General geometric correction method for push-broom and wobble-broom linear array imaging

Abstract

The application provides a general system geometry correction method for push-broom and swing-broom linear array imaging, and belongs to the technical field of data processing.The method comprises the following steps: establishing a strict imaging model, wherein the coordinate conversion process of the strict imaging model comprises general coordinate conversion and image coordinate system to camera coordinate system conversion, the general coordinate conversion comprises conversion among a camera coordinate system, a satellite body coordinate system, a space inertial coordinate system J2000 and a geocentric geodetic coordinate system WGS84, and the image coordinate system to camera coordinate system conversion comprises the coordinate conversion process of swing-broom camera data and push-broom camera data, the swing-broom camera data is standardized into push-broom camera data based on a bias matrix and a rotation matrix, and then subsequent general coordinate conversion is performed; and target image data is subjected to system geometry correction based on the strict imaging model.The application realizes unified correction of two types of imaging data, i.e.push-broom and swing-broom imaging data.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a general system geometric correction method for pushbroom and swing-broom linear array imaging. Background Technology

[0002] With the rapid development of aerospace remote sensing technology, high-resolution linear array imaging cameras are increasingly widely used in surveying, resource exploration, and environmental monitoring. Pushbroom and swingbroom cameras, as two mainstream systems of linear array imaging, meet the observation needs of different scenarios due to their high geometric stability and flexible scanning capabilities, respectively. However, the two types of cameras differ significantly in their imaging mechanisms and data geometric characteristics. Traditional processing methods often require the construction of independent geometric correction systems for different systems, leading to complex data processing workflows. Therefore, establishing a universal system geometric correction method compatible with both pushbroom and swingbroom linear array imaging data is of crucial engineering application value for simplifying ground processing system architecture, improving the efficiency of collaborative applications of multi-source remote sensing data, and achieving high-precision unified geolocation.

[0003] Currently, the system geometric correction for pushbroom and swing-broom line scan images generally employs discrete modeling or approximate unification based on empirical formulas. Existing technologies typically construct correction methods based on rigorous imaging models for pushbroom line scan cameras, while designing separate geometric models including scan angle compensation for swing-broom line scan cameras. These two methods are independent in terms of coordinate transformation logic and parameter solving. Some processing methods often directly embed the scan angle parameters of the swing-broom data into the pushbroom model to establish the swing-broom model, or eliminate system differences through empirical methods such as polynomial fitting, lacking a unified description of the imaging physical processes of the two types of cameras.

[0004] In summary, how to unify the coordinate system and physical model of push-broom and swing-broom linear array imaging to achieve high-precision universal processing has become an urgent technical problem to be solved. Summary of the Invention

[0005] This application provides a general system geometric correction method for pushbroom and swingbroom linear array imaging to solve the problem of high system development and maintenance costs caused by inconsistent models in the prior art.

[0006] This application provides a general system geometric correction method for pushbroom and swingbroom linear array imaging, including the following steps: A rigorous imaging model is established. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. Based on the aforementioned rigorous imaging model, systematic geometric correction is performed on the target image data.

[0007] According to the general system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application, the rigorous imaging model is as follows: ; in: ; k The scale factor is obtained by calculating the intersection of the view vector and the Earth reference ellipsoid in the WGS84 coordinate system. t For satellite camera imaging time parameters, This is the rotation matrix from the satellite camera coordinate system to the satellite body coordinate system. This is the rotation matrix from the satellite body coordinate system to the J2000 coordinate system. Let be the rotation matrix from the J2000 coordinate system to the WGS84 coordinate system, 𝑅(𝛼(t)) be the rotation matrix for the sweep angle, and 𝛼(t) be the sweep angle. The bias matrix, To obtain the WGS84 coordinates of GPS satellites, These are the WGS84 coordinates corresponding to the image points. For the camera's physical layer matrix, x For the pixel sequence number, p Principal point coordinates, d p For pixel size, f Main distance, , This refers to pixel distortion.

[0008] According to the general system geometric correction method for pushbroom and swing-broom linear array imaging provided in this application, when the target image data of the swing-broom type is a satellite ascending orbit image, the target image data of the swing-broom type is rotated by 0° and then matrix transposed before storage. When the target image data of the sweeping type is satellite deorbiting image, the sweeping target image data is rotated by 180° and then matrix transposed before storage.

