Space-space-based infrared imaging simulation platform based on CUDA (Compute Unified Device Architecture)

Abstract

The invention relates to an air-space-based infrared imaging simulation platform based on CUDA (Compute Unified Device Architecture). The air-space-based infrared imaging simulation platform comprises a global calculation class, a three-dimensional position calculation module, a camera generation module, a sun calculation module, a background data extraction module, a target calculation module and an imaging effect calculation module. The platform meets the requirement of high-precision signal-level space-space-based infrared detection imaging simulation, the simulation image data generation efficiency is greatly improved under the condition that the signal-level imaging simulation precision in a space scene is guaranteed, and rapid generation of imaging simulation data of a space platform joint observation target in a multi-observation track scene is achieved. And the universality of the platform is improved by using modular data and development.

Description

Technical Field

[0001] The present invention belongs to the technical field of space exploration, and particularly relates to a space-based and airborne infrared imaging simulation platform based on CUDA. Background Art

[0002] In recent years, with the maturity of space development technology, the development of the stratosphere and above has gradually become the scientific and technological development direction of countries around the world. It has important application values in many fields such as space science research, space remote sensing, and atmospheric data collection. Therefore, the importance of space exploration systems has been continuously increasing. Among them, space-based infrared detection systems have advantages such as a wide observation range, good timeliness, and outstanding observation capabilities. Target detection based on space-based infrared detection systems is an important means for detecting high-speed flying targets. Major countries around the world are applying relevant systems. Therefore, it is very meaningful to establish a complete complex infrared background radiation model based on the detection band, analyze the influence of complex environmental factors, detector parameters, optical parameters, etc. in the space-based optical imaging link on the target detection accuracy, and compare the similarities and differences between the target imaging characteristics and the background imaging characteristics, which is beneficial to improving the accuracy of space-based detection targets and providing an important reference for building an efficient full-scene collaborative detection platform. At the same time, by generating simulation image sequences under different conditions through the simulation platform, it provides sufficient and accurate data support for studying target detection and recognition algorithms.

[0003] In order to establish a space-based and airborne infrared imaging simulation platform for high-speed targets, the existing technical methods are as follows:

[0004] In 2023, Tian Hao et al. conducted a model quantitative analysis on the target detection performance of space-based infrared in the critical detection mode using a near-earth orbit infrared detection platform with an orbital altitude of 1500 km. They analyzed the aerodynamic characteristics and infrared radiation characteristics, and combined with the atmospheric transmission model and detector model to analyze the space-based infrared early warning detection capabilities at different limit detection distances and limit signal-to-noise ratios in the space-based infrared critical observation mode. Xue Chaofan studied the influence of solid particles on radiation in high-temperature and high-pressure media based on the low-altitude tail flame flow field using the reverse Monte Carlo method, established an analysis model for the spectral radiance of high-altitude high-temperature targets and the radiance of the ground background, studied the wavebands of space-based targets detectable by the earth background radiation, and analyzed the infrared radiation wavebands suitable for detection. In 2022, Li Peize conducted a flow field analysis of high-speed aircraft, established models for the aircraft skin, atmospheric field, and reflected radiation, studied complex backgrounds, and used the OSG rendering engine to build a three-dimensional scene to generate simulation results. In 2021, Li Jianbing discussed the physical advantages of the ray tracing method in infrared simulation, established an accurate model for the space-based infrared radiation imaging link coupling multiple elements at the global scale. Based on data such as emissivity, temperature, and elevation measured by remote sensing, the surface radiation texture was designed, and the non-uniform BRDF radiation reflection calculation model of the surface and ocean was studied; a space-based infrared detection ray tracing simulation system was constructed, and the rendering result images of the earth background and target movement processes under different conditions were simulated and analyzed.

[0005] However, in the existing technologies, some methods mainly conduct detailed modeling for high-speed targets, and for the background, they mostly qualitatively analyze the detection capabilities of space-based platforms based on equivalent background models, without completely establishing a task profile combined with the detection of space-based infrared platforms, lacking a refined model for the radiation characteristics in the spectral dimension of the target background. Some methods have established an integrated simulation platform for the full scene, but their scenes are mostly simulations based on open-source 3D rendering engines, and the generated results are all grayscale images. The coupling with the physical model and the refined simulation of radiation signals are not perfect enough. Some models consider the full scene of the target background in detail, but the imaging simulation uses complex ray algorithms, resulting in low efficiency in generating simulation images; and their scenes are all preset and determined scenes, unable to cover the problems of simultaneous imaging in multiple regions, various orbits, multiple imaging platforms, and other spectral wavebands, which is inconsistent with the actual required scenes, resulting in low simulation accuracy. Summary of the Invention

[0006] To solve the above problems existing in the prior art, the present invention provides a space-air-based infrared imaging simulation platform based on CUDA. The technical problems to be solved by the present invention are realized through the following technical solutions:

[0007] An embodiment of the present invention provides a space-air-based infrared imaging simulation platform based on CUDA, including:

[0008] The global calculation class is used to initialize input and output parameters and preprocess data, and then call different modules to calculate each frame until the frame loop ends, obtaining the imaging simulation results of all frames;

[0009] The three-dimensional position calculation module is used to obtain the satellite position and satellite orbit state at the current moment according to the incoming preset satellite data during each frame loop;

[0010] The camera generation module is used to generate a camera ray matrix according to the camera pose, satellite pose, and camera matrix parameters at the current moment during each frame loop;

[0011] The sun calculation module is used to update the sun vector calculation information according to the simulation time information in the global calculation class during each frame loop;

[0012] The background data extraction module is used to generate the background radiation information received by the detector at the current moment by combining the sun vector calculation information and the camera ray matrix during each frame loop;

[0013] The target calculation module is used to calculate the target radiation information received by the detector at the current moment by combining the sun vector calculation information and the target position information during each frame loop;

[0014] The imaging effect calculation module is used to fuse the background radiation information and the target radiation information, and combine the detector effect to convert the fused radiation information into a grayscale matrix to generate the imaging simulation result of each frame during each frame loop.

