Polarization sensor modeling method considering sunlight interference

Abstract

The invention relates to a polarization sensor modeling method considering sunlight interference, which comprises the following steps of: dividing light incident to a polarization sensor into three parts, namely linearly polarized light, vertically incident unpolarized light and sunlight, and representing sunlight interference transmission light intensity by considering the influence of sunlight interference on the polarization sensor on the basis of the Malus law; based on a long short-term memory network, polarized light intensity information, a sun vector, polarization sensor calibration parameters, a polarization sensor field angle, an inner wall reflection coefficient and a diffuse reflection coefficient are used as inputs, sunlight interference transmission light intensity and a sunlight proportion are used as outputs, and a polarization angle after sunlight interference compensation is used as a training target. Learning a sunlight interference transmission light intensity equation; influences of linearly polarized light, vertical incidence unpolarized light and sunlight on the polarization sensor are comprehensively considered, a polarization sensor model considering sunlight interference is established by combining a light intensity gain coefficient of the polarization sensor, and polarization information such as a polarization angle and a polarization degree is solved.

Description

Technical Field

[0001] The present invention relates to a polarization sensor modeling method considering sunlight interference, which realizes the learning of sunlight interference by using a long short-term memory network, establishes a polarization sensor model considering sunlight interference, and can realize estimation and compensation of sunlight interference at any time and in any observation direction, thereby improving the angular measurement accuracy of the polarization sensor. Background Art

[0002] Inspired by the navigation mechanisms of animals such as desert ants, bees and butterflies, bionic polarization navigation technology has received widespread attention. This technology has the advantages of anti-electromagnetic interference and no error accumulation, and has been successfully applied to fields such as drones, ground robots and underwater robots. In the bionic polarization navigation system, polarization sensors play a key role. They can sense the polarization information in the sky and realize the inversion interpretation of navigation information. As the "soul" of the polarization sensor, the polarization sensor model can realize the characterization of the sensor input / output and key parameters, and is an important guarantee for high-precision polarization sensors. Therefore, how to design a fine modeling method for polarization sensors in complex interference environments is the key to improving sensor performance.

[0003] At present, researchers have conducted a lot of research on polarization sensor modeling. For example, Chinese invention patent CN201810129371.8 ("Multi-source error calibration method of biomimetic polarization sensor based on unscented Kalman filter") conducted an in-depth analysis of multi-source errors such as installation error, measurement noise and proportional factor of polarization sensor, constructed a polarization sensor model, and realized the calibration and compensation of multi-source errors by using unscented Kalman filter technology without relying on high-precision instruments and equipment. Chinese invention patent CN202310212296.2 ("A polarization zero position calibration system for polarization sensor") proposes a polarization zero position calibration system for polarization sensor, which realizes the precise calibration of zero-degree polarization direction of polarization sensor by integrating components such as integrating sphere, polarizer and turntable, eliminating installation errors in navigation applications. Chinese invention patent CN202010475084.X ("A bionic polarization compass calibration method based on polarization two-dimensional residual information") proposes a bionic polarization compass calibration method based on polarization two-dimensional residual information. This method introduces the extinction ratio coefficient and the polarizer installation error, and improves the accuracy and stability of polarization solution through parameter iteration estimation. However, the above-mentioned polarization sensor modeling method only considers the internal parameter errors of the sensor, and has not fully considered the influence of external interference factors. In actual environments, sunlight interference, as a kind of unpolarized light interference, has a particularly significant impact on polarization sensors. When sunlight is incident on the polarization sensor, there are complex processes such as inner wall reflection and diffuse reflection, which seriously affect the angular measurement accuracy of the polarization sensor. Therefore, how to achieve a fine characterization of sunlight interference, establish a polarization sensor model considering sunlight interference, and improve the angular measurement accuracy of the polarization sensor needs to be studied urgently. Summary of the invention

