Optical-laser-SAR multi-source data uncontrolled joint positioning method and system

Abstract

This invention belongs to the technical field of remote sensing image processing and discloses a method and system for uncontrolled joint positioning of multi-source optical-laser-SAR data. This method uses satellite-borne laser altimetry data as an elevation control constraint to improve the elevation positioning accuracy of optical stereo image pairs, providing a high-precision elevation benchmark for SAR satellite imagery. It also uses SAR satellite imagery as a planar constraint to improve the planar positioning accuracy of optical stereo image pairs, achieving high-precision joint positioning of satellite-borne optical-laser-SAR multi-source remote sensing data without ground control, and improving the geometric positioning accuracy of optical satellite remote sensing imagery without ground control.

Description

Technical Field

[0001] The present invention belongs to the technical field of remote sensing image processing, and more specifically, relates to an optical-laser-SAR multi-source data uncontrolled joint positioning method and system. Background Art

[0002] Using ground control points to eliminate geometric errors in optical satellite remote sensing images and improve the positioning accuracy of optical satellite remote sensing images is an effective means to achieve high-precision mapping of optical satellite remote sensing images. However, in uninhabited areas such as deserts and forests, as well as in areas such as sensitive military areas abroad, the deployment of ground control points is extremely difficult, resulting in limited application of traditional methods for improving the positioning accuracy of optical satellite remote sensing images based on ground control points. With the launch of optical satellites, SAR (Synthetic Aperture Radar) satellites and satellites equipped with spaceborne laser altimeters at home and abroad, a wealth of global multi-source earth observation data has been accumulated. Among them, the elevation measurement accuracy of spaceborne laser altimetry data can reach the decimeter level, and the plane positioning accuracy of SAR satellite images can reach the pixel level.

[0003] Due to differences in observation mechanisms between SAR and optical satellite imagery, when matching optical and SAR satellite imagery, the matching results are susceptible to differences in two influencing modalities, making it difficult to achieve pixel-level matching accuracy, which in turn affects the accuracy of the joint adjustment results. Furthermore, although the planar positioning accuracy of SAR satellite imagery after precise orbit determination can theoretically reach pixel-level accuracy, its actual positioning accuracy is highly dependent on the accuracy of the elevation datum due to its large observation angle. Spaceborne laser altimetry data can achieve sub-meter elevation accuracy and is often used as elevation control data to improve the elevation positioning accuracy of optical stereo image pairs. After joint adjustment of optical stereo image pairs with spaceborne laser altimetry data, the elevation accuracy of the optical stereo image pairs can be better than 1 meter. However, due to the large spot diameter of spaceborne laser altimetry data, the accuracy of the joint adjustment results is significantly affected by the consistency of the horizontal accuracies of the two data. When the planar positioning accuracy of the optical stereo image pairs is poor, the elevation accuracy of the joint adjustment results is also difficult to achieve. Summary of the Invention

[0004] The present invention solves the problem of low geometric positioning accuracy of optical satellite remote sensing images without ground control in the prior art by providing an optical-laser-SAR multi-source data uncontrolled joint positioning method and system.

[0005] The present invention provides an optical-laser-SAR multi-source data uncontrolled joint positioning method, comprising the following steps:

[0006] Step 1: Construct an optical-laser-SAR multi-source satellite image regional network, extract image points with the same name from optical satellite stereo image pairs and SAR satellite images, and optical satellite stereo image pairs and spaceborne laser altimetry data; obtain optical-SAR satellite image connection points based on the extracted image points with the same name from optical satellite stereo image pairs and SAR satellite images; obtain optical stereo image and spaceborne laser altimetry data connection points based on the extracted image points with the same name from optical satellite stereo image pairs and spaceborne laser altimetry data;

[0007] Step 2: Using the satellite-borne laser altimetry data as the elevation reference, the systematic elevation error of the optical satellite stereo image pair is detected based on the connection points between the optical stereo image and the satellite-borne laser altimetry data. The elevation error of the geometric model of the optical satellite stereo image pair is corrected to obtain the geometric model of the optical satellite stereo image pair after elevation error compensation.

[0008] Step 3: Based on the geometric model of the optical satellite stereo image pair after elevation error compensation, the forward intersection method is used to calculate the elevation coordinates of the object point corresponding to the optical-SAR satellite image connection point;

[0009] Step 4: Based on the elevation coordinates obtained in step 3 and the SAR satellite image geometric model, calculate the object plane coordinates of the optical-SAR satellite image connection point.

[0010] Step 5: Using the object points corresponding to the optical-SAR satellite image connection points in steps 3 and 4 as horizontal and vertical control points, the planar errors in the optical satellite stereo image pair regional network are detected and compensated.

[0011] Step 6: Generate evenly distributed virtual control points for all optical satellite stereo image pairs in the survey area, and add an affine error image cube compensation model. Establish a block adjustment model for a single optical satellite stereo image pair without ground control based on the virtual control points, and solve the adjustment model parameters to achieve positioning without ground control.

