Circularly polarized light image enhancement method for electric field vector estimation

Abstract

This invention belongs to the field of optical image processing technology, and in particular to a method for enhancing circularly polarized light images by estimating the electric field vector, comprising the following steps: S1: The acquired left-handed circularly polarized light image I LCP Image I of right-hand circularly polarized light RCP The total light intensity image I was obtained by performing Stokes parametric calculation. Sum And circular polarization image I DoCP The method of this invention can calculate the circularly polarized electric field intensity information, make reasonable use of the advantages of polarization images and light intensity images, improve the contrast between the target and the background, and provide clearer structural information around the target edges, which is more in line with human visual habits. The final circularly polarized enhanced image has outstanding contrast and accurate edge detail information, and has important application prospects in a variety of advanced computer vision tasks.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical image processing, in particular to a circularly polarized light image enhancement method based on electric field vector estimation. BACKGROUND

[0002] In recent years, China's polarization imaging technology has developed steadily, and has made significant progress in military detection, industrial production, medical services and other fields. However, for some complex scenes such as low light, underwater and the like, linearly polarized light images often exhibit a lot of noise, which interferes with target detection and seriously degrades image quality. Therefore, it is necessary to conduct in-depth and effective exploration of polarization information. Since the scattered light after the interaction between polarized light and the target contains polarization information determined by the characteristics of the target itself, in-depth mining of this polarization information is the key to solving these problems. Compared with linearly polarized light, the electric field vector of circularly polarized light rotates at a constant angular velocity in the vertical plane of the propagation direction and only has two rotation directions. Therefore, circularly polarized light has good polarization maintaining characteristics and can still retain effective polarization information in the reflected light of the target in some complex environments. However, the fewer polarization states not only bring stability but also increase the difficulty of polarization information mining. Therefore, in-depth mining of effective polarization information is crucial to improving the quality of polarization images and highlighting target information.

[0003] Although the traditional image enhancement method based on light intensity image can improve the texture details and edge information of the target to a certain extent, it is not effective for polarization images with more information sources. Therefore, while mining more dimensional polarization information, an effective calculation method is also needed to use more information sources to enhance the image. Therefore, the present application provides a circularly polarized light image enhancement method based on electric field vector estimation to solve the above problems. SUMMARY

[0004] (I) Technical problems solved

[0005] In view of the deficiencies in the prior art, the present application provides a circularly polarized light image enhancement method based on electric field vector estimation, which solves the problems raised in the background art.

[0006] (II) Technical solutions

[0007] In order to achieve the above purpose, the present application specifically adopts the following technical solutions:

[0008] A circularly polarized light image enhancement method based on electric field vector estimation, comprising the following steps:

[0009] S1: performing Stokes parameter calculation on the collected left-handed circularly polarized light image I LCP and the right-handed circularly polarized light image I RCP to obtain a total light intensity image I Sumand the circular polarization degree image I DoCP ;

[0010] S2: calculating the left circularly polarized light image I LCP and the right circularly polarized light image I RCP , the horizontal vibration direction electric field intensity matrix E H , the vertical vibration direction electric field intensity matrix E V , the horizontal electric field vibration phase offset ε H , the vertical electric field vibration phase offset ε V , and the circular polarization electric field intensity image I CPE ;

[0011] S3: based on the calculated images I CPE and I DoCP , calculating the circular polarization electric field intensity edge component Γ CPE and the circular polarization degree edge component Γ DoCP ;

[0012] S4: based on the calculated images I Sum , I DoCP , I CPE , and the edge components Γ CPE , Γ DoCP , calculating the contrast information weight w JCS and the edge information weight w JES , and finally calculating the circular polarization enhancement image I ECP .