[0009] According to the general system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application, the step of performing system geometric correction on the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. In the corrected image plane, for each pixel, its position on the target image plane is calculated based on the strict imaging model, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

[0010] According to the general system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application, the step of performing system geometric correction on the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. Several sampling points are selected from the target image data, and the ground coordinates corresponding to each sampling point are calculated based on the rigorous imaging model to obtain several sets of image point-ground coordinate data pairs; Based on the aforementioned sets of image point-ground coordinate data pairs, the coefficients of the rational function model are calculated. In the corrected image plane, for each pixel, its position on the target image plane is calculated based on the rational function model, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

[0011] According to the general system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application, the step of performing system geometric correction on the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. Several sampling points are selected from the corrected image plane, and the image coordinates corresponding to each sampling point are calculated based on the strict imaging model to obtain several sets of image point-ground coordinate data pairs. The corrected image plane is divided into several small blocks. For each small block, the coefficients of the corresponding polynomial are calculated based on several sets of image point-ground coordinate data pairs within the small block. In the corrected image plane, for each pixel, its position on the target image plane is obtained by inverse polynomial calculation based on the small block corresponding to the pixel, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

[0012] This application also provides a universal system geometric correction device for pushbroom and swingbroom linear array imaging, comprising the following modules: The model building module is used to: build a rigorous imaging model. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. The data correction module is used to perform systematic geometric correction on the target image data based on the rigorous imaging model.

[0013] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a general system geometric correction method for push-broom and swing-broom linear array imaging as described above.

[0014] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a general system geometric correction method for push-broom and swing-broom linear array imaging as described above.

[0015] This application also provides a computer program product, including a computer program that, when executed by a processor, implements a general system geometric correction method for pushbroom and swing-broom linear array imaging as described above.

[0016] This application provides a general system geometric correction method for pushbroom and swing-broom linear array imaging. It introduces a processing logic that normalizes swing-broom camera data to pushbroom camera data based on bias and rotation matrices. First, the offset of the swing-broom camera's scanning center is compensated using the bias matrix. Then, the rotation matrix is ​​used to convert the dynamic pointing of the scanning mirror into a static coordinate frame consistent with that of the pushbroom camera. Mathematically, this reduces the imaging geometry of the swing-broom data to the standard pushbroom model. This standardization process eliminates the essential differences in image coordinate system definitions between the two types of cameras, allowing for a unified use of the pushbroom camera's coordinate transformation process to achieve a rigorous transformation from image coordinates to camera coordinates, and then to the geocentric coordinate system. Because this method unifies the system geometric correction methods for both types of data based on a rigorous imaging model, it avoids the systematic errors introduced by parameter fitting in traditional empirical methods. Furthermore, by sharing a common coordinate system transformation, it significantly reduces model complexity. Ultimately, both pushbroom-based raw data and standardized swing-broom-based raw data can achieve high-precision system geometric correction under the same rigorous imaging model. This ensures consistency in coordinate system transformation across different image formats and improves the processing efficiency and positioning accuracy of the general system, effectively solving the problem of high development and maintenance costs caused by model fragmentation in existing technologies. In summary, this application achieves a unified system geometric correction method for two types of imaging data by establishing a rigorous imaging model that includes general coordinate system transformation and image-to-camera coordinate system transformation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the general system geometric correction method for push-broom and swing-broom linear array imaging provided in this application. Figure 2 This is a schematic diagram of the structure of the rigorous imaging model provided in this application; Figure 3 This is a flowchart illustrating the L1A-level image geometric correction method provided in this application; Figure 4 This is a flowchart illustrating the geometric correction example provided in this application; Figure 5 This is a schematic diagram of the structure of the universal system geometric correction device for push-broom and swing-broom linear array imaging provided in this application; Figure 6 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] The following is combined Figures 1 to 6 This application describes a general system geometric correction method for push-broom and swing-broom linear array imaging.

[0021] Figure 1 This is a flowchart illustrating the general system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application, as shown below. Figure 1 As shown, the method includes the following: S110, establish a rigorous imaging model; S120, Based on the aforementioned rigorous imaging model, perform systematic geometric correction on the target image data.

[0022] In this embodiment, the coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation.

[0023] It should be noted that the execution entity of the universal system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application embodiment can be a server or computer device, such as a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA). The execution entity of this universal system geometric correction method for pushbroom and swingbroom linear array imaging can also be a geometric correction system, which belongs to the aforementioned server or computer device.

[0024] Here, the linear array detectors of the pushbroom camera are arranged perpendicular to the flight direction. As the satellite flies forward, it continuously acquires images. Its image coordinate system is relatively stationary. The acquisition of each row of images depends only on the forward motion of the satellite, and the geometric relationship is simple and stable. The swing-broom camera changes the line of sight by mechanically swinging the reflector, thereby performing a wide scan perpendicular to the flight direction. Its imaging process introduces a high-frequency changing scanning angle, resulting in a more complex geometric model than the pushbroom.

[0025] In S110, the rigorous imaging model is a rigorous mathematical model based on the physical mechanism of satellite imaging. It does not rely on empirical formulas (such as polynomial fitting), but uses the satellite's orbital parameters, attitude parameters, and the camera's interior and exterior orientation elements to accurately describe the physical process of how each pixel in the image reflects light from the ground and projects it onto the detector.

