A method for self-adapting correction of a pixel area of a thermal infrared image of a Mars rover

Abstract

The application belongs to the field of image processing, and discloses a kind of pixel area adaptive correction method of Mars car thermal infrared image, the method includes: obtaining thermal infrared image and binocular stereo image pair, carries out polar rectification processing and sparse point cloud calculation to binocular stereo image pair, obtains target sparse point cloud, and constructs high-resolution digital elevation model;The pixel coordinates and internal parameters of thermal infrared image are obtained, the sight ray parameter equation is constructed, and the target intersection between the sight ray parameter equation and the high-resolution digital elevation model is calculated;According to the theoretical spot radius of target intersection, the actual ground projection area is calculated according to the theoretical spot radius;According to the actual ground projection area, the real ground radiation brightness is calculated, and the real ground radiation brightness is corrected and thermal inertia inversion processing, and the thermal inertia distribution map is obtained.The application can effectively solve the pixel area distortion problem in oblique imaging, and effectively improve the accuracy of terrain passability interpretation.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method, system, terminal, and computer-readable storage medium for adaptive correction of pixel area in thermal infrared images of a Mars rover. Background Technology

[0002] Thermal inertia is a physical parameter characterizing a material's ability to resist temperature changes. For the regolith covering the Martian surface, a high or low thermal inertia directly reflects the soil's density. High thermal inertia corresponds to compacted rock or hard soil, while low thermal inertia corresponds to loose sand or dust. Therefore, by measuring the thermal inertia distribution of the terrain in front of the Mars rover, soft sandy areas and hard surfaces can be identified, thereby assessing the terrain's traversability. Thermal inertia, as a physical parameter reflecting soil density, can be obtained by inverting surface temperature information acquired through a thermal infrared imaging system.

[0003] However, in practical applications, this thermal infrared imaging system faces a long-neglected problem that seriously affects data quality: pixel area distortion. This distortion causes a systematic deviation in the thermal inertia values ​​retrieved from the near and far fields of the same soil type, which seriously affects the accuracy of terrain accessibility interpretation.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main objective of this invention is to provide a pixel area adaptive correction method, system, terminal, and computer-readable storage medium for thermal infrared images of Mars rovers. This invention aims to solve the problem that existing thermal infrared imaging systems suffer from pixel area distortion, which leads to systematic deviations in the thermal inertia values ​​retrieved from the near and far fields of images for the same soil type, resulting in low accuracy in terrain passability interpretation.

[0006] To achieve the above objectives, the present invention provides a pixel area adaptive correction method for thermal infrared images of Mars rovers, the pixel area adaptive correction method for thermal infrared images of Mars rovers comprising the following steps: Acquire thermal infrared images and binocular stereo images, perform epipolar correction processing and sparse point cloud computing on the binocular stereo images to obtain target sparse point cloud, and construct a high-resolution digital elevation model based on the target sparse point cloud. Obtain the pixel coordinates and intrinsic parameters of the thermal infrared image, construct the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model. The theoretical spot radius is calculated based on the target intersection point, and the actual surface projection area is calculated based on the theoretical spot radius. The true surface radiance is calculated based on the actual surface projection area, and the true surface radiance is then corrected for radiance and subjected to thermal inertia inversion to obtain a thermal inertia distribution map.

[0007] Optionally, the pixel area adaptive correction method for the Mars rover's thermal infrared image, wherein acquiring the thermal infrared image and the binocular stereo image pair, performing epipolar correction processing and sparse point cloud computing on the binocular stereo image pair to obtain the target sparse point cloud, and constructing a high-resolution digital elevation model based on the target sparse point cloud, specifically includes: Thermal infrared images were acquired using the thermal infrared camera on the Mars rover, and stereo image pairs were acquired using the binocular stereo navigation camera. The stereo image pairs are subjected to epipolar correction to obtain row-aligned image pairs, and the disparity value of each pixel in the row-aligned image pairs is calculated using the semi-global matching (SGM) algorithm. The target distance from the target spatial point to the camera baseline is calculated based on the disparity value, and the three-dimensional coordinates of the target spatial point are calculated using the camera projection model based on the target distance. A target sparse point cloud is generated based on the three-dimensional coordinate points, and the target sparse point cloud is divided into grids and its elevation values ​​are calculated to obtain a high-resolution digital elevation model.

[0008] Optionally, the adaptive correction method for pixel area of ​​the Mars rover's thermal infrared image, wherein obtaining the pixel coordinates and intrinsic parameters of the thermal infrared image, constructing a line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculating the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model, specifically includes: The elevation values ​​in the high-resolution digital elevation model are obtained, and gradient calculation and normalization are performed based on the elevation values ​​to obtain the unit normal vector. Obtain the pixel coordinates, intrinsic parameters, and extrinsic parameters of the thermal infrared image, and calculate the first unit vector of the viewing direction in the camera coordinate system based on the pixel coordinates and the intrinsic parameters; Based on the rotation matrix in the extrinsic parameters, the viewing direction corresponding to the first unit vector is transformed into a world coordinate system to obtain a second unit vector of the viewing direction in the world coordinate system. Determine the target position of the camera's optical center in the world coordinate system, construct the line-of-sight ray parameter equation based on the target position and the second unit vector, and use a ray projection algorithm to calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model.

[0009] Optionally, the adaptive correction method for pixel area of ​​the Mars rover thermal infrared image, wherein the step of calculating the theoretical spot radius based on the target intersection point and calculating the actual surface projection area based on the theoretical spot radius specifically includes: The observation distance is calculated based on the target location and the intersection point of the target, and the theoretical spot radius is calculated based on the observation distance; The search radius in the high-resolution digital elevation model is determined based on the theoretical spot radius, and the search range is determined based on the search radius; The projection parameters are determined based on the search range, and the radius of the cone at the vertical distance and depth is calculated based on the projection parameters. The high-resolution digital elevation model is divided into meshes to obtain multiple triangular patches. The actual three-dimensional area is calculated based on the vertex positions of the multiple triangular patches, the vertical distances, and the radius of the cone at the depth, to obtain the actual surface projection area.