[0013] Further, the calculation formula of the circular polarization electric field intensity image I CPE in S2 is as follows:

[0014]

[0015]

[0016] In the formula, E Total is the electric field intensity superposition matrix, is the left circular polarization electric field intensity matrix in the horizontal vibration direction, is the right circular polarization electric field intensity matrix in the horizontal vibration direction, is the left circular polarization electric field intensity matrix in the vertical vibration direction, and -i represents phase lag is the right circular polarization electric field intensity matrix in the vertical vibration direction, and +i represents phase advance E RCP is the right circular polarization light electric field intensity matrix, and E LCP is the left circular polarization light electric field intensity matrix.

[0017] Furthermore, the horizontal electric field vibration phase shift ε in S2 H and the phase shift ε of the vertical electric field vibration V The calculation formula is as follows:

[0018] ε H =μ H +σ H ·Z H

[0019] ε V =μ V +σ V ·Z V

[0020] Where μ H Represents Ω H The mean of the sample space, μ V Represents Ω V The mean of the sample space, Represents Ω H Variance of the sample space Represents Ω V The variance of the sample space, Z H It is a random variable that follows a normal distribution, with mean and variance of (μ) and (μ) respectively. H , σ H 2 Z V It is a random variable that follows a normal distribution, with mean and variance of (μ) and (μ) respectively. V , σ V 2 );

[0021] Ω H The sample space representing the phase compensation in the horizontal direction, Ω V The sample space representing vertical phase compensation is calculated using the following formula:

[0022]

[0023] The max(·) function is the maximum value function.

[0024] Furthermore, the edge component Γ of the circularly biased electric field intensity in S3 CPE and the edge component of circular polarization Γ DoCP The specific formula for the calculation method is as follows:

[0025]

[0026] In the formula, Γ(n) represents an edge component image calculation formula, n is a current pixel point, NMS(G(n), θ(n)) represents non-maximum suppression, for each pixel point, according to the edge gradient angle parameter, the edge gradient amplitude parameter is detected for local maximum value, only the local maximum value, that is, the edge pixel, G(n) represents an edge gradient parameter, θ(n) represents an edge gradient angle parameter, T is a hyper parameter, the value is in (0-255), used for edge pixel value correction, χ(n) represents an input image, GS(·) represents a Gaussian blur function, S x represents a convolution operation in a vertical direction y represents a convolution operation in a horizontal direction

[0027] Γ CPE is obtained by substituting I CPE into the edge component image calculation formula DoCP is obtained by substituting I DoCP into the edge component image calculation formula.

[0028] Further, the calculation method of the edge information weight w JES and the contrast information weight w JCS in the S4 is as follows:

[0029]

[0030]

[0031] In the formula, w(I1(i,j), I2(i,j)) represents a joint weight calculation formula, I1(i,j), I2(i,j) represent two image inputs respectively, (i,j) represents a pixel position, μ1 and μ2 are obtained by the calculation formula of μ i , N represents the total number of pixels, and n represents a current pixel point.

[0032] The edge information weight w CPE is calculated according to Γ DoCP (i,j), Γ JES (i,j):

[0033] w JES = w(Γ CPE (i,j), Γ DoCP (i,j))

[0034] The contrast information weight w Sum is calculated according to I ECP (i,j), I JCS (i,j):

[0035] w JCS = w(I Sum (i,j), I ECP(i,j)).

[0036] Further, the circular polarization enhanced image I ECP in S4 is calculated as follows:

[0037] I ECP = (a CS -w JCS )(a ES -w JES )I Sum +a CS w JCS I CPE +a ES w JES I DoCP

[0038]

[0039] where a CS is the contrast weight scaling factor, a ES is the edge weight scaling factor, μ JCS is the mean of the w JCS weight matrix, and μ JES is the mean of the w JES weight matrix.