[0012] Preferably, step 1 includes the following sub-steps:

[0013] Step 1.1: Based on the geodetic coordinates of the satellite-borne laser altimeter data and the RFM model of the optical satellite stereo image pair, calculate the coordinates of the initial image points of the satellite-borne laser altimeter data in the downward optical image, and use the tie point automatic matching algorithm to obtain the tie points of the laser image points in the downward optical image in the front and rear optical images;

[0014] Step 1.2: Use the optical-SAR satellite image tie point automatic matching algorithm to obtain the tie points between the optical satellite stereo image pair and the SAR satellite image;

[0015] Step 1.3: Use the optical satellite image tie point automatic matching algorithm to obtain the tie points between optical satellite images.

[0016] Preferably, step 2 includes the following sub-steps:

[0017] Step 2.1: Based on the geometric error characteristics of the single-view optical satellite image, an image square error compensation model is added to the initial RFM model of the single-view image;

[0018] Step 2.2: Based on the initial RFM model of the single-view image with the additional image square error compensation model, the basic adjustment model is established with the connection points between optical satellite images as observation values;

[0019] Step 2.3: According to the principle of least squares adjustment, the constructed observation error equation is normalized to construct a modified normal equation that only contains the parameters of the image additional image square error compensation model to be solved;

[0020] Step 2.4: Solve the modified equations. Based on the current iterative RFM model of each image and the obtained image additional image square error compensation model parameters, solve the geographic coordinates of the object points corresponding to each tie point through spatial forward intersection and substitute them into the next adjustment solution until the results of the two adjustments are consistent, or the iteration is stopped when the number of iterations exceeds the set number.

[0021] Step 2.5: For the connection points between the acquired optical stereo images and the satellite-borne laser altimetry data, refine the RFM model and the image point coordinates of the connection points using the adjusted images of each scene. Obtain the geographic coordinates of the corresponding object points through spatial forward intersection, and calculate the elevation difference with the satellite-borne laser altimetry data. When multiple satellite-borne laser altimetry data exist in the survey area, calculate the mean elevation difference of all satellite-borne laser altimetry data. Using the downward-looking image of all optical satellite stereo image pairs in the survey area as the reference, correct the geometric error of the corresponding fore- and fore-looking images along the epipolar line according to the mean elevation difference, and regenerate the RFM model of each scene image.

[0022] Preferably, in step 2.5, the compensation model used to correct the geometric error along the epipolar line direction is expressed as follows:

[0023] l+Δl=F x (Lat,Lon,Height)

[0024] Where l is the image plane coordinate of the image point, (Lat, Lon, Height) is the longitude, latitude and elevation geographic coordinates of the object point corresponding to the image point, and Δl is the additional system error translation model compensation parameter.

[0025] Preferably, the step 5 comprises the following steps:

[0026] Step 5.1: Based on steps 2.1 and 2.2, construct the error equation for the block adjustment of the optical satellite stereo image pair with additional plane control points;

[0027] Step 5.2: According to the principle of least squares adjustment, normalize the observation error equation to obtain the normal equation;

[0028] Step 5.3: Solve the normal equation and regenerate the refined RFM model for each scene image.

[0029] Preferably, in step 5, the plane error includes rotation error, scaling error and attitude drift error.

[0030] Preferably, step 6 includes the following sub-steps:

[0031] Step 6.1: For all optical satellite images within the survey area, generate a uniform grid at set intervals and take the center point of each grid. Using the image-space coordinates of each grid center point and the initial RFM model of the corresponding image, and taking the mean elevation surface of the geodetic coordinates in the RFM model as the elevation datum, obtain the object-space geographic coordinates of the center point through spatial forward intersection. Construct a pair of virtual control points between each grid center image point and the corresponding object point.

[0032] Step 6.2: Use the connecting point image points and virtual control point image points between all optical satellite images to construct the observation error equation;

[0033] Step 6.3: Iteratively solve the adjustment parameters in the observation error equation.

[0034] Preferably, the observation error equation in step 6.2 is expressed as follows:

[0035]

[0036] Where V vc is the residual vector of the error equation corresponding to the virtual control point, V tp is the residual vector of the error equation corresponding to the tie point; x and t are the adjustment parameters to be solved, x is the image space additional parameter vector of the RFM image to be adjusted, t is the object space coordinate correction vector of the tie point, A vc is the partial derivative coefficient matrix of the image square additional parameters corresponding to the virtual control points, A tp is the partial derivative coefficient matrix of the image square additional parameters corresponding to the connection points, B tp is the partial derivative coefficient matrix of the object space geographic coordinate correction number corresponding to the connection point, L vc is the constant vector corresponding to the virtual control point, L tp is the constant vector corresponding to the connection point, P VC is the weight matrix corresponding to the virtual control point, P tp is the weight matrix corresponding to the connection point.