[0015] In an embodiment of the present invention, the three-dimensional position calculation module is specifically used for:

[0016] During each frame loop, call the external interface of the STK tool library or the AeroSpaceToolBox toolkit, and extract the three-dimensional position, orientation, and velocity vector of the satellite in the J2000 coordinate system at the current moment from the incoming simulation time information and the six orbital elements of the satellite orbit, where the six orbital elements of the satellite orbit include the orbital inclination, right ascension of the ascending node, eccentricity, semi-major axis, argument of perigee, and mean anomaly;

[0017] Convert the three-dimensional position, orientation, and velocity vector of the satellite in the J2000 coordinate system into the three-dimensional position, orientation, and velocity vector of the satellite in the Earth-centered Earth-fixed coordinate system to obtain the satellite position and satellite orbit state at the current moment.

[0018] In an embodiment of the present invention, the camera generation module is specifically used for:

[0019] During each frame loop, generate a camera ray matrix according to the camera pose, satellite pose, and camera matrix parameters at the current moment through an algorithm for traversing pixels.

[0020] In one embodiment of the present invention, the light direction vector of each pixel in the camera light matrix is:

[0021] LosToIJ = [realX_Cam, realY_Cam, -focal]

[0022] realX_Cam = RealX - Dx

[0023] realY_Cam = RealY - Dy

[0024] Dx = RealX × (X0(1) × l + X0(2) × l 2 + X0(3) × l 3 ) + X0(4) × (l + 2RealX 2 )

[0025] + 2X0(5) × RealX × RealY + X0(6) × l

[0026] Dy = RealY × (X0(1) × l + X0(2) × l 2 + X0(3) × l 3 ) + X0(5) × (l + 2RealY 2 )

[0027] + 2X0(4) × RealX × RealY + X0(7) × l

[0028] l = RealX 2 + RealY 2

[0029] Wherein, LosToIJ is the light direction vector of the pixel, focal is the focal length of the camera, realX_Cam and realY_Cam represent the xy coordinates of the pixel after correcting the distortion, X0(1 - 7) are the distortion coefficients in the camera parameters, RealX and RealY are the physical coordinates of the pixel points, l is the square of the distance from the pixel point to the image center, and Dx and Dy are the distortion correction amounts in the xy pixel directions respectively.

[0030] In one embodiment of the present invention, the background data extraction module is specifically used for:

[0031] In each frame cycle, the background type of the light is obtained by intersecting the camera light matrix with the earth and atmosphere models, and the background types of the light include ground object background, limb background, and deep space background;

[0032] When the background type of the light is a ground object background, data for calculating the intrinsic radiation of the ground object is retrieved from an external ground object background database according to the latitude, longitude, and altitude position information at the ground object background. The intrinsic radiation of the ground object is calculated using the data for calculating the intrinsic radiation of the ground object according to Planck's formula. The reflected radiation brightness of the ground object is calculated based on the solar irradiance and the surface reflectivity of the ground, and the intrinsic radiation of the ground object and the reflected radiation brightness of the ground object are superimposed. The superimposed result is multiplied by the atmospheric transmittance to obtain the ground object background radiation data; wherein, the solar irradiance is extracted from a solar radiation data table according to the solar zenith angle, and the solar zenith angle is calculated from the solar ray direction vector and the normal direction of the Earth's surface;

[0033] When the background type of the light is a limb background, the atmospheric radiation texture data is retrieved according to the tangent height value of the limb observation, and the radiation data at the limb is obtained according to the atmospheric radiation texture data;

[0034] When the background type of the light is a deep space background, the three-dimensional spatial position of the intersection point between the enclosing sphere surrounding the Earth and the atmosphere and the light is obtained, and the three-dimensional spatial position is converted into right ascension and declination values. The radiation data of the deep space background is extracted from an infrared star catalog according to the right ascension and declination values;

[0035] The ground object background radiation data, the radiation data at the limb, and the radiation data of the deep space background are superimposed to obtain the background radiation information received by the detector at the current moment.

[0036] In an embodiment of the present invention, the target calculation module is specifically configured to:

[0037] In each frame cycle, according to the simulated relative time, the target position information and the temperature change of the target are extracted from a preset target orbit state data file, wherein the target position information includes three-dimensional position coordinates and attitude information;

[0038] According to Planck's formula, the intrinsic radiation brightness of the target in the spectral range is obtained using the temperature change of the target;

[0039] The solar zenith angle is calculated from the solar vector calculation information and the normal direction of the target surface. The solar irradiance at the current position is extracted from the solar radiation data table according to the target position information and the solar zenith angle. The atmospheric transmittance at the current position is obtained from the local atmospheric data texture. The solar irradiance reaching the target surface, the Earth irradiance reaching the target surface, and the radiation of the Earth's reflected solar radiation reaching the target are calculated according to the solar irradiance and the atmospheric transmittance. After superimposing the solar irradiance reaching the target surface, the Earth irradiance reaching the target surface, and the radiation of the Earth's reflected solar radiation reaching the target, the superimposed result is multiplied by the BRDF bidirectional reflection radiation model to obtain the reflected radiation brightness of the target;

[0040] Superimpose the intrinsic radiation luminance and the reflected radiation luminance, and multiply the superimposed result by the atmospheric transmittance to obtain the target radiation information received by the detector at the current moment.

[0041] In an embodiment of the present invention, the solar irradiance E reaching the target surface s-t is:

[0042]

[0043] where R s is the solar radius, l st is the distance between the sun and the target, M sun is the solar radiant emittance, and θ st is the angle of incidence of the sun on the target surface;

[0044] The terrestrial irradiance E reaching the target surface j-e is:

[0045]

[0046] where U is the total set of the earth's surface where radiation can reach the target, L earj is the spectral radiance from the j-th surface of the earth to the target, A j is the area of the j-th grid, ψ j is the angle between the outer normal of the j-th grid of the earth and the line connecting the center of the j-th grid and the target, Ω j is the solid angle from the j-th grid of the earth to the target surface, A m is the area of the target surface;

[0047] The radiation E of the earth-reflected solar radiation reaching the target is:

[0048]

[0049] where τ atm is the atmospheric transmittance in the specified band, ρ e is the surface reflectivity, l e-t is the distance from the ground to the satellite, E se is the irradiance generated by solar radiation on the earth's surface, θ se is the angle between the solar radiation and the normal of the surface element on the ground, θ er is the angle between the normal of the surface element on the ground and the line connecting the target and the surface element on the ground, θ et is the angle between the normal of the target surface and the line connecting the target surface and the surface element on the ground, dA e is the area of the surface element on the ground.

