Image production method for laser height measurement data points of footprint-free camera of high-resolution satellite

Abstract

The application discloses a kind of laser height measurement data point of high-resolution satellite footprintless camera, including S1, carry out high-resolution camera and laser height measurement device geometric calibration, estimate the installation deviation between high-resolution camera and laser height measurement device;S2, utilize the position, pose of imaging, calculate the ground coordinates corresponding to laser point, the object side coordinates are projected to high-resolution satellite image by geometric model, obtain the coordinates of laser point on high-resolution image;When image is stereo camera, further through laser auxiliary adjustment optimization the relative deviation between the two;S3, utilize high-resolution satellite image to construct pyramid file, and realize ground object identification by multispectral, as quality mark record to point file.The application does not need footprint camera to record laser emission direction, utilizes same platform pose, track error consistency to eliminate the deviation between laser and image, simultaneously provides panchromatic, multispectral image file, can be used in other satellite height auxiliary adjustment.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing and photogrammetry technology, specifically relating to a method for creating images of laser altimetry data points from a high-resolution satellite footprint-free camera. Background Technology

[0002] Laser altimeters, as a crucial payload for satellite Earth observation, can acquire sub-meter-level surface elevation information. Using laser altimeter points as elevation control data to assist in the joint adjustment of high-resolution stereo imagery is a core technological approach to improving the elevation accuracy of satellite imagery in the absence of ground control. Especially in global mapping applications, where ground control data is difficult to obtain in many areas, laser altimeters, due to their high elevation accuracy, have become an important alternative to traditional ground control points.

[0003] In engineering practice, to reduce satellite development costs or save payload resources, many high-resolution satellites are not equipped with footprint cameras. The lack of footprint cameras makes it impossible to directly record the laser emission direction and its spot position on the ground through imagery, making accurate image-side positioning of laser altimetry points on high-resolution images extremely difficult. If projection transformation is performed solely based on the satellite's initial attitude and orbit parameters and payload installation relationships, due to undetected system installation errors and drift in attitude and orbit measurements themselves, the coordinates of the laser points on the high-resolution images will have a deviation of several or even tens of pixels. This deviation not only directly affects the accuracy of the laser points as elevation control during joint adjustment, but also introduces equivalent elevation errors due to image-side errors in undulating terrain such as hills and mountains, thus severely limiting the application potential of laser altimetry data from satellites without footprint cameras in high-precision mapping. Furthermore, even if the problem of accurate laser point positioning is solved, traditional methods often only output an isolated coordinate value, lacking image slices and ground feature attribute information for visual interpretation. This makes it difficult for users to intuitively determine whether a point falls on a non-ideal surface such as water or vegetation when conducting joint adjustment using laser altimeters, thus hindering the efficient selection of reliable elevation control points.

[0004] Therefore, overcoming the positioning difficulties caused by footprintless cameras, eliminating the systematic deviation between laser altimeters and high-resolution cameras, and recording multispectral information in the point record file that can be visually interpreted and analyzed has become an urgent technical problem to be solved. Summary of the Invention

[0005] The main technical problem addressed by this invention is to solve the issues of inaccurate laser altimetry point positioning on imagery and the lack of ground feature markers, which hinder efficient subsequent surveying applications due to the absence of footprint cameras on high-resolution satellites. This invention fully utilizes the consistency of attitude and orbit errors experienced by high-resolution cameras and laser altimeters on the same satellite platform, as well as the stability of system installation deviations, to propose a method for creating laser altimetry data point recording images on high-resolution satellites without footprint cameras. This method not only achieves accurate positioning of laser altimetry points on high-resolution imagery but also simultaneously generates standardized point recording files containing multispectral ground feature markers and quality markers, thus providing a set of high-quality control data that can be viewed and analyzed for subsequent elevation-assisted adjustment.

[0006] To solve the above-mentioned technical problems, the objective of this invention is achieved through the following technical solution:

[0007] In a first aspect, the present invention provides a method for creating images of laser altimetry data points from a high-resolution satellite footprint-free camera, the method comprising the following steps:

[0008] Step 1: Based on the ground calibration field data, conduct on-orbit geometric joint calibration of the high-resolution camera and the laser altimeter, and estimate and calculate the installation deviation parameters between the high-resolution camera and the laser altimeter.

[0009] Step 2: For the laser altimetry data to be processed, use the position and attitude data of the satellite imaging time, the laser ranging value, and the installation deviation parameters to calculate the ground three-dimensional coordinates of the laser spot; use the rational function model (RFM) of the high-resolution camera to project the ground three-dimensional coordinates onto the high-resolution image acquired in the same imaging session to obtain the precise image point coordinates of the laser altimetry point on the high-resolution image.

[0010] Step 3: Using precise image point coordinates as the center, crop image slices with preset pixel sizes from the high-resolution image to construct a multi-level resolution pyramid file; based on the multispectral data of the high-resolution image, calculate the Normalized Difference Vegetation Index (NDVI) and identify the land cover type as a quality identifier; package the image point coordinates, pyramid file, and quality identifier in a standardized format to generate a point record file.

[0011] Furthermore, step 1 specifically includes:

[0012] Step 1.1: Acquire high-resolution imagery and full waveform data of laser altimetry synchronously transmitted during satellite transit over the ground geometric calibration field. The ground geometric calibration field has high-precision ground control point (GCP) and digital elevation model (DEM) data.

[0013] Step 1.2: Perform waveform decomposition processing on the full waveform data of laser altimetry. By separating the ground echo and the multiple scattering signals from targets such as vegetation, extract the accurate laser ranging value.