[0010] Optionally, the pixel area adaptive correction method for the Mars rover's thermal infrared image, wherein the step of calculating the actual three-dimensional area based on the vertex positions and vertical distances of the multiple triangular patches and the radius of the cone at the depth to obtain the actual surface projection area specifically includes: Calculate the vertical distance corresponding to each vertex position and the radius of the cone at the depth in each of the triangular facets. If the vertical distance corresponding to each vertex position and the radius of the cone at the depth both satisfy that the vertical distance is less than or equal to the radius of the cone at the depth, then calculate the actual three-dimensional area of ​​each triangular facet. The actual three-dimensional areas corresponding to all the triangular facets are summed to obtain the actual surface projection area.

[0011] Optionally, the pixel area adaptive correction method for the Mars rover's thermal infrared image, wherein calculating the true surface radiance based on the actual surface projection area specifically includes: Calculate the target angle between the line of sight direction and the surface normal based on the unit normal vector and the first unit vector, and calculate the vertical observation area based on the target angle and the actual surface projection area. The physical area and original radiance of the pixels are obtained, and the true radiance of the Earth's surface is calculated based on the physical area of ​​the pixels, the original radiance, and the vertical observation area.

[0012] Optionally, the pixel area adaptive correction method for the Mars rover's thermal infrared image, wherein the step of performing radiance correction and thermal inertia inversion processing on the true surface radiance to obtain a thermal inertia distribution map specifically includes: The area correction coefficient is calculated based on the physical area of ​​the pixel, the actual surface projection area, and the target angle. The actual surface radiance is then corrected based on the original radiance and the area correction coefficient to obtain the initial corrected radiance. Determine the current surface slope. If the current surface slope exceeds a preset slope threshold, perform a secondary correction on the initial corrected radiance to obtain the secondary corrected radiance. The secondary corrected radiance is processed by surface temperature conversion and thermal inertia inversion to obtain a thermal inertia distribution map.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a pixel area adaptive correction system for Mars rover thermal infrared images, wherein the pixel area adaptive correction system for Mars rover thermal infrared images includes: A high-resolution digital elevation model construction module is used to acquire thermal infrared images and binocular stereo image pairs, perform epipolar correction processing and sparse point cloud computing on the binocular stereo image pairs to obtain target sparse point cloud, and construct a high-resolution digital elevation model based on the target sparse point cloud. The target intersection calculation module is used to obtain the pixel coordinates and intrinsic parameters of the thermal infrared image, construct the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model. The actual surface projection area calculation module is used to calculate the theoretical spot radius based on the target intersection point, and to calculate the actual surface projection area based on the theoretical spot radius. The thermal inertia distribution map generation module is used to calculate the true surface radiance based on the actual surface projection area, and to perform radiance correction and thermal inertia inversion processing on the true surface radiance to obtain the thermal inertia distribution map.

[0014] Furthermore, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a pixel area adaptive correction program for a Mars rover thermal infrared image stored in the memory and executable on the processor, wherein when the pixel area adaptive correction program for the Mars rover thermal infrared image is executed by the processor, the steps of the pixel area adaptive correction method for the Mars rover thermal infrared image as described above are implemented.

[0015] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a pixel area adaptive correction program for a Mars rover thermal infrared image, and when the pixel area adaptive correction program for the Mars rover thermal infrared image is executed by a processor, it implements the steps of the pixel area adaptive correction method for the Mars rover thermal infrared image as described above.

[0016] In this invention, a pair of thermal infrared images and binocular stereo images are acquired. Epipolar correction and sparse point cloud computing are applied to the binocular stereo image pair to obtain a target sparse point cloud. A high-resolution digital elevation model is then constructed based on the target sparse point cloud. The pixel coordinates and intrinsic parameters of the thermal infrared images are acquired. A line-of-sight ray parameter equation is constructed based on the pixel coordinates and intrinsic parameters, and the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model is calculated. The theoretical spot radius is calculated based on the target intersection point, and the actual surface projection area is calculated based on the theoretical spot radius. The true surface radiance is calculated based on the actual surface projection area, and radiance correction and thermal inertia inversion processing are applied to the true surface radiance to obtain a thermal inertia distribution map. This invention achieves pixel-by-pixel area adaptive correction of thermal infrared images by integrating visible light 3D reconstruction technology. This effectively solves the pixel area distortion problem in strabismic imaging and improves the spatial consistency and evaluation accuracy of thermal inertia inversion. Simultaneously, it also effectively improves the accuracy of terrain passability interpretation. Attached Figure Description

[0017] Figure 1 This is a flowchart of a preferred embodiment of the pixel area adaptive correction method for thermal infrared images of Mars rovers according to the present invention; Figure 2 This is a schematic diagram of the overall implementation process of a preferred embodiment of the pixel area adaptive correction method for thermal infrared images of Mars rovers according to the present invention. Figure 3 This is a structural diagram of a preferred embodiment of the pixel area adaptive correction system for thermal infrared images of Mars rovers according to the present invention; Figure 4 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] Currently, Mars rover mobility assessment methods are mainly divided into three categories. The first category is based on visual geometry, which uses navigation and obstacle avoidance cameras to acquire the three-dimensional shape of the terrain and judges traversability by identifying geometric features such as rock size, slope, and roughness. This type of method can only perceive the surface appearance and cannot distinguish the differences in mechanical properties between loose sand and compacted soil; visually similar terrain may have vastly different load-bearing capacities. The second category is based on wheel-ground mechanical parameter identification, which analyzes signals such as wheel drive current and slip rate to infer parameters such as soil cohesion and internal friction angle, thereby assessing the risk of getting stuck. This type of method requires actual contact with the soil and a certain degree of slippage to obtain information, resulting in a passive response lag, and may exacerbate the getting stuck due to active slippage. The third category is based on remote sensing assessment methods using thermal inertia. This method uses thermal infrared cameras to observe the diurnal temperature changes of the surface to invert thermal inertia, distinguishing between soft and compact surfaces from a thermophysical perspective. This method can achieve non-contact, long-distance, and forward-looking assessment, and has unique advantages compared to the first two categories.

[0020] Thermal inertia is a physical parameter characterizing a material's ability to resist temperature changes. For the regolith covering the Martian surface, the level of thermal inertia directly reflects the density of the soil. High thermal inertia corresponds to compacted rock or hard soil, while low thermal inertia corresponds to loose sand or dust. Therefore, by measuring the distribution of thermal inertia of the terrain in front of the Mars rover, soft sandy areas and hard surfaces can be identified, thereby assessing the terrain's traversability.