[0050] In one embodiment of the present invention, the BRDF bidirectional reflection radiation model is as follows:

[0051]

[0052] where / is the incident direction, v is the observation direction, ρ d is the diffuse reflectance, n is the normal direction, M is a coefficient to be determined, ρ s is the specular reflectance, F is the Fresnel factor, D is the normal distribution function, and G is the geometric function.

[0053] In one embodiment of the present invention, the imaging effect calculation module is specifically configured to:

[0054] Calculate the position of the target on the image plane, and use the detector integration time and the velocity vector of the target to generate the image plane coordinates of multiple target passing position points within the integration time, and obtain the radiation information of each target passing position point;

[0055] Fuse the radiation information of each target passing position point with the background radiation information to obtain the radiation distribution on the detector incident surface;

[0056] Superimpose the radiation distribution on the detector incident surface with the detector effect to obtain the imaging simulation result.

[0057] In one embodiment of the present invention, the three-dimensional position calculation module, the camera generation module, the sun calculation module, the background data extraction module, the target calculation module, and the imaging effect calculation module perform parallel computing using a unified computing device architecture.

[0058] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0059] The present invention considers the simultaneous detection and imaging of targets and backgrounds by multiple cameras at different orbital altitudes, and divides the entire link of signal transmission into physical models of several modules such as a three-dimensional position calculation module, a camera generation module, a sun calculation module, a background data extraction module, a target calculation module, and an imaging effect calculation module, which meets the needs of high-precision signal-level space-based infrared detection imaging simulation. Under the condition of ensuring the signal-level imaging simulation accuracy in the space-air scenario, the simulation image data generation efficiency is greatly improved, the rapid generation of imaging simulation data for the joint observation of targets by space platforms in a multi-observation orbit scenario is realized, and the modular data and development are used to improve the versatility of the platform. Description of the Drawings

[0060] Figure 1 It is a schematic diagram of a space-based infrared imaging simulation platform based on CUDA provided by an embodiment of the present invention;

[0061] Figure 2 Schematic diagram of the operation process of a background data extraction module provided by an embodiment of the present invention;

[0062] Figure 3 Schematic diagram of the operation of a target calculation module provided by an embodiment of the present invention;

[0063] Figure 4 Schematic diagram of the process of an imaging effect calculation module provided by an embodiment of the present invention;

[0064] Figure 5 Object-image space relationship diagram of the imaging scene of a spatial point target provided by an embodiment of the present invention. Detailed implementation manners

[0065] The present invention will be further described in detail below in conjunction with specific embodiments, but the implementation manners of the present invention are not limited thereto.

[0066] Embodiment 1

[0067] Please refer to Figure 1 , Figure 1 Schematic diagram of a space-based infrared imaging simulation platform based on CUDA provided by an embodiment of the present invention.

[0068] The space-based infrared imaging simulation platform based on CUDA in this embodiment includes: a global calculation class, a three-dimensional position calculation module, a camera generation module, a solar calculation module, a background data extraction module, a target calculation module, and an imaging effect calculation module. Among them, the simulation calculation of the frame loop and the scheduling of global data are both completed under the global calculation class, and the specific calculation algorithms required for the simulation respectively enter the three-dimensional position calculation module, the background data extraction module, the camera generation module, the target calculation module, the solar calculation module, and the imaging effect calculation module for processing, and finally generate the radiation data reaching the detector or the imaging simulation image sequence.

[0069] The global calculation class is used to initialize the input and output parameters and perform data preprocessing, and then call different modules to calculate each frame until the frame loop ends, and obtain the imaging simulation results of all frames.

[0070] Specifically, the global calculation class is mainly responsible for the progress of the frame loop and the transmission scheduling control of data algorithms. Its main process is as follows: Initialize various input and output parameters, including the location of data files, detector parameters, camera orbit parameters, target location parameters, simulation time, etc. After that, some data preprocessing and preparation work will be carried out, such as calling the camera generation module to generate the initial light matrix of the camera; Another example is to calculate the camera platform orbit information in advance for direct use during the frame loop to improve the smoothness of the loop. After the preparation is completed, the simulation loop will start in the frame loop. In the frame loop, the following steps will be mainly taken. First, update the spatial positions of the target model, background model, and detector model for each frame. Then, transmit data such as satellite position, orientation, and detector parameters to the camera module to generate the camera light matrix, and return this matrix to the total control class for storage; Transmit the time information to the sun calculation module to update the sun light direction information for each frame and send it back to the total control for storage; Transmit time, sun, position and attitude, camera light matrix, etc. to the background calculation module. This module is responsible for the extraction, calculation, and invocation of background data, generates the background radiation matrix, and outputs the background radiation brightness matrix; Transmit time, sun, satellite position and attitude, target orbit data, camera light matrix, etc. to the background calculation module, retrieve the target position and target temperature data according to the simulation relative time, calculate the target radiation brightness, calculate the target image plane coordinates, simulate the trailing, and generate the target radiation distribution matrix; Transmit the radiation matrix to the imaging class, fuse the target and background to generate simulation data, and then form a simulation image through photoelectric conversion and gray quantization. After imaging, continue the frame loop until the entire simulation process ends, generating a simulation image sequence dataset for the entire process.

[0071] Furthermore, the algorithm processing of the three-dimensional position calculation module, camera generation module, sun calculation module, background data extraction module, target calculation module, and imaging effect calculation module uses the unified computing device architecture (CUDA) for parallel computing acceleration. During this period, all detectors carried by all space carrier platforms set will detect the target and generate simulation results.

[0072] The three-dimensional position calculation module is used to obtain the satellite position and satellite orbit state at the current moment according to the preset satellite data passed in during each frame loop.

[0073] Specifically, the three-dimensional position calculation module is used to call the external interface set by the Satellite Tool Kit (STK) toolkit or the AeroSpaceToolBox toolkit from the program during each frame loop, input the preset satellite data according to its data requirements, and obtain the satellite position and satellite orbit state at the current moment based on the input preset satellite data. The preset satellite data mainly includes: simulation time information and the six orbital elements of the satellite orbit. The six orbital elements of the satellite orbit include the orbital inclination, the right ascension of the ascending node, the eccentricity, the semi-major axis, the argument of perigee, and the mean anomaly. The simulation time information and the six orbital elements of the satellite orbit are returned in the form of strings. After parsing the simulation time information and the six orbital elements of the satellite orbit, the three-dimensional position, orientation, and velocity vector of the satellite in the J2000 coordinate system at the current moment are extracted. Then, matlab is used to convert the three-dimensional position, orientation, and velocity vector of the satellite in the J2000 coordinate system into the Earth-centered Earth-fixed coordinate system used in the system, and the three-dimensional position, orientation, and velocity vector of the satellite in the Earth-centered Earth-fixed coordinate system are obtained as the satellite position and satellite orbit state at the current moment. Exemplarily, the satellite situation includes that the satellite is located in the geosynchronous orbit, low Earth orbit, or large elliptical orbit.