[0026] like Figure 2 As shown, universal coordinate system transformation refers to the chain of mapping data from the camera coordinate system to the geocentric and geofixed coordinate system in remote sensing data processing. The origin of the camera coordinate system is located at the center of the camera lens, and the Z-axis points in the direction of the optical axis. It is used to describe the local spatial position of the image point on the camera. The satellite body coordinate system usually takes the satellite's center of mass or the camera's mounting reference point as its farthest point. The X-axis is perpendicular to the orbit and roughly parallel to the orbital plane, the Y-axis is the direction of satellite flight, and the Z-axis is determined by the right-hand rule and usually points to the Earth. The space inertial coordinate system J2000 is centered on the Earth's center of mass, with the X-axis pointing to the vernal equinox, the Z-axis pointing to the North Pole, and the Y-axis determined by the right-hand rule. The geocentric and geofixed coordinate system WGS84 (World Geodetic Coordinate System 1984) is the currently universally used geocentric and geofixed coordinate system. The X-axis points to the intersection of the Greenwich Meridian and the equator, the Z-axis points to the North Pole, and the Y-axis is determined by the right-hand rule. It can be used to convert to latitude, longitude, and elevation coordinates.

[0027] Here, the bias matrix is ​​used to compensate for the field of view center offset caused by mirror scanning of the oscillating camera or the mechanical eccentricity of the detector installation; the rotation matrix is ​​used to describe the angular transformation between the coordinate axes, which means converting the dynamic scanning angle of the oscillating mirror into an equivalent camera attitude angle, making it mathematically equivalent to the observation vector of the pushbroom camera.

[0028] In this embodiment, for pushbroom camera data, the observation vector in the camera coordinate system is calculated directly based on the detector's focal length and pixel position. For swing-broom camera data, since swing-broom has dynamic mirror scanning action, the scanning angle parameter at the scanning moment is first extracted. Then, the mechanical error of the scanning center is corrected using the offset matrix. Finally, the reflected light path vector of the scanning mirror is mathematically transformed into an equivalent virtual pushbroom observation vector using the rotation matrix. This makes the swing-broom camera data completely consistent with the pushbroom camera data in mathematical structure, so that there is no need to distinguish the imaging system afterward, and the same set of coordinate transformation algorithms can be reused.

[0029] In the specific implementation process, in addition to the real-time sweep angle generated by the continuous swinging of the reflector of the sweeping camera during operation, which generates a sweep angle rotation matrix that changes with time, the linear array arrangement direction of the sweeping detector is inconsistent with that of the pushbroom detector. The arrangement direction can be transformed by the rotation matrix. At this time, the sweep angle rotation matrix and the rotation matrix are mathematically merged. Through the merged total rotation matrix, the original pixel coordinates acquired by the sweeping camera at any time are directly mapped to the standard pushbroom camera coordinate system in one step. This simplifies the code architecture of the ground processing system and fundamentally eliminates the systematic geometric distortion caused by the mismatch between the two system models.

[0030] It is understandable that, since the coordinate mapping method from linear array pixels to the physical layer of the camera is the same for pushbroom and swingbroom linear array satellite cameras, a unified mapping model is used for conversion, including but not limited to mapping models based on parameters such as principal point, principal distance, and pixel distortion, as well as methods such as pointing angle model for coordinate mapping.

[0031] In S120, the target image data to be processed (whether push-broom or swing-broom) and its corresponding satellite auxiliary data (ephemeris, attitude quaternions, etc.) are input into the model. Using the established rigorous imaging model and combined with the digital elevation model, the ground latitude and longitude corresponding to each pixel on the image are calculated. Based on the calculated geographic coordinates, the boundary and resolution of the corrected image are determined. The original distorted image is resampled into a standard orthorectified map product through grayscale interpolation.

[0032] The general system geometric correction method for pushbroom and swing-broom linear array imaging provided in this application introduces a processing logic for standardizing swing-broom camera data to pushbroom camera data based on an offset matrix and a rotation matrix. First, the offset matrix compensates for the scanning center offset of the swing-broom camera. Then, the rotation matrix converts the dynamic pointing of the scanning mirror into a static coordinate frame consistent with that of the pushbroom camera. Mathematically, this reduces the imaging geometry of the swing-broom data to the standard pushbroom model. This standardization process eliminates the essential differences in the definition of the image coordinate system between the two types of cameras, allowing for a unified use of the coordinate transformation process of the pushbroom camera to complete a rigorous transformation from image coordinates to camera coordinates, and then to the geocentric coordinate system. Because this method unifies the system geometric correction method for the two types of data based on a rigorous imaging model, it avoids the systematic errors introduced by parameter fitting in traditional empirical methods. Furthermore, by sharing a common coordinate system transformation, it significantly reduces model complexity. Ultimately, both pushbroom-based raw data and standardized swing-broom-based raw data can achieve high-precision system geometric correction under the same rigorous imaging model. This ensures consistency in coordinate system transformation across different image formats and improves the processing efficiency and positioning accuracy of the general system, effectively solving the problem of high development and maintenance costs caused by model fragmentation in existing technologies. In summary, this application achieves a unified system geometric correction method for two types of imaging data by establishing a rigorous imaging model that includes general coordinate system transformation and image-to-camera coordinate system transformation.