[0004] In order to solve the above problems and overcome the shortcomings of the prior art, the present invention proposes a polarization sensor modeling method considering sunlight interference, which divides the light incident on the polarization sensor into three parts: linear polarized light, vertical incident unpolarized light and sunlight. On the basis of Malus's law, the influence of sunlight interference on the polarization sensor is further considered to characterize the sunlight interference transmission light intensity; based on the long short-term memory network, the polarization intensity information, solar vector, polarization sensor calibration parameters, polarization sensor field angle, inner wall reflection coefficient and diffuse reflection coefficient are used as input, and the sunlight interference transmission light intensity and sunlight proportion are used as output. The polarization angle after sunlight interference compensation is used as the training target to learn the sunlight interference transmission light intensity equation and sunlight proportion; then, the influence of linear polarized light, vertical incident unpolarized light and sunlight on the polarization sensor is comprehensively considered, and the light intensity gain coefficient of the polarization sensor is further combined to establish a polarization sensor model considering sunlight interference. Finally, based on the established polarization sensor model, the estimation and compensation of the sunlight interference transmission light intensity are completed, and the polarization angle and polarization degree are solved. It can estimate and compensate for the intensity of sunlight interference transmitted at any time and in any observation direction, thus improving the angle measurement accuracy of the polarization sensor.

[0005] The technical solution of the present invention is: a polarization sensor modeling method considering sunlight interference, and the implementation steps are as follows:

[0006] Step (1), the light incident on the polarization sensor is divided into three parts: linear polarized light, vertically incident unpolarized light and sunlight. The output light intensity of the polarization sensor is expressed as: is the linear polarized light transmission intensity of the ith optical path channel, is the vertical incident unpolarized light transmission intensity of the i-th optical path channel, is the sunlight interference transmission intensity of the i-th optical path channel, i=1,2,...,N represents the i-th channel of the polarization sensor, and N is the total number of channels. Characterized by I in Indicates the input light intensity, represents the extinction ratio coefficient of the polarizer, d represents the degree of polarization, represents the polarization angle, θ i Indicates the polarizer installation angle; Characterized by k μ is the proportion of sunlight in the incident light. Further considering the impact of sunlight on the polarization sensor, Characterized by Where T i 、T r and T s They represent the transmittance of direct sunlight, reflected light from the inner wall of the polarization sensor, and diffuse reflected light in the polarizer, respectively.μd and k μr They represent the proportion of direct sunlight and reflected light in sunlight interference, K r and K s They represent the inner wall reflection and diffuse reflection loss coefficients of the polarization sensor respectively.

[0007] Step (2), for the sunlight interference transmitted light intensity in step (1) The long short-term memory network is used for learning. i 、Sun vectors n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , polarization sensor inner wall reflection coefficient τ r and diffuse reflectance τ d As training input. Use sunlight to interfere with the transmitted light intensity and the proportion of sunlight in the incident light k μ As output, the polarization angle of sunlight interfering with the transmitted light intensity will be compensated As a training target. Use the reference navigation information to infer the ideal polarization angle information And use it as a label to learn and k μ , where Sg(·) represents the established sunlight interference transmission light intensity equation.

[0008] Step (3), combining step (1) and step (2), considering the influence of linear polarized light, vertical incident unpolarized light and sunlight on the polarization sensor, combining the light intensity gain coefficient of the polarization sensor Build a polarization sensor model that takes sunlight interference into account:

[0009] Step (4): Based on the polarization sensor model established in step (3), the polarized light intensity y at any time and in any observation direction is calculated. i 、Sun vector information n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, the transmitted light intensity of sunlight interference is realized and sunlight ratio k μ The output light intensity of the polarization sensor is used to compensate for the interference of sunlight on the transmitted light intensity, and the polarization information such as the polarization angle and degree of polarization is calculated.