[0074] The camera generation module is used to generate a camera ray matrix according to the camera pose, satellite pose, and camera matrix parameters at the current moment during each frame loop.

[0075] Specifically, the camera generation module is used to generate a camera ray matrix by traversing the pixel algorithm according to the camera pose, satellite pose (including the three-dimensional position and orientation of the satellite in the Earth-centered Earth-fixed coordinate system), and camera matrix parameters (including the imaging effects caused by the distortion of the camera) at the current moment during each frame loop. The camera ray matrix includes the three-dimensional point positions of the detector pixels and the ray direction vectors in the Earth-fixed coordinate system. The camera ray matrix is the ray direction vector starting from the center of each pixel of the camera, and its calculation steps are as follows:

[0076] First, calculate the distortion correction amounts Dx and Dy in the xy pixel directions:

[0077] Dx = RealX × (X0(1) × l + X0(2) × l 2 + X0(3) × l 3 ) + X0(4) × (l + 2RealX 2 )

[0078] + 2X0(5) × RealX × RealY + X0(6) × l

[0079] Dy = RealY × (X0(1) × l + X0(2) × l 2 + X0(3) × l 3 ) + X0(5) × (l + 2RealY 2 )

[0080] +2X0(4)×RealX×RealY+X0(7)×l

[0081] l = RealX 2 +RealY 2

[0082] Among them, X0(1 - 7) are the distortion coefficients in the camera parameters, RealX and RealY are the physical coordinates of the pixel points, l is the square of the distance from the pixel point to the image center, and Dx and Dy are the distortion correction amounts in the xy pixel directions respectively.

[0083] Through Dx and Dy, the pixel position can be corrected to obtain more accurate physical coordinates. The corrected pixel coordinates are:

[0084] realX_Cam = RealX - Dx

[0085] realY_Cam = RealY - Dy

[0086] Among them, realX_Cam and realY_Cam represent the xy coordinates of the pixel after correcting the distortion.

[0087] The light direction vector of the pixel can be expressed as:

[0088] LosToIJ = [realX_Cam, realY_Cam, -focal]

[0089] Among them, focal is the focal length of the camera.

[0090] Furthermore, the camera light matrix is composed of the light direction vectors of each pixel.

[0091] The sun calculation module is used to update the sun vector calculation information according to the simulation time information in the global calculation class during each frame loop.

[0092] Specifically, the sun calculation module is mainly used to update the vector calculation information related to sunlight stored according to the simulation time information in the global calculation class. The sun vector calculation information includes: the sun position and the three-dimensional direction vector of sunlight in the Earth-centered Earth-fixed coordinate system. These information will be repeatedly used in the target and background calculations, so they are saved as global variables in the global calculation class.

[0093] Specifically, the calculation formula for the sun position is:

[0094] Px = R·cos(δ)·cos(ω)

[0095] Py = R·cos(δ)·sin(ω)

[0096] Pz = R·sin(δ)

[0097]

[0098] ω = 15°·(T - 12)

[0099] Where N is the day of the year, 23.45° is the tilt angle of the Earth's axis of rotation, δ is the solar declination angle, i.e., the deviation angle of the sun relative to the equator, ω is the solar hour angle, i.e., the deviation angle of the sun relative to the local noon, T is the current Greenwich Mean Time, Px, Py, and Pz are the x, y, and z coordinate positions of the sun in the ECEF coordinate system, and R is the distance from the Earth to the sun.

[0100] Furthermore, the three-dimensional direction vector pointing from the sun position to the origin of the coordinate system is the sunlight direction vector in the Earth-Centered Earth-Fixed coordinate system.

[0101] A background data extraction module, which is used to generate the background radiation information received by the detector at the current moment by combining the sun vector calculation information and the camera ray matrix during each frame loop;

[0102] The background data extraction module is mainly responsible for generating the radiation data of the background in the imaging simulation. For its flowchart, please refer to Figure 2 , Figure 2 which is a schematic diagram of the operation process of a background data extraction module provided by an embodiment of the present invention.

[0103] Specifically, the background data extraction module is used to obtain the background type of the light by intersecting the incoming camera ray matrix with the Earth and atmosphere model during each frame loop. The background types of the light include the ground object background, the limb background, and the deep space background.

[0104] When the camera ray matrix intersects with the Earth and atmosphere model and the background type of the light is the ground object background, the longitude, latitude, and altitude position information at the ground object background is obtained. At this time, the background data extraction module retrieves the data for calculating the intrinsic radiation of the ground object from the external ground object background database according to the longitude, latitude, and altitude position information at the ground object background. The data for calculating the intrinsic radiation of the ground object is a data table distributed according to longitude and latitude, which can be in the.tif format or other formats, and mainly includes the radiation brightness information of the ground object background in a specific radiation band, or the surface temperature information, reflectivity information, and cloud data in this area. Then, the background data extraction module uses the gdal tool library to assist in data reading, and calculates the intrinsic radiation of the ground object according to Planck's formula using the data read for calculating the intrinsic radiation of the ground object. At the same time, due to the influence of factors such as atmospheric transmittance and solar irradiance, the background data extraction module calculates the reflected radiation brightness of the ground object according to the solar irradiance and the surface reflectivity in the ground object data, and its calculation formula is:

[0105] L = E sun · ρ / π

[0106] Where L is the reflected radiance, E sun is the solar irradiance, and ρ is the surface albedo.

[0107] After that, the intrinsic radiation of the object and the reflected radiance of the object are superimposed, and the superimposed result is multiplied by the atmospheric transmittance to obtain the object background radiation data before entering the detector. Among them, the solar irradiance is extracted from the solar radiation data table according to the solar zenith angle. The solar zenith angle is calculated from the direction of the solar ray vector and the normal direction of the earth's surface. The solar radiation data table is local data that has been calculated using the MODTRAN software before the simulation.