[0040] (III) Beneficial Effects

[0041] Compared with the prior art, the present application provides a circularly polarized light image enhancement method for electric field vector estimation, which has the following beneficial effects:

[0042] The method of the present application can calculate the circular polarization electric field intensity information, reasonably utilize the advantage information of the polarization image and the light intensity image, improve the contrast of the target and the background, have more clear structure information around the target edge, be more in line with human visual habits, and finally obtain a circular polarization enhanced image with outstanding contrast and accurate edge detail information, which has important application prospects in subsequent various advanced computer vision tasks. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 is a flow chart of the circularly polarized light image enhancement method for electric field vector estimation of the present application;

[0044] Figure 2 is an architectural principle diagram of the circularly polarized light image enhancement method for electric field vector estimation of the present application;

[0045] Figure 3 is a circularly polarized light intensity image I Sum of the present application;

[0046] Figure 4is a DoCP circular polarization degree image I of the present application DoCP ;

[0047] Figure 5 is a circular polarization electric field intensity image I of the present application CPE ;

[0048] Figure 6 is a circular polarization enhancement image I of the present application ECP . DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0050] EMBODIMENT

[0051] As shown in Figures 1-6 , a circularly polarized light image enhancement method for electric field vector estimation according to one embodiment of the present application comprises the following steps:

[0052] S1: performing Stokes parameter calculation on the acquired left-circularly polarized light image I LCP and right-circularly polarized light image I RCP to obtain a total light intensity image I Sum and a circular polarization degree image I DoCP ; the calculated images are as shown in Figure 3 and Figure 4 .

[0053] S2: calculating the acquired left-circularly polarized light image I LCP and right-circularly polarized light image I RCP , calculating the horizontal vibration direction electric field intensity matrix E H , the vertical vibration direction electric field intensity matrix E V , the horizontal electric field vibration phase offset ε H , the vertical electric field vibration phase offset ε V , and the circular polarization electric field intensity image I CPE .

[0054] The calculation formula of the circular polarization electric field intensity image I CPE is as follows:

[0055]

[0056] In the formula, E Total is an electric field intensity superposition matrix, Eh= -iEi Eh= -iEi Eh= -iEi Eh= -iEi Eh= -iEi RCP Eh= -iEi LCP Eh= -iEi Eh= -iEi

[0057] Eh= -iEi H Eh= -iEi V Eh= -iEi Eh= -iEi

[0058] Eh= -iEi H Eh= -iEi H Eh= -iEi H Eh= -iEi H Eh= -iEi Eh= -iEi

[0059] Eh= -iEi V Eh= -iEi V Eh= -iEi V Eh= -iEi V Eh= -iEi Eh= -iEi

[0060] Eh= -iEi H Eh= -iEi H Eh= -iEi V Eh= -iEi V Eh= -iEi Eh= -iEi H Eh= -iEi Eh= -iEi V Eh= -iEi H Eh= -iEi H Eh= -iEi H Eh= -iEi 2 Eh= -iEi V Eh= -iEi V Eh= -iEi V Eh= -iEi 2 Eh= -iEi Eh= -iEi

[0061] Eh= -iEi H Eh= -iEi V Eh= -iEi Eh= -iEi

[0062] Eh= -iEi Eh= -iEi Eh= -iEi

[0063] Eh= -iEi Eh= -iEi

[0064] This invention innovatively proposes the above calculation formula, and its technical function is to calculate a new information source image, namely, the circularly polarized electric field intensity image I. CPE ,like Figure 5 As shown, by decomposing and calculating the horizontal and vertical electric field intensity information, then calculating the offset matrix, and finally synthesizing the intensity image, this optical image calculation method can obtain a new information source image, effectively extract the intensity information of vibrating photons, and is more conducive to extracting the contrast information between the target and the background in different images.