[0014] Step 1.3: Use high-precision ground control points to perform on-orbit geometric calibration of the high-resolution camera, eliminate internal distortion and installation errors of the camera, and obtain a compensated and accurate imaging model;

[0015] Step 1.4: Substitute the initial installation matrix of the laser altimeter, satellite attitude and orbit data, and accurate laser ranging values ​​into the geometric positioning model to calculate the predicted ground coordinates of the laser spot;

[0016] Step 1.5: Project the predicted ground coordinates onto the image plane. By comparing the systematic deviation between the projected coordinates and the known theoretical positions generated by the high-precision DOM / DEM of the calibration field, construct error equations point by point. The form of the error equations is as follows:

[0017] (1)

[0018] In the formula, v is the residual vector between the coordinates of the projected image point and the coordinates of the theoretical image point, A is the partial derivative coefficient matrix of the installation deviation matrix correction, ΔM is the correction vector of the installation deviation matrix, and l is the constant term vector.

[0019] The fixed installation deviation matrix of the laser altimeter relative to the high-resolution camera is obtained by using the least squares adjustment method, and is used as the installation deviation parameter.

[0020] Furthermore, step 2 specifically includes:

[0021] Step 2.1: Based on the observation time corresponding to the laser altimeter point to be processed, interpolate to obtain the satellite GPS orbit position, star-sensor attitude quaternion, and laser ranging value at that time.

[0022] Step 2.2: Using the installation deviation parameters calculated in Step 1, the outgoing pointing vector of the laser altimeter is corrected. Based on the satellite orbit position, the corrected outgoing pointing vector, and the laser ranging value, the geodetic latitude and longitude coordinates and elevation values ​​of the laser footprint on the Earth ellipsoid are calculated through the geometric positioning model, that is, the three-dimensional ground coordinates of the laser spot.

[0023] Step 2.3: Using the RFM (Reasonable Frequency Analysis) of the images captured simultaneously by the high-resolution camera and the laser altimeter, the three-dimensional coordinates of the ground are projected backward onto the image plane to obtain the precise image point coordinates of the laser altimeter points on the high-resolution image. The RFM uses rational polynomials to establish the mathematical mapping relationship between the pixel coordinates of the image and the geodetic coordinates of their corresponding ground points. The basic equations are defined as follows:

[0024]

[0025] In the formula, These are the coordinates of the ground points. and image coordinates The normalized coordinates after translation and scaling take values ​​between [-1, 1]. Translation parameters for image coordinate regularization The translation parameters for regularizing the ground point coordinates; The scaling factor for image coordinate regularization, The scaling factor for regularizing ground point coordinates; Let each represent a general polynomial, where the variables are... The powers of all variables do not exceed three, and the sum of the powers of all variables also does not exceed three, as follows:

[0026] The ground three-dimensional coordinates calculated in step 2.2 are used as input. The corresponding normalized image coordinates are calculated by formulas (1) and (2). After inverse normalization, the accurate image point coordinates are obtained.

[0027] Step 2.4: When the high-resolution camera is a stereo camera, an adjustment regional network is constructed from multiple stereo images on the same satellite imaging orbit as a unit. The image-side affine transformation compensation-based RFM is used as the regional network adjustment model. The image-side affine transformation supplementary model is used for the error compensation of the image RFM. Then, the relationship between the image point coordinates and the ground point coordinates described by RFM in formula (2) is corrected to:

[0028] Connecting points are set up within the stereo image and between adjacent stereo images. The precise image point coordinates of the laser altimeter obtained in step 2.3 are used as elevation control points to participate in the joint adjustment. The elevation values ​​of the laser altimeter are fixed as known quantities. The affine transformation parameters of the image RFM and the ground latitude and longitude coordinates corresponding to the laser altimeter image points are used as unknowns to construct error equations point by point. The affine transformation parameters are solved by least squares adjustment iteration. The solved affine transformation parameters are used to perform image-side affine transformation compensation on the image RFM, generating a new RFM for each image, thereby optimizing the relative deviation between the laser altimeter and the stereo image.

[0029] Furthermore, step 3 specifically includes:

[0030] Step 3.1: Using the precise image point coordinates obtained in Step 2 as the center, crop an image slice with a preset pixel size on the panchromatic or multispectral band of the high-resolution image. The preset pixel size is determined according to the ratio of the laser spot ground diameter to the image ground resolution to ensure complete coverage of the laser spot and its surrounding ground features.

[0031] Step 3.2: Use the resampling algorithm to generate pyramid files of different resolution layers for the image slices. The generated pyramid files are used to provide scalable and smoothly loadable visual image data in the point record file, which is convenient for visual interpretation in subsequent joint adjustment operations.

[0032] Step 3.3: Obtain the reflectance of each band of the multispectral image within the laser spot coverage area, calculate the Normalized Difference Vegetation Index (NDVI) or extract other spectral features. The formula for calculating NDVI is:

[0033] (7)

[0034] In the formula, For near-infrared reflectivity, The reflectance is in the red light band; the land cover type is automatically identified according to the preset spectral classification rules, and the land cover type includes at least one of the following: vegetation, water body, bare soil and snow.

[0035] Step 3.4: Encapsulate the unique identifier of the laser altimeter point, the precise image point coordinates, the pyramid file, the identified land cover type, the imaging time, and the corresponding sensor parameters into a standardized format and write them into the point record file.

[0036] Furthermore, when the high-resolution camera is a stereo camera, after step 2, the following steps are performed: the relative deviation between the laser altimeter and the stereo image is optimized by laser-assisted regional network adjustment to improve the elevation accuracy of the stereo image.