[0108] When the camera ray matrix is tangent to the earth and atmosphere model and the background type of the ray is the limb background, the tangent height value of the limb observation will be returned. At this time, the module retrieves the atmospheric radiation texture data according to the tangent height value of the limb observation, and obtains the radiation data at the limb by looking up the table according to the atmospheric radiation texture data. Among them, the atmospheric radiation texture data is a two-dimensional texture distributed according to the tangent height and the solar zenith angle (the angle between the sun and the normal at the tangent height), which is generated by the atmospheric radiation calculation software modtran. The table to be queried is a data table generated by the atmospheric radiation calculation software modtran and indexed according to the zenith angle and the tangent point height.

[0109] When the camera ray matrix does not intersect with the earth and atmosphere model and the background type of the ray is the deep space background, the ray will intersect with a larger enclosing sphere that surrounds the earth and the atmosphere, obtain the three-dimensional spatial position of the intersection point, and convert it into right ascension and declination values. Among them, the declination δ is the angle between the point and the equatorial plane:

[0110]

[0111] The right ascension α is the angle between the projection of the point on the equatorial plane and the direction of the vernal equinox. The calculation formula is:

[0112] α = arctan2(y, x)

[0113] Where x, y, and z are the three-dimensional coordinate positions.

[0114] Furthermore, the radiation data of the deep space background can be extracted from the infrared star catalog according to the right ascension and declination values. The infrared star catalog can download the publicly available data of various observatories.

[0115] Finally, the object background radiation data, the radiation data at the limb, and the radiation data of the deep space background are superimposed to obtain the background radiation information received by the detector at the current moment, and it is transmitted to the control class.

[0116] A target calculation module, which is used to calculate the target radiation information received by the detector at the current moment by combining the solar vector calculation information and the target position information during each frame loop;

[0117] The target calculation module is mainly responsible for obtaining and updating the target position and radiation information. The process is as Figure 3 shown, Figure 3 which is a schematic diagram of the operation of the target calculation module provided by an embodiment of the present invention.

[0118] Specifically, during each frame loop, the target calculation module extracts the target position information and the temperature change of the target from a preset target orbit state data file according to the simulated relative time. Among them, the target position information includes three-dimensional position coordinates and attitude information, and transmits them to the control class for the next target simulation calculation. The target simulation calculation process is as follows: First, according to Planck's formula, the intrinsic radiation luminance of the target in the spectral range is obtained by using the temperature change of the target; then, according to the target position information and the solar zenith angle, the solar radiation irradiance at the current position is extracted from the solar radiation data table, and the atmospheric transmittance at the current position is obtained from the local atmospheric data texture. Among them, the solar zenith angle is calculated from the direction of the solar ray vector and the normal direction of the target surface, and the solar radiation data table is local data that has been calculated using MODTRAN software before the simulation; then, using these data, the influence of solar, earth radiation, earth albedo radiation on the target and the attenuation influence of radiation during atmospheric propagation can be calculated. Specifically, it includes: calculating the solar radiation irradiance reaching the target surface, the earth radiation irradiance reaching the target surface, and the radiation of the earth-reflected solar radiation reaching the target according to the solar radiation irradiance and the atmospheric transmittance, and after superimposing the solar radiation irradiance reaching the target surface, the earth radiation irradiance reaching the target surface, and the radiation of the earth-reflected solar radiation reaching the target, multiplying the superimposed result by the BRDF bidirectional reflection radiation model to obtain the reflected radiation luminance of the target; and superimposing the intrinsic radiation luminance and the reflected radiation luminance, and multiplying the superimposed result by the atmospheric transmittance to obtain the target radiation information received by the detector at the current moment. Among them, the reflection radiation influence caused by solar and earth radiation will be calculated using the BRDF bidirectional reflection radiation model.

[0119] Specifically, the solar radiation irradiance E s-t reaching the target surface is:

[0120]

[0121] where R s is the solar radius, l st is the distance between the sun and the target, M sun is the solar radiant emittance, which can be calculated by Planck's formula, and θ st is the angle of incidence of the sun on the target surface;

[0122] The terrestrial irradiance E reaching the target surface j-e is given by:

[0123]

[0124] where U is the total set of the terrestrial surfaces to which radiation can reach the target, and L earj is the spectral radiance from the j-th terrestrial surface to the target, with the unit of W / m 2 ·μm·sr; A j is the area of the j-th grid, ψ j is the angle between the outer normal of the j-th grid on the earth and the line connecting the center of the j-th grid to the target, Ω j is the solid angle from the j-th grid on the earth to the target surface, A m is the area of the target surface;

[0125] The radiation E of the reflected solar radiation from the earth reaching the target is:

[0126]

[0127] where τ atm is the atmospheric transmittance in the specified band (the spectral band in the input detector parameters), ρ e is the surface reflectivity, l e-t is the distance from the ground to the satellite, E se is the irradiance generated by the solar radiation on the earth's surface, θ se is the angle between the solar radiation and the normal of the surface element on the ground, θ er is the angle between the normal of the surface element on the ground and the line connecting the target and the surface element on the ground, θ et is the angle between the normal of the target surface and the line connecting the target surface and the surface element on the ground, dA e is the area of the surface element on the ground.

[0128] The BRDF bidirectional reflection radiation model is:

[0129]

[0130] where / is the incident direction, v is the observation direction, ρ d is the diffuse reflectance, n is the normal direction, M is a coefficient to be determined and its value is adjusted according to the actual situation, ρ s is the specular reflectance, F is the Fresnel factor, D is the normal distribution function, and D adopts the Trowbridge-Reitz distribution, G is the geometric function.

[0131] The imaging effect calculation module is used to fuse the background radiation information and the target radiation information during each frame cycle, and combine the detector effect to convert the fused radiation information into a grayscale matrix, generating the imaging simulation result for each frame.

[0132] The imaging effect calculation module is mainly responsible for the fusion of the target background radiation matrix, and simulating the change process of the signal in the detector and the generated effect when outputting the imaging simulation result. Its process is as Figure 4 shown, Figure 4 which is the schematic diagram of the imaging effect calculation module process provided by the embodiment of the present invention.