[0010] Furthermore, in step (1), the light incident on the polarization sensor is divided into three parts: linearly polarized light, vertically incident unpolarized light, and sunlight. The output light intensity of the polarization sensor is expressed as: is the linear polarized light transmission intensity of the ith optical path channel, is the vertical incident unpolarized light transmission intensity of the i-th optical path channel, is the sunlight interference transmission intensity of the i-th optical path channel, i=1,2,...,N represents the i-th channel of the polarization sensor, and N is the total number of channels. Characterized by I in Indicates the input light intensity, represents the extinction ratio coefficient of the polarizer, d represents the degree of polarization, represents the polarization angle, θ i Indicates the polarizer installation angle; Characterized by k μ is the proportion of sunlight in the incident light. Further considering the impact of sunlight on the polarization sensor, Characterized by Where T i 、T r and T s They represent the transmittance of direct sunlight, reflected light from the inner wall of the polarization sensor, and diffuse reflected light in the polarizer, respectively. μd and k μr They represent the proportion of direct sunlight and reflected light in sunlight interference, K r and K s They represent the inner wall reflection and diffuse reflection loss coefficients of the polarization sensor respectively.

[0011] Furthermore, in step (2), the sunlight interference transmission light intensity in step (1) is The long short-term memory network is used for learning. i 、Sun vectors n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , polarization sensor inner wall reflection coefficient τ r and diffuse reflectance τ d As training input. Use sunlight to interfere with the transmitted light intensity and sunlight accounts for k of the incident light μ As output, the polarization angle of sunlight interfering with the transmitted light intensity will be compensated As a training target. Use the heading angle, pitch angle and roll angle information provided by high-precision inertial navigation to infer the ideal polarization angle information And use it as a label to learn and kμ , where Sg(·) represents the established sunlight interference transmission light intensity equation.

[0012] Using polarization sensors to obtain polarized light intensity information i , using the solar calendar to calculate the solar vector s n . Calculate the field of view angle of the inner wall based on the optical path parameters of the polarization sensor: Where γ is the bottom radius of the inner wall of the polarization sensor, and f is the vertical distance from the top of the inner wall to the photosensitive chip. Based on the material of the polarization sensor, the inner wall reflection coefficient is τ r , diffuse reflectance is τ d .

[0013] To obtain the standard polarization angle First, based on the Rayleigh scattering model, the relationship between the polarization vector and the solar vector is established:

[0014]

[0015] The sun vector is represented by s n =[sinA s oeLh s cosA s oeLh s s1H s ] T 3×1 , the attitude transformation matrix from the carrier coordinate system to the geographic coordinate system is expressed as Polarization vector e b It is expressed as:

[0016]

[0017] Furthermore, the standard polarization angle can be obtained through the polarization vector The specific expression is as follows:

[0018]

[0019] in, Represents the attitude transformation matrix The value of the i-th row and j-th column, i, j = 1, 2, 3, A s and H s Represent the solar azimuth and altitude respectively.

[0020] In the study of sunlight interference with transmitted light intensity, polarized light intensity information is used i 、Sun vectors n , polarization sensor calibration parameters C p and the polarization sensor field of view v a , inner wall reflection coefficient τ rand diffuse reflectance τ d As training input, predict the intensity of sunlight interfering with the transmitted light and obtain the polarization angle after interference compensation Use the polarization angle obtained by inversion The sunlight interference transmission intensity equation is trained as the true value label. During the training process, local features are extracted through a one-dimensional convolutional layer, and the long short-term memory network is used to capture the time dependency in the data; finally, the estimation ability of the sunlight interference transmission intensity is further enhanced through a fully connected layer. The loss function of the training can be expressed as:

[0021]

[0022] Among them, σ and ρ are weights used to balance the mean square error and standard deviation, and RMSE and STD represent the mean square error and standard deviation of the predicted value and the true value, respectively.

[0023] Use polarization intensity information i 、Sun vectors n , polarization sensor calibration parameters C p and the polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, based on the label value of formula (3) and the loss function of formula (4), the equation for the sunlight interference transmission intensity is learned and the proportion of sunlight in the incident light k μ .

