi-feature objective function |
| CN106846289A (en) * | 2017-01-17 | 2017-06-13 | 中北大学 | A kind of infrared light intensity and polarization image fusion method based on conspicuousness migration with details classification |
| 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 |
| CN111462011A (en) * | 2020-04-01 | 2020-07-28 | 电子科技大学 | Method and system for removing infrared polarization angle image noise |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114066759A (en) * | 2021-11-18 | 2022-02-18 | 电子科技大学 | FPGA-based infrared image real-time distortion correction method and system |
| CN114066759B (en) * | 2021-11-18 | 2023-08-01 | 电子科技大学 | FPGA-based infrared image real-time distortion correction method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112927166B (en) | 2023-02-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108492274B (en) | Long-wave infrared polarization feature extraction and fusion image enhancement method | |
| EP4109392A1 (en) | Image processing method and image processing device | |
| CN104796582B (en) | Video image denoising and Enhancement Method and device based on random injection retinex | |
| CN110322485B (en) | Rapid image registration method of heterogeneous multi-camera imaging system | |
| CN112927166B (en) | Infrared polarization angle image combined noise reduction method and device based on phase correction | |
| CN113421206B (en) | Image enhancement method based on infrared polarization imaging | |
| CN114529593A (en) | Infrared and visible light image registration method, system, equipment and image processing terminal | |
| CN116883235A (en) | Distributed photoelectric oriented image stitching method and device | |
| CN111462011B (en) | Method and system for removing image noise of infrared polarization angle | |
| CN110009575B (en) | Infrared image stripe noise suppression method based on sparse representation | |
| Kim et al. | Under display camera quad bayer raw image restoration using deep learning | |
| CN116071259A (en) | Infrared image enhancement method based on secondary guide filtering | |
| CN115937021A (en) | Polarization defogging method based on frequency domain feature separation and iterative optimization of atmospheric light | |
| Natarajan | A review on underwater image enhancement techniques | |
| CN109829899B (en) | Background suppression algorithm for steel coil end surface defect detection | |
| Chen et al. | CDP-GAN: Near-infrared and visible image fusion via color distribution preserved GAN | |
| WO2021178172A1 (en) | Joint rolling shutter image stitching and rectification | |
| Wang et al. | Image edge detection algorithm based onwavelet fractional differential theory | |
| CN111652820A (en) | Method and device for correcting infrared image stripe nonuniformity and image processing system | |
| Qi et al. | Modified adaptive color plane based color interpolation algorithm for digital image sensors | |
| CN112837312B (en) | Method and system for improving image quality of polarization infrared thermal imager | |
| Li et al. | Robust video denoising for mixed Poisson, Gaussian and impule noise | |
| CN115018721A (en) | Heterogeneity correction method based on multi-scale analysis and singular value decomposition | |
| Angitha et al. | A review on image dehazing techniques | |
| CN117372485B (en) | One-dimensional phase correlation image registration method based on phase spectrum and amplitude spectrum recombination |
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 |