Edge-based background radiation two-dimensional dynamic scene modeling method based on autoregressive technique

Abstract

The application discloses a method for modeling a two-dimensional dynamic scene of an edge-of-sight background radiation based on an autoregressive technique, and relates to the technical field of atmospheric edge-of-sight imaging. According to sensor parameters and a radiation transmission integral equation, the method calculates the deterministic background radiation of an edge-of-sight observation. The method deforms a radiation transmission equation and combines Taylor expansion to obtain statistical information of the background radiation. The method generates random background fluctuations by two-dimensional step filtering, including fitting statistics, generating a Gaussian random matrix, solving an autocorrelation sequence by inverse Fourier transformation, obtaining autoregressive coefficients by solving a Yule-Walker equation, and generating a two-dimensional random fluctuation structure by longitudinal and horizontal filtering. The deterministic background and the random background are superimposed to obtain a two-dimensional scene radiation image of the edge-of-sight observation. The application solves the problems of large modeling deviation and weak pertinence, and improves the authenticity and accuracy of scene modeling by precisely coupling two background components.

Description

Technical Field

[0001] This invention relates to the field of atmospheric edge imaging technology, and in particular to a two-dimensional dynamic scene modeling method for edge background radiation based on autoregressive technology. Background Technology

[0002] The importance of atmospheric limb environments in basic science, climate and environmental monitoring, and optoelectronic engineering is becoming increasingly prominent. Understanding the characteristics of atmospheric limb environments is crucial for the development of key technologies such as target optical detection and satellite remote sensing. Limb observation, as an important method of atmospheric detection, can obtain information on the vertical structure of the middle and upper atmosphere. The two-dimensional scene characteristics of its background radiation directly affect the detection accuracy and the precision of data interpretation.

[0003] Currently, existing methods for modeling background radiation at the edge of the atmosphere mostly focus on solving a single component. They either only consider the distribution of deterministic background radiation, ignoring the influence of random background fluctuations, resulting in a large deviation between the modeling results and the real atmospheric scene; or they use the spatial mean and variance of the radiation scene to generate perturbation images for modeling random background fluctuations, without being based on physical mechanisms. This makes it difficult to describe the actual physical origin of background fluctuations and cannot accurately reflect the random changes caused by factors such as temperature perturbation and molecular number density perturbation, thus failing to meet the requirements of high-precision detection.

[0004] Furthermore, existing methods employ relatively simple coupling mechanisms between deterministic and random backgrounds, lacking a systematic modeling process. This results in generated two-dimensional scene radiation images that cannot accurately match the actual needs of edge observation, limiting their application in fields such as target optical detection and satellite remote sensing. Therefore, there is an urgent need for a two-dimensional scene modeling method for edge observation background radiation that can accurately couple deterministic and random backgrounds and closely reflect real atmospheric characteristics. Summary of the Invention

[0005] To overcome the shortcomings of the existing technology, this invention provides a two-dimensional dynamic scene modeling method for edge background radiation based on autoregressive technology. By solving for deterministic background radiation and random background fluctuations, the method achieves precise coupling between the two, generating a two-dimensional scene radiation image that closely matches the characteristics of the real atmosphere, thereby improving the interpretation accuracy and application value of edge observation data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution, including: A two-dimensional dynamic scene modeling method based on autoregressive background radiation includes: Step 1: Calculate the deterministic background radiation in the two-dimensional scene at the edge: Based on the sensor position and height, the tangent point position and height of the edge observation, and the sensor's field of view, the background radiation characteristics of different pixels within the sensor's field of view under different observation zenith angles and azimuth angles are calculated iteratively using the atmospheric radiative transfer integral equation to obtain the deterministic background radiation. Step 2: Calculate the statistical information of the lateral background radiation: Transform the atmospheric radiative transfer integral equation in Step 1 to obtain the background radiance expression within the spectral bandwidth. Perform Taylor expansion on the radiation fluctuation according to the vibrational temperature and dynamic temperature of atmospheric molecules to obtain the covariance of any two points on the sensor. Then, fit the spectral coefficients, standard deviation and correlation length of the background radiation. Step 3: Generate the random background undulation structure in the two-dimensional scene at the edge: Combine the statistical information of the edge background radiation and use the inverse Fourier transform to perform power spectrum filtering on the horizontal direction of Gaussian white noise, and use the statistical information of the background radiation and combine the autoregressive moving average model to perform dynamic spatial sequence filtering on the vertical direction of Gaussian white noise to obtain the horizontal one-dimensional static radiation undulation structure and the vertical one-dimensional dynamic radiation undulation structure. Then couple the two to obtain the random background undulation structure. Step 4: Synthesize the two-dimensional scene radiation image of the edge observation: The deterministic background radiation obtained in Step 1 is superimposed with the random background undulation structure obtained in Step 3 to obtain the two-dimensional scene radiation image of the edge observation.

