Matching method between satellite-borne laser altimetry data and stereo image pairs based on minimum height difference
By matching the satellite-based laser height measurement data and stereo image pairs through the minimum height difference method, the problem of geometric accuracy deviation of remote sensing images is solved, more accurate data matching and terrain modeling is achieved, and higher accuracy is provided for high-precision geographic information. sex.
Patent Information
- Application Number
- CN202411173814.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2044-08-26
AI Technical Summary
The matching problem of onboard laser altitude measurement data with high-resolution remote sensing images leads to a deviation in geometric accuracy of remote sensing images, limiting its potential in high-precision applications.
Using the satellite-borne laser height measurement data and stereo image pair matching method based on the minimum height difference, the optimized stereo image geometry model is generated through format conversion, accuracy filtering, matching point of the same name, free network adjustment and error equation solution, and the optimized stereo image geometric model is generated and passed The buffer generates candidate subpoints to calculate the minimum height difference to determine the best matching point.
It effectively improves the matching accuracy of remote sensing images and satellite-based laser altitude measurement data, improves the accuracy of terrain modeling and the accuracy of photogrammetry models, and provides higher-precision basic geographical information for land and resource surveys and disaster monitoring.
Smart Images
Figure CN118710941B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of remote sensing and land resources surveying and mapping, and in particular to a method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference. Background Art
[0002] High-resolution remote sensing images have detailed surface information and the ability to generate high-precision digital elevation models (DEMs). These data capture subtle changes and details of the surface through high-resolution imaging technology, providing rich three-dimensional geographic information. The high spatial resolution of stereo image pair data enables subtle surface structures and terrain features to be clearly recorded, ensuring the accuracy and comprehensiveness of surface information, and has important application value in the fields of geographic information systems, precision mapping, and environmental monitoring. However, due to the errors of the sensors themselves, the errors of the orbital parameters, and the lack of ground control points, remote sensing images have certain deviations in geometric accuracy, which limits their potential in high-precision applications. Therefore, improving the geometric accuracy of remote sensing images has become a key research topic.
[0003] Satellite-borne laser altimetry data provide accurate ground elevation measurements through laser pulse ranging technology. These data record surface elevation values in point form with high accuracy and high confidence. The laser altimetry data has a large spot diameter and can cover a large surface area, providing reliable ground control point information. The accuracy and consistency of these data provide important support for surface elevation measurements, ensuring the accuracy and precision of elevation data.
[0004] Satellite-borne laser altimetry data can also provide valuable ground control point information for geometric correction of stereo image pairs. However, since laser altimetry data usually exists in point form and lacks continuous footprint information, and the ground spot (footprint) diameter of laser altimetry data is large, often at the level of 10 meters, while high-precision stereo images usually have sub-meter resolution. How to match satellite-borne laser point data with 10-meter footprints with sub-meter high-resolution images and correctly mark the position of the laser points in the images is a major challenge in this field. In order to overcome these challenges, it is still necessary to develop new data matching methods to fully utilize the high-precision advantages of laser altimetry data and achieve accurate matching with remote sensing image pixels, thereby significantly improving the geometric accuracy of remote sensing images. Summary of the invention
[0005] In order to solve the above technical problems, the present invention provides a method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference, comprising the following steps:
[0006] S1. Obtain satellite-borne laser altimetry data of the study area, convert the format of the satellite-borne laser altimetry data and perform precision screening processing to obtain satellite-borne laser elevation point data containing point elevation information;
[0007] S2, obtain the study area high-resolution stereo image pair image data, the stereo image pair image data is carried out to the same name point matching and free network adjustment to calculate the connection point; Then the satellite-borne laser elevation point data obtained in step S1 is used as the control point on the stereo image pair, and the stereo image pair is combined with the adjustment to solve the error equation, the optimized stereo image imaging geometric model is obtained and the digital elevation model data is generated;
[0008] S3, taking the plane coordinates of each satellite-borne laser elevation point data obtained as the center of the circle and the spot diameter of the satellite-borne laser point as the diameter to make a buffer zone; then within the buffer zone, with the spatial resolution of the stereo image as the spacing, generate multiple candidate sub-points, that is, the position of each sub-point corresponds to a pixel position of the stereo image;
[0009] S4, assigning the elevation value of each satellite-borne laser elevation point data to all candidate sub-points within its buffer range, recorded as laser elevation value; at the same time, using all sub-points to extract the elevation values of their corresponding positions on the digital elevation model generated in step S2, recorded as digital elevation model elevation value;
[0010] S5. For each sub-point of the satellite-borne laser elevation point, the difference between its laser elevation value and the elevation value of the digital elevation model is calculated, and all satellite-borne laser elevation points in the sample area are traversed to calculate the overall elevation difference of the buffer zone;
[0011] S6. Traverse all satellite-borne laser elevation points and their sub-points, calculate the overall elevation difference of the buffer zone when different sub-point combinations are used, and when the overall elevation difference of the buffer zone is the minimum value among all combinations, it is considered that the position of the sub-point at this time is the best matching point between the satellite-borne laser elevation point and the stereo image.