[0024] Furthermore, in step (3), in combination with step (1) and step (2), the influence of linear polarized light, vertically incident unpolarized light and sunlight on the polarization sensor is comprehensively considered, and the relationship between the output light intensity and the incident light intensity after the action of the polarization device is established:

[0025]

[0026] Further combined with the light intensity gain coefficient of the polarization sensor Build a polarization sensor model that takes sunlight interference into account:

[0027]

[0028] Furthermore, in step (4), based on the polarization sensor model established in step (3), the polarized light intensity y at any time and in any observation direction is converted into i 、Sun vector information n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τd As input, the transmitted light intensity of sunlight interference is realized and sunlight accounts for k of the incident light μ The output light intensity of the polarization sensor is compensated for the interference of sunlight on the transmitted light intensity, and the polarization information such as polarization angle and polarization degree are calculated. After compensation, the output light intensity of the polarization sensor is

[0029]

[0030] Combining formulas (6) and (7), the polarization sensor model considering sunlight interference is expressed in matrix form: in They represent the output light intensity of the four channels of the sensor respectively, k represents the kth sampling point, and n represents the total number of sampling points. is the polarization sensor parameter matrix, and the polarization angle is The degree of polarization is d o =[d o1 ,d o2 ,...,d on ] n×1 , is a vector containing the polarization angle and degree of polarization. The least squares estimate of x is as follows:

[0031]

[0032] Polarization angle after sunlight interference compensation Degree of polarization d o As follows:

[0033]

[0034] in, and Respectively represent vector Items 1, 2 and 3 of .

[0035] The advantages of the present invention over the prior art are: a polarization sensor modeling method taking sunlight interference into consideration is proposed, a long short-term memory network is used to achieve fine characterization and learning of sunlight interference, and estimation and compensation of sunlight interference can be achieved at any time and in any observation direction, thereby improving the angular measurement accuracy of the polarization sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of a polarization sensor modeling method considering sunlight interference of the present invention; DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the protection scope of the present invention.

[0038] Polarization sensors mainly sense polarized light after sunlight is scattered by the atmosphere, but since polarization sensors have a certain viewing angle, sunlight interference will also be incident on the sensor optical path. The present invention proposes a polarization sensor modeling method that takes sunlight interference into account. By using the long short-term memory network method, an equation characterizing the intensity of sunlight interference transmission is constructed, and the proportion of sunlight in the incident light is predicted. A polarization sensor model that takes sunlight interference into account is established, which realizes the estimation and compensation of the intensity of sunlight interference transmission at any time and in any observation direction, thereby improving the angle measurement accuracy of the polarization sensor.

[0039] The specific implementation steps of the present invention are as follows:

[0040] Step (1), the light incident on the polarization sensor is divided into three parts: linear polarized light, vertically incident unpolarized light and sunlight. The output light intensity of the polarization sensor is expressed as: is the linear polarized light transmission intensity of the ith optical path channel, is the vertical incident unpolarized light transmission intensity of the i-th optical path channel, is the sunlight interference transmission intensity of the i-th optical path channel, i=1,2,...,N represents the i-th channel of the polarization sensor, and N is the total number of channels. Characterized by I in Indicates the input light intensity, represents the extinction ratio coefficient of the polarizer, d represents the degree of polarization, represents the polarization angle, θ i Indicates the polarizer installation angle; Characterized by k μ is the proportion of sunlight in the incident light. Further considering the impact of sunlight on the polarization sensor, Characterized by Where T i 、T r and T s They represent the transmittance of direct sunlight, reflected light from the inner wall of the polarization sensor, and diffuse reflected light in the polarizer, respectively. μd and k μr They represent the proportion of direct sunlight and reflected light in sunlight interference, K r and K sThey represent the inner wall reflection and diffuse reflection loss coefficients of the polarization sensor respectively.

[0041] Step (2), for the sunlight interference transmitted light intensity in step (1) The long short-term memory network is used for learning. i 、Sun vectors n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , polarization sensor inner wall reflection coefficient τ r and diffuse reflectance τ d As training input. Use sunlight to interfere with the transmitted light intensity and the proportion of sunlight in the incident light k μ As output, the polarization angle of sunlight interfering with the transmitted light intensity will be compensated As a training target. Use the heading, pitch and roll angle information provided by the high-precision inertial navigation system (INS900A) to infer the ideal polarization angle information. And use it as a label to learn and k μ , where Sg(·) represents the established sunlight interference transmission light intensity equation.