[0037] Preferably, in step 6.3, after solving the adjustment parameters, the method further includes: updating and obtaining the refined RFM models of all optical satellite images in the survey area.

[0038] In another aspect, the present invention provides an optical-laser-SAR multi-source data uncontrolled joint positioning system, comprising:

[0039] A regional network construction unit is used to construct an optical-laser-SAR multi-source satellite image regional network, extract image points with the same name from optical satellite stereo pairs and SAR satellite images, and from optical satellite stereo pairs and spaceborne laser altimetry data; obtain optical-SAR satellite image connection points based on the extracted image points with the same name from optical satellite stereo pairs and SAR satellite images; and obtain optical stereo images and spaceborne laser altimetry data connection points based on the extracted image points with the same name from optical satellite stereo pairs and spaceborne laser altimetry data;

[0040] An elevation error compensation unit is used to detect the systematic elevation error of the optical satellite stereo image pair based on the connection points between the optical stereo image and the satellite-borne laser altimetry data, using the satellite-borne laser altimetry data as an elevation reference, and to correct the elevation error of the geometric model of the optical satellite stereo image pair, thereby obtaining the geometric model of the optical satellite stereo image pair after elevation error compensation;

[0041] The coordinate calculation unit includes an elevation coordinate calculation module and a plane coordinate calculation module; the elevation coordinate calculation module is used to calculate the elevation coordinates of the object point corresponding to the optical-SAR satellite image connection point using the forward intersection method based on the optical satellite stereo image pair geometric model after elevation error compensation; the plane coordinate calculation module is used to calculate the plane coordinates of the object point corresponding to the optical-SAR satellite image connection point based on the elevation coordinates and the SAR satellite image geometric model;

[0042] The plane error compensation unit is used to detect and compensate the plane error in the optical satellite stereo image pair regional network by using the object point corresponding to the optical-SAR satellite image connection point as the plane height control point;

[0043] The adjustment parameter calculation unit is used to generate evenly distributed virtual control points for all optical satellite stereo image pairs in the survey area, and add an affine error image square compensation model to establish a single optical satellite stereo image pair without ground control area block adjustment model based on the virtual control points, and solve the adjustment model parameters;

[0044] The optical-laser-SAR multi-source data uncontrolled joint positioning system is used to execute the steps in the above-mentioned optical-laser-SAR multi-source data uncontrolled joint positioning method.

[0045] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0046] The present invention replaces ground control points with SAR satellite images and satellite-borne laser altimetry data, and uses SAR satellite images and satellite-borne laser altimetry data as plane control constraints and elevation control constraints, respectively, to improve the reliability of the optical satellite image block adjustment system under conditions without ground control. The present invention uses SAR satellite images as plane control constraints to eliminate the accumulation of system errors such as rotation error, scaling error, attitude and track drift error, etc. in the optical satellite image block, and eliminates accidental errors through large-scale block adjustment of homologous optical satellite images, thereby improving the accuracy of the optical satellite image block adjustment results under conditions without ground control. The present invention adopts a step-by-step compensation strategy for plane error and elevation error, which can avoid the cross-influence and transmission accumulation problems of plane error and elevation error in the satellite-borne optical-laser-SAR multi-source remote sensing data joint adjustment system. In summary, the present invention proposes a ground-free joint positioning method for spaceborne optical-laser-SAR multi-source remote sensing data combined with plane-elevation step-by-step compensation. Spaceborne laser altimetry data is used as an elevation control constraint to improve the elevation positioning accuracy of optical stereo image pairs, provide a high-precision elevation benchmark for SAR satellite images, and then use SAR satellite images as plane constraints to improve the plane positioning accuracy of optical stereo image pairs, thereby achieving the purpose of high-precision joint positioning of spaceborne optical-laser-SAR multi-source remote sensing data without ground control conditions, and improving the geometric positioning accuracy of optical satellite remote sensing images without ground control conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of an uncontrolled joint positioning method for optical-laser-SAR multi-source data provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0048] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0049] Example 1:

[0050] Example 1 provides an optical-laser-SAR multi-source data uncontrolled joint positioning method, see Figure 1 , which includes the following steps:

[0051] Step 1: Construct an optical-laser-SAR multi-source satellite image regional network, extract the same-name image points from the optical satellite stereo image pair-SAR satellite image and the optical satellite stereo image pair-spaceborne laser altimetry data; obtain the optical-SAR satellite image connection points based on the extracted same-name image points from the optical satellite stereo image pair-SAR satellite image; obtain the connection points between the optical stereo image and the spaceborne laser altimetry data based on the extracted same-name image points from the optical satellite stereo image pair-spaceborne laser altimetry data.