[0133] Specifically, the process of the imaging effect calculation module mainly includes:

[0134] First, according to the position and velocity information of the target, calculate the position of the target on the image plane. Specifically, convert the three-dimensional position coordinates of the target in the Earth-centered Earth-fixed coordinate system into the position coordinates in the camera body three-dimensional coordinate system. Among them, in the camera body three-dimensional coordinate system, the line-of-sight direction is the z-axis direction, and the pixel arrangement direction is the x and y-axis directions; then, project the converted position coordinates onto the image plane position of the detector, and use the projected x and y coordinates as the coordinates of the target on the image plane. Further, due to the characteristics of high-speed targets, trailing will occur during imaging. Therefore, use the detector integration time and the velocity vector of the target to generate the image plane coordinates of multiple target passing positions within the integration time. The image plane coordinates of multiple positions form a line, thus simulating the high-speed trailing characteristic. After simulating the trailing characteristic, the radiation information of each target passing position point is obtained simultaneously.

[0135] Then, superimpose and fuse the radiation information of each target passing position point with the background radiation information to obtain the radiation distribution on the incident surface of the detector.

[0136] Finally, superimpose the radiation distribution on the incident surface of the detector with the detector effect to obtain the imaging simulation result and output it. The detector effect includes noise, the influence of the optical system transmittance, the point spread effect caused by the optical effect of the detector system, etc. The signal transmission and conversion process in the detector includes processes such as the photoelectric conversion from the radiation signal to the electrical signal and the grayscale quantization from the electrical signal to the grayscale.

[0137] Specifically, the calculation expression of the radiation power of the radiation generated by the radiation source and propagating to the entrance pupil (before entering the optical system) in the object space:

[0138] P = I·A opt / l 2 = L·A s ·A opt / l 2

[0139] Among them, P is the radiation power before the entrance pupil, Aopt is the entrance pupil area of the optical system, l is the distance between the background or target and the entrance pupil, L is the radiance of the background or target, A s is the surface area of the radiation source, and I is the radiation intensity in front of the entrance pupil. The expressions for background radiation information and target radiation information both adopt the above formula.

[0140] Superimpose the radiation distribution on the incident surface of the detector with the detector effect, obtain the imaging simulation result and output it, specifically including:

[0141] Please refer to Figure 5 , Figure 5 is the object-image space relationship diagram of the spatial point target imaging scenario provided by the embodiment of the present invention. According to the object-image relationship, the imaging size on the detector image plane can be obtained through the following formula:

[0142] A spix / f 2 =A s / l 2

[0143] where A spix is the imaging size on the image plane, and f is the focal length of the optical system.

[0144] According to the above formula, it can be deduced that after the fused imaging simulation radiation signal passes through the optical system, the radiation power value received on the detector surface can be approximately obtained through the following formula:

[0145] P spix =τ opt ×P opt =τ opt ×L×A pix ×A opt / f 2

[0146] where P spix is the radiation power value received on the pixel, τ opt is the radiation transmittance of the optical system, P opt is, L is the radiance value of the radiation source corresponding to the pixel, A pix is the pixel area, A opt is the entrance pupil area of the optical system, and f is the focal length of the optical system.

[0147] S52. Convert the radiation power value on the detector surface into the response voltage on a single pixel of the detector.

[0148] Specifically, after the signal light enters the detector pixel, the photoelectric effect will occur to form a signal voltage. The signal carrier is then converted from light to electricity and finally quantized to form a grayscale result, generating the final image result. In this embodiment, by analyzing the detector imaging effect that occurs during the signal process, the conversion from the radiation signal distribution matrix to the two-dimensional grayscale distribution image is achieved.

[0149] For the photoelectric conversion process on a single pixel of the detector, its expression is as follows:

[0150]

[0151] Among them, U pixel is the response voltage on the pixel, with the unit V (volt), P in (λ) is the input radiation power on the detector pixel, with the unit W (watt), R(λ) is the detector responsivity, λ is the spectral wavelength, with the unit μm (micrometer), τ int is the detector imaging integration time, with the unit s (second).

[0152] S53. Perform a Fourier transform on the response voltage from the spatial domain to obtain the response voltage in the frequency domain distribution, multiply the response voltage in the frequency domain distribution by the modulation transfer function to obtain the response voltage frequency domain distribution with the spatial transfer effect added, and then perform an inverse Fourier transform on the response voltage frequency domain distribution with the spatial transfer effect added to obtain the response voltage with the spatial transfer effect superimposed.

[0153] In actual situations, due to the non-uniform response of the detector pixels, noise generated by the irregular changes in the circuit signals, etc., detector effects such as imaging blur and picture noise will occur.

[0154] For the spatial transfer effect of the detection system that causes imaging blur, the degradation of the imaging pixels it causes is mainly described using the Modulation Transfer Function (MTF). According to the timing of physical effects, considering the spatial modulation characteristics of each physical effect module in the frequency domain, the total MTF of the system is equal to the product of the MTF values of each subsystem, and the MTF value of each module is less than 1. The MTF modulation transfer function expression is:

[0155] MTF sys = MTF optic ·MTF dect

[0156] Among them, is the modulation transfer function of the optical system, used to add the spatial transfer effect of the optical system, and is the modulation transfer function of the detector system, used to add the spatial transfer effect of the detector system. The MTF is mainly obtained through parameters such as the spatial frequency, cut-off frequency, and detector field of view in the detector parameters.

[0157] The process of using the MTF method for spatial transfer effect is mainly processed in the frequency domain. Therefore, the response voltage matrix received by the detector can be Fourier-transformed from the spatial domain to obtain the frequency-domain distribution of the response voltage matrix. Multiplying the response voltage matrix in the frequency domain by the MTF modulation transfer function can obtain the frequency-domain distribution of the response voltage matrix after adding the spatial transfer effect, and then performing an inverse Fourier transform on it can obtain the response voltage matrix with the spatial transfer effect superimposed.

[0158] S54. Add noise to the response voltage with the spatial transfer effect superimposed to obtain the response voltage after adding noise.

[0159] The noise is mainly manifested as irregularly distributed noise particles on the screen. The main reason is the thermal effect of the internal molecules of the material. The charged carriers inside the detector will have a fluctuating movement of breaking away from the electric potential, also known as thermal current noise, which is not related to the signal and is also called additive noise. During the simulation process, a function with a random number seed is usually used to simulate the random noise distribution of the detector. In this embodiment, a Gaussian noise (Gauss Noise) simulation model is used. The probability distribution function of Gaussian white noise can be expressed as:

[0160]

[0161] where x represents the gray value of the imaging noise, μ is the average value of the noise gray value, and σ is the variance of the noise gray value.