[0065] S3: Image I obtained based on calculation CPE and I DoCP Calculate the edge component Γ of the circularly deflected electric field intensity. CPE and the edge component of circular polarization Γ DoCP ;

[0066] Edge component Γ of circularly biased electric field intensity CPE and the edge component of circular polarization Γ DoCP The specific formula for the calculation method is as follows:

[0067]

[0068]

[0069] In the formula, Γ(n) represents the edge component image calculation formula, n is the current pixel, NMS(G(n),θ(n)) represents non-maximum suppression. For each pixel, local maximum detection is performed on the edge gradient magnitude parameter based on its edge gradient angle parameter, and only local maximum values ​​are retained, i.e., edge pixels. G(n) represents the edge gradient parameter, θ(n) represents the edge gradient angle parameter, T is a hyperparameter with a value in (0~255) used for edge pixel value correction, χ(n) represents the input image, GS(·) represents the Gaussian blur function, and S x S represents the convolution operation in the vertical direction. y Represents convolution operations in the horizontal direction;

[0070] Γ CPE by I CPE Substituting into the edge component image calculation formula, we get Γ DoCP by I DoCP The edge component image calculation formula is used to obtain the result.

[0071] S4: Image I obtained based on calculation Sum I DoCP I CPE and edge component Γ CPE , Γ DoCP Calculate the contrast information weight w JCS and edge information weight w JESFinally, the circularly polarized enhanced image I was calculated. ECP ,like Figure 6 As shown;

[0072] Edge information weight w JES and contrast information weight w JCS The formula for calculating is as follows:

[0073]

[0074] In the formula, w(I1(i,j),I2(i,j)) represents the joint weight calculation formula, I1(i,j) and I2(i,j) represent the two image inputs respectively, (i,j) represents the pixel position, and μ1 and μ2 are both derived from μ i The calculation formula is obtained, where N represents the total number of pixels and n represents the current pixel.

[0075] By Γ CPE (i,j),Γ DoCP Calculate the edge information weight w for (i,j) JES :

[0076] w JES =w(Γ CPE (i,j),Γ DoCP (i,j))

[0077] by I Sum (i,j),I ECP Calculate the contrast information weight w for (i,j) JCS :

[0078] w JCS =w(I Sum (i,j),I ECP (i,j));

[0079] Circular polarization enhanced image I ECP The calculation formula is as follows:

[0080] I ECP =(a CS -w JCS (a) ES -w JES )I Sum +a CS w JCS I CPE +a ES w JES I DoCP

[0081]

[0082] Where a CS a is the contrast weighting factor.ES is an edge weight proportion factor, μ JCS is w JCS is a mean value of the weight matrix, μ JES is w JES is a mean value of the weight matrix.

[0083] The present application proposes a weight coefficient calculation formula and a circular polarization image enhancement calculation formula, and finally obtains a circular polarization enhanced image. The technical effect is that the circular polarization image enhancement can reasonably extract and balance the dominant information in different source images. Therefore, the circular polarization image enhancement highlights the contrast information of the target and the background in the image, strengthens the edge features of the image, and is helpful for further analysis and use of the image.

[0084] Finally, it should be noted that: the above only for the preferred embodiments of the present application, and not for the purpose of limiting the present application, although the foregoing detailed description of the present application is made with reference to the foregoing embodiments, for those skilled in the art, it still can be modified to the technical solutions recorded in the foregoing embodiments, or equivalent replacement of some of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A method for enhancing circularly polarized light images by estimating the electric field vector, characterized in that: Includes the following steps: S1: The total light intensity image I is obtained by performing Stokes parametric calculations on the acquired left-handed and right-handed circularly polarized light images. Sum And circular polarization image I DoCP ; S2: Calculate the acquired left-handed circularly polarized light image I LCP Image I of right-hand circularly polarized light RCP Calculate the electric field intensity matrix E in the horizontal vibration direction. H The electric field intensity matrix E perpendicular to the vibration direction V Horizontal electric field vibration phase shift ε H Vertical electric field vibration phase shift ε V and circularly polarized electric field intensity image I CPE ; S3: Image I obtained based on calculation CPE and I DoCP Calculate the edge component Γ of the circularly deflected electric field intensity. CPE and the edge component of circular polarization Γ DoCP ; S4: Image I obtained based on calculation Sum I DoCP I CPE and edge component Γ CPE , Γ DoCP Calculate the contrast information weight w JCS and edge information weight w JES Finally, the circularly polarized enhanced image I was calculated. ECP ; The edge information weight w in S4 JES and contrast information weight w JCS The formula for calculating is as follows: In the formula, w(I1(i,j),I2(i,j)) represents the joint weight calculation formula, I1(i,j) and I2(i,j) represent the two image inputs respectively, (i,j) represents the pixel position, and μ1 and μ2 are both derived from μ i The calculation formula is obtained, where N represents the total number of pixels and n represents the current pixel. By Γ CPE (i,j),Γ DoCP Calculate the edge information weight w for (i,j) JES : w JES =w(Γ CPE (i,j),Γ DoCP (i,j)) byI Sum (i,j),I ECP Calculate the contrast information weight w for (i,j) JCS : w JCS =w(I Sum (i,j),I ECP (i,j)); The circular polarization enhancement image I in S4 ECP The calculation formula is as follows: I ECP =(a CS -w JCS )(a ES -w JES )I Sum +a CS w JCS I CPE +a ES w JES I DoCP Where a CS a is the contrast weighting factor. ES μ is the edge weight scaling factor. JCS For w JCS The mean of the weight matrix, μ JES For w JES The mean of the weight matrix.