[0052] Step 1.1: Based on the geodetic coordinates of the satellite-borne laser altimeter data and the RFM (Rational Function Model) model of the optical satellite stereo image pair, calculate the coordinates of the initial image points of the satellite-borne laser altimeter data in the downward optical image, and use the tie point automatic matching algorithm to obtain the tie points of the laser image points in the downward optical image in the front and rear optical images.

[0053] It should be noted that the RFM model that appears in step 1.1 is a general term, and the initial RFM model, iterative RFM model, and refined RFM model that appear in subsequent steps are specific RFM models.

[0054] The RFM model is a mapping relationship between the image coordinates of satellite images and the geodetic coordinates of corresponding ground points based on rational polynomials. Its form is as follows:

[0055]

[0056]

[0057] Where, (l n , s n ) is the normalized coordinate corresponding to the image coordinate (l, s), ranging from [-1, 1]; LineOff, SampleOff are the translation values ​​of the image coordinates; LineScale, SampleScale are the scaling values ​​of the image coordinates; (U, V, W) are the normalized coordinates of the corresponding ground point geodetic coordinates (Lon, Lat, Height), ranging from [-1, 1]; LonOff, LatOff, HeiOff are the translation values ​​of the geodetic coordinates, LonScale, LatScale, HeiScale are the scaling values ​​of the geodetic coordinates; a i 、b i 、c i d i (i=1~20) is the RFM model parameter.

[0058] Step 1.2: Use the optical-SAR satellite image tie point automatic matching algorithm to obtain the tie points between the optical satellite stereo image pair and the SAR satellite image.

[0059] The present invention requires that each pair of optical satellite stereo images have at least three tie points with the SAR satellite imagery, and all optical-SAR tie points must have corresponding image point coordinates in all images of the optical satellite stereo pair. If the automatic matching results do not meet these requirements, manual measurement can be used to obtain the optical-SAR satellite image tie points.

[0060] Step 1.3: Use the optical satellite image tie point automatic matching algorithm to obtain the tie points between optical satellite images.

[0061] The present invention requires that the connection points within each pair of optical satellite stereo images in the area should be evenly distributed, with a number of no less than 25; the number of connection points between adjacent stereo images should be no less than 5; and the overlap of 90% of the optical image connection points in the area (i.e., the number of images where the connection point is located) should be equal to the total number of images in the overlapping area.

[0062] Step 2: Using the satellite-borne laser altimetry data as the elevation reference, the systematic elevation error of the optical satellite stereo image pair is detected based on the connection points between the optical stereo image and the satellite-borne laser altimetry data. The elevation error of the geometric model of the optical satellite stereo image pair is corrected to obtain the geometric model of the optical satellite stereo image pair after elevation error compensation.

[0063] Step 2.1: Based on the geometric error characteristics of single-view optical satellite images, an image square error compensation model is added to the initial RFM model of the single-view image. Combining formulas (1) and (2), it can be expressed as follows:

[0064]

[0065] It should be noted that F in the above formula x (Lat, Lon, Height) and F y (Lat, Lon, Height) is only a mathematical expression of the functional relationship and does not have any special physical meaning.

[0066] Among them, the image space compensation models Δl and Δs can be expressed as polynomials of the row and column coordinates of the image points, namely:

[0067]

[0068] Where a l0 ,a l1 ,a l2 ,…and b s0 ,b s1 ,b s2 ,…are all correction parameters that can be used to express or define different physical meanings. For example, a l0 Indicates that the image point l coordinate shift caused by various errors in the scanning direction is absorbed, including the ephemeris error in the scanning direction, the roll angle error, the calibration error of the principal point l coordinate, and the CCD physical distortion calibration error; b s0 Indicates that the image point s coordinate shift caused by various errors in the flight direction is absorbed, including the ephemeris error in the flight direction, the pitch error, the calibration error of the principal point s coordinate, and the CCD physical distortion calibration error. Parameter a l1 and b s1 Indicates that the slight error caused by gyro drift during imaging is absorbed. Parameter a l2 and bs2 It means that internal orientation errors such as lens focal length and lens distortion errors can be absorbed.

[0069] Step 2.2: Based on the initial RFM model of the single-view image with the additional image square error compensation model established by formula (5), the basic flat model is established with the optical satellite image connection points as observation values:

[0070]

[0071] It should be noted that G in the above formula x and G y It is only a mathematical expression of the functional relationship and does not have any special physical meaning.

[0072] For tie points, since their unknown parameters include not only the RFM image-space additional parameters of the image where the image point is located, but also its corresponding object-space coordinates, the error equation constructed by the tie points is a nonlinear equation and needs to be linearized. The error equation after linearization is:

[0073]

[0074] Where, v l is the residual error of the image point row coordinate observation value, v s is the residual of the observation value of the image point column coordinate, (Lat, Lon, Height) 0 It represents the initial value of the object coordinate of the connection point, and d(Lat,Lon,Height) represents the correction value of the object coordinate of the connection point.

[0075] According to the indirect adjustment principle, the error equation is expressed in formula (8).