[0021] During Mars rover exploration, terrain accessibility assessment requires accurate identification of soft sandy areas and rocky regions. Thermal inertia, a physical parameter reflecting soil compaction, can be derived by inverting surface temperature information obtained through a thermal infrared imaging system. The rover's onboard thermal infrared camera quantitatively inverts the thermal inertia value of the target area by measuring the periodic changes in surface temperature, thereby distinguishing between soft and compact terrain. However, this thermal infrared imaging system faces a long-neglected problem in practical applications that seriously affects data quality: pixel area distortion.

[0022] Specifically, thermal infrared cameras employ focal plane array detectors with a fixed instantaneous field of view (IFOV). The instantaneous field of view refers to the imaging angle corresponding to a single detector element in space, influenced by the size of the individual detector element and the lens focal length. When the camera observes the Earth's surface at an angle of tilt, the projected area of ​​a pixel on the horizontal plane varies significantly with the observation distance and angle: near-field pixels have a smaller projected area, while far-field pixels have a larger projected area, and the ground coverage area corresponding to pixels from near to far in the same image can differ by tens of times. Typical thermal infrared camera parameters include an instantaneous field of view (IFOV) of 1.91 mRad, a field of view of 42°×30°, and an infrared resolution of 384×288 pixels. When observing from an oblique angle, the area of ​​the elliptical region covered by the intersection of the far-field pixel IFOV cone and the terrain surface differs greatly from that of the near-field pixels, directly causing the radiance values ​​between different pixels to lose physical comparability.

[0023] This distortion of the projected area leads to two serious problems: First, different pixels correspond to different degrees of surface mixing. Far-field pixels may contain a mixture of various land features (such as sand, rocks, and shadows), causing the radiance value to not represent pure land features. This breaks the physical basis of thermal inertia inversion—the assumption that each pixel corresponds to a homogeneous substance.

[0024] Secondly, there is spatial inconsistency in thermal inertia inversion. Thermal inertia inversion typically assumes that the soil thermophysical properties corresponding to each pixel are uniform. However, projection area distortion breaks this assumption, resulting in systematic deviations in the thermal inertia values ​​retrieved for the same soil type in the near and far fields of the image. For example, the thermal inertia value retrieved in the near field (approximately 1 m) of flat sandy land might be 200 J·m. - ²·K - ¹·s - ¹ / ², while the thermal inertia of the same sandy land inverted in the far field (about 5 m away) may drop to 150 J·m. - ²·K - ¹·s - The error rate is 1 / 2, reaching 25%, which seriously affects the accuracy of terrain passability assessment.

[0025] In summary, the existing technology has the following gaps: there is currently no method that can utilize the 3D terrain information reconstructed by the Mars rover's binocular vision to calculate the actual surface projection area of ​​each pixel in the thermal infrared image and perform adaptive radiance correction based on the projection area ratio to eliminate pixel area distortion caused by distance and slope in oblique imaging, thereby improving the spatial consistency and accuracy of thermal inertia inversion. Therefore, there is an urgent need to develop a pixel area adaptive correction method tailored to the characteristics of Mars rover oblique thermal infrared imaging.

[0026] To address the aforementioned problems, this invention discloses an adaptive correction method for pixel area in Mars rover thermal infrared images that integrates visual three-dimensional terrain information, belonging to the field of environmental perception technology for deep space exploration mobile robots. This invention aims to solve the technical problem in existing Mars rover thermal infrared imaging systems where the tilted angle of the thermal infrared camera observation of the surface leads to inconsistent actual surface projection areas for different pixels, resulting in increased spatial non-uniformity in thermal inertia inversion and improved passability assessment errors.

[0027] The technical process of this invention includes: acquiring binocular stereo images of the terrain ahead using a Mars rover-borne binocular stereo vision system to generate a high-resolution digital elevation model (DEM); acquiring thermal infrared images of the same field of view using a thermal infrared camera; for each pixel in the thermal infrared image, calculating the intersection area of ​​the instantaneous field of view (IFOV) cone of the pixel with the DEM based on the intrinsic parameters of the thermal infrared camera and the line of sight, and determining the actual projected area of ​​the pixel on the ground surface; calculating an area correction coefficient based on the actual projected area and the nominal area of ​​the pixel on the focal plane; correcting the measured radiance of the pixel using the area correction coefficient to obtain the radiance per unit projected area; converting the corrected radiance into temperature and inverting the thermal inertia to obtain a thermal inertia distribution map that eliminates pixel area distortion.

[0028] This invention achieves pixel-by-pixel area adaptive correction of thermal infrared images by integrating visible light 3D reconstruction technology, effectively solving the pixel area distortion problem caused by distance and slope in strabismus imaging, improving the spatial consistency and evaluation accuracy of thermal inertia inversion, and can be widely applied in the field of thermal infrared remote sensing data processing for unmanned mobile exploration platforms such as deep space exploration Mars rovers and lunar rovers.

[0029] The preferred embodiment of the present invention describes a pixel area adaptive correction method for Mars rover thermal infrared images, such as... Figure 1 and Figure 2 As shown, the pixel area adaptive correction method for the thermal infrared image of the Mars rover includes the following steps: Step S10: Acquire thermal infrared image and binocular stereo image pair, perform epipolar correction processing and sparse point cloud computing on the binocular stereo image pair to obtain target sparse point cloud, and construct a high-resolution digital elevation model based on the target sparse point cloud.

[0030] This invention belongs to the field of environmental perception technology for deep space exploration mobile robots, specifically relating to a method for preprocessing thermal infrared images of Mars rovers, and in particular a method for adaptive correction of pixel projection area and radiance compensation of thermal infrared images using three-dimensional terrain information reconstructed by binocular vision.

[0031] Specifically, thermal infrared images are acquired using a thermal infrared camera on the Mars rover, and stereo image pairs are acquired using a binocular stereo navigation camera. Epipolar correction is applied to the stereo image pairs to obtain row-aligned image pairs, and the disparity value of each pixel in the row-aligned image pairs is calculated using a semi-global matching (SGM) algorithm. The target distance from the target spatial point to the camera baseline is calculated based on the disparity value, and the three-dimensional coordinates of the target spatial point are calculated using a camera projection model based on the target distance. A sparse point cloud of the target is generated based on the three-dimensional coordinates, and the sparse point cloud is then meshed and its elevation values ​​are calculated to obtain a high-resolution digital elevation model.