[0007] Preferably, in step one, the lateral background radiation characteristics of the sensor pixel at a certain observation zenith angle and azimuth angle are as follows: ;

[0008] in, Indicates the sensor observation position At this location, the wave number is Background radiance; Let be the integral height variable along the observation path, representing a point on the adjacent path. and sensor observation position A certain position between; Indicates position Location, wave number The source function term characterizes the radiation emitted by the atmosphere itself; Indicates the spatial position (height) of the sensor; Indicates wave number; Indicates the height of the tangent point observed at the edge; Indicates position Location, wave number Atmospheric absorption coefficient; Indicates position Atmospheric molecular density at that location; Indicates from position Arrive at the location , wave number Atmospheric transmittance.

[0009] Preferably, atmospheric transmittance Specifically: ;

[0010] in, Indicates from position Arrive at the location A certain position; Indicates position Location, wave number Atmospheric absorption coefficient; Indicates position Atmospheric molecular density at that location.

[0011] Preferably, in step two, the atmospheric radiative transfer integral equation in step one is transformed to obtain the expression for the radiance of the adjacent background within the spectral bandwidth, specifically: ;

[0012] in, Indicates the observed spectral bandwidth; Indicates spectral bandwidth Internal, Location Background radiance at the edge; Indicates the wavelength of radiation; Indicates position Location, wavelength The source function term.

[0013] Preferably, for radiation fluctuations A Taylor expansion is performed based on the vibrational and kinetic temperatures of atmospheric molecules to obtain the covariance between any two points on the sensor, specifically: ;

[0014] in, Indicates position Radiation covariance; This indicates the height of two different spatial positions within the sensor's field of view; This represents the integral height variable along the two observation paths; Indicates position Covariance of temperature fluctuations; Representing positions respectively The weighting function for radiation fluctuation response.

[0015] Preferably, the radiation fluctuation response weighting function is as follows: ; ; ;

[0016] in, The radiation response function representing the temperature correlation of vibration; The radiation response function directly related to the dynamic temperature; Indicates the different vibrational states of atmospheric molecules; The first molecule representing atmospheric molecules Vibrational temperature; Indicates the dynamic temperature of atmospheric molecules; This represents the rate of change of the vibrational temperature with respect to the dynamic temperature.

[0017] Preferably, step three is as follows: S1. Fit the statistical information of the background radiation at the edge within the sensor's field of view to obtain the spectral coefficients, standard deviations, and correlation lengths at different observation zenith angles; S2. Set the number of pixels used to overcome the relaxation time of the autoregressive filter. According to the number of sensor pixels The number of generated pixels is A Gaussian random matrix with a mean of zero and a variance of 1; where, The sensor is configured with vertical pixel count and horizontal pixel count respectively; S3. Perform an inverse Fourier transform on the atmospheric power spectral density to obtain the autocorrelation function; input the spectral coefficients, standard deviations, and correlation lengths corresponding to the maximum observed zenith angle within the field of view to obtain the autocorrelation sequence used for the relaxation time of the autoregressive filter; S4. Set the order of the autoregressive filter and, in conjunction with the autocorrelation sequence under the maximum observed zenith angle, solve the Yule-Walker equation under the autoregressive moving average model to obtain the coefficients of the autoregressive spatial sequence. S5. Using autoregressive spatial sequences to apply the first Gaussian random matrix The rows are filtered to obtain a spatial sequence used to overcome the relaxation time of the autoregressive filter. S6, the first part of the Gaussian random matrix The row replacement is a spatial sequence used to overcome relaxation time, and the Gaussian random matrix is ​​then subjected to a subsequent... Okay, using the spectral coefficients, standard deviations, and correlation lengths corresponding to different observed zenith angles within the field of view, the coefficients of the spatial sequence at different observed zenith angles are obtained by solving the Yule-Walker equation; S7. Using spatial sequences at different observed zenith angles to refine the Gaussian random matrix. Perform autoregressive filtering on the rows and extract the filtered spatial sequence. Arriving Okay, a one-dimensional dynamic radiation fluctuation structure is obtained in the vertical direction of the Gaussian random matrix; S8. Construct different power spectral density functions based on the spectral coefficients, standard deviations, and correlation lengths under different observation zenith angles within the field of view. S9. Use the power spectral density function to perform spatial filtering on the row direction of the Gaussian random matrix obtained in step S7, and obtain a one-dimensional static radiation fluctuation structure in the horizontal direction of the Gaussian random matrix.