[0033] In an optional embodiment, the rigorous imaging model is as follows: ; in: ; k The scale factor is obtained by calculating the intersection of the view vector and the Earth reference ellipsoid in the WGS84 coordinate system. t For satellite camera imaging time parameters, This is the rotation matrix from the satellite camera coordinate system to the satellite body coordinate system. This is the rotation matrix from the satellite body coordinate system to the J2000 coordinate system. Let be the rotation matrix from the J2000 coordinate system to the WGS84 coordinate system, 𝑅(𝛼(t)) be the rotation matrix for the sweep angle, and 𝛼(t) be the sweep angle. The bias matrix, To obtain the WGS84 coordinates of GPS satellites, These are the WGS84 coordinates corresponding to the image points. For the camera's physical layer matrix, x For the pixel sequence number, p Principal point coordinates, d p For pixel size, f Main distance, , This refers to pixel distortion.

[0034] Here, It is the rotation matrix from the satellite camera coordinate system to the satellite body coordinate system, i.e., the installation matrix, which is used to describe the fixed installation angle deviation of the camera on the satellite platform; It is the rotation matrix from the satellite body coordinate system to the J2000 space inertial coordinate system, which is usually provided by attitude measurement equipment such as star sensors, reflecting the real-time attitude of the satellite in inertial space. It is the rotation matrix from the J2000 coordinate system to the WGS84 geocentric-earth-fixed coordinate system, mainly used to compensate for coordinate differences caused by the Earth's rotation. It is a dynamic rotation matrix specifically designed for sweeping cameras, in which It is the real-time sweep angle that changes with time t. The function of this matrix is ​​to incorporate the angle changes introduced by dynamic scanning into a unified coordinate transformation chain. The offset matrix is ​​used to compensate for mechanical eccentricity or fixed line-of-sight offset that occurs during the installation of the oscillating scanning mechanism or detector; Describes the normalized 3D coordinates of a single pixel on an ideal focal plane; size factor k In the WGS84 coordinate system, the scaling factor extends along the line-of-sight vector until it intersects with the Earth's reference ellipsoid (whose elevation values ​​are described by the Digital Elevation Model, DEM). It determines the actual length of the line of sight, thereby calculating the absolute three-dimensional coordinates of the ground. .

[0035] The general system geometric correction method for pushbroom and swing-broom linear array imaging provided in this application embodiment performs the following on swing-broom data: The operation defaults to pushbroom data. As an identity matrix with [D] as the zero vector, the observation vectors of the two types of cameras are unified into a completely consistent expression at the mathematical level. All subsequent coordinate transformations and ground intersection calculations can reuse the same core algorithm module, simplifying the software architecture of the ground data processing system, reducing the cost of code development, testing, and maintenance, and avoiding the risk of logical inconsistencies that may arise from the parallel operation of the two systems. A sweep angle rotation matrix that varies with time t is explicitly introduced into the rigorous imaging model. The bias matrix [D] strictly follows the physical optical path laws of light scanning through the mirror, accurately restoring the true line-of-sight direction of the scanning mirror at different times, row by row and pixel by pixel. This effectively compensates for the complex geometric distortions caused by large-angle side views and rapid oscillation scanning, enabling oscillation-scanning images to achieve a high-precision geometric positioning level comparable to push-broom images, meeting the requirements for high-quality, large-area imaging. Furthermore, all the camera's internal physical properties are encapsulated in the underlying layer... Within the matrix, external spatial attitude transformations such as the sweep angle are encapsulated in a subsequent rotation matrix chain. This makes the calibration of internal physical parameters completely independent of the processing of external attitude and trajectory. When the orientation elements within the camera experience slight drifts, only individual updates are needed. The calculation parameters can be obtained without changing the coordinate transformation framework of the entire rigorous imaging model, thereby improving the robustness of the system.

[0036] In an optional embodiment, when the sweeping target image data is a satellite ascending orbit image, the sweeping target image data is rotated by 0° and matrix transposed before storage. When the target image data of the sweeping type is satellite deorbiting image, the sweeping target image data is rotated by 180° and then matrix transposed before storage.

[0037] In this embodiment, the storage method for L1A-level image pixels is based on the physical arrangement order of the linear array pixels and the increasing imaging time order. By setting a rotation matrix and matrix transpose, the storage method for L1A-level image data of pushbroom and swing-broom linear array satellite cameras is unified. L1A-level image data is the raw image data that has undergone basic processing but has not yet undergone system geometric correction, i.e., target image data.