[0012] The technical solution further defined in the present invention is:
[0013] Furthermore, in step S1, ICESat-2 satellite-borne laser altimetry data and GF-7 satellite stereo image pair remote sensing images of the study area are obtained, and the ICESat-2 laser altimetry data are format converted and precision screened to obtain satellite-borne laser elevation point data containing point elevation information.
[0014] As described above, in the method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference, in step S1, the format conversion method is specifically: converting the H5 data format into the CSV data format.
[0015] As described above, in the method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference, in step S1, the method for precision screening is specifically as follows: referring to the elevation attributes of satellite-borne laser elevation points, according to the principles of quality mark == 0; slope mark < 0.01; cloud cover mark < 3; city mark == 0; quantity mark ≥ 20; day and night mark = 1, satellite-borne laser elevation points whose precision is impaired due to environmental factors are eliminated.
[0016] As described above, in the method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference, in step S2, the matching method for same-name points is a semi-global dense matching algorithm, as shown in the following formula:
[0017]
[0018] in, yes Pixels in direction In the parallax Cost aggregation result; for a given stereo image pair, the same-name points differ only in column coordinates: ,in and are the column coordinates in the left and right images; It's pixels In the parallax The matching cost when It is the direction The previous pixel; It is parallax. And not for and Other parallaxes when and is the penalty term for the disparity change between adjacent pixels, and .
[0019] As described above, in the method for matching satellite-borne laser altimetry data with stereo image pairs based on minimum height difference, in step S2, the free network adjustment method is rank-deficient free network adjustment, and its equation is:
[0020]
[0021] in, is the observation vector, which contains all the measured observation values; is the design matrix used to describe the coefficients of the unknown parameters in each observation equation; is the unknown parameter vector, and is the parameter calculated by adjustment; is the covariance matrix of the observations, which is used to describe the distribution of observation errors; is the residual vector representing the difference between the actual observed values and the calculated values.
[0022] As described above, the method for matching satellite-borne laser altimetry data with a stereo image pair based on minimum height difference, in step S2, the method for puncturing the satellite-borne laser elevation point data on the stereo image pair is specifically as follows: according to the coordinates of the satellite-borne laser elevation point Positioning on the stereo image pair, the corresponding pixel on the stereo image pair is considered to be the best matching point of the satellite-borne laser elevation point.
[0023] As described above, in the method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference, in step S2, the error equation is an error equation based on a rational function model, and the formula is:
[0024]
[0025] in, , is the coefficient matrix, is the image affine transformation parameter correction matrix, is the object coordinate correction matrix of the connection point, is the initial value constant matrix, is the weight matrix.
[0026] As described above, in the method for matching satellite-borne laser altimetry data with stereo image pairs based on minimum height difference, in step S5, the formula for calculating the overall height difference of the buffer zone is as follows:
[0027]
[0028] in, is the overall height difference of the buffer zone, is the number of satellite-borne laser elevation points, Number the candidate sub-points of each satellite-borne laser elevation point, is the total number of satellite-borne laser elevation points, is the total number of candidate sub-points for each satellite-borne laser elevation point, It is the difference between the laser elevation value of the sub-point and the elevation value of the digital elevation model.
[0029] As described above, in the method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference, in step S6, the formula for calculating the overall minimum height difference of the buffer zone is as follows:
[0030]
[0031] in, is the overall minimum height difference of the buffer zone, is the number of possible combinations of all sub-points.