[0037] Furthermore, the systematic bias in step 1.5 stems from the fact that the laser altimeter and the high-resolution camera are fixed to the same satellite platform, and the attitude and orbit measurement errors experienced by both are strongly correlated. That is, when there is systematic drift in the satellite position and attitude measurement data, the direction and magnitude of the influence on the high-resolution camera and the laser altimeter are consistent. By eliminating the installation bias matrix, the positioning accuracy of laser point projection can be improved to the sub-pixel level under ground control conditions. The error equation constructed in step 1.5 is normalized, and the normal equation is as follows:

[0038] (8)

[0039] Recorded as:

[0040] (9)

[0041] Furthermore, in step 2.2, when calculating the three-dimensional coordinates of the ground, if a high-precision Earth ellipsoid elevation is lacking, it is necessary to use publicly available global DEM data to perform iterative elevation calculations, and determine the three-dimensional spatial position of the laser spot on the terrain surface represented by the DEM by successive approximations.

[0042] Furthermore, the quality label in step 3.3 also includes the calculated laser echo waveform characteristic parameters, which include echo broadening and backscattering coefficient. Echo broadening reflects the complexity of the terrain undulation and vertical structure of features within the laser footprint, while backscattering coefficient reflects the surface reflectivity characteristics. The waveform characteristic parameters are used to assist in assessing the confidence level of the laser altimeter as an elevation control point. When the echo broadening exceeds a preset threshold or the backscattering coefficient is lower than a preset minimum value, the laser altimeter is marked as a low-confidence point, reminding the user to use it cautiously or remove it in subsequent joint adjustment.

[0043] Furthermore, step 2.4, laser-assisted regional network adjustment, specifically includes:

[0044] For each laser altimeter image point, the image-side affine transformation parameters of the image RFM and the corresponding ground latitude and longitude coordinates of the laser altimeter image point are used as unknowns, and the elevation value of the laser altimeter point is used as a known quantity. Error equations are constructed point by point. For each connection point, the image-side affine transformation parameters of the image RFM and the corresponding ground three-dimensional coordinates of the connection point are used as unknowns. Error equations are constructed point by point. The affine transformation parameters of each image RFM are obtained according to the least squares adjustment principle. The image-side affine transformation compensation is performed on the image RFM to generate a new RFM.

[0045] Secondly, the present invention also provides a system for creating high-resolution satellite laser altimetry data point recording images that performs the above-described method, comprising:

[0046] The geometric calibration module is configured to perform on-orbit geometric joint calibration of the high-resolution camera and the laser altimeter based on ground calibration field data, and to estimate and calculate installation deviation parameters.

[0047] The coordinate projection positioning module is configured to use installation deviation parameters and attitude data at the time of satellite imaging to perform laser emission pointing vector correction, ground three-dimensional coordinate calculation and RFM-based back projection calculation, and output the precise image-side coordinates of the laser point on the high-resolution image.

[0048] The Point Record encapsulation and generation module is configured to perform image slice cropping, multi-level pyramid file generation, multispectral land cover type identification, and quality labeling assessment centered on precise image point coordinates, and encapsulate all results in a standardized format and write them into the Point Record file.

[0049] The beneficial effects of this invention are as follows: (1) This invention does not rely on a footprint camera to record the laser emission direction. It accurately estimates the installation deviation between the high-resolution camera and the laser altimeter through geometric calibration, and makes full use of the strong consistency of attitude and orbit measurement errors caused by the two being fixed on the same satellite platform. This effectively eliminates residual system deviations and achieves sub-pixel-level accurate positioning of laser altimeter points on high-resolution images under conditions without footprint images. This method breaks through the limitations of payload hardware configuration and provides a practical and feasible high-precision image-side positioning scheme for laser points on satellites without footprint cameras.

[0050] (2) The point record file generated by this invention provides accurate image point coordinates, integrates high-resolution panchromatic and multispectral image pyramid slices, and automatically identifies the land cover type of the laser spot landing point through spectral features. This transforms the laser altimeter point from a single coordinate value in traditional methods into a standardized information unit that integrates visual interpretation, land cover analysis, and quality traceability. This facilitates the rapid identification and removal of abnormal laser points that fall on non-ideal land surfaces such as water bodies and vegetation before joint adjustment, improving the efficiency and reliability of the use of elevation control points and ensuring the quality of adjustment results from the source.

[0051] (3) This method has a high degree of automation. Only one geometric calibration is needed in the early stage of orbit to obtain installation deviation parameters. Then, standardized point-based image products can be mass-produced from the laser altimetry data of the entire orbit and even the entire life cycle of the satellite. For high-resolution satellite payloads equipped with stereo cameras, the method further supports using laser altimetry points as elevation control constraints to carry out laser-assisted stereo image regional network adjustment. This effectively transfers the elevation accuracy of the laser altimetry points to the stereo image geometric model, achieving a significant improvement in the elevation accuracy of stereo images under ground control conditions. The entire method provides a stable and reliable source of high-quality control data for global ground control mapping, reduces the cost of field measurements in overseas and difficult areas, and balances accuracy improvement with production efficiency. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This invention relates to a method for creating images based on laser altimetry data points from a high-resolution satellite footprintless camera.

[0054] flow chart.

[0055] Figure 2 This is a schematic diagram of the installation deviation detection and image positioning principle of the present invention.

[0056] Figure 3 This is a schematic diagram of the laser height measurement point recording product structure of the present invention.