[0038] Here, ascending orbit refers to the satellite's orbital process from south to north; descending orbit refers to the satellite's orbital process from north to south. Because the flight directions differ by approximately 180°, the pixel arrangement order along the orbital direction of the raw image data acquired by the same linear array camera during ascending and descending orbits is naturally reversed.

[0039] In the specific implementation process, auxiliary metadata of the target image data, such as header files or XML files, is read to extract the satellite's flight trajectory information and automatically determine whether the current image belongs to ascending or descending orbit data. If the flight direction is determined to be in the positive direction of the reference, i.e., ascending orbit data, the original image is rotated 0°, keeping the original scan sequence unchanged, and then directly matrix transpose is performed. If the flight direction is determined to be opposite to the reference, i.e., descending orbit data, the original image is rotated 180°, and then matrix transpose is also performed. After the above processing, the matrix arrangement logic in the computer memory is consistent for both ascending and descending orbit data. This standardized matrix is ​​written to the storage device for subsequent use by the rigorous imaging model.

[0040] The general system geometric correction method for pushbroom and swing-broom linear array imaging provided in this application provides a method that forcibly rotates the deorbiting data by 180° before storage, so that all images present a uniform physical orientation in the first dimension of the matrix. The subsequent transpose operation maps this uniform orbital orientation to a specific dimension of the matrix. The subsequent geometric correction algorithm no longer needs to care whether the satellite is ascending or descending, but only needs to read the data according to the fixed matrix dimension, thereby eliminating the branch judgment logic inside the algorithm and significantly reducing the coupling of the code and the probability of error.

[0041] In an optional embodiment, the systematic geometric correction of the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. In the corrected image plane, for each pixel, its position on the target image plane is calculated based on the strict imaging model, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

[0042] like Figure 3 As shown in this embodiment, the latitude and longitude of the image corners and the spatial resolution of the L1A-level image with a unified storage method are calculated using a rigorous imaging model that takes into account both pushbroom and swing-broom linear array imaging. This determines the latitude and longitude range of the image and the number of rows and columns of pixels in the L2A-level image. Then, the pixel values ​​are back-calculated from the L1A-level image to the resampled pixel values ​​in the L2A-level image point by point in the L2A-level image plane using the rigorous imaging model, thus obtaining the L2A-level image. L2A-level image refers to an image product with accurate geographic coordinates that has undergone system geometric correction, i.e., corrected image data.

[0043] Here, spatial resolution refers to the actual sampling interval of the target image on the ground, usually measured in degrees. The corrected image plane is a virtual, regular digital grid, its four corner points defined by calculated latitude and longitude ranges, and its interior composed of equally spaced rows and columns of pixels. It serves as a blank standard map, awaiting filling in the texture of the original image. Resampling calculates the grayscale or color value that the current pixel location should have based on neighboring integer pixel values ​​using specific mathematical interpolation algorithms, such as bilinear interpolation or cubic convolution.

[0044] In the specific implementation process, the rigorous imaging model established earlier is first called to project the four corner points (upper left, upper right, lower left, and lower right) of the original image into the WGS84 geocentric geofixed coordinate system and transform them to obtain their corresponding latitude and longitude coordinates. Then, combined with the coverage of these corner points and the calculated spatial resolution, a regular rectangular area is delineated. The length and width of this area are determined by the latitude and longitude span and the set spatial resolution, thereby generating the corrected image plane.

[0045] Furthermore, the process iterates through every pixel on the corrected image plane. For each pixel, its geographic coordinates are substituted into the rigorous imaging model for inverse calculation to determine its row and column numbers on the original target image plane. These row and column numbers are not integers but floating-point coordinates. Based on these floating-point coordinates, the neighboring pixels in the original image are found, for example, the four or sixteen surrounding pixels. The grayscale value of the new pixel is calculated using a resampling algorithm and then filled into its current position on the corrected image plane. Once all pixels on the corrected image plane have undergone the inverse calculation and resampling, the output image is the corrected image data with accurate geographic coordinates and free from geometric distortion.

[0046] The general system geometric correction method for push-broom and swing-broom linear array imaging provided in this application embodiment ensures, from an algorithmic logic perspective, that each pixel of the output image has one and only one definite grayscale value by performing inverse calculation on each pixel of the corrected image plane, thereby eliminating the loss or redundancy of image information and ensuring the visual continuity and integrity of the corrected image.

[0047] In an optional embodiment, the systematic geometric correction of the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. Several sampling points are selected from the target image data, and the ground coordinates corresponding to each sampling point are calculated based on the rigorous imaging model to obtain several sets of image point-ground coordinate data pairs; Based on the aforementioned sets of image point-ground coordinate data pairs, the coefficients of the rational function model are calculated. In the corrected image plane, for each pixel, its position on the target image plane is calculated based on the rational function model, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

[0048] Here, the Rational Function Model (RFM) is a mathematical model that uses rational polynomials (i.e., the ratio of two polynomials) to approximate the mapping relationship between image pixel coordinates and ground geographic coordinates. The coefficients of the RFM include 80 core coefficients and 10 scaling and offset parameters used for coordinate normalization.