[0032] The beneficial effects of the present invention are:
[0033] In the present invention, firstly, satellite-borne laser elevation point data containing point elevation information is obtained by performing format conversion, precision screening and other processing on satellite-borne laser altimetry data, and the satellite-borne laser elevation point data is used as a control point for joint adjustment with a stereo image pair, thereby generating a DEM; secondly, a buffer zone is made with the plane coordinates of the satellite-borne laser elevation point as the center of a circle and the spot diameter of the laser elevation point as the diameter, candidate sub-points are generated with the spatial resolution of the stereo image as the spacing, and then laser elevation values and DEM elevation values are added to all sub-points; finally, the overall elevation difference of the buffer zone is calculated, and then all satellite-borne laser elevation points and their sub-points are traversed, and then the overall elevation difference of the buffer zone when different sub-points are combined is calculated, and the best matching point between the satellite-borne laser elevation point and the stereo image is obtained, thereby effectively improving the matching accuracy between the remote sensing image and the satellite-borne laser altimetry data, especially when the satellite-borne laser altimetry data and the stereo image pair are jointly used for terrain modeling, more accurate point puncture can be performed, and a more accurate photogrammetric model can be obtained, thereby providing a more accurate basic geographic information product for land and resources surveys and disaster monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 It is a schematic diagram of the overall process of the present invention;
[0035] Figure 2 It is a schematic diagram of a stereoscopic image pair of the Gaofen-7 satellite of a sample area in an embodiment of the present invention;
[0036] Figure 3 A schematic diagram of a DEM generated before the ICESat-2 control point is subjected to the minimum height difference method in an embodiment of the present invention;
[0037] Figure 4 A candidate sub-point map is generated for traversing ICESat-2 satellite-borne laser elevation points in an embodiment of the present invention;
[0038] Figure 5 The following is a comparison diagram of the matching points of the ICESat-2 satellite-borne laser elevation points and the stereo images in the embodiment of the present invention and the optimal points obtained by this method: Figure 5 (a) is a schematic diagram of the rear view image of the Gaofen-7 stereo image pair; Figure 5 (b) is a schematic diagram of the front view image of the Gaofen-7 stereo image pair. DETAILED DESCRIPTION
[0039] This embodiment provides a method for matching satellite-borne laser altimetry data with stereo image pairs based on minimum height difference, such as Figure 1 As shown, the following steps are included:
[0040] S1. For example Figure 2The GF-7 satellite stereo image pair images of the sample area shown in the figure are used to obtain the ICESat-2 satellite-borne laser altimetry data and the GF-7 satellite stereo image pair remote sensing images of the study area. The ICESat-2 laser altimetry data are converted from the H5 data format to the CSV data format, and the elevation attributes of the ICESat-2 satellite-borne laser elevation points are referred to. According to the principle of "quality mark == 0; slope mark < 0.01; cloud cover mark < 3; city mark == 0; quantity mark ≥ 20; day and night mark = 1", the satellite-borne laser elevation points with impaired accuracy caused by environmental factors such as clouds, aerosols, and terrain slope are eliminated to obtain the satellite-borne laser elevation point data containing point elevation information.
[0041] S2, then the GF-7 stereo image pair is subjected to radiation correction and geometric correction preprocessing, and the stereo image pair image data is subjected to homonymous point matching and free network adjustment to calculate the connection points; the homonymous point matching method is a semi-global dense matching algorithm, as shown in the following formula:
[0042]
[0043] in, yes Pixels in direction In the parallax Cost aggregation result; for a given stereo image pair, the same-name points differ only in column coordinates: ,in and are the column coordinates in the left and right images; It's pixels In the parallax The matching cost when It is the direction The previous pixel; It is parallax. And not for and Other parallaxes when and is the penalty term for the disparity change between adjacent pixels, and .
[0044] The free network adjustment method is rank-deficient free network adjustment, and its equation is:
[0045]
[0046] in, is the observation vector, which contains all the measured observation values; is the design matrix used to describe the coefficients of the unknown parameters in each observation equation; is the unknown parameter vector, and is the parameter calculated by adjustment; is the covariance matrix of the observations, which is used to describe the distribution of observation errors; is the residual vector representing the difference between the actual observed values and the calculated values.
[0047] Then the satellite-borne laser elevation point data obtained in step S1 is used as a control point to puncture points on the Gaofen-7 stereo image pair, and the coordinates of the satellite-borne laser elevation point are calculated. Positioning on the stereo image pair, the corresponding pixel on the stereo image pair is considered to be the best matching point of the satellite-borne laser elevation point.
[0048] Then, the best matching points are adjusted together with the stereo image pair to solve the error equation, obtain the optimized stereo image geometric model and generate the digital elevation model (DEM) data. The DEM generated before the minimum height difference method is obtained by the ICESat-2 control points. Figure 3 shown.
[0049] The error equation is based on the rational function model (RFM), and the formula is:
[0050]
[0051] in, , is the coefficient matrix, is the image affine transformation parameter correction matrix, is the object coordinate correction matrix of the connection point, is the initial value constant matrix, is the weight matrix.