[0057] Figure 4 This is a flowchart of the dual quality identification collaborative determination process of the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in further detail below with reference to the embodiments and accompanying drawings.

[0059] Reference Figure 1 , Figure 2 , Figure 3 and Figure 4 The image shown is a flowchart of the method for creating images based on laser altimetry data points from a high-resolution satellite's footprint-free camera, including the following steps:

[0060] Step 1: Based on the ground calibration field data, conduct on-orbit geometric joint calibration of the high-resolution camera and the laser altimeter, and estimate and calculate the installation deviation parameters between the high-resolution camera and the laser altimeter.

[0061] This is the foundational step of the entire method, its purpose being to establish a precise relative installation relationship between the laser altimeter and the high-resolution camera, providing systematic error compensation parameters for the subsequent precise positioning of the laser point. This step further includes:

[0062] Step 1.1: Acquire high-resolution imagery and full-waveform laser altimetry data synchronously transmitted during satellite transit over the ground geometric calibration field. The ground geometric calibration field contains high-precision ground control point (GCP) and digital elevation model (DEM) data to ensure the reliability of the calibration results. Acquire high-resolution imagery and full-waveform laser altimetry data synchronously transmitted during satellite transit over this area as calibration input.

[0063] Step 1.2: Extract the laser altimeter waveform data, and remove noise interference through waveform decomposition and other methods to calculate the accurate laser ranging value. Waveform decomposition is crucial for ensuring ranging accuracy; it separates ground echoes from multiple scattering signals from targets such as vegetation, thereby accurately extracting the ranging value corresponding to the ground spot. The original full waveform signal received by the laser altimeter can be represented as the superposition of multiple Gaussian components, and its mathematical model is as follows:

[0064] (1)

[0065] In the formula, Let be the intensity of the echo signal received at time t. This represents the total number of Gaussian components contained in the waveform. Let be the amplitude of the i-th Gaussian component. Let i be the peak time of the i-th Gaussian component. Let be the half-width and height of the i-th Gaussian component. The background noise is represented by the waveform. The model's parameters are solved using nonlinear least-squares fitting or expectation-maximization algorithms to separate the last Gaussian component representing the ground echo. The peak time of this component is the arrival time of the laser pulse at the ground. Combined with the speed of light, the accurate laser ranging value can be calculated.

[0066] (2)

[0067] In the formula, R is the laser ranging value, and c is the speed of light. This represents the peak time of the Gaussian component of the ground echo. This represents the emission time of the laser pulse.

[0068] Step 1.3 involves using high-precision ground control points to perform on-orbit geometric calibration of the high-resolution camera, eliminating internal camera distortion and installation errors, and obtaining accurate on-orbit exterior orientation elements and imaging models. This step ensures that the imaging geometry of the high-resolution camera itself is sufficiently accurate, avoiding coupling camera-specific errors to subsequent installation deviation calculations. The on-orbit geometric calibration employs a rigorous imaging model based on collinearity conditions. For a linear pushbroom high-resolution camera, its basic equations are defined as follows:

[0069] (3)

[0070] In the formula, Here are the three-dimensional coordinates of the ground control point, and here are the row and column coordinates of the corresponding image point. Let f be the satellite orbital position vector at the moment of imaging, f be the focal length of the high-resolution camera, and m be the imaging scale. This is the rotation matrix from the satellite body coordinate system to the WGS84 coordinate system. Let be the pointing angle of the camera sensor in the satellite's body coordinate system. By simultaneously solving the collinearity condition equations of multiple high-precision ground control points, the precise installation matrix and internal distortion parameters of the camera are calculated using the least squares adjustment method.

[0071] Step 1.4 involves incorporating the initial installation matrix of the laser altimeter, satellite attitude and orbit data, and the precise laser ranging values ​​calculated in Step 1.2 into the geometric positioning model to forward calculate the predicted ground coordinates of the laser spot. The initial installation matrix here is derived from ground laboratory measurements taken before satellite launch; however, due to launch vibrations and changes in the on-orbit thermal environment, this value has a non-negligible deviation.

[0072] Step 1.5: Project the predicted ground coordinates onto the image plane. By comparing the systematic deviation between the projected coordinates and the known theoretical positions generated by the high-precision DOM / DEM of the calibration field, error equations are constructed point by point. Since the high-resolution camera and the laser altimeter are fixed to the same satellite platform, their attitude and orbit errors are strongly correlated—meaning that measurement errors in satellite position and attitude have almost identical effects on both. Utilizing this characteristic, by comparing the systematic deviation between the projected position of the laser predicted coordinates calculated in Step 1.4 on the image and the known theoretical position generated by the high-precision DEM of the area, error equations are constructed point by point. The form of the error equation is:

[0073] (4)

[0074] In the formula, v is the residual vector between the coordinates of the projected image point and the coordinates of the theoretical image point, and A is the partial derivative coefficient matrix of the installation deviation matrix correction. Let l be the correction vector of the installation deviation matrix, and l be the constant term vector obtained by substituting the initial value.

[0075] Based on the principle of least squares adjustment, the above error equations are normalized to obtain the following normal equation form:

[0076] (5)

[0077] Recorded as:

[0078] (6)

[0079] In the formula, N is the coefficient matrix of the normal equation, and U is the vector of constant terms in the normal equation. Solving this normal equation yields the fixed installation deviation matrix of the laser altimeter relative to the high-resolution camera. This matrix is ​​used as a system error compensation parameter and can be applied to all subsequent laser point projections, fundamentally eliminating projection coordinate deviations caused by installation errors, and achieving sub-pixel level positioning accuracy for laser point projection on the high-resolution image.