[0076] V=Ax+Bt-LP (8)

[0077]

[0078] Where V represents the residual vector of the image point coordinate observation value; x represents the error compensation parameter vector to be solved, t=[T1…T j …T n ] T (j=1,2…n) represents the coordinate correction value vector of each connection point, T j =d(Lat,Lon,Hei) j Represents the geographic coordinate correction number of the jth connection point, n represents the number of connection points, A and B are the partial derivative coefficient matrices corresponding to the unknown numbers, L is a constant vector, P is a weight matrix, Lat is latitude, Lon is longitude, and Hei is elevation.

[0079] Step 2.3: According to the principle of least squares adjustment, the constructed observation error equation is normalized to construct a modified normal equation that only contains the parameters of the image additional image square error compensation model to be solved. The resulting equation is shown in formula (13):

[0080]

[0081] During the block adjustment process, the dimension of the object coordinates t of the tie points is usually much higher than the image additional parameters x. In the field of photogrammetry, the tie point coordinates t are often eliminated first, and the modified equation containing only the additional model parameters x is constructed, as shown in formula (14):

[0082] [A T AA T B(B T B) -1 B T A]x=A T LA T B(B T B) -1 B T L (14)

[0083] Step 2.4: Solve the modified equations. According to the current iterative RFM model of each image and the obtained image additional image square error compensation model parameters (i.e., image additional parameters x), solve the geographic coordinates of the object points corresponding to each connection point through spatial forward intersection, and substitute them into the next adjustment solution until the results of the two adjustments are consistent, or the iteration is stopped when the number of iterations is greater than the set number.

[0084] Step 2.5: For the connection points between the acquired optical stereo images and the satellite-borne laser altimetry data, refine the RFM model and the image point coordinates of the connection points using the adjusted images of each scene. Obtain the geographic coordinates of the corresponding object points through spatial forward intersection, and calculate the elevation difference with the satellite-borne laser altimetry data. When multiple satellite-borne laser altimetry data exist in the survey area, calculate the mean elevation difference of all satellite-borne laser altimetry data. Using the downward-looking image of all optical satellite stereo image pairs in the survey area as the reference, correct the geometric error of the corresponding fore- and fore-looking images along the epipolar line according to the mean elevation difference, and regenerate the RFM model of each scene image.

[0085] The compensation model used to correct the geometric error along the epipolar line is expressed as follows:

[0086] l+Δl=F x (Lat,Lon,Height) (15)

[0087] Where l is the image plane coordinate of the image point, (Lat, Lon, Height) is the longitude, latitude and elevation geographic coordinates of the object point corresponding to the image point, and Δl is the additional system error translation model compensation parameter.

[0088] That is, in step 2, the satellite-borne laser altimetry data is used as the elevation control constraint, and the systematic error translation model is added to the optical satellite stereo image pair geometric model to construct a geometric model for systematic elevation error compensation, which is in the form of (15).

[0089] Step 3: Based on the geometric model of the optical satellite stereo image pair after elevation error compensation, the forward intersection method is used to calculate the elevation coordinates of the object point corresponding to the optical-SAR satellite image connection point.

[0090] Step 4: Based on the elevation coordinates obtained in step 3 and the SAR satellite image geometric model, calculate the object plane coordinates of the optical-SAR satellite image connection point.

[0091] Step 5: Using the object points corresponding to the optical-SAR satellite image connection points obtained in steps 3 and 4 as horizontal height control points, detect and compensate for possible planar errors (such as rotation error, scaling error, and attitude and orbit drift error) in the optical satellite stereo image pair regional network.

[0092] Step 5.1: Based on steps 2.1 and 2.2, construct the error equation for the block adjustment of optical satellite stereo images with additional plane control points.

[0093] The form of the equation is still as shown in formula (8). For the image point of the level control point, since the geographic coordinates of the corresponding object point are precisely known, the unknown parameters in the constructed error equation only include the additional parameters of the RFM image square of the image where the image point is located. At this time, formula (8) can be expressed as:

[0094] V=Ax-L P (16)

[0095] Step 5.2: According to the principle of least squares adjustment, normalize the observation error equation to obtain the normal equation.

[0096] The normal equation is as follows:

[0097] A T PAx=A T PL (17)

[0098] According to the solution of the normal equation, we have x=(A T PA) -1 (A T PL).

[0099] Step 5.3: Solve the normal equation and regenerate the refined RFM model for each scene image.

[0100] Step 6: Generate evenly distributed virtual control points for all optical satellite stereo image pairs in the survey area, and add an affine error image cube compensation model. Establish a block adjustment model for a single optical satellite stereo image pair without ground control based on the virtual control points, and solve the adjustment model parameters to achieve positioning without ground control.