[0042] Using polarization sensors to obtain polarized light intensity information i , using the solar calendar to calculate the solar vector s n . Calculate the field of view angle of the inner wall based on the optical path parameters of the polarization sensor: Where γ is the bottom radius of the inner wall of the polarization sensor, and f is the vertical distance from the top of the inner wall to the photosensitive chip. Based on the material of the polarization sensor, the inner wall reflection coefficient is τ r , diffuse reflectance is τ d .

[0043] To obtain the standard polarization angle First, based on the Rayleigh scattering model, the relationship between the polarization vector and the solar vector is established:

[0044]

[0045] The sun vector is represented by s n =[sinA s oeLh s cosA s oeLh s s1H s ] T 3×1 , the attitude transformation matrix from the carrier coordinate system to the geographic coordinate system is expressed as Polarization vector e b It is expressed as:

[0046]

[0047] Furthermore, the standard polarization angle can be obtained through the polarization vector The specific expression is as follows:

[0048]

[0049] in, Represents the attitude transformation matrix The value of the i-th row and j-th column, i, j = 1, 2, 3, A s and H s Represent the solar azimuth and altitude respectively.

[0050] In the study of sunlight interference with transmitted light intensity, polarized light intensity information is used i 、Sun vectors n , polarization sensor calibration parameters C p and the polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As training input, predict the intensity of sunlight interfering with the transmitted light and obtain the polarization angle after interference compensation Use the polarization angle obtained by inversion The sunlight interference transmission intensity equation is trained as the true value label. During the training process, local features are extracted through a one-dimensional convolutional layer, and the long short-term memory network is used to capture the time dependency in the data; finally, the estimation ability of the sunlight interference transmission intensity is further enhanced through a fully connected layer. The loss function of the training can be expressed as:

[0051]

[0052] Among them, σ and ρ are weights used to balance the mean square error and standard deviation, and RMSE and STD represent the mean square error and standard deviation of the predicted value and the true value, respectively.

[0053] Use polarization intensity information i 、Sun vectors n , polarization sensor calibration parameters C p and the polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, based on the label value of formula (3) and the loss function of formula (4), the equation for the sunlight interference transmission intensity is learned and the proportion of sunlight in the incident light k μ .

[0054] Step (3), combining step (1) and step (2), comprehensively considering the influence of linear polarized light, vertically incident unpolarized light and sunlight on the polarization sensor, and establishing the relationship between the output light intensity and the incident light intensity after the action of the polarization device:

[0055]

[0056] Further combined with the light intensity gain coefficient of the polarization sensor Build a polarization sensor model that takes sunlight interference into account:

[0057]

[0058] Step (4): Based on the polarization sensor model established in step (3), the polarized light intensity y at any time and observation direction is calculated. i 、Sun vector information n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, the transmitted light intensity of sunlight interference is realized and sunlight accounts for k of the incident light μ The output light intensity of the polarization sensor is compensated for the interference of sunlight on the transmitted light intensity, and the polarization angle, polarization degree and other information are calculated. After compensation, the output light intensity of the polarization sensor is

[0059]

[0060] Combining formulas (6) and (7), the polarization sensor model considering sunlight interference is expressed in matrix form: in They represent the output light intensity of the four channels of the sensor respectively, k represents the kth sampling point, and n represents the total number of sampling points. is the polarization sensor parameter matrix, and the polarization angle is The degree of polarization is d o =[d o1 ,d o2 ,...,d on ] n×1 , is a vector containing the polarization angle and degree of polarization. The least squares estimate of x is as follows:

[0061]

[0062] Polarization angle after sunlight interference compensation Degree of polarization d o As follows:

[0063]

[0064] in, and Respectively represent vector Items 1, 2 and 3 of .

[0065] Although the above describes the illustrative specific embodiments of the present invention to facilitate those skilled in the art to understand the present invention, and it should be clear that the present invention is not limited to the scope of the specific embodiments, for those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations using the concept of the present invention are protected.