[0018] The present invention also provides a readable storage medium having a computer program stored thereon, wherein the computer program, when executed, implements the two-dimensional dynamic scene modeling method for adjacent background radiation based on autoregressive technology.

[0019] The present invention also provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned two-dimensional dynamic scene modeling method for adjacent background radiation based on autoregressive technology.

[0020] The present invention also provides a computer program product, which includes a computer program / instruction that, when executed by a processor, implements the aforementioned two-dimensional dynamic scene modeling method for adjacent background radiation based on autoregressive technology.

[0021] The advantages of this invention are: (1) This invention considers both deterministic limb background radiation and random limb background fluctuations. It accurately solves the deterministic background through the atmospheric radiation transfer equation and generates the random background by combining statistical analysis and two-dimensional step-by-step filtering. This achieves the system coupling of the two and solves the problem of large modeling deviation caused by the existing method only considering a single component. This makes the generated two-dimensional limb scene radiation image more in line with the real atmospheric limb environment.

[0022] (2) The modeling of random edge background fluctuations adopts a two-dimensional stepwise approach of longitudinal autoregressive filtering and transverse power spectrum filtering. In the longitudinal direction, dynamic sequences are generated by combining the autoregressive moving average model with the Yule-Walker equation to fit the atmospheric turbulence evolution characteristics. In the transverse direction, static structures are generated by Fourier power spectrum filtering to match horizontal spatial correlation. At the same time, a buffer row is set to overcome filter relaxation distortion and improve the modeling accuracy of random background fluctuations.

[0023] (3) The entire modeling process is complete, from deterministic background solution and statistical calculation to random background generation and two-dimensional scene synthesis. The steps are clear and the logic is rigorous. It can be directly applied to fields such as target optical detection and satellite remote sensing detection, providing technical support for the accurate interpretation of edge observation data and enhancing the application value of edge observation technology.

[0024] (4) During the modeling process, filtering parameters were determined for different observation zenith angles so that the model could be adapted to the atmospheric characteristics of different observation angles, thereby enhancing the versatility and applicability of the method. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of a two-dimensional scene for edge observation.

[0026] Figure 2 A schematic diagram illustrating the process of generating random background undulations in a two-dimensional scene.

[0027] Figure 3 This is a two-dimensional scene radiation image observed at the edge. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] The two-dimensional scene with lateral background radiation is determined by two parts: deterministic lateral background radiation and random lateral background undulation.

[0030] For deterministic lateral background radiation, it usually refers to the background radiation distribution observed by the sensor within a certain field of view. It is deterministic radiation produced by stable distributed substances such as atmospheric molecules, and has a spatial distribution law that can be accurately calculated. Generally, if the position and height of the sensor, the position and height of the tangent point of the lateral observation, and the field of view of the sensor are known, the background radiation characteristics of the sensor pixel under a certain observation zenith angle and azimuth angle can be calculated by the following formula (1): (1) in, Indicates the sensor observation position At this location, the wave number is Background radiance; Let be the integral height variable along the observation path, representing a point on the adjacent path. and sensor observation position A certain position between; Indicates position Location, wave number The source function term characterizes the radiation emitted by the atmosphere itself; Indicates the spatial position (height) of the sensor; Indicates wave number; Indicates the height of the tangent point observed at the edge; Indicates position Location, wave number Atmospheric absorption coefficient; Indicates position Atmospheric molecular density at that location; Indicates from position Arrive at the location , wave number Atmospheric transmittance.

[0031] It can be expressed as the following formula (2): (2) in, Indicates from position Arrive at the location A certain position; Indicates position Location, wave number Atmospheric absorption coefficient; Indicates position Atmospheric molecular density at that location.

[0032] By combining equations (1) and (2), the tangent height corresponding to different pixels within the sensor's field of view and the background radiance value under different azimuth angles are calculated iteratively, and a deterministic background in the two-dimensional scene can be obtained.

[0033] like Figure 1 As shown, Figure 1 This is a schematic diagram of a two-dimensional scene for edge observation. Figure 1 middle, Represents the Earth's center. Indicates the sensor's observation position. Indicates the satellite's orbital altitude. This indicates the surface thickness of the detection path corresponding to different detection pixels within the field of view.