[0101] Step 6.1: For all optical satellite images within the survey area, generate uniform grids at regular intervals and take the center point of each grid. Using the image-space coordinates of each grid center point and the initial RFM model of the corresponding image, and taking the mean elevation surface of the geodetic coordinates in the RFM model as the elevation datum, obtain the object-space geographic coordinates of the center point through spatial forward intersection. Construct a pair of virtual control points between each grid center image point and the corresponding object point.

[0102] Step 6.2: Use the connecting point image points and virtual control point image points between all optical satellite images to construct the observation error equation.

[0103] Specifically, the virtual control points are regarded as real control points with a certain accuracy, and their block adjustment error equations can be directly constructed according to formula (6). The observation error equations constructed by all the connection point image points and the virtual control point image points are written into matrix form, as shown in formula (18).

[0104] That is, the observation error equation in step 6.2 is expressed as follows:

[0105]

[0106] The above equations represent the matrix of the observation error equations constructed for all virtual control point and tie point image points. The subscripts vc and tp represent virtual control points and tie points, respectively. x and t are the adjustment parameters to be solved, representing the image-space additional parameter vector and the object-space coordinate correction vector of the RFM image to be adjusted, respectively. A and B are the partial derivative coefficient matrices of the corresponding unknowns, while L and P are the corresponding constant vector and weight matrix, respectively.

[0107] Specifically, V vc is the residual vector of the error equation corresponding to the virtual control point, V tp is the residual vector of the error equation corresponding to the tie point; x and t are the adjustment parameters to be solved, x is the image space additional parameter vector of the RFM image to be adjusted, t is the object space coordinate correction vector of the tie point, A vc is the partial derivative coefficient matrix of the image square additional parameters corresponding to the virtual control points, A tp is the partial derivative coefficient matrix of the image square additional parameters corresponding to the connection points, B tp is the partial derivative coefficient matrix of the object space geographic coordinate correction number corresponding to the connection point, L vcis the constant vector corresponding to the virtual control point, L tp is the constant vector corresponding to the connection point, P VC is the weight matrix corresponding to the virtual control point, P tp is the weight matrix corresponding to the connection point.

[0108] That is, for all optical satellite stereo image pairs to be adjusted in the regional network, their image planes are divided into uniform regular grids, the coordinates of the central image points of each grid are taken (l, s), and the corresponding object point coordinates (Lat, Lon, HeiOFF) are obtained by forward intersection on the regional average elevation surface with the geometric model that has been compensated for systematic errors, and a pair of virtual control points are obtained. Then, the image points of the connecting points between all optical satellite stereo image pairs and the image points of the virtual control points are used to construct a single optical satellite stereo image pair without ground control regional network adjustment model based on virtual control points. The error equation is as follows (18).

[0109] Step 6.3: Iteratively solve the adjustment parameters in the observation error equation.

[0110] For the optical satellite stereo image pair with optical-SAR satellite image connection points, that is, the optical satellite stereo image pair processed in step 5, its virtual control point corresponds to the weight matrix P VC It can be set according to the geometric positioning accuracy of the processed optical satellite stereo image pair without ground control to ensure that it plays a control constraint role in the overall regional block adjustment process of all optical satellite images in the survey area.

[0111] Follow the methods shown in steps 2.3 and 2.4 to iteratively solve the adjustment parameters x and t.

[0112] After the iteration is completed (i.e. after the adjustment parameters are solved), the refined RFM models of all optical satellite images in the survey area are updated.

[0113] Example 2:

[0114] Example 2 provides an optical-laser-SAR multi-source data uncontrolled joint positioning system, including:

[0115] A regional network construction unit is used to construct an optical-laser-SAR multi-source satellite image regional network, extract image points with the same name from optical satellite stereo pairs and SAR satellite images, and from optical satellite stereo pairs and spaceborne laser altimetry data; obtain optical-SAR satellite image connection points based on the extracted image points with the same name from optical satellite stereo pairs and SAR satellite images; and obtain optical stereo images and spaceborne laser altimetry data connection points based on the extracted image points with the same name from optical satellite stereo pairs and spaceborne laser altimetry data;

[0116] An elevation error compensation unit is used to detect the systematic elevation error of the optical satellite stereo image pair based on the connection points between the optical stereo image and the satellite-borne laser altimetry data, using the satellite-borne laser altimetry data as an elevation reference, and to correct the elevation error of the geometric model of the optical satellite stereo image pair, thereby obtaining the geometric model of the optical satellite stereo image pair after elevation error compensation;

[0117] The coordinate calculation unit includes an elevation coordinate calculation module and a plane coordinate calculation module; the elevation coordinate calculation module is used to calculate the elevation coordinates of the object point corresponding to the optical-SAR satellite image connection point using the forward intersection method based on the optical satellite stereo image pair geometric model after elevation error compensation; the plane coordinate calculation module is used to calculate the plane coordinates of the object point corresponding to the optical-SAR satellite image connection point based on the elevation coordinates and the SAR satellite image geometric model;

[0118] The plane error compensation unit is used to detect and compensate the plane error in the optical satellite stereo image pair regional network by using the object point corresponding to the optical-SAR satellite image connection point as the plane height control point;

[0119] The adjustment parameter calculation unit is used to generate evenly distributed virtual control points for all optical satellite stereo image pairs in the survey area, and add an affine error image square compensation model to establish a single optical satellite stereo image pair without ground control area block adjustment model based on the virtual control points, and solve the adjustment model parameters;

[0120] The optical-laser-SAR multi-source data uncontrolled joint positioning system is used to execute the steps in the optical-laser-SAR multi-source data uncontrolled joint positioning method as described in Example 1.