Claims

1. A polarization sensor modeling method considering sunlight interference, characterized in that: The implementation steps are as follows: Step (1), the light incident on the polarization sensor is divided into three parts: linear polarized light, vertically incident unpolarized light and sunlight. The output light intensity of the polarization sensor is expressed as: is the linear polarized light transmission intensity of the ith optical path channel, is the vertical incident unpolarized light transmission intensity of the i-th optical path channel, is the sunlight interference transmission intensity of the i-th optical path channel, i=1,2,...,N represents the i-th channel of the polarization sensor, and N is the total number of channels; where, Characterized by I in Indicates the input light intensity, represents the extinction ratio coefficient of the polarizer, d represents the degree of polarization, represents the polarization angle, θ i Indicates the polarizer installation angle; Characterized by k μ is the proportion of sunlight in the incident light; further considering the impact of sunlight on the polarization sensor, Characterized by Where T i , T r and T s They represent the transmittance of direct sunlight, reflected light from the inner wall of the polarization sensor, and diffuse reflected light in the polarizer, respectively. μd and k μr They represent the proportion of direct sunlight and reflected light in sunlight interference, K r and K s They represent the inner wall reflection and diffuse reflection loss coefficients of the polarization sensor respectively; Step (2), for the sunlight interference transmitted light intensity in step (1) Using long short-term memory network for learning; using polarization intensity information y i 、Sun vectors n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , polarization sensor inner wall reflection coefficient τ r and diffuse reflectance τ d As training input; use sunlight to interfere with the transmitted light intensity and the proportion of sunlight in the incident light k μ As output, the polarization angle of sunlight interfering with the transmitted light intensity will be compensated As a training target; use the reference navigation information to infer the ideal polarization angle information And use it as a label to learn and k μ , where Sg(·) represents the established solar interference transmission light intensity equation; Step (3), combining step (1) and step (2), considering the influence of linear polarized light, vertical incident unpolarized light and sunlight on the polarization sensor, and combining the light intensity gain coefficient K of the polarization sensor Ii , establish a polarization sensor model considering sunlight interference: Step (4): Based on the polarization sensor model established in step (3), the polarized light intensity y at any time and in any observation direction is calculated. i 、Sun vector information n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, the transmitted light intensity of sunlight interference is realized and sunlight ratio k μ The prediction of the polarization sensor output light intensity is used to compensate for the interference of sunlight on the transmitted light intensity, and the polarization information such as the polarization angle and degree of polarization is calculated.

2. The polarization sensor modeling method considering sunlight interference according to claim 1 is characterized in that: In the step (1), the light incident on the polarization sensor is divided into three parts: linearly polarized light, vertically incident unpolarized light, and sunlight. The output light intensity of the polarization sensor is expressed as: is the linear polarized light transmission intensity of the ith optical path channel, is the vertical incident unpolarized light transmission intensity of the i-th optical path channel, is the sunlight interference transmission intensity of the i-th optical path channel, i=1,2,...,N represents the i-th channel of the polarization sensor, and N is the total number of channels; where, Characterized by I in Indicates the input light intensity, represents the extinction ratio coefficient of the polarizer, d represents the degree of polarization, represents the polarization angle, θ i Indicates the polarizer installation angle; Characterized by k μ is the proportion of sunlight in the incident light. Further considering the impact of sunlight on the polarization sensor, Characterized by Where T i , T r and T s They represent the transmittance of direct sunlight, reflected light from the inner wall of the polarization sensor, and diffuse reflected light in the polarizer, respectively. μd and k μr They represent the proportion of direct sunlight and reflected light in sunlight interference, K r and K s They represent the inner wall reflection and diffuse reflection loss coefficients of the polarization sensor respectively.