2. The method for enhancing circularly polarized light images based on electric field vector estimation according to claim 1, characterized in that: The circularly polarized electric field intensity image I in S2 CPE The calculation formula is as follows: In the formula E Total The electric field intensity superposition matrix is... The matrix of the left-hand circularly polarized electric field intensity in the horizontal vibration direction is shown. The matrix of the right-hand circularly polarized electric field intensity in the horizontal vibration direction is shown. Here is the matrix of the left-hand circularly polarized electric field intensity perpendicular to the vibration direction, and -i represents the phase lag. Here is the right-hand circularly polarized electric field intensity matrix perpendicular to the vibration direction, and +i indicates phase lead. E RCP E is the right-hand circularly polarized photoelectric field intensity matrix. LCP This is the left-hand circularly polarized photoelectric field intensity matrix.

3. The method for enhancing circularly polarized light images based on electric field vector estimation according to claim 2, characterized in that: The horizontal electric field vibration phase shift ε in S2 H and the phase shift ε of the vertical electric field vibration V The calculation formula is as follows: e H =μ H +s H ·Z H e V =μ V +s V ·Z V Where μ H Represents Ω H The mean of the sample space, μ V Represents Ω V The mean of the sample space, Represents Ω H Variance of the sample space Represents Ω V The variance of the sample space, Z H It is a random variable that follows a normal distribution, with mean and variance of (μ) and (μ) respectively. H , σ H 2 Z V It is a random variable that follows a normal distribution, with mean and variance of (μ) and (μ) respectively. V , σ V 2 ); Ω H The sample space representing the phase compensation in the horizontal direction, Ω V The sample space representing vertical phase compensation is calculated using the following formula: The max(·) function is the maximum value function.

4. The method for enhancing circularly polarized light images based on electric field vector estimation according to claim 2, characterized in that: The edge component Γ of the circular electric field intensity in S3 CPE and the edge component of circular polarization Γ DoCP The specific formula for the calculation method is as follows: In the formula, Γ(n) represents the edge component image calculation formula, n is the current pixel, NMS(G(n),θ(n)) represents non-maximum suppression. For each pixel, local maximum detection is performed on the edge gradient magnitude parameter based on its edge gradient angle parameter, and only local maximum values ​​are retained, i.e., edge pixels. G(n) represents the edge gradient parameter, θ(n) represents the edge gradient angle parameter, T is a hyperparameter with a value in (0~255) used for edge pixel value correction, χ(n) represents the input image, GS(·) represents the Gaussian blur function, and S x S represents the convolution operation in the vertical direction. y Represents convolution operations in the horizontal direction; Γ CPE byI CPE Substituting into the edge component image calculation formula, we get Γ DoCP byI DoCP The edge component image calculation formula is used to obtain the result.

Images