[0032] like Figure 2 As shown, firstly, thermal infrared images and synchronized stereoscopic vision images are acquired. The specific implementation steps are as follows: While the Mars rover is stationary, both the rigidly mounted thermal infrared camera and the stereo navigation camera on its mast simultaneously acquire images. Thermal infrared images. Used for radiometric measurements, the pixel position in its two-dimensional image coordinate system is represented by coordinates. express, The x-coordinate of the pixel position. Let be the ordinate of the pixel position, where , , and These are the image width and height, respectively; binocular stereo image pairs Used for 3D reconstruction The image is for left correction. The image is right-corrected; record the mast's pitch angle at this point. and yaw angle In addition, the intrinsic and extrinsic parameters of the thermal infrared camera and the binocular navigation camera, including the focal length. Principal point coordinates Radial distortion coefficient Tangential distortion coefficient The extrinsic parameters include the translation vector of the camera coordinate system relative to the Mars body coordinate system. and rotation matrix .

[0033] The stereo image pair was acquired using a stereo navigation camera rigidly fixed to the rover's mast. First, when the rover stopped, the thermal infrared camera and the stereo camera were activated for imaging. The left and right cameras of the stereo navigation camera each captured an image with a fixed baseline constraint at the same time. Because the left and right cameras have a fixed baseline distance in space, the pixel coordinates of the same point on the ground will have a lateral positional difference in the left and right images. These two images with parallax constitute the stereo image pair used for 3D reconstruction.

[0034] Furthermore, a high-resolution digital elevation model (DEM) is generated based on a binocular stereo vision system. The specific implementation steps are as follows: 1. Epipolar correction: For binocular stereo images... Perform epipolar correction to align the epipolar lines of the left and right images horizontally, resulting in row-aligned image pairs. , A left-corrected image with row alignment. This is a right-corrected image with row alignment. After correction, the pixel coordinates of the same spatial point in the left and right images satisfy... That is, the vertical coordinates are the same, and the parallax exists only in the horizontal direction. Among them, and This indicates that after epipolar correction, the left and right images are aligned, so the ordinate of the same spatial point is equal in both images, and parallax exists only in the horizontal direction. This represents the ordinate of a spatial point in the left-corrected image; This represents the ordinate of the same spatial point in the right-corrected image.

[0035] The specific steps of epipolar correction are as follows: Input the intrinsic and extrinsic parameters of the binocular camera. The extrinsic parameters include the translation vector and the rotation matrix. Decompose the rotation matrix into two matrices that rotate half of each of the left and right cameras. Construct a correction rotation matrix with the baseline direction as the new X-axis. Calculate the overall rotation matrix of the left and right cameras respectively, that is, multiply the correction rotation matrix by the half rotation matrix of the left and right cameras respectively, and calculate the new intrinsic parameter matrix. Perform pixel-by-pixel reprojection mapping on the left and right images respectively. Finally, generate row-aligned image pairs through bilinear interpolation so that the ordinate of the same spatial point is equal in the left and right images.

[0036] 2. Disparity Calculation: The disparity value of each pixel is calculated using the semi-global matching (SGM) algorithm. The unit is pixels, where disparity is defined as: ; in, and These are the pixel coordinates of the spatial point in the left and right images, respectively. Let x be the x-coordinate of the spatial point in the left-corrected image. The x-coordinate of the same spatial point in the right-corrected image is obtained by performing a pixel-by-pixel search and matching on the epipolar-corrected left and right images using the SGM semi-global matching algorithm. A pixel is selected in the left-corrected image, and its x-coordinate is... The vertical axis is Since the left and right images are already row-aligned, the matching points for this point in the right corrected image are in the same row. The matching algorithm searches from left to right in this row of the right image for the position most similar to the grayscale feature of this pixel. After finding it, the x-coordinate of this position in the right image is recorded. Subtracting the two values ​​gives the disparity value of that pixel. These are the ordinates of the pixels in the left and right images. The larger the parallax, the closer the object is to the camera; the smaller the parallax, the farther away the object is.

[0037] 3. Calculating distance from parallax: Based on the principle of triangulation, spatial points... P Distance to camera baseline Z Calculated by the following formula: ; in, The baseline length of the binocular camera (unit: meters), which is the distance between the optical centers of the left and right cameras; Focal length of the navigation camera (unit: pixels); This is the disparity value of that pixel (in pixels).

[0038] 4. Calculate three-dimensional coordinates : at a known distance Z Based on this, the three-dimensional coordinates of spatial points are calculated using a camera projection model: ; in, The x-coordinate of the three-dimensional coordinate system. The vertical coordinate of the three-dimensional coordinate system. The coordinates of the principal point of the navigation camera (unit: pixels). The x-coordinate of the principal point. The vertical coordinates of the main point coordinates form a sparse point cloud (i.e., the target sparse point cloud).

[0039] 5. Generate Digital Elevation Model (DEM): Interpolate the target sparse point cloud according to a regular grid to generate a high-resolution digital elevation model (DEM). The specific implementation steps are as follows: Divide all the obtained three-dimensional point clouds according to a regular planar grid. For each grid cell, calculate the elevation value of the grid point based on the elevation values ​​of the point cloud falling within or adjacent to the grid, thereby generating a continuous elevation surface.

[0040] Furthermore, the grid spacing of the DEM satisfy: ; in, This represents the ground sampling distance (in meters per pixel) corresponding to one pixel observed by the thermal infrared camera during vertical observation. This condition ensures that the resolution of the DEM is better than three times that of the thermal infrared image, thus supporting pixel-level area calculations.

[0041] Step S20: Obtain the pixel coordinates and intrinsic parameters of the thermal infrared image, construct the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model.