[0080] Step 2: For the laser altimetry data to be processed, the ground three-dimensional coordinates of the laser spot are calculated using the position and attitude data of the satellite imaging time, the laser ranging value, and the installation deviation parameters. Using the rational function model (RFM) of the high-resolution camera, the ground three-dimensional coordinates are projected onto the high-resolution image acquired in the same imaging session to obtain the precise image point coordinates of the laser altimetry point on the high-resolution image. When the high-resolution camera is a stereo camera, the relative deviation between the laser altimetry point and the stereo image is further optimized through laser-assisted regional network adjustment to improve the elevation accuracy of the stereo image.

[0081] After completing the calibration work in step 1, the core task of this step is to accurately map the compensated laser altimetry data onto the image space of the high-resolution image, achieving high-precision image point positioning. This step further includes:

[0082] Step 2.1: Based on the recorded time of the laser altimeter data to be processed, obtain the satellite GPS orbital position, star-sensor attitude quaternion, and corresponding laser ranging value at that time through time interpolation. Since the data sampling frequency of GPS and star-sensor is usually higher than the observation frequency of the laser altimeter, a linear interpolation method is used to ensure interpolation accuracy.

[0083] Step 2.2: Correct the actual laser altimeter pointing vector using the installation deviation compensation parameters obtained in Step 1. Combine the satellite orbital position, the corrected laser pointing direction, and the laser ranging value, and calculate the geodetic latitude and longitude coordinates and elevation values ​​of the laser footprint on the Earth's ellipsoid using a rigorous geometric positioning model.

[0084] For elevation, if a high-precision Earth ellipsoid elevation is lacking as a constraint, it is necessary to use publicly available global DEM data (such as the Global 30-arc-second Digital Elevation Model provided by the U.S. Geological Survey, or AW3D30, Copernicus DEM, etc.) to iteratively solve for elevation, gradually approximating and determining the true three-dimensional spatial position of the laser spot.

[0085] Step 2.3: Using the RFM (Reasonable Function Model) of the images captured simultaneously by the high-resolution camera and the laser altimeter, the three-dimensional coordinates of the ground are projected backward onto the image plane to obtain the precise image point coordinates of the laser altimeter points on the high-resolution image. The RFM uses rational polynomials to establish a mathematical mapping relationship between the pixel coordinates of the image and the geodetic coordinates of their corresponding ground points. To achieve this backward projection, this invention employs the RFM model of the high-resolution camera image. RFM refers to establishing a mathematical mapping relationship between the pixel coordinates of the image and the geodetic coordinates of their corresponding ground points using rational polynomials. It has the advantages of being simple in form and independent of sensor type. Its basic equations are defined as follows:

[0086] In the formula, These are the coordinates of the ground points. and image coordinates After translation and scaling, the normalized coordinates take values ​​between [-1, 1]. The purpose of normalization is to avoid rounding errors caused by large differences in the magnitude of parameter values ​​during the calculation process, thereby enhancing the stability of parameter solution. Translation parameters for image coordinate regularization The translation parameters are used for regularizing the coordinates of ground points. The scaling factor for image coordinate regularization, This is the scaling factor for regularizing the ground point coordinates. Let each represent a general polynomial, where the variables are... The powers of all variables do not exceed 3, and the sum of the powers of all variables also does not exceed 3, as follows:

[0087] Using the RFM model of the high-resolution camera image acquired in the same laser imaging session, the calculated three-dimensional coordinates of the laser ground are projected backwards onto the image surface. Since the installation deviation was eliminated in step 1 by utilizing the consistency of the same platform error, the image point coordinates obtained by projection are now the precise image-side coordinates of the laser altimeter point on the high-resolution image.

[0088] Step 2.4: When the high-resolution camera is a three-line array stereo camera, in order to eliminate the possible residual relative deviation between the laser altimeter and the stereo image, the elevation value of the laser altimeter is further used as a known constraint, and laser-assisted regional network adjustment is carried out with the stereo image acquired on the same track to optimize the geometric correspondence between the two. Specifically, it includes: taking the satellite imaging and photography track as a unit, constructing an adjustment regional network for multiple stereo images on the same track, and using RFM based on image-side affine transformation compensation as the regional network adjustment model. The error compensation of image RFM adopts the image-side affine transformation supplementary model, then the relationship between the image point coordinates and the ground point coordinates described by RFM in formula (4) is corrected to:

[0089] In the formula, This represents the affine transformation parameters used to compensate for RFM system errors. Connecting points are established within the stereo image and between adjacent stereo images. Using the elevation values ​​of the laser altimeters as known quantities, the affine transformation parameters of the image's RFM, and the ground latitude and longitude coordinates corresponding to the laser altimeter image points as unknowns, an error equation is constructed. The least squares adjustment method is used to solve the equation globally, transferring the elevation accuracy of the laser altimeters to the stereo image. After the adjustment iteration converges, the calculated affine transformation parameters are used to compensate for the RFM in the image, generating a new RFM for each image. This optimizes the relative deviation between the laser altimeters and the stereo image, improving the elevation accuracy of the stereo image.

[0090] Step 3: Using precise image point coordinates as the center, crop image slices with preset pixel sizes from the high-resolution image to construct a multi-level resolution pyramid file; based on the multispectral data of the high-resolution image, calculate the Normalized Difference Vegetation Index (NDVI) and identify the land cover type as a quality identifier; package the image point coordinates, pyramid file, and quality identifier in a standardized format to generate a point record file.