3. The polarization sensor modeling method considering sunlight interference according to claim 1 is characterized in that: In the step (2), the sunlight interference transmitted light intensity in step (1) Using long short-term memory network for learning; using polarization intensity information y i 、Sun vectors n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , polarization sensor inner wall reflection coefficient τ r and diffuse reflectance τ d As training input; Interference of transmitted light intensity with sunlight and sunlight accounts for k of the incident light μ As output, the polarization angle of sunlight interfering with the transmitted light intensity will be compensated As a training target; use the heading angle, pitch angle and roll angle information provided by high-precision inertial navigation to infer the ideal polarization angle information And use it as a label to learn and k μ , where Sg(·) represents the established solar interference transmission light intensity equation; Using polarization sensors to obtain polarized light intensity information i , using the solar calendar to calculate the solar vector s n ; Calculate the field of view angle of the inner wall based on the optical path parameters of the polarization sensor: Where γ is the bottom radius of the inner wall of the polarization sensor, and f is the vertical distance from the top of the inner wall to the photosensitive chip. Based on the material of the polarization sensor, the inner wall reflection coefficient is τ r , diffuse reflectance is τ d ; To obtain the standard polarization angle First, based on the Rayleigh scattering model, the relationship between the polarization vector and the solar vector is established: The sun vector is represented by s n =[sinA s oeLh s cosA s oeLh s sI s ] T 3×1 , the attitude transformation matrix from the carrier coordinate system to the geographic coordinate system is expressed as Polarization vector e b It is expressed as: Furthermore, the standard polarization angle can be obtained through the polarization vector The specific expression is as follows: in, Represents the attitude transformation matrix The value of the i-th row and j-th column, i, j = 1, 2, 3, A s and H s Respectively represent the solar azimuth and altitude; In the study of sunlight interference with transmitted light intensity, polarized light intensity information is used i 、Sun vectors n , polarization sensor calibration parameters C p and the polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As training input, predict the intensity of sunlight interfering with the transmitted light and obtain the polarization angle after interference compensation Use the polarization angle obtained by inversion As the true value label, the equation of sunlight interference transmission intensity is trained; during the training process, the local features are extracted through the one-dimensional convolution layer, and the long short-term memory network is used to capture the time dependency in the data; finally, the estimation ability of sunlight interference transmission intensity is further enhanced through the fully connected layer; the loss function of the training can be expressed as: Among them, σ and ρ are weights used to balance the mean square error and standard deviation, RMSE and STD represent the mean square error and standard deviation of the predicted value and the true value respectively; Use polarization intensity information i 、Sun vectors n , polarization sensor calibration parameters C p and the polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, based on the label value of formula (3) and the loss function of formula (4), the equation for the sunlight interference transmission intensity is learned and the proportion of sunlight in the incident light k μ .

4. The polarization sensor modeling method considering sunlight interference according to claim 1 is characterized in that: In the step (3), in combination with the step (1) and the step (2), the influence of linear polarized light, vertically incident unpolarized light and sunlight on the polarization sensor is comprehensively considered, and the relationship between the output light intensity and the incident light intensity after the action of the polarization device is established: Further combined with the light intensity gain coefficient of the polarization sensor Build a polarization sensor model that takes sunlight interference into account:

5. The polarization sensor modeling method considering sunlight interference according to claim 1 is characterized in that: The step (4) is based on the polarization sensor model established in step (3), and the polarization intensity y at any time and in any observation direction is calculated. i 、Sun vector information n , polarization sensor calibration parameters C p 、Polarization sensor field of view v a , inner wall reflection coefficient τ r and diffuse reflectance τ d As input, the transmitted light intensity of sunlight interference is realized and sunlight accounts for k of the incident light μ The prediction of the output light intensity of the polarization sensor is that the sunlight interferes with the transmitted light intensity, and the polarization information such as the polarization angle and polarization degree are calculated. After compensation, the output light intensity of the polarization sensor is Combining formulas (6) and (7), the polarization sensor model considering sunlight interference is expressed in matrix form: in They represent the output light intensity of the four channels of the sensor, k represents the kth sampling point, and n represents the total number of sampling points; is the polarization sensor parameter matrix, and the polarization angle is The degree of polarization is d o =[d o1 ,d o2 ,...,d on ] n×1 , is a vector containing the polarization angle and degree of polarization; the least squares estimate of x is as follows: Polarization angle after sunlight interference compensation Degree of polarization d o As follows: in, and Respectively represent vector Items 1, 2 and 3 of .

Images