[0049] Here, the sampling points are virtual control points selected to generate RFM. By selecting a large number of pixels on the original image according to certain rules (such as uniform grid), and using a high-precision rigorous imaging model to calculate their corresponding real ground coordinates, an image point-ground point dataset for fitting RFM is constructed.

[0050] In the specific implementation process, firstly, based on the established rigorous imaging model, the latitude and longitude of the four corner points of the original image are calculated. Combined with the spatial resolution of the image, the range of the final corrected image, i.e. the corrected image plane, is defined. On the original image, several sampling points are selected evenly at a certain density, for example, every few tens of pixels. The precise ground coordinates corresponding to these sampling points are calculated one by one using the rigorous imaging model, thereby obtaining a large number of image point-ground coordinate data pairs. Then, these data pairs are used as input, and numerical optimization algorithms such as the least squares method are used to fit rational polynomial coefficients that can best express these point-positional relationships.

[0051] Furthermore, the algorithm iterates through each pixel on the corrected image plane, calls the RFM model calculated above, and quickly calculates the row and column positions on the original image corresponding to the geographic coordinates. Based on the calculated positions, the original image is resampled to obtain pixel values, which are then filled into the corrected image plane. Once all pixels are filled, the system-corrected standard image data can be output.

[0052] The general system geometric correction method for push-broom and swing-broom linear array imaging provided in this application improves the overall processing efficiency of system geometric correction by pre-sampling and fitting RFM and using rational polynomial model formulas to replace strict imaging models for fast calculation in the pixel-by-pixel correction stage.

[0053] In an optional embodiment, the systematic geometric correction of the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. Several sampling points are selected from the corrected image plane, and the image coordinates corresponding to each sampling point are calculated based on the strict imaging model to obtain several sets of image point-ground coordinate data pairs. The corrected image plane is divided into several small blocks. For each small block, the coefficients of the corresponding polynomial are calculated based on several sets of image point-ground coordinate data pairs within the small block. In the corrected image plane, for each pixel, its position on the target image plane is obtained by inverse polynomial calculation based on the small block corresponding to the pixel, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

[0054] Here, a small patch refers to a number of rectangular sub-regions logically divided from the entire remote sensing image. The polynomial coefficients are mathematical formula parameters fitted using the least squares method for each small patch, using the sampling point data pairs within that region. These parameters may include the 6 parameters of an affine transformation or the 12 parameters of a quadratic polynomial. These coefficients reflect the mathematical patterns of image distortion within that local region.

[0055] In the specific implementation process, based on the rigorous imaging model, the latitude and longitude of the four corner points of the original image are calculated. Combined with the spatial resolution of the image, the range of the final corrected image, i.e. the corrected image plane, is defined. The original image of the whole scene is divided into several small blocks. For each small block, points on the boundary are selected evenly as sampling points. The precise ground coordinates corresponding to them are calculated using the rigorous imaging model to form an image point-ground coordinate data pair dataset.

[0056] Furthermore, for each small block, the polynomial coefficients of that block are independently calculated using the least squares method. Then, every pixel on the corrected image plane is traversed. First, it is determined which small block in the original image the pixel corresponds to. Then, the polynomial coefficients corresponding to that small block are used to quickly calculate the row and column position of the pixel in the original image. Subsequently, a resampling algorithm is used to obtain the grayscale value and fill it into the corrected image plane. After all pixels have completed the inverse calculation and assignment, the image data after system geometric correction is obtained.

[0057] The general system geometric correction method for pushbroom and swingbroom linear array imaging provided in this application divides the image into several small blocks. Within each local area, complex nonlinear distortions can be approximated as smooth linear or low-order surface changes. Low-order polynomials can approximate the local true imaging geometry with extremely high accuracy. This avoids the instability of global high-order models and ensures consistent and high-precision correction results whether in the image center or at the edge. Furthermore, reducing complex physical relationships to simple polynomial formulas significantly reduces the computational load of system geometric correction, improves data processing efficiency, and meets the needs of large-scale data production. In addition, under the block-based mechanism, the polynomial of each block is solved independently. Abnormal data within a block only affects the correction effect of that local area and does not affect the entire image, thereby improving the system's fault tolerance and enhancing the algorithm's robustness to heterogeneous data and local anomalies.

[0058] The following is combined Figure 4 Taking the multi-track data from the wide-swath thermal infrared imager (GF5A-WTI) of the Hyperspectral Integrated Observation Satellite as an example, the system geometric correction method of this application is explained.