[0034] For random background fluctuations in a two-dimensional scene, they are usually caused by fluctuations in atmospheric molecular number density, population, and dynamic temperature along the line of sight, exhibiting non-steady-state and random disturbance characteristics, and playing an important role in the observation of the edge environment of the middle and upper atmosphere. Solving for random background fluctuations usually requires statistical information on background radiation. In order to obtain the variance and covariance information of the background radiation at the sensor location, equation (1) needs to be transformed to obtain the following equation (3): (3) in, Indicates the observed spectral bandwidth; Indicates spectral bandwidth Internal, Location Background radiance at the edge; Indicates the wavelength of radiation; Indicates position Location, wavelength The source function term.

[0035] Equation (3) considers the background radiance value within a certain spectral band. Based on equation (3), and considering the radiance fluctuation... (Indicates radiance) The small changes) can be obtained by performing a Taylor expansion based on the vibrational temperature and dynamic temperature of atmospheric molecules, and the covariance of any two points on the sensor can be expressed as follows (4): (4) in, Indicates position Radiation covariance; This indicates the height of two different spatial positions within the sensor's field of view; This represents the integral height variable along the two observation paths; Indicates position Covariance of temperature fluctuations; Representing positions respectively The weighting function for radiation fluctuation response.

[0036] This represents the radiation fluctuation response function. .

[0037] and The expression can be determined by the following equations (5) and (6): (5) (6) in, The radiation response function representing the temperature correlation of vibration; The radiation response function directly related to the dynamic temperature; Indicates the different vibrational states of atmospheric molecules; The first molecule representing atmospheric molecules Vibrational temperature; Indicates the dynamic temperature of atmospheric molecules; This represents the rate of change of the vibrational temperature with respect to the dynamic temperature.

[0038] The radiation statistics obtained through calculation can then be used to... Figure 2 The processing procedure described herein calculates the random background undulation structure in a two-dimensional scene. Its main idea is to combine statistical information and employ inverse Fourier transform to perform power spectral filtering on the horizontal static direction of Gaussian white noise, and to utilize radiation statistics and combine them with an autoregressive moving average model to perform dynamic spatial sequence filtering on the vertical direction of Gaussian white noise. Specifically, it includes the following steps: S1. The statistical information of the background radiation at the edge within the sensor's field of view is fitted to obtain information such as the spectral coefficient, standard deviation, and correlation length at different observation zenith angles.

[0039] S2. Set the number of pixels used to overcome the relaxation time of the autoregressive filter. Based on the number of sensor pixels (assuming the sensor has 1000 pixels) , (The sensor has vertical and horizontal pixel counts respectively), and the generated pixel size is... A Gaussian random matrix with a mean of zero and a variance of 1.

[0040] S3. Perform an inverse Fourier transform on the atmospheric power spectral density to obtain the autocorrelation function. By inputting the spectral coefficients, standard deviations, and correlation lengths corresponding to the maximum observed zenith angle within the field of view, an autocorrelation sequence for the relaxation time of the autoregressive filter can be obtained.

[0041] S4. Set the order of the autoregressive filter and, in conjunction with the autocorrelation sequence under the maximum observed zenith angle, solve the Yule-Walker equation under the autoregressive moving average model to obtain the coefficients of the autoregressive spatial sequence.

[0042] The following is a brief introduction to the specific principles of solving the Yule-Walker equation under the autoregressive moving average model. Generally speaking, in the autoregressive moving average model, the spatial sequence can usually be expressed as follows (7): (7) in, Represents the spatial sequence number 1 Random fluctuation values ​​at each point; Indicates the order of the autoregressive filter; Represents the autoregressive coefficients to be solved; A numerical value representing the preceding time (spatial position) of a spatial sequence; This represents a Gaussian random matrix with a mean of zero and a variance of 1.

[0043] Therefore, to obtain the coefficients of a spatial sequence, the core problem is equivalent to how to solve for the coefficients of the spatial sequence based on the autocorrelation sequence. A simple method is to multiply both sides of equation (7). By taking the spatial average of both sides of equation (7), the autocorrelation function can be constructed, which can be expressed as equation (8) below: (8) in, Represents the autocorrelation function, the autocorrelation statistic of a spatial sequence; This represents the autocorrelation value under different delays; ; This represents the spatial average. By solving the Yule-Walker equation for the above spatial sequence, the spatial sequence determined by the autocorrelation function can be obtained.