[0042] Specifically, the elevation values ​​in the high-resolution digital elevation model are obtained, and gradient calculation and normalization are performed based on the elevation values ​​to obtain a unit normal vector; the pixel coordinates, intrinsic parameters, and extrinsic parameters of the thermal infrared image are obtained, and a first unit vector of the line-of-sight direction in the camera coordinate system is calculated based on the pixel coordinates and the intrinsic parameters; the line-of-sight direction corresponding to the first unit vector is transformed into a world coordinate system based on the rotation matrix in the extrinsic parameters to obtain a second unit vector of the line-of-sight direction in the world coordinate system; the target position of the camera optical center in the world coordinate system is determined, and a line-of-sight ray parameter equation is constructed based on the target position and the second unit vector; and a ray projection algorithm is used to calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model.

[0043] Furthermore, the terrain normal vector and slope are calculated. The specific implementation steps are as follows: The DEM contains each grid point. 3D coordinates The three-dimensional coordinates are set to elevation values, and the gradient is calculated using the central difference method: ; in, For elevation in X Rate of change of direction; For elevation in Y Rate of change of direction; Indicates the first [number] in the DEM mesh i+1 line, number j The elevation value at the grid point, that is, the elevation value along the current point. X The elevation of the next adjacent grid point in the direction; Indicates the first [number] in the DEM mesh i -1 line, number j The elevation value at the grid point, that is, the elevation value along the current point. X The elevation of the previous adjacent grid point in the direction; This represents the difference between the elevation values ​​of the adjacent grid points behind the current grid point and the elevation values ​​of the adjacent grid points in front of it, reflecting the following: X The change in elevation between two adjacent grid points in the direction. Indicates the first [number] in the DEM mesh i line, number j The elevation value at column +1, i.e., the elevation value along the current point. Y The elevation of the next adjacent grid point in the direction; Indicates the first [number] in the DEM mesh i line, number j The elevation value at the -1 column grid point, i.e., the elevation value along the current point. Y The elevation of the previous adjacent grid point in the direction; This represents the difference between the elevation values ​​of the adjacent grid points behind the current grid point and the elevation values ​​of the adjacent grid points in front of it, reflecting the following: Y The change in elevation between two adjacent grid points in the direction.

[0044] The unnormalized normal vector components are: ; in, For the unnormalized normal vector in X Component of direction; For the unnormalized normal vector in Y Component of direction; This represents the component of the unnormalized normal vector in the Z direction.

[0045] The normalized unit normal vector is: ; Furthermore, slope angle Calculated using the following formula: .

[0046] Furthermore, a precise geometric correspondence is established between the thermal infrared camera and the DEM. The specific implementation steps are as follows: 1. Obtain the intrinsic parameters of the thermal infrared camera: These intrinsic parameters include focal length. Principal point coordinates Radial distortion coefficient Tangential distortion coefficient The extrinsic parameters include the translation vector of the camera coordinate system relative to the Mars body coordinate system. and rotation matrix .

[0047] 2. Convert pixel coordinates to normalized image coordinates: For any pixel in a thermal infrared image Calculate its coordinates on the normalized image plane: ; in, For pixels on the normalized image plane X Direction coordinates; For pixels on the normalized image plane Y Direction coordinates.

[0048] 3. Calculate the unit vector of the viewing direction in the camera coordinate system: The unit vector of the viewing direction (i.e., the direction from the camera optical center to the corresponding spatial point of the pixel) in the camera coordinate system. (The first unit vector) is: .

[0049] 4. Transform the line of sight to the world coordinate system: Use the obtained rotation matrix R Transform the line of sight to the world coordinate system to obtain the second unit vector of the line of sight in the world coordinate system. : ; in, It is the unit vector (i.e., the second unit vector) of the line of sight in the world coordinate system.

[0050] 5. Establish the parametric equations for the line-of-sight ray. Let the position of the camera's optical center in the world coordinate system be... O Then the parametric equation of the line of sight ray is: ; in, t This is the distance parameter (in meters) from the optical center along the line of sight.

[0051] 6. Find the intersection of the ray and the DEM: Use the ray projection algorithm to project the line-of-sight ray. The specific steps for finding the intersection with the DEM are as follows: using a variable step size Proceed along the ray and calculate the horizontal coordinates of the current point of the ray at each step. The surface elevation at that location is obtained by interpolation from the DEM. When the ray height With surface elevation The difference is less than the preset threshold At that time, it was assumed that the rays intersected with the Earth's surface, and the intersection point was recorded. .

[0052] Among them, variable step size The determination of the step size usually takes into account both terrain complexity and computational efficiency: that is, a larger step size is used in flat areas to advance quickly and reduce the amount of computation, while in areas with drastic terrain undulations or large slope changes, the step size is automatically reduced for fine search to avoid skipping the surface intersection due to excessive step size; at the same time, the DEM grid spacing is used as the upper limit of the step size and multiplied by a scaling factor (0~1) to ensure that the sampling density is sufficient to capture terrain details, thereby balancing computational speed and intersection accuracy.

[0053] 7. Output intersection coordinates: For each cell Obtain the intersection of its line of sight and the target on the ground. The intersection point is the center point that the pixel is aiming at on the ground, which is used to calculate the actual projected area later.

[0054] Step S30: Calculate the theoretical spot radius based on the target intersection point, and calculate the actual surface projection area based on the theoretical spot radius.

[0055] Specifically, the observation distance is calculated based on the target location and the intersection point of the target, and the theoretical spot radius is calculated based on the observation distance; the search radius in the high-resolution digital elevation model is determined based on the theoretical spot radius, and the search range is determined based on the search radius; the projection parameters are determined based on the search range, and the vertical distance and the radius of the cone at the depth are calculated based on the projection parameters; the grid of the high-resolution digital elevation model is divided into multiple triangular patches, and the vertical distance and the radius of the cone at the depth corresponding to each vertex position in each triangular patch are calculated. If the vertical distance and the radius of the cone at the depth corresponding to each vertex position satisfy that the vertical distance is less than or equal to the radius of the cone at the depth, then the actual three-dimensional area of ​​each triangular patch is calculated; the actual three-dimensional areas corresponding to all the triangular patches are accumulated to obtain the actual surface projection area.

[0056] To accurately correct pixel area distortion in thermal infrared images caused by changes in observation distance and angle, it is necessary to first determine the size of the actual surface area covered by each pixel. Since each pixel of a thermal infrared camera does not intersect the surface with an ideal ray, but rather projects onto the terrain surface as a cone with a fixed instantaneous field of view (IFOV), the irregular area formed by the intersection of this cone and the real undulating terrain surface is the actual observed surface area of ​​that pixel. This invention accurately calculates the actual projected surface area of ​​each pixel by fusing a high-resolution digital elevation model (DEM) with the camera imaging geometry model.