[0091] After accurately acquiring the image-side coordinates of the laser points, the purpose of this step is to construct a visualized point tracking product, enabling visual interpretation and automated quality assessment, thereby serving the selection of control points in subsequent joint adjustment operations. This step further includes:

[0092] Step 3.1: Using the precise image point coordinates obtained in Step 2 as the center, crop an image slice with a preset pixel size from the panchromatic or multispectral bands of the high-resolution image. The preset pixel size is determined based on the ratio of the laser spot's ground diameter to the image's ground resolution, ensuring complete coverage of the laser spot and its surrounding features. The slice size is usually determined by a combination of the laser spot diameter and the image resolution. For example, for a laser altimeter with a ground spot diameter of approximately 50 meters, a 500-pixel × 500-pixel slice can be cropped from a 2-meter resolution high-resolution image to ensure that the laser spot and its surrounding features are completely included.

[0093] Step 3.2: Generate the pyramid layered file for this slice using a resampling algorithm. The pyramid file is stored at multiple resolutions, downsampling from the original resolution to a thumbnail layer, which facilitates quick loading, smooth scaling, and visual interpretation of the terrain and feature characteristics of the area where the laser point is located during subsequent joint adjustment operations.

[0094] Step 3.3: After acquiring the image slices, further extract the average reflectance of each band of the multispectral image within the laser spot area, and calculate spectral indices such as the Normalized Difference Vegetation Index (NDVI) or the Normalized Difference Water Index (NDWI). The formula for calculating NDVI is as follows:

[0095] (12)

[0096] In the formula, For near-infrared reflectivity, Reflectivity in the red light band;

[0097] The system automatically identifies the land cover type of laser landing points based on preset spectral classification rules. Classification results can include typical land cover categories such as vegetation, water bodies, bare soil, snow and ice, and buildings. This automatic identification result will serve as the core content of quality labeling, because laser points landing on water bodies often have unreliable echo signals, and laser points landing on dense vegetation may experience reduced elevation accuracy due to multiple scattering. Users can selectively remove outliers during adjustment based on this labeling.

[0098] Furthermore, the quality identifier also includes laser echo waveform characteristic parameters calculated from the waveform decomposition process in step 1.2. These waveform characteristic parameters include echo broadening and backscattering coefficient. Echo broadening reflects the complexity of terrain undulations and vertical structures of features within the laser footprint, while the backscattering coefficient reflects the surface reflectivity characteristics. These waveform characteristic parameters are used to assist in assessing the confidence level of the laser altimeter point as an elevation control point. When the echo broadening exceeds a preset threshold or the backscattering coefficient is lower than a preset minimum value, the laser altimeter point is marked as a low-confidence point, reminding the user to use it cautiously or remove it from subsequent joint adjustment.

[0099] Step 3.4 encapsulates the unique identifier of the laser altimeter point, precise image coordinates, pyramid file, identified land cover type, imaging time, and corresponding sensor parameters into a standardized format and writes it into the point record file. Specifically, the unique identifier of the laser point, the precise image coordinates calculated in Step 2, the image tile pyramid file generated in Step 3.2, the light spot land cover type identified in Step 3.3, and auxiliary metadata such as imaging time, sensor ID, and solar altitude angle are encapsulated and written into the point record file according to a standard format. This file can be stored in a structured text format such as XML or JSON, and also includes the binary data file of the image tile, forming a self-describing, directly usable standardized point record product.

[0100] Through the step-by-step progression described above, this invention first completes the on-orbit geometric calibration of the high-resolution camera and laser altimeter in step 1 to obtain installation deviation parameters; then, in step 2, it uses these parameters to correct the laser pointing, accurately projects the laser point onto the high-resolution image, and performs laser-assisted adjustment to optimize the geometric accuracy of the stereo image as needed; finally, in step 3, it generates a standardized point-of-care product, thus establishing a complete technical chain from load calibration to product output.

[0101] Compared with the prior art, one or more embodiments of the present invention may have the following advantages:

[0102] (1) This invention does not rely on a footprint camera to record the laser emission direction. It accurately estimates the installation deviation between the high-resolution camera and the laser altimeter through geometric calibration, and fully utilizes the strong consistency of attitude and orbit measurement errors caused by the two being fixed on the same satellite platform to effectively eliminate residual system deviations. This achieves sub-pixel-level precise positioning of laser altimeter points on high-resolution images under conditions without footprint images. This method breaks through the limitations of payload hardware configuration and provides a practical and feasible high-precision image-side positioning scheme for laser points on satellites without footprint cameras.

[0103] (2) The point record file generated by this invention provides accurate image point coordinates, integrates high-resolution panchromatic and multispectral image pyramid slices, and automatically identifies the land cover type of the laser spot landing point through spectral features. This transforms the laser altimeter point from a single coordinate value in traditional methods into a standardized information unit that integrates visual interpretation, land cover analysis, and quality traceability. This facilitates the rapid identification and removal of abnormal laser points that fall on non-ideal land surfaces such as water bodies and vegetation before joint adjustment, improving the efficiency and reliability of the use of elevation control points and ensuring the quality of adjustment results from the source.

[0104] (3) This method has a high degree of automation. Only one geometric calibration is needed in the early stage of orbit to obtain installation deviation parameters. Then, standardized point-based image products can be mass-produced from the laser altimetry data of the entire orbit and even the entire life cycle of the satellite. For high-resolution satellite payloads equipped with stereo cameras, the method further supports using laser altimetry points as elevation control constraints to carry out laser-assisted stereo image regional network adjustment. This effectively transfers the elevation accuracy of the laser altimetry points to the stereo image geometric model, achieving a significant improvement in the elevation accuracy of stereo images under ground control conditions. The entire method provides a stable and reliable source of high-quality control data for global ground control mapping, reduces the cost of field measurements in overseas and difficult areas, and balances accuracy improvement with production efficiency.