[0044] S5. Using autoregressive spatial sequences to apply the first Gaussian random matrix The rows are filtered to obtain a spatial sequence used to overcome the relaxation time of the autoregressive filter.

[0045] S6, the first part of the Gaussian random matrix The row replacement is a spatial sequence used to overcome relaxation time, and the Gaussian random matrix is ​​then subjected to a subsequent... Okay, using the spectral coefficients, standard deviations, and correlation lengths corresponding to different observed zenith angles within the field of view, the coefficients of the spatial sequence under different observed zenith angles are obtained by solving the Yule-Walker equation.

[0046] S7. Using spatial sequences at different observed zenith angles to refine the Gaussian random matrix. Perform autoregressive filtering on the rows, and truncate the filtered spatial sequence (size: )of Arriving If so, a one-dimensional dynamic radiation fluctuation structure can be obtained in the vertical direction of the Gaussian random matrix.

[0047] S8. Based on the spectral coefficients, standard deviations, and correlation lengths under different observation zenith angles within the field of view, construct different power spectral density functions.

[0048] S9. By using the power spectral density function to perform spatial filtering on the row direction of the Gaussian random matrix obtained in step S7, a one-dimensional static radiation fluctuation structure can be obtained in the horizontal direction of the Gaussian random matrix.

[0049] Based on the above steps, the random background undulation structure in the 2D scene at the edge can be calculated from the statistical information of the edge background radiation. By synthesizing the deterministic background and the random background, a 2D scene radiation image under edge observation conditions can finally be obtained. Figure 3The paper presents a 2D radiative image of a lateral scene during a mid-latitude summer daytime, with a sensor altitude of 500km, an instantaneous field of view of 0.005 degrees, and a central tangent point height of 80km within the field of view. The left side represents the random background undulations in the 2D scene, while the right side represents the deterministic background radiance.

[0050] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A two-dimensional dynamic scene modeling method based on adjacent background radiation using autoregressive techniques, characterized in that, include: Step 1: Calculate the deterministic background radiation in the two-dimensional scene at the edge: Based on the sensor position and height, the tangent point position and height of the edge observation, and the sensor's field of view, the background radiation characteristics of different pixels within the sensor's field of view under different observation zenith angles and azimuth angles are calculated iteratively using the atmospheric radiative transfer integral equation to obtain the deterministic background radiation. Step 2: Calculate the statistical information of the lateral background radiation: Transform the atmospheric radiative transfer integral equation in Step 1 to obtain the background radiance expression within the spectral bandwidth. Perform Taylor expansion on the radiation fluctuation according to the vibrational temperature and dynamic temperature of atmospheric molecules to obtain the covariance of any two points on the sensor. Then, fit the spectral coefficients, standard deviation and correlation length of the background radiation. Step 3: Generate the random background undulation structure in the two-dimensional scene at the edge: Combine the statistical information of the edge background radiation and use the inverse Fourier transform to perform power spectrum filtering on the horizontal direction of Gaussian white noise, and use the statistical information of the background radiation and combine the autoregressive moving average model to perform dynamic spatial sequence filtering on the vertical direction of Gaussian white noise to obtain the horizontal one-dimensional static radiation undulation structure and the vertical one-dimensional dynamic radiation undulation structure. Then couple the two to obtain the random background undulation structure. Step 4: Synthesize the two-dimensional scene radiation image of the edge observation: The deterministic background radiation obtained in Step 1 is superimposed with the random background undulation structure obtained in Step 3 to obtain the two-dimensional scene radiation image of the edge observation. Step three is detailed below: S1. Fit the statistical information of the background radiation at the edge within the sensor's field of view to obtain the spectral coefficients, standard deviations, and correlation lengths at different observation zenith angles; S2. Set the number of pixels used to overcome the relaxation time of the autoregressive filter. According to the number of sensor pixels The number of generated pixels is A Gaussian random matrix with a mean of zero and a variance of 1; where, The sensor is configured with vertical pixel count and horizontal pixel count respectively; S3. Perform an inverse Fourier transform on the atmospheric power spectral density to obtain the autocorrelation function; input the spectral coefficients, standard deviations, and correlation lengths corresponding to the maximum observed zenith angle within the field of view to obtain the autocorrelation sequence used for the relaxation time of the autoregressive filter; S4. Set the order of the autoregressive filter and, in conjunction with the autocorrelation sequence under the maximum observed zenith angle, solve the Yule-Walker equation under the autoregressive moving average model to obtain the coefficients of the autoregressive spatial sequence. S5. Using autoregressive spatial sequences to apply the first Gaussian random matrix The rows are filtered to obtain a spatial sequence used to overcome the relaxation time of the autoregressive filter. S6, the first part of the Gaussian random matrix The row replacement is a spatial sequence used to overcome relaxation time, and the Gaussian random matrix is ​​then subjected to a subsequent... Okay, using the spectral coefficients, standard deviations, and correlation lengths corresponding to different observed zenith angles within the field of view, the coefficients of the spatial sequence at different observed zenith angles are obtained by solving the Yule-Walker equation; S7. Using spatial sequences at different observed zenith angles to refine the Gaussian random matrix. Perform autoregressive filtering on the rows and extract the filtered spatial sequence. Arriving Okay, a one-dimensional dynamic radiation fluctuation structure is obtained in the vertical direction of the Gaussian random matrix; S8. Construct different power spectral density functions based on the spectral coefficients, standard deviations, and correlation lengths under different observation zenith angles within the field of view. S9. Use the power spectral density function to perform spatial filtering on the row direction of the Gaussian random matrix obtained in step S7, and obtain a one-dimensional static radiation fluctuation structure in the horizontal direction of the Gaussian random matrix.