[0057] like Figure 2 As shown, calculate the actual projected area of ​​the Earth's surface for each pixel in the thermal infrared image. The specific implementation steps are as follows: 1. Obtaining the IFOV parameters of the thermal infrared camera: The instantaneous field of view (IFOV) of the thermal infrared camera (unit: radians) is determined by the detector pixel size. d (Unit: meters) and lens focal length f (Unit: meter) is determined, and its expression is: ; 2. Determine the intersection point of the pixel line of sight and the DEM, i.e., the observation distance: from the pixel... The corresponding surface intersection (i.e., the target intersection) and camera optical center O Calculate the observation distance, where the camera optical center is located. O The position in the world coordinate system is given by the extrinsic parameters.

[0058] 3. Calculate the theoretical spot radius: at the target intersection point Q At point, the radius of the cone's cross-section perpendicular to the line of sight. for: ; in, Intersection of the camera's optical center and the target on the ground surface Q The distance; It is the instantaneous field of view (IFOV), which is the angle subtended by a single pixel.

[0059] 4. Determine the local search range: with Q Centered on the DEM, take the search radius. The expression is: ; in, k For safety factors, the value ranges from 2 to 4, and all DEM grid points within this area will be used for subsequent judgment.

[0060] 5. Determine if a DEM grid point falls within the IFOV cone: For each DEM grid point within the search window P Perform the following steps: A. Calculation P Projection parameters in the line of sight : ; in, Let be the unit vector of the line of sight. If If the point is located behind the camera, it can be eliminated directly.

[0061] B. Calculation P Vertical distance to line of sight : ; Calculate the radius of the cone at that depth. : ; Judgment condition: If ,but P It is covered by the IFOV cone of that pixel.

[0062] 6. Accumulate the area of ​​the covered triangular patches: Divide the DEM mesh into triangular patches (each rectangular mesh is divided into two triangles along its diagonal). For each triangular patch, if all three vertices satisfy... If the criteria are met, then its entire area will be included. If some vertices satisfy the condition, the area of ​​the sub-region falling within the cone is calculated by polygon clipping or centroid coordinate integration and then accumulated.

[0063] Formula for calculating the actual 3D area of ​​a triangular facet (the vertex of the triangular facet is...) A , B , C ): ; A , B , C Let be the three-dimensional coordinate vectors of the three vertices of the triangular facet in the world coordinate system; This represents the actual three-dimensional area of ​​the triangular facet.

[0064] The final result is the actual surface projection area of ​​the pixel. : .

[0065] Step S40: Calculate the true surface radiance based on the actual surface projection area, and perform radiance correction and thermal inertia inversion processing on the true surface radiance to obtain a thermal inertia distribution map.

[0066] Specifically, the target angle between the line of sight and the surface normal is calculated based on the unit normal vector and the first unit vector, and the vertical observation area is calculated based on the target angle and the actual surface projection area.

[0067] Furthermore, the observed zenith angle is calculated. and vertical projected area: using the DEM normal vector Calculate the angle between the line of sight and the normal to the ground surface. : ; The equivalent projected area under vertical observation conditions is derived according to the following steps: According to Lambert's cosine law, the radiance of a Lambert surface in any direction is... L Regardless of the observation direction, the radiation intensity per unit area in the observation direction is proportional to the cosine of the angle between the observation direction and the normal. For an area of... In the surface area, under oblique observation (i.e., the angle between the line of sight and the normal), When the radiant flux received by the sensor is proportional to the projected area of ​​the region perpendicular to the line of sight, that is... Vertical observation ( When the projection area of ​​the same region in the direction perpendicular to the line of sight is itself, that is... .

[0068] To make the radiation flux under tilted observations equivalent to that under vertical observations, a virtual vertical observation area is introduced. This makes their projected areas equal, as expressed by: ; Right now: .

[0069] The physical area and original radiance of each pixel are obtained, and the true radiance of the land surface is calculated based on the physical area of ​​the pixel, the original radiance, and the vertical observation area. An area correction coefficient is calculated based on the physical area of ​​the pixel, the actual projected area of ​​the land surface, and the target angle. The true radiance of the land surface is then corrected based on the original radiance and the area correction coefficient to obtain the initial corrected radiance.

[0070] Furthermore, radiance correction is performed based on the projection area ratio, and the specific implementation steps are as follows: After completing the actual surface projection area of ​​each pixel And observing the zenith angle After calculation, the Lambert cosine law in radiative transmission is used to establish a quantitative relationship between the original sensor radiance and the corrected radiance, thereby eliminating pixel area distortion caused by observation distance and angle.

[0071] Let the physical area of ​​the detector pixel be ( d (where the pixel area is), the original radiance is According to Lambert's cosine law, the actual surface radiance... The law of conservation of energy is satisfied, and its expression is: ; Solve for the true radiance of the Earth's surface : ; To obtain a standardized brightness independent of the observation angle, a vertical observation is taken ( Equivalent value under ) Define area correction factor : ; Then the corrected radiance for: ; The current surface slope is determined. If the current surface slope exceeds a preset slope threshold, the initial corrected radiance is subjected to a secondary correction process to obtain the secondary corrected radiance.

[0072] Furthermore, the radiation direction caused by the slope is corrected, and the specific implementation steps are as follows: When the surface slope exceeds a preset threshold, a secondary correction is performed on the corrected radiance. The angle between the line of sight and the normal to the horizontal plane is calculated. The angle between the line of sight and the normal to the slope (i.e., the observed zenith angle). Based on the Lambert assumption, the corrected formula is: ; Revised Replace the original Used for subsequent calculations.

[0073] The secondary corrected radiance is processed by surface temperature conversion and thermal inertia inversion to obtain a thermal inertia distribution map.

[0074] Furthermore, the corrected thermal inertia distribution map is output. The specific implementation steps are as follows: After completing the radiance correction, the corrected radiance of each pixel is obtained. The radiance is converted into surface temperature, and then the thermal inertia is retrieved to ultimately generate a thermal inertia distribution map that eliminates pixel area distortion and the effects of slope.