[0059] The GF5A-WTI features a thermal infrared spectral band, a 1500km swath width, and a spatial resolution of hundreds of meters. A single sweeping cycle of strip imagery covers a range of 1500km in the vertical direction and 102km in the rail direction. Each L2A-level image of the GF5A-WTI is stitched together from multiple adjacent sweeping strip images after geometric correction according to geographic information. Its geometric positioning accuracy was verified using 12 L2A-level images of the GF5A-WTI.

[0060] The rigorous imaging model for both pushbroom and swing-broom linear array imaging is divided into a general coordinate system transformation and an image-to-camera coordinate system transformation. The general coordinate system transformation includes transformations between the camera coordinate system, satellite body coordinate system, J2000 coordinate system, and WGS84 coordinate system. The image-to-camera coordinate system transformation is performed separately for pushbroom and swing-broom types, depending on the imaging method. Since the coordinate mapping method from linear array pixels to the camera physical layer is consistent for both imaging methods, it is unified. An offset matrix is ​​used to represent errors such as principal point offset in swing-broom linear array cameras. Because the arrangement orientation of pushbroom and swing-broom linear array pixels in the image coordinate system is different, a rotation matrix can be used to transform the arrangement orientation of swing-broom linear array pixels. This rotation matrix can be combined with the swing-broom angle rotation matrix of the swing-broom linear array camera.

[0061] A single L1A-level image from a GF5A-WTI camera contains 15 images captured during a sweep cycle. This image is separated into 15 L1A-level image data points according to the sweep cycle, and their storage method is converted: when the image is an ascending track image, the rotation matrix angle is set to 0° and the matrix is ​​transposed; when the image is a descending track image, the rotation matrix angle is set to 180° and the matrix is ​​transposed similarly. Through this processing, the image data storage method is converted according to the physical arrangement order of the linear array pixels and the ascending order of the imaging time.

[0062] The latitude and longitude coordinates and spatial resolution of the corner points of 15 L1A-level images were calculated to determine the latitude and longitude range of the entire L2A-level image. The number of rows and columns of pixels was determined by using the spatial resolution of the intermediate images. A strict imaging model was fitted in the plane of the entire L2A-level image using a block polynomial fitting method. The pixel values ​​were resampled from the 15 L1A-level images for each image point to obtain an L2A-level image of GF5A-WTI.

[0063] Geometric positioning accuracy was checked on 12 L2A-level images from GF5A-WTI. The geometric positioning accuracy is shown in Table 1. The planar mean square error of the images is below 200 meters. The maximum planar mean square error is 171.13 meters, the maximum vertical mean square error is 135.98 meters, and the maximum mean square error along the track is 109.10 meters. This meets the requirement that the geometric positioning accuracy of L2A-level images from GF5A-WTI is less than 200 meters.

[0064] Table 1. Geometric positioning accuracy of L2A level images for GF5A-WTI Image number Number of checkpoints Vertical track error (m) Track error (m) Planar mean square error (m) 2025.4.11-12450-004 276 64.75 77.21 100.77 2025.4.13-12480-001 615 95.72 109.10 145.14 2025.5.17-12974-002 344 113.04 66.27 131.04 2025.5.17-12974-003 228 120.64 54.03 132.18 2025.6.8-13295-003 1367 61.70 77.99 99.45 2025.6.9-13302-004 1391 108.76 90.05 141.20 2025.6.12-13353-001 643 59.83 105.01 120.86 2025.6.18-13441-003 353 70.57 89.68 114.12 2025.6.19-13455-003 282 125.56 74.41 145.95 2025.6.19-13455-005 415 135.98 103.91 171.13 2025.6.22-13491-001 1006 131.79 80.21 154.28 2025.7.30-14052-004 415 130.42 96.90 162.48 Maximum value 135.98 109.10 171.13 The general system geometric correction device for pushbroom and swing-broom linear array imaging provided in this application is described below. The general system geometric correction device for pushbroom and swing-broom linear array imaging described below can be referred to in correspondence with the general system geometric correction method for pushbroom and swing-broom linear array imaging described above.

[0065] Figure 5 This is a schematic diagram of the geometric correction device for a universal system of pushbroom and swingbroom linear array imaging provided in this application, as shown below. Figure 5 As shown, the universal system geometric correction device for pushbroom and swingbroom linear array imaging can include, but is not limited to: The model building module 510 is used to: build a rigorous imaging model. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. The data correction module 520 is used to perform systematic geometric correction on the target image data based on the rigorous imaging model.

[0066] It should be noted that the universal system geometric correction device for push-broom and swing-broom linear array imaging provided in this embodiment of the invention can execute the universal system geometric correction method for push-broom and swing-broom linear array imaging described in any of the above embodiments during specific operation, which will not be elaborated in this embodiment.

[0067] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communications interface 620, and the memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a general system geometric correction method for pushbroom and swing-broom linear array imaging, the method including: A rigorous imaging model is established. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. Based on the aforementioned rigorous imaging model, systematic geometric correction is performed on the target image data.