[0105] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and variations in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection for this invention shall still be determined by the scope defined in the appended claims.

Claims

1. A method for creating images based on laser altimetry data points from a high-resolution satellite footprint-free camera, characterized in that, The method includes the following steps: Step 1: Based on ground calibration field data, conduct on-orbit geometric joint calibration of the high-resolution camera and the laser altimeter, and estimate and calculate the installation deviation parameters between the high-resolution camera and the laser altimeter. Step 2: For the laser altimetry data to be processed, the ground three-dimensional coordinates of the laser spot are calculated using the position and attitude data of the satellite imaging time, the laser ranging value, and the installation deviation parameters. Using the rational function model (RFM) of the high-resolution camera, the ground three-dimensional coordinates are projected onto the high-resolution image acquired in the same imaging session to obtain the precise image point coordinates of the laser altimetry point on the high-resolution image. When the high-resolution camera is a stereo camera, the relative deviation between the laser altimetry point and the stereo image is further optimized through laser-assisted regional network adjustment to improve the elevation accuracy of the stereo image. Step 3: Using the precise image point coordinates as the center, crop image slices with a preset pixel size from the high-resolution image to construct a multi-level resolution pyramid file; calculate the Normalized Difference Vegetation Index (NDVI) and identify the land cover type based on the multispectral data of the high-resolution image as a quality identifier; package the image point coordinates, the pyramid file, and the quality identifier in a standardized format to generate a point record file.

2. The method for creating images from laser altimetry data points of a high-resolution satellite footprintless camera according to claim 1, characterized in that, Step 1 specifically includes: Step 1.1: Acquire high-resolution imagery and full waveform data of laser altimetry synchronously transmitted during satellite transit over the ground geometric calibration field. The ground geometric calibration field has high-precision ground control point (GCP) and digital elevation model (DEM) data. Step 1.2: Perform waveform decomposition processing on the full waveform data of the laser altimeter, and extract accurate laser ranging values ​​by separating the ground echo and the multiple scattering signals of targets such as vegetation; Step 1.3: Use the high-precision ground control points to perform on-orbit geometric calibration of the high-resolution camera, eliminate internal distortion and installation errors of the camera, and obtain a compensated and accurate imaging model; Step 1.4: Substitute the initial installation matrix of the laser altimeter, the satellite attitude and orbit data, and the precise laser ranging value into the geometric positioning model to calculate the predicted ground coordinates of the laser spot; Step 1.5: Project the predicted ground coordinates onto the image plane. By comparing the systematic deviation between the projected coordinates and the known theoretical positions generated by the high-precision DOM / DEM of the calibration field, construct error equations point by point. The form of the error equations is: (1) In the formula, v is the residual vector between the coordinates of the projected image point and the coordinates of the theoretical image point, A is the partial derivative coefficient matrix of the installation deviation matrix correction, ΔM is the correction vector of the installation deviation matrix, and l is the constant term vector. The fixed installation deviation matrix of the laser altimeter relative to the high-resolution camera is obtained by using the least squares adjustment method, and is used as the installation deviation parameter.

3. The method for creating images from laser altimetry data points of a high-resolution satellite footprintless camera according to claim 1, characterized in that, Step 2 specifically includes: Step 2.1: Based on the observation time corresponding to the laser altimeter point to be processed, interpolate to obtain the satellite GPS orbit position, star-sensor attitude quaternion, and laser ranging value at that time. Step 2.2: Using the installation deviation parameters calculated in Step 1, the outgoing pointing vector of the laser altimeter is corrected. Based on the satellite orbit position, the corrected outgoing pointing vector, and the laser ranging value, the geodetic latitude and longitude coordinates and elevation values ​​of the laser footprint on the Earth ellipsoid are calculated through the geometric positioning model, that is, the ground three-dimensional coordinates of the laser spot. Step 2.3: Using the RFM (Reasonable Frequency Analysis) of the images captured simultaneously by the high-resolution camera and the laser altimeter, the three-dimensional coordinates of the ground are projected backward onto the image plane to obtain the precise image point coordinates of the laser altimeter point on the high-resolution image. The RFM utilizes rational polynomials to establish a mathematical mapping relationship between the pixel coordinates of the image and the geodetic coordinates of the corresponding ground points. The basic equation is defined as follows: ； In the formula, These are the coordinates of the ground points. and image coordinates The normalized coordinates after translation and scaling take values ​​between [-1, 1]. Translation parameters for image coordinate regularization The translation parameters for regularizing the ground point coordinates; The scaling factor for image coordinate regularization, The scaling factor for regularizing ground point coordinates; Let each represent a general polynomial, where the variables are... The powers of all variables do not exceed three, and the sum of the powers of all variables also does not exceed three, as follows: The ground three-dimensional coordinates calculated in step 2.2 are used as input, and the corresponding normalized image coordinates are calculated by formulas (1) and (2). Then, the precise image point coordinates are obtained by inverse normalization. Step 2.4: When the high-resolution camera is a stereo camera, an adjustment regional network is constructed from multiple stereo images on the same satellite imaging orbit as a unit. The image-side affine transformation compensation-based RFM is used as the regional network adjustment model. The image-side affine transformation supplementary model is used for the error compensation of the image RFM. Then, the relationship between the image point coordinates and the ground point coordinates described by RFM in formula (2) is corrected to: In the formula, , The image-side compensation value for the systematic error of the image point coordinates is: In the formula, The affine transformation parameters represent the compensation for RFM system errors; Connecting points are set up within the stereo image and between adjacent stereo images. The precise image point coordinates of the laser altimeter obtained in step 2.3 are used as elevation control points to participate in the joint adjustment. The elevation values ​​of the laser altimeter are fixed as known quantities. The affine transformation parameters of the image RFM and the ground latitude and longitude coordinates corresponding to the laser altimeter image points are used as unknowns to construct error equations point by point. The affine transformation parameters are solved by least squares adjustment iteration. The solved affine transformation parameters are used to perform image-side affine transformation compensation on the image RFM, generating a new RFM for each image, thereby optimizing the relative deviation between the laser altimeter and the stereo image.