2. The two-dimensional dynamic scene modeling method based on autoregressive background radiation according to claim 1, characterized in that, In step one, the lateral background radiation characteristics of a sensor pixel at a certain observation zenith angle and azimuth angle are specifically as follows: in, Indicates the sensor observation position At this location, the wave number is Background radiance; Let be the integral height variable along the observation path, representing a point on the adjacent path. and sensor observation position A certain position between; Indicates position Location, wave number The source function term characterizes the radiation emitted by the atmosphere itself; Indicates the spatial position (height) of the sensor; Indicates wave number; Indicates the height of the tangent point observed at the edge; Indicates position Location, wave number Atmospheric absorption coefficient; Indicates position Atmospheric molecular density at that location; Indicates from position Arrive at the location , wave number Atmospheric transmittance.

3. The two-dimensional dynamic scene modeling method based on autoregressive technology for adjacent background radiation according to claim 2, characterized in that, Atmospheric transmittance Specifically: in, Indicates from position Arrive at the location A certain position; Indicates position Location, wave number Atmospheric absorption coefficient; Indicates position Atmospheric molecular density at that location.

4. The two-dimensional dynamic scene modeling method based on autoregressive background radiation according to claim 2 or 3, characterized in that, In step two, the atmospheric radiative transfer integral equation from step one is transformed to obtain the expression for the lateral background radiance within the spectral bandwidth, specifically: in, Indicates the observed spectral bandwidth; Indicates spectral bandwidth Internal, Location Background radiance at the edge; Indicates the wavelength of radiation; Indicates position Location, wavelength The source function term.

5. The two-dimensional dynamic scene modeling method based on autoregressive technology for adjacent background radiation according to claim 4, characterized in that, radiation fluctuation A Taylor expansion is performed based on the vibrational and kinetic temperatures of atmospheric molecules to obtain the covariance between any two points on the sensor, specifically: in, Indicates position Radiation covariance; This indicates the height of two different spatial positions within the sensor's field of view; This represents the integral height variable along the two observation paths; Indicates position Covariance of temperature fluctuations; Representing positions respectively The weighting function for radiation fluctuation response.

6. The two-dimensional dynamic scene modeling method based on autoregressive background radiation according to claim 5, characterized in that, The specific weighting function for radiation fluctuation response is as follows: in, The radiation response function representing the temperature correlation of vibration; The radiation response function directly related to the dynamic temperature; Indicates the different vibrational states of atmospheric molecules; The first molecule representing atmospheric molecules Vibrational temperature; Indicates the dynamic temperature of atmospheric molecules; This represents the rate of change of the vibrational temperature with respect to the dynamic temperature.

7. A readable storage medium, characterized in that, It stores a computer program, which, when executed, implements the two-dimensional dynamic scene modeling method based on autoregressive technology for adjacent background radiation as described in any one of claims 1 to 6.

8. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the two-dimensional dynamic scene modeling method based on autoregressive technology for adjacent background radiation as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, It includes a computer program / instruction that, when executed by a processor, implements the two-dimensional dynamic scene modeling method based on autoregressive technology for adjacent background radiation as described in any one of claims 1 to 6.

Images