[0068] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0069] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the general system geometric correction method for push-broom and swing-broom linear array imaging provided by the above methods. The method includes: A rigorous imaging model is established. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. Based on the aforementioned rigorous imaging model, systematic geometric correction is performed on the target image data.

[0070] Furthermore, this application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a general system geometric correction method for pushbroom and swing-broom linear array imaging provided by the methods described above, the method comprising: A rigorous imaging model is established. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. Based on the aforementioned rigorous imaging model, systematic geometric correction is performed on the target image data.

[0071] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0072] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A universal system geometric correction method for pushbroom and swingbroom linear array imaging, characterized in that, include: A rigorous imaging model is established. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. Based on the aforementioned rigorous imaging model, systematic geometric correction is performed on the target image data.

2. The universal system geometric correction method for pushbroom and swingbroom linear array imaging according to claim 1, characterized in that, The rigorous imaging model is as follows: ； in: ; k The scale factor is obtained by calculating the intersection of the view vector and the Earth reference ellipsoid in the WGS84 coordinate system. t For satellite camera imaging time parameters, This is the rotation matrix from the satellite camera coordinate system to the satellite body coordinate system. This is the rotation matrix from the satellite body coordinate system to the J2000 coordinate system. Let be the rotation matrix from the J2000 coordinate system to the WGS84 coordinate system, 𝑅(𝛼(t)) be the rotation matrix for the sweep angle, and 𝛼(t) be the sweep angle. The bias matrix, To obtain the WGS84 coordinates of GPS satellites, The WGS84 coordinates corresponding to the image points. For the camera's physical layer matrix, x For the pixel sequence number, p Principal point coordinates, d p For pixel size, f Main distance, , This refers to pixel distortion.

3. The universal system geometric correction method for pushbroom and swingbroom linear array imaging according to claim 1, characterized in that, When the target image data of the sweeping type is satellite imagery, the sweeping target image data is rotated by 0° and then matrix transposed before storage. When the target image data of the sweeping type is satellite deorbiting image, the sweeping target image data is rotated by 180° and then matrix transposed before storage.

4. The universal system geometric correction method for pushbroom and swingbroom linear array imaging according to any one of claims 1-3, characterized in that, The systematic geometric correction of the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. In the corrected image plane, for each pixel, its position on the target image plane is calculated based on the strict imaging model, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

5. The general system geometric correction method for pushbroom and swingbroom linear array imaging according to any one of claims 1-3, characterized in that, The systematic geometric correction of the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. Several sampling points are selected from the target image data, and the ground coordinates corresponding to each sampling point are calculated based on the rigorous imaging model to obtain several sets of image point-ground coordinate data pairs; Based on the aforementioned sets of image point-ground coordinate data pairs, the coefficients of the rational function model are calculated. In the corrected image plane, for each pixel, its position on the target image plane is calculated based on the rational function model, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

6. The universal system geometric correction method for pushbroom and swingbroom linear array imaging according to any one of claims 1-3, characterized in that, The systematic geometric correction of the target image data based on the rigorous imaging model includes: Based on the rigorous imaging model, the latitude and longitude coordinates of the four corner points of the target image data in the geocentric geofixed coordinate system and the spatial resolution of the target image data are calculated to determine the corrected image plane. Several sampling points are selected from the corrected image plane, and the image coordinates corresponding to each sampling point are calculated based on the strict imaging model to obtain several sets of image point-ground coordinate data pairs. The corrected image plane is divided into several small blocks. For each small block, the coefficients of the corresponding polynomial are calculated based on several sets of image point-ground coordinate data pairs within the small block. In the corrected image plane, for each pixel, its position on the target image plane is obtained by inverse polynomial calculation based on the small block corresponding to the pixel, and then the value of the pixel is obtained by resampling based on the pixel values ​​of the neighboring pixels. After all pixels have been inversely calculated and resampled, the corrected image data corresponding to the target image data is obtained.

7. A universal system geometric correction device for pushbroom and swingbroom linear array imaging, characterized in that, include: The model building module is used to: build a rigorous imaging model. The coordinate transformation process of the rigorous imaging model includes a general coordinate system transformation and an image coordinate system to camera coordinate system transformation. The general coordinate system transformation includes the transformation between the camera coordinate system, the satellite body coordinate system, the space inertial coordinate system J2000, and the geocentric coordinate system WGS84. The image coordinate system to camera coordinate system transformation includes the coordinate transformation process between oscillating camera data and pushbroom camera data. The oscillating camera data is normalized to pushbroom camera data based on the bias matrix and rotation matrix before subsequent general coordinate system transformation. The data correction module is used to perform systematic geometric correction on the target image data based on the rigorous imaging model.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the general system geometric correction method for pushbroom and swing-broom linear array imaging as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the general system geometric correction method for pushbroom and swing-broom linear array imaging as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the general system geometric correction method for pushbroom and swing-broom linear array imaging as described in any one of claims 1 to 6.

Images