4. The method for creating images from laser altimetry data points of a high-resolution satellite footprintless camera according to claim 1, characterized in that, Step 3 specifically includes: Step 3.1: Using the precise image point coordinates obtained in Step 2 as the center, crop an image slice with a preset pixel size on the panchromatic or multispectral band of the high-resolution image. The preset pixel size is determined according to the ratio of the laser spot ground diameter to the image ground resolution to ensure complete coverage of the laser spot and its surrounding ground features. Step 3.2: Use a resampling algorithm to generate pyramid files of different resolution layers for the image slices. The generated pyramid files are used to provide scalable and smoothly loadable visual image data in the point record file, which is convenient for visual interpretation in subsequent joint adjustment operations. Step 3.3: Obtain the reflectance of each band of the multispectral image within the coverage area of ​​the laser spot, calculate the Normalized Difference Vegetation Index (NDVI) or extract other spectral features. The formula for calculating the NDVI is: (7) In the formula, For near-infrared reflectivity, The reflectance is in the red light band; the land cover type is automatically identified according to the preset spectral classification rules, and the land cover type includes at least one of vegetation, water body, bare soil and snow; Step 3.4: Encapsulate the unique identifier of the laser altimeter point, the precise image point coordinates, the pyramid file, the identified land cover type, the imaging time, and the corresponding sensor parameters into a standardized format and write them into the point record file.

5. The method according to claim 1, characterized in that, When the high-resolution camera is a stereo camera, the following steps are performed after step 2: the relative deviation between the laser altimeter and the stereo image is optimized by laser-assisted regional network adjustment to improve the elevation accuracy of the stereo image.

6. The method according to claim 2, characterized in that, The system deviation mentioned in step 1.5 stems from the fact that the laser altimeter and the high-resolution camera are fixed to the same satellite platform, and the attitude and orbit measurement errors experienced by both are strongly correlated. That is, when there is system drift in the satellite position and attitude measurement data, the high-resolution camera and the laser altimeter are affected in the same direction and magnitude. By eliminating the installation deviation matrix, the positioning accuracy of laser point projection can be improved to the sub-pixel level under ground control conditions. The error equation constructed in step 1.5 is normalized, and the normal equation is as follows: 。 7. The method according to claim 3, characterized in that, In step 2.2, when calculating the three-dimensional coordinates of the ground, if a high-precision Earth ellipsoid elevation is lacking, it is necessary to use publicly available global DEM data to perform iterative elevation calculations. The three-dimensional spatial position of the laser spot on the terrain surface represented by the DEM is determined by successive approximations.

8. The method according to claim 4, characterized in that, The quality labeling in step 3.3 also includes the calculated laser echo waveform characteristic parameters, which include echo broadening and backscattering coefficient. Echo broadening reflects the complexity of terrain undulation and vertical structure of features within the laser footprint, while backscattering coefficient reflects the surface reflectivity characteristics. These waveform characteristic parameters are used to assist in assessing the confidence level of the laser altimeter as an elevation control point. When the echo broadening exceeds a preset threshold or the backscattering coefficient is lower than a preset minimum value, the laser altimeter is marked as a low-confidence point, reminding the user to use it cautiously or remove it in subsequent joint adjustment.

9. The method according to claim 3, characterized in that, The laser-assisted regional network adjustment described in step 2.4 specifically includes: For each laser altimeter image point, the image-side affine transformation parameters of the image RFM and the corresponding ground latitude and longitude coordinates of the laser altimeter image point are used as unknowns, and the elevation value of the laser altimeter point is used as a known quantity. Error equations are constructed point by point. For each connection point, the image-side affine transformation parameters of the image RFM and the corresponding ground three-dimensional coordinates of the connection point are used as unknowns. Error equations are constructed point by point. The affine transformation parameters of each image RFM are obtained according to the least squares adjustment principle. The image-side affine transformation compensation is performed on the image RFM to generate a new RFM.

10. A system for creating images from high-resolution satellite laser altimetry data points, performing the method as described in any one of claims 1 to 9, characterized in that, include: The geometric calibration module is configured to perform on-orbit geometric joint calibration of the high-resolution camera and the laser altimeter based on ground calibration field data, and to estimate and calculate installation deviation parameters. The coordinate projection positioning module is configured to use the installation deviation parameters and the attitude and orbit data at the time of satellite imaging to perform laser emission pointing vector correction, ground three-dimensional coordinate calculation and RFM-based back projection calculation, and output the precise image-side coordinates of the laser point on the high-resolution image. The Point Record Encapsulation and Generation Module is configured to perform image slicing, multi-level pyramid file generation, multispectral land cover type identification, and quality labeling assessment centered on the precise image point coordinates, and encapsulate all results in a standardized format and write them into the Point Record file.

Images