[0075] First, the corrected radiance is converted into equivalent surface temperature. T The thermal infrared camera was calibrated using a blackbody radiation source under laboratory conditions to obtain the response function between radiance and blackbody temperature. For the thermal infrared camera onboard the Mars rover, which typically operates in the 8–14 μm band, its response characteristics can be expressed using a linear approximation. ; in, T Equivalent surface temperature, unit: K, which is the true surface temperature value corresponding to each pixel in the thermal infrared image after radiance correction. Where, a and b These are the radiation calibration coefficients, with units of [missing information]. and It was obtained by least-squares fitting of blackbody radiation data.

[0076] Then, the thermal inertia is retrieved using surface temperature information. Thermal inertia Defined as ,in, Thermal conductivity, For density, Specific heat capacity, unit: Thermal inertia reflects the ability of surface materials to resist temperature changes. High thermal inertia corresponds to dense or moist materials (such as rock), while low thermal inertia corresponds to loose materials (such as fine sand). On the surface of Mars, thermal inertia can be estimated by the diurnal amplitude of surface temperature using a simplified model based on a one-dimensional heat conduction equation. P : ; in: The solar constant at Mars' orbit has an average value of approximately ; It is the surface albedo; It is the Stefan-Boltzmann constant; and These are the highest and lowest temperatures at the location of that pixel, respectively, in Kelvin (K). This represents the angular velocity of Mars' rotation.

[0077] Thermal inertia is calculated pixel by pixel. This generates a thermal inertia distribution map that eliminates the effects of area distortion and slope. This map can be used to identify soft terrain (low thermal inertia) and hard surfaces (high thermal inertia), assisting in Mars rover path planning.

[0078] Beneficial effects: Compared with existing technologies, this invention is the first to utilize binocular vision 3D terrain information to correct pixel area distortion in Mars rover thermal infrared images pixel by pixel, thereby improving the spatial consistency of thermal inertia inversion and the reliability of terrain passability assessment. This method relies entirely on the existing hardware of the Mars rover, requires no additional sensors, and can be extended to other deep space exploration platforms such as lunar rovers.

[0079] Furthermore, such as Figure 3 As shown, based on the above-mentioned adaptive correction method for pixel area of ​​Mars rover thermal infrared images, the present invention also provides a corresponding adaptive correction system for pixel area of ​​Mars rover thermal infrared images, wherein the adaptive correction system for pixel area of ​​Mars rover thermal infrared images includes: The high-resolution digital elevation model construction module 51 is used to acquire thermal infrared images and binocular stereo image pairs, perform epipolar correction processing and sparse point cloud computing on the binocular stereo image pairs to obtain target sparse point cloud, and construct a high-resolution digital elevation model based on the target sparse point cloud. The target intersection calculation module 52 is used to obtain the pixel coordinates and intrinsic parameters of the thermal infrared image, construct the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model. The actual surface projection area calculation module 53 is used to calculate the theoretical spot radius based on the target intersection point, and to calculate the actual surface projection area based on the theoretical spot radius. The thermal inertia distribution map generation module 54 is used to calculate the true surface radiance based on the actual surface projection area, and to perform radiance correction and thermal inertia inversion processing on the true surface radiance to obtain the thermal inertia distribution map.

[0080] Furthermore, such as Figure 4 As shown, based on the above-mentioned adaptive correction method and system for pixel area of ​​Mars rover thermal infrared images, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 4 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0081] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a pixel area adaptive correction program 40 for Mars rover thermal infrared images, which can be executed by the processor 10 to implement the pixel area adaptive correction method for Mars rover thermal infrared images in this application.

[0082] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing a pixel area adaptive correction method for the thermal infrared image of the Mars rover.

[0083] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display 30 is used to display information on the terminal and to display a visual user interface.

[0084] In one embodiment, when the processor 10 executes the pixel area adaptive correction program 40 for the Mars rover thermal infrared image in the memory 20, it implements the steps of the pixel area adaptive correction method for the Mars rover thermal infrared image as described above.

[0085] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a pixel area adaptive correction program for a Mars rover thermal infrared image, and when the pixel area adaptive correction program for the Mars rover thermal infrared image is executed by a processor, it implements the steps of the pixel area adaptive correction method for the Mars rover thermal infrared image as described above.

[0086] In summary, this invention provides a method, system, terminal, and storage medium for adaptive correction of pixel area in thermal infrared images of a Mars rover. The method includes: acquiring a thermal infrared image and a pair of binocular stereo images; performing epipolar correction and sparse point cloud computing on the pair of binocular stereo images to obtain a target sparse point cloud; and constructing a high-resolution digital elevation model based on the target sparse point cloud; acquiring the pixel coordinates and intrinsic parameters of the thermal infrared image; constructing a line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters; and calculating the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model; calculating the theoretical spot radius based on the target intersection point; and calculating the actual surface projection area based on the theoretical spot radius; calculating the true surface radiance based on the actual surface projection area; and performing radiance correction and thermal inertia inversion processing on the true surface radiance to obtain a thermal inertia distribution map. This invention achieves pixel-by-pixel area adaptive correction of thermal infrared images by integrating visible light 3D reconstruction technology, which can effectively solve the pixel area distortion problem in strabismus imaging, improve the spatial consistency and evaluation accuracy of thermal inertia inversion, and also effectively improve the accuracy of terrain passability interpretation.