[0121] Since the functions of the various units included in the system provided in Example 2 correspond to the various steps in the method provided in Example 1, Example 2 can be understood by referring to the description of Example 1, and will not be repeated here.

[0122] Finally, it should be noted that the above specific implementation methods are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for uncontrolled joint positioning of optical-laser-SAR multi-source data, characterized in that: The following steps are involved: Step 1: Construct an optical-laser-SAR multi-source satellite image regional network, extract image points with the same name from optical satellite stereo image pairs and SAR satellite images, and optical satellite stereo image pairs and spaceborne laser altimetry data; obtain optical-SAR satellite image connection points based on the extracted image points with the same name from optical satellite stereo image pairs and SAR satellite images; obtain optical stereo image and spaceborne laser altimetry data connection points based on the extracted image points with the same name from optical satellite stereo image pairs and spaceborne laser altimetry data; Step 2: Using the satellite-borne laser altimetry data as the elevation reference, the systematic elevation error of the optical satellite stereo image pair is detected based on the connection points between the optical stereo image and the satellite-borne laser altimetry data. The elevation error of the geometric model of the optical satellite stereo image pair is corrected to obtain the geometric model of the optical satellite stereo image pair after elevation error compensation. The step 2 includes the following sub-steps: Step 2.1: Based on the geometric error characteristics of single-view optical satellite images, an image square error compensation model is added to the initial RFM model of the single-view image; Step 2.2: Based on the initial RFM model of the single-view image with the additional image square error compensation model, the basic adjustment model is established with the connection points between optical satellite images as observation values; Step 2.3: According to the principle of least squares adjustment, the constructed observation error equation is normalized to construct a modified normal equation that only contains the parameters of the image additional image square error compensation model to be solved; Step 2.4: Solve the modified equations. Based on the current iterative RFM model of each image and the obtained image additional image square error compensation model parameters, solve the geographic coordinates of the object points corresponding to each tie point through spatial forward intersection and substitute them into the next adjustment solution until the results of the two adjustments are consistent, or the iteration is stopped when the number of iterations exceeds the set number. Step 2.5: For the connection points between the acquired optical stereo images and the satellite-borne laser altimeter data, refine the RFM model and the coordinates of the connection points using the adjusted imagery. Obtain the geographic coordinates of the corresponding object points through spatial forward intersection, and calculate the elevation difference with the satellite-borne laser altimeter data. If multiple satellite-borne laser altimeter data exist within the survey area, calculate the mean elevation difference of all satellite-borne laser altimeter data. Using the downward-looking image of all optical satellite stereo image pairs within the survey area as the reference, correct the geometric errors of the corresponding fore- and fore-looking images along the epipolar line based on the mean elevation difference, and regenerate the RFM model for each image. Step 3: Based on the geometric model of the optical satellite stereo image pair after elevation error compensation, the forward intersection method is used to calculate the elevation coordinates of the object point corresponding to the optical-SAR satellite image connection point; Step 4: Based on the elevation coordinates obtained in step 3 and the SAR satellite image geometric model, calculate the object plane coordinates of the optical-SAR satellite image connection point. Step 5: Using the object points corresponding to the optical-SAR satellite image connection points in steps 3 and 4 as horizontal and vertical control points, the planar errors in the optical satellite stereo image pair regional network are detected and compensated. The step 5 comprises the following steps: Step 5.1: Based on steps 2.1 and 2.2, construct the error equation for the block adjustment of the optical satellite stereo image pair with additional plane control points; Step 5.2: According to the principle of least squares adjustment, normalize the observation error equation to obtain the normal equation; Step 5.3: Solve the normal equation and regenerate the refined RFM model of each scene image; Step 6: Generate evenly distributed virtual control points for all optical satellite stereo image pairs in the survey area, and add an affine error image cube compensation model. Establish a block adjustment model for a single optical satellite stereo image pair without ground control based on the virtual control points, and solve the adjustment model parameters to achieve positioning without ground control. Step 6 includes the following sub-steps: Step 6.1: For all optical satellite images within the survey area, generate a uniform grid at set intervals and take the center point of each grid. Using the image-space coordinates of each grid center point and the initial RFM model of the corresponding image, and taking the mean elevation surface of the geodetic coordinates in the RFM model as the elevation datum, obtain the object-space geographic coordinates of the center point through spatial forward intersection. Construct a pair of virtual control points between each grid center image point and the corresponding object point. Step 6.2: Use the connecting point image points and virtual control point image points between all optical satellite images to construct the observation error equation; Step 6.3: Iteratively solve the adjustment parameters in the observation error equation.