[0052] S3, take the plane coordinates of each satellite-borne laser elevation point data as the center of the circle, and the ICESat-2 satellite-borne laser spot diameter of 17.5m as the diameter to make a buffer zone; then within the buffer zone, with the spatial resolution of the GF-7 stereo image of 0.8m as the spacing, generate 373 candidate sub-points, that is, the position of each sub-point corresponds to a pixel position of the stereo image; ICESat-2 satellite-borne laser elevation points traverse to generate candidate sub-points as shown in the figure Figure 4 As shown, R is 8.75m and d is 0.8m.
[0053] S4. Assign the elevation value of each ICESat-2 satellite-borne laser elevation point data to all candidate sub-points within its buffer range, recorded as laser elevation value; at the same time, use all sub-points to extract the elevation values of their corresponding positions on the digital elevation model generated in step S2, recorded as digital elevation model elevation value.
[0054] S5. For each sub-point of the ICESat-2 satellite laser elevation point, calculate the difference between its laser elevation value and the digital elevation model elevation value, and traverse all the satellite laser elevation points in the sample area to calculate the overall elevation difference of the buffer zone. The specific formula is as follows:
[0055]
[0056] in, is the overall height difference of the buffer zone, is the number of satellite-borne laser elevation points, Number the candidate sub-points of each satellite-borne laser elevation point, is the total number of satellite-borne laser elevation points, is the total number of candidate sub-points for each satellite-borne laser elevation point, It is the difference between the laser elevation value of the sub-point and the elevation value of the digital elevation model.
[0057] S6. Traverse all ICESat-2 satellite laser elevation points and their sub-points, calculate the overall elevation difference of the buffer zone when different sub-point combinations are used, and when the overall elevation difference of the buffer zone is the minimum value among all combinations, the position of the sub-point (or the pixel where it is located) is considered to be the best matching point between the satellite laser elevation point and the stereo image; the formula for calculating the overall minimum elevation difference of the buffer zone is as follows:
[0058]
[0059] in, is the overall minimum height difference of the buffer zone, is the number of possible combinations of all sub-points.
[0060] In this embodiment, ICESat-2 satellite-borne laser elevation points are traversed to generate a candidate sub-point map as shown in Figure 4 As shown by Figure 4 It can be seen that after the calculation by the minimum height difference method, the best matching point of an ICESat-2 satellite-borne laser elevation point is changed from the original center position to the point position of its lower right corner, with a horizontal offset of nearly 8m; Figure 5 The comparison between the position of the ICESat-2 satellite laser elevation point and the stereo image matching point and the best position obtained by this method in the embodiment is shown in FIG. Figure 5 Figure (a) is the rear view image of the Gaofen-7 stereo image pair. Figure 5 Figure (b) is the front view image of the GF-7 stereo image pair. The number 23 in the figure is the number of the ICESat-2 satellite-borne laser elevation point. The five-pointed star represents the position of the point before the minimum height difference method is used for calculation, and the triangle represents the best matching point after the minimum height difference method is used for calculation.
[0061] The method of this embodiment, firstly, obtains satellite-borne laser elevation point data containing point elevation information by performing format conversion, accuracy screening and other processing on satellite-borne laser altimetry data, and uses the satellite-borne laser elevation point data as control points for joint adjustment with stereo image pairs, thereby generating a DEM; secondly, uses the plane coordinates of the satellite-borne laser elevation point as the center of the circle and the spot diameter of the laser elevation point as the diameter to make a buffer zone, generates candidate sub-points with the spatial resolution of the stereo image as the spacing, and then adds laser elevation values and DEM elevation values to all sub-points; finally, calculates the overall elevation difference of the buffer zone, then traverses all satellite-borne laser elevation points and their sub-points, and then calculates the overall elevation difference of the buffer zone when different sub-points are combined, and obtains the best matching point between the satellite-borne laser elevation point and the stereo image, thereby effectively improving the matching accuracy between the remote sensing image and the satellite-borne laser altimetry data, especially when the satellite-borne laser altimetry data and the stereo image pairs are jointly used for terrain modeling, more accurate point puncture can be performed, and a more accurate photogrammetric model can be obtained, so as to provide a more accurate basic geographic information product for land and resources surveys and disaster monitoring.
[0062] In addition to the above embodiments, the present invention may also have other implementation modes. Any technical solution formed by equivalent replacement or equivalent transformation falls within the protection scope required by the present invention.