[0087] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0088] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0089] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A pixel area adaptive correction method for thermal infrared images of Mars rovers, characterized in that, The pixel area adaptive correction method for the Mars rover's thermal infrared image includes: Acquire thermal infrared images and binocular stereo images, perform epipolar correction processing and sparse point cloud computing on the binocular stereo images to obtain target sparse point cloud, and construct a high-resolution digital elevation model based on the target sparse point cloud. Obtain the pixel coordinates and intrinsic parameters of the thermal infrared image, construct the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model. The steps of acquiring the pixel coordinates and intrinsic parameters of the thermal infrared image, constructing the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculating the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model specifically include: The elevation values ​​in the high-resolution digital elevation model are obtained, and gradient calculation and normalization are performed based on the elevation values ​​to obtain the unit normal vector. Obtain the pixel coordinates, intrinsic parameters, and extrinsic parameters of the thermal infrared image, and calculate the first unit vector of the viewing direction in the camera coordinate system based on the pixel coordinates and the intrinsic parameters; Based on the rotation matrix in the extrinsic parameters, the viewing direction corresponding to the first unit vector is transformed into a world coordinate system to obtain a second unit vector of the viewing direction in the world coordinate system. Determine the target position of the camera optical center in the world coordinate system, construct the line-of-sight ray parameter equation based on the target position and the second unit vector, and use a ray projection algorithm to calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model; The theoretical spot radius is calculated based on the target intersection point, and the actual surface projection area is calculated based on the theoretical spot radius. The step of calculating the theoretical spot radius based on the target intersection point and calculating the actual surface projection area based on the theoretical spot radius specifically includes: The observation distance is calculated based on the target location and the intersection point of the target, and the theoretical spot radius is calculated based on the observation distance; The search radius in the high-resolution digital elevation model is determined based on the theoretical spot radius, and the search range is determined based on the search radius; The projection parameters are determined based on the search range, and the radius of the cone at the vertical distance and depth is calculated based on the projection parameters. The high-resolution digital elevation model is divided into meshes to obtain multiple triangular patches. The actual three-dimensional area is calculated based on the vertex positions of the multiple triangular patches, the vertical distance, and the radius of the cone at the depth, to obtain the actual surface projection area. The true surface radiance is calculated based on the actual surface projection area, and the true surface radiance is then corrected for radiance and subjected to thermal inertia inversion to obtain a thermal inertia distribution map.

2. The pixel area adaptive correction method for Mars rover thermal infrared images according to claim 1, characterized in that, The process of acquiring thermal infrared images and binocular stereo images, performing epipolar correction and sparse point cloud computing on the binocular stereo images to obtain a target sparse point cloud, and constructing a high-resolution digital elevation model based on the target sparse point cloud specifically includes: Thermal infrared images were acquired using the thermal infrared camera on the Mars rover, and stereo image pairs were acquired using the binocular stereo navigation camera. The stereo image pairs are subjected to epipolar correction to obtain row-aligned image pairs, and the disparity value of each pixel in the row-aligned image pairs is calculated using the semi-global matching (SGM) algorithm. The target distance from the target spatial point to the camera baseline is calculated based on the disparity value, and the three-dimensional coordinates of the target spatial point are calculated using the camera projection model based on the target distance. A target sparse point cloud is generated based on the three-dimensional coordinate points, and the target sparse point cloud is divided into grids and its elevation values ​​are calculated to obtain a high-resolution digital elevation model.

3. The pixel area adaptive correction method for Mars rover thermal infrared images according to claim 1, characterized in that, The step of calculating the actual three-dimensional area based on the vertex positions and vertical distances of the multiple triangular facets and the radius of the cone at the depth, to obtain the actual surface projection area, specifically includes: Calculate the vertical distance corresponding to each vertex position and the radius of the cone at the depth in each of the triangular facets. If the vertical distance corresponding to each vertex position and the radius of the cone at the depth both satisfy that the vertical distance is less than or equal to the radius of the cone at the depth, then calculate the actual three-dimensional area of ​​each triangular facet. The actual three-dimensional areas corresponding to all the triangular facets are summed to obtain the actual surface projection area.

4. The pixel area adaptive correction method for Mars rover thermal infrared images according to claim 1, characterized in that, The calculation of the true surface radiance based on the actual surface projection area specifically includes: Calculate the target angle between the line of sight direction and the surface normal based on the unit normal vector and the first unit vector, and calculate the vertical observation area based on the target angle and the actual surface projection area. The physical area and original radiance of the pixels are obtained, and the true radiance of the Earth's surface is calculated based on the physical area of ​​the pixels, the original radiance, and the vertical observation area.

5. The pixel area adaptive correction method for Mars rover thermal infrared images according to claim 4, characterized in that, The process of performing radiance correction and thermal inertia inversion on the actual surface radiance to obtain a thermal inertia distribution map specifically includes: The area correction coefficient is calculated based on the physical area of ​​the pixel, the actual surface projection area, and the target angle. The actual surface radiance is then corrected based on the original radiance and the area correction coefficient to obtain the initial corrected radiance. Determine the current surface slope. If the current surface slope exceeds a preset slope threshold, perform a secondary correction on the initial corrected radiance to obtain the secondary corrected radiance. The secondary corrected radiance is processed by surface temperature conversion and thermal inertia inversion to obtain a thermal inertia distribution map.

6. A pixel area adaptive correction system for thermal infrared images of a Mars rover, characterized in that, The pixel area adaptive correction system for the Mars rover's thermal infrared image is used to implement the pixel area adaptive correction method for the Mars rover's thermal infrared image as described in any one of claims 1-5. The pixel area adaptive correction system for the Mars rover's thermal infrared image includes: A high-resolution digital elevation model construction module is used to acquire thermal infrared images and binocular stereo image pairs, perform epipolar correction processing and sparse point cloud computing on the binocular stereo image pairs to obtain target sparse point cloud, and construct a high-resolution digital elevation model based on the target sparse point cloud. The target intersection calculation module is used to obtain the pixel coordinates and intrinsic parameters of the thermal infrared image, construct the line-of-sight ray parameter equation based on the pixel coordinates and intrinsic parameters, and calculate the target intersection point between the line-of-sight ray parameter equation and the high-resolution digital elevation model. The actual surface projection area calculation module is used to calculate the theoretical spot radius based on the target intersection point, and to calculate the actual surface projection area based on the theoretical spot radius. The thermal inertia distribution map generation module is used to calculate the true surface radiance based on the actual surface projection area, and to perform radiance correction and thermal inertia inversion processing on the true surface radiance to obtain the thermal inertia distribution map.

7. A terminal, characterized in that, The terminal includes: a memory, a processor, and a pixel area adaptive correction program for the Mars rover's thermal infrared image stored in the memory and executable on the processor. When the pixel area adaptive correction program for the Mars rover's thermal infrared image is executed by the processor, it implements the steps of the pixel area adaptive correction method for the Mars rover's thermal infrared image as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a pixel area adaptive correction program for a Mars rover thermal infrared image, which, when executed by a processor, implements the steps of the pixel area adaptive correction method for a Mars rover thermal infrared image as described in any one of claims 1-5.

Images