[0162] In the simulation, according to the Noise Equivalent Flux Density (NEFD) of the detector, the root mean square of the noise can be obtained and the simulation of Gaussian random noise can be carried out. The noise equivalent flux is the minimum detectable signal for the infrared system to detect a point source target, and its expression is as follows:

[0163]

[0164] In the formula, U noise is the root mean square value of the response voltage of the system noise, and U Δsignal is the output signal value when the difference in irradiance between the target and the background is E Δsignal . When the temperature of the background increases, the NEFD of the system will also increase accordingly.

[0165] The conversion formula between NEFD and the root mean square of the noise voltage is as follows:

[0166]

[0167] In the formula, τ opt is the transmittance of the optical system, D is the aperture of the optical system, and R is the detector responsivity.

[0168] Using the above formula, the voltage distribution of the noise can be obtained, and thus the voltage distribution of the noise is added to the response voltage with the spatial transfer effect superimposed to obtain the response voltage after adding the noise.

[0169] S55. Convert the response voltage after adding the noise into a grayscale value to obtain the imaging simulation result.

[0170] Specifically, there are upper and lower bounds for the linear quantization of the voltage value of the output signal. The lower bound of the quantization is expressed as the noise equivalent voltage value of the detector, i.e., U min = U noise , and the upper bound of the voltage calculation is related to the full well charge number n of the detector unit and the integration capacitance C of the detector unit:

[0171] U max = n·q / C

[0172] where q is the absolute value of the charge of an electron, q = 1.6×10 -19 C.

[0173] Taking the conversion of a common 8-bit grayscale image as a reference, the upper limit G max = 255 of the grayscale quantization, and the lower limit is 0. Then the formula for the grayscale voltage conversion is as follows:

[0174]

[0175] where G is the grayscale value of the imaging image.

[0176] From this formula, the grayscale value of the imaging image can be finally obtained to obtain the imaging simulation result of each frame.

[0177] The method of this embodiment integrates all-scenario factors, and can simulate multiple space-borne platform detectors under different scenario factors within the whole time to detect and image the target and the background simultaneously. Moreover, it tracks and processes the radiation signal throughout the whole process, improving both the authenticity of the scenario setting and the accuracy of the simulated signal. At the same time, this method is developed based on CUDA, utilizes the powerful parallel computing ability of the GPU to accelerate large-scale computing tasks, and adopts a modular simulation algorithm development. The efficient parallel computing and overall data scheduling make the simulation more real-time and efficient. The simulation platform developed using CUDA can perform parallel thread operations for each pixel surface of the detector, and multiple sub-threads can be set up for ray simulation on its surface. The system can allocate and regulate all thread resources, enabling more refined or upgraded development for some special requirements, with good flexibility. In addition, the background data flexibly uses the data of an external database, which is convenient for data extraction, replacement, and upgrade. The set parameters of each module can be replaced, the external database can be replaced, and the parallel thread resources can be regulated, facilitating the upgrade and maintenance of the platform and enhancing the versatility of the platform. Furthermore, the imaging mechanism does not use the grayscale mapping of an open-source rendering engine, but adopts the method of finding the intersection of the detector rays and each model, and analyzes directly from the perspective of the signal transmission amount, which not only improves the authenticity of the algorithm but also enhances the operation efficiency of the algorithm.

[0178] Therefore, the simulation platform of this embodiment integrates all the simulation elements of the signal from emission by the target and the background to imaging grayscale. The simulation accuracy is better, and the scenario setting has multiple satellites imaging simultaneously, which is more in line with the real situation. Based on the modular development of CUDA, the simulation efficiency is higher, and the universality is better, meeting the needs of high-precision signal-level space-based infrared detection imaging simulation. Under the condition of ensuring the signal-level imaging simulation accuracy in the space scenario, it greatly improves the simulation image data generation efficiency, realizes the rapid generation of imaging simulation data for the joint observation of the target by the space platform under multiple observation orbit scenarios, and uses modular data and development to enhance the versatility of the platform.

[0179] The above content is a further detailed description of the present invention in combination with specific preferred implementation manners. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or replacements can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A CUDA-based space-based infrared imaging simulation platform, characterized in that: include: The global calculation class is used to initialize input and output parameters and preprocess data, and then call different modules to calculate each frame until the frame cycle ends to obtain the imaging simulation results of all frames; The three-dimensional position calculation module is used to obtain the current satellite position and satellite orbit status according to the input preset satellite data in each frame cycle; The camera generation module is used to generate the camera light matrix according to the current camera pose, satellite pose and camera matrix parameters in each frame loop; The sun calculation module is used to update the sun vector calculation information according to the simulation time information in the global calculation class in each frame loop; A background data extraction module is used to combine the solar vector calculation information and the camera light matrix in each frame cycle to generate the background radiation information received by the detector at the current moment; A target calculation module is used to calculate the target radiation information received by the detector at the current moment by combining the solar vector calculation information and the target position information in each frame cycle; The imaging effect calculation module is used to fuse the background radiation information and the target radiation information in each frame cycle, and combine the detector effect to convert the fused radiation information into a gray matrix to generate an imaging simulation result for each frame.

2. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The three-dimensional position calculation module is specifically used for: In each frame loop, the external interface of the STK tool library or the AeroSpaceToolBox toolkit is called to extract the three-dimensional position, orientation and velocity vector of the satellite in the J2000 coordinate system at the current moment from the incoming simulation time information and the six satellite orbit numbers, wherein the six satellite orbit numbers include orbital inclination, right ascension of ascending node, eccentricity, semi-major axis, argument of perigee and mean anomaly; The three-dimensional position, orientation and velocity vector of the satellite in the J2000 coordinate system are converted into the three-dimensional position, orientation and velocity vector of the satellite in the Earth-centered Earth-fixed coordinate system to obtain the satellite position and satellite orbit state at the current moment.

3. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The camera generation module is specifically used for: In each frame loop, the camera light matrix is ​​generated according to the current camera pose, satellite pose and camera matrix parameters through the algorithm of traversing pixels.

4. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The light direction vector of each pixel in the camera light matrix is: LosToIJ=[realX_Cam,realY_Cam,-focal] realX_Cam=RealX-Dx realY_Cam=RealY-Dy Dx=RealX×(X0(1)×l+X0(2)×l 2 +X0(3)×l 3 )+X0(4)×(l+2RealX 2 ) +2X0(5)×RealX×RealY+X0(6)×l Dy=RealY×(X0(1)×l+X0(2)×l 2 +X0(3)×l 3 )+X0(5)×(l+2RealY 2 ) +2X0(4)×RealX×RealY+X0(7)×l l=RealX 2 +RealY 2 Among them, LosToIJ is the light direction vector of the pixel, focal is the focal length of the camera, realX_Cam and realY_Cam represent the xy coordinates of the pixel after correction of distortion, X0(1-7) is the distortion coefficient in the camera parameters, RealX and RealY are the physical coordinates of the pixel points, l is the square of the distance from the pixel point to the center of the image, and Dx and Dy are the distortion correction amounts in the xy pixel directions respectively.

5. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The background data extraction module is specifically used for: In each frame loop, the background type of the light is obtained by intersecting the camera light matrix with the earth and atmosphere model, and the background type of the light includes ground object background, edge background and deep space background; When the background type of the light is a ground object background, data for calculating ground object intrinsic radiation is retrieved from an external ground object background database according to the latitude and longitude altitude position information at the ground object background, ground object intrinsic radiation is calculated using the data for calculating ground object intrinsic radiation according to Planck's formula, the reflected radiation brightness of the ground object is calculated according to the solar radiation illuminance and the surface reflectivity, and the ground object intrinsic radiation and the reflected radiation brightness of the ground object are superimposed, and the superposition result is multiplied by the atmospheric transmittance to obtain ground object background radiation data; wherein the solar radiation illuminance is extracted from a solar radiation data table according to the solar zenith angle, and the solar zenith angle is calculated by the direction vector of the sun's rays and the direction of the normal to the earth's surface; When the background type of the light is an edge background, atmospheric radiation texture data is retrieved according to the cut height value observed at the edge, and radiation data at the edge is obtained according to the atmospheric radiation texture data; When the background type of the light is a deep space background, obtaining the three-dimensional spatial position of the intersection point between the light and the enclosing sphere enclosing the earth and the atmosphere, converting the three-dimensional spatial position into right ascension and declination values, and extracting the radiation data of the deep space background from the infrared star catalog according to the right ascension and declination values; The background radiation data of the ground object, the radiation data at the edge and the radiation data of the deep space background are superimposed to obtain the background radiation information received by the detector at the current moment.

6. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The target calculation module is specifically used for: In each frame cycle, according to the relative time of the simulation, the target position information and the temperature change of the target are extracted from the preset target orbit state data file, wherein the target position information includes three-dimensional position coordinates and attitude information; According to Planck's formula, the temperature change of the target is used to obtain the intrinsic radiation brightness of the target in the spectral range; The solar zenith angle is calculated based on the solar vector calculation information and the normal direction of the target surface, and the solar radiation illuminance at the current position is extracted from the solar radiation data table based on the target position information and the solar zenith angle, and the atmospheric transmittance at the current position is obtained from the local atmospheric data texture; the solar radiation illuminance reaching the target surface, the earth radiation illuminance reaching the target surface, and the radiation of the solar radiation reflected by the earth reaching the target are calculated based on the solar radiation illuminance and the atmospheric transmittance, and after superimposing the solar radiation illuminance reaching the target surface, the earth radiation illuminance reaching the target surface, and the radiation of the solar radiation reflected by the earth reaching the target, the superposition result is multiplied by the BRDF bidirectional reflection radiation model to obtain the reflected radiation brightness of the target; The intrinsic radiation brightness and the reflected radiation brightness are superimposed, and the superposition result is multiplied by the atmospheric transmittance to obtain the target radiation information received by the detector at the current moment.

7. The CUDA-based space-based infrared imaging simulation platform according to claim 6, characterized in that: The solar radiation intensity E reaching the target surface s-t for: Among them, R s is the radius of the sun, l st is the distance between the sun and the target, M sun is the solar radiation emittance, θ st is the angle of incidence of the sun on the target surface; The earth irradiance E reaching the target surface j-e for: Among them, U is the total set of the earth's surface where radiation can reach the target, L earj is the spectral radiance from the jth surface of the Earth to the target, A j is the area of ​​the jth grid, ψ j is the angle between the outer normal of the jth grid of the earth and the line connecting the center of the jth grid and the target, Ω j is the solid angle from the jth grid of the earth to the target surface, A m is the area of ​​the target surface; The radiation E at the target where the earth reflects the solar radiation is: Among them, τ atm is the atmospheric transmittance in the specified band, ρ e is the surface reflectivity, l e-t is the distance from the ground to the satellite, E se is the irradiance generated by solar radiation on the earth’s surface, θ se is the angle between the solar radiation and the normal line of the ground surface, θ er is the angle between the normal line of the ground element and the line connecting the target and the ground element, θ et is the angle between the target surface normal and the line connecting the target surface and the ground surface element, dA e is the area of ​​the ground surface element.

8. The CUDA-based space-based infrared imaging simulation platform according to claim 6, characterized in that: The BRDF bidirectional reflection radiation model is: Where / is the incident direction, v is the observation direction, ρ d is the diffuse reflectivity, n is the normal direction, M is the coefficient to be determined, ρ s is the specular reflectivity, F is the Fresnel factor, D is the normal distribution function, and G is the geometric function.

9. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The imaging effect calculation module is specifically used for: The position of the target on the image plane is calculated, and the detector integration time and the target's velocity vector are used to generate the image plane coordinates of multiple target passing points within the integration time to simulate the high-speed imaging tailing effect, and obtain the radiation information of each target passing point; The radiation information of each target passing position point is integrated with the background radiation information to obtain the radiation distribution of the incident surface of the detector; The radiation distribution on the incident surface of the detector is superimposed on the detector effect to obtain an imaging simulation result.

10. The CUDA-based space-based infrared imaging simulation platform according to claim 1, characterized in that: The three-dimensional position calculation module, the camera generation module, the sun calculation module, the background data extraction module, the target calculation module, and the imaging effect calculation module use a unified computing device architecture to perform parallel calculations.

Images