2. The optical-laser-SAR multi-source data uncontrolled joint positioning method according to claim 1, characterized in that: The step 1 includes the following sub-steps: Step 1.1: Based on the geodetic coordinates of the satellite-borne laser altimeter data and the RFM model of the optical satellite stereo image pair, calculate the coordinates of the initial image points of the satellite-borne laser altimeter data in the downward optical image, and use the tie point automatic matching algorithm to obtain the tie points of the laser image points in the downward optical image in the front and rear optical images; Step 1.2: Use the optical-SAR satellite image tie point automatic matching algorithm to obtain the tie points between the optical satellite stereo image pair and the SAR satellite image; Step 1.3: Use the optical satellite image tie point automatic matching algorithm to obtain the tie points between optical satellite images.

3. The optical-laser-SAR multi-source data uncontrolled joint positioning method according to claim 1, characterized in that: In step 2.5, the compensation model used to correct the geometric error along the epipolar line direction is expressed as follows: Where, l is the image plane coordinate of the image point, are the longitude, latitude and elevation geographic coordinates of the object point corresponding to the image point, Shift the model compensation parameters for additional systematic errors.

4. The optical-laser-SAR multi-source data uncontrolled joint positioning method according to claim 1, characterized in that: In step 5, the plane error includes rotation error, scaling error and attitude drift error.

5. The optical-laser-SAR multi-source data uncontrolled joint positioning method according to claim 1, characterized in that: The observation error equation in step 6.2 is expressed as follows: Where, is the residual vector of the error equation corresponding to the virtual control point, is the residual vector of the error equation corresponding to the connection point; x and t are adjustment parameters to be solved, is the additional parameter vector of the image square of the image RFM to be adjusted, is the object space coordinate correction vector of the connection point, is the partial derivative coefficient matrix of the image-side additional parameters corresponding to the virtual control points, is the partial derivative coefficient matrix of the image-side additional parameters corresponding to the connection points, is the partial derivative coefficient matrix of the object space geographic coordinate correction number corresponding to the connection point, is the constant vector corresponding to the virtual control point, is the constant vector corresponding to the connection point, is the weight matrix corresponding to the virtual control point, is the weight matrix corresponding to the connection point.

6. The optical-laser-SAR multi-source data uncontrolled joint positioning method according to claim 1, characterized in that: In the step 6.3, after solving the adjustment parameters, the step further includes: updating and obtaining the refined RFM models of all optical satellite images in the survey area.

7. An optical-laser-SAR multi-source data uncontrolled joint positioning system, characterized in that: include: A regional network construction unit is used to construct an optical-laser-SAR multi-source satellite image regional network, extract image points with the same name from optical satellite stereo pairs and SAR satellite images, and from optical satellite stereo pairs and spaceborne laser altimetry data; obtain optical-SAR satellite image connection points based on the extracted image points with the same name from optical satellite stereo pairs and SAR satellite images; and obtain optical stereo images and spaceborne laser altimetry data connection points based on the extracted image points with the same name from optical satellite stereo pairs and spaceborne laser altimetry data; An elevation error compensation unit is used to detect the systematic elevation error of the optical satellite stereo image pair based on the connection points between the optical stereo image and the satellite-borne laser altimetry data, using the satellite-borne laser altimetry data as an elevation reference, and to correct the elevation error of the geometric model of the optical satellite stereo image pair, thereby obtaining the geometric model of the optical satellite stereo image pair after elevation error compensation; Coordinate calculation unit, including elevation coordinate calculation module and plane coordinate calculation module; The elevation coordinate calculation module is used to calculate the elevation coordinates of the object point corresponding to the optical-SAR satellite image connection point using the forward intersection method based on the optical satellite stereo image pair geometric model after elevation error compensation; the plane coordinate calculation module is used to calculate the plane coordinates of the object point corresponding to the optical-SAR satellite image connection point based on the elevation coordinates and the SAR satellite image geometric model; The plane error compensation unit is used to detect and compensate the plane error in the optical satellite stereo image pair regional network by using the object point corresponding to the optical-SAR satellite image connection point as the plane height control point; The adjustment parameter calculation unit is used to generate evenly distributed virtual control points for all optical satellite stereo image pairs in the survey area, and add an affine error image square compensation model to establish a single optical satellite stereo image pair without ground control area block adjustment model based on the virtual control points, and solve the adjustment model parameters; The optical-laser-SAR multi-source data uncontrolled joint positioning system is used to perform the steps in the optical-laser-SAR multi-source data uncontrolled joint positioning method according to any one of claims 1 to 6.

Images