Claims
1. A method for matching satellite-borne laser altimetry data with stereo image pairs based on minimum height difference, characterized in that: The following steps are involved: S1. Obtain satellite-borne laser altimetry data of the study area, convert the format of the satellite-borne laser altimetry data and perform precision screening processing to obtain satellite-borne laser elevation point data containing point elevation information; S2, obtain the study area high-resolution stereo image pair image data, the stereo image pair image data is carried out to the same name point matching and free network adjustment to calculate the connection point; Then the satellite-borne laser elevation point data obtained in step S1 is used as the control point on the stereo image pair, and the stereo image pair is combined with the adjustment to solve the error equation, the optimized stereo image imaging geometric model is obtained and the digital elevation model data is generated; S3, taking the plane coordinates of each satellite-borne laser elevation point data obtained as the center of the circle and the spot diameter of the satellite-borne laser point as the diameter to make a buffer zone; then within the buffer zone, with the spatial resolution of the stereo image as the spacing, generate multiple candidate sub-points, that is, the position of each sub-point corresponds to a pixel position of the stereo image; S4, assigning the elevation value of each satellite-borne laser elevation point data to all candidate sub-points within its buffer range, recorded as laser elevation value; at the same time, using all sub-points to extract the elevation values of their corresponding positions on the digital elevation model generated in step S2, recorded as digital elevation model elevation value; S5. For each sub-point of the satellite-borne laser elevation point, the difference between its laser elevation value and the elevation value of the digital elevation model is calculated, and all satellite-borne laser elevation points in the sample area are traversed to calculate the overall elevation difference of the buffer zone; S6. Traverse all satellite-borne laser elevation points and their sub-points, calculate the overall elevation difference of the buffer zone when different sub-point combinations are used, and when the overall elevation difference of the buffer zone is the minimum value among all combinations, it is considered that the position of the sub-point at this time is the best matching point between the satellite-borne laser elevation point and the stereo image.
2. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S1, ICESat-2 satellite-borne laser altimetry data and GF-7 satellite stereo image pair remote sensing images of the study area are obtained, and the ICESat-2 laser altimetry data are format converted and precision screened to obtain satellite-borne laser elevation point data containing point elevation information.
3. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S1, the format conversion method is specifically: converting the H5 data format into the CSV data format.
4. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S1, the precision screening method is specifically as follows: referring to the elevation attributes of the satellite-borne laser elevation points, according to the principles of quality mark==0; slope mark<0.01; cloud cover mark<3; city mark==0; quantity mark≥20; day and night mark=1, the satellite-borne laser elevation points whose precision is impaired due to environmental factors are eliminated.
5. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S2, the same-name point matching method is a semi-global dense matching algorithm, as shown in the following formula: in, yes Pixels in direction In the parallax Cost aggregation result; for a given stereo image pair, the same-name points differ only in column coordinates: ,in and are the column coordinates in the left and right images; It's pixels In the parallax The matching cost when It is the direction The previous pixel; It is parallax. And not for and Other parallaxes when and is the penalty term for the disparity change between adjacent pixels, and .
6. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S2, the free network adjustment method is rank-deficient free network adjustment, and its equation is: in, is the observation vector, which contains all the measured observation values; is the design matrix, which is used to describe the coefficients of the unknown parameters in each observation equation; is the unknown parameter vector, and is the parameter calculated by adjustment; is the covariance matrix of the observations, which is used to describe the distribution of observation errors; is the residual vector representing the difference between the actual observed values and the calculated values.
7. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S2, the method of puncturing the satellite-borne laser elevation point data on the stereo image pair is specifically as follows: according to the coordinates of the satellite-borne laser elevation point Positioning on the stereo image pair, the corresponding pixel on the stereo image pair is considered to be the best matching point of the satellite-borne laser elevation point.
8. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S2, the error equation is an error equation based on a rational function model, and the formula is: in, , is the coefficient matrix, is the image affine transformation parameter correction matrix, is the object coordinate correction matrix of the connection point, is the initial value constant matrix, is the weight matrix.
9. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S5, the formula for calculating the overall height difference of the buffer zone is as follows: in, is the overall height difference of the buffer zone, is the number of satellite-borne laser elevation points, Number the candidate sub-points of each satellite-borne laser elevation point, is the total number of satellite-borne laser elevation points, is the total number of candidate sub-points for each satellite-borne laser elevation point, It is the difference between the laser elevation value of the sub-point and the elevation value of the digital elevation model.
10. The method for matching satellite-borne laser altimetry data and stereo image pairs based on minimum height difference according to claim 1, characterized in that: In step S6, the formula for calculating the overall minimum height difference of the buffer zone is as follows: in, is the overall minimum height difference of the buffer zone, is the number of possible combinations of all sub-points.