A method for constructing an image control point library based on stereo mapping satellite images

Abstract

The present invention relates to a method for constructing an image control point library based on stereo mapping satellite images, and belongs to the field of remote sensing mapping technology. The present invention improves the overall ground positioning accuracy of stereo mapping satellite image data by extracting feature points and performing regional network adjustment processing on the acquired stereo mapping satellite images; and extracts feature points from the stereo mapping satellite images, and uses image control points generated by the intersection of feature points, combined with image control points generated by matching and intersection of ground regular grid points, to obtain an image control point database with strong positioning capability. Based on the image control point database, a quantitative analysis can be performed on the ground positioning capability of the stereo mapping satellite. When there is non-stereo observation remote sensing satellite data, the image control points are retrieved from the image control point database and image control point matching is performed, thereby realizing the optimization of non-stereo observation remote sensing satellite image positioning parameters assisted by control information.

Description

Technical Field

[0001] The invention relates to a method for constructing an image control point library based on stereo mapping satellite images, and belongs to the technical field of remote sensing mapping. Background Art

[0002] Stereo mapping satellites refer to optical remote sensing satellites that can perform stereoscopic imaging of the earth and have high positioning accuracy, such as the meter-level imaging resolution satellites such as "Tianhui-1" and "Ziyuan-3", and sub-meter-level imaging resolution satellites such as "Gaofen-7" and "Tianhui-3". Since stereo mapping satellites have high earth positioning accuracy, the image data acquired by the satellites can usually meet the requirements of 1:50,000, 1:10,000 or even 1:5,000 topographic map revision and measurement, and play an important role in remote sensing and mapping-related industry applications.

[0003] Throughout their lifecycles, stereo mapping satellites acquire a large amount of stereo image data. This image data (including positioning parameters, the same below) can be used as raw data to participate in special missions and produce various types of remote sensing mapping products, such as small- and medium-scale orthophotos, scene change detection, and ground object interpretation. It can also serve as a benchmark for other non-stereo observation remote sensing satellites, helping to improve their ground positioning capabilities. For example, stereo mapping satellite data can be jointly processed with data from non-stereo observation remote sensing satellites such as Gaofen-1 and Gaofen-2 to improve the ground positioning accuracy of non-stereo observation remote sensing satellite data. Therefore, from an application perspective, stereo mapping satellite image data actually represents the overall application capabilities of the satellite.

[0004] Although the life cycle of stereo mapping satellites is limited, satellite image data can be stored and applied for a long time. If the satellite's ground positioning capability is extracted from the stereo mapping satellite image data and this positioning capability is quantitatively expressed and stored to achieve the "data instead of satellite" function, the application performance of the satellite can continue to be utilized through satellite image data after the life cycle of the stereo mapping satellite ends. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for constructing an image control point library based on stereo mapping satellite images, so as to optimize the positioning parameters of non-stereo observation remote sensing satellite images.

[0006] In order to solve the above technical problems, the present invention provides a method for constructing an image control point library based on stereo mapping satellite images, which method comprises the following steps:

[0007] 1) Obtain multi-temporal stereo mapping satellite images and their positioning parameter data for a certain area, obtain the same-name feature points from the multi-temporal stereo mapping satellite images, and obtain the image square correction parameters and the overall image positioning accuracy of the satellite images through regional block adjustment processing;

[0008] 2) Gridding the areas corresponding to the feature points with the same name, matching and obtaining the elevation coordinates of the plane grid points, and obtaining the image points with the same name on multiple overlapping images;

[0009] 3) The obtained feature points and image points with the same name are used as candidate points, and the local image centered on the candidate points is selected as the control image. The image coordinates, ground coordinates, ground positioning accuracy and control image of the candidate points are used as a description file to form the image control points;

[0010] 4) The image control points, the stereo mapping satellite image obtained in step 1), and the corresponding image correction parameters are stored in a set database to form an image control point database.

[0011] The present invention extracts feature points and performs block adjustment processing on acquired stereo mapping satellite images to obtain image square correction parameters of satellite images and the overall ground positioning accuracy of the images, thereby improving the overall ground positioning accuracy of the image data. Feature points are then extracted from the stereo mapping satellite images, and image control points generated by the intersection of the feature points are combined with image control points generated by matching and intersecting regular ground grid points to obtain an image control point database with strong positioning capabilities. Based on this image control point database, a quantitative analysis of the ground positioning capabilities of the stereo mapping satellite can be performed. When non-stereo observation remote sensing satellite data is available, image control points are retrieved from the image control point database and image control point matching is performed to achieve optimization of non-stereo observation remote sensing satellite image positioning parameters with the assistance of control information.

[0012] Furthermore, when new stereo mapping satellite image data is added, the geographic coordinate range corresponding to the image data is calculated and a retrieval area is constructed. The constructed retrieval area is used to search in the image control point database. If an image control point exists in the retrieval area, the image control point is updated according to the positioning accuracy of the new stereo mapping satellite image data; if not, the new stereo mapping satellite image data is directly used to form a new image control point and added to the image control point database.

[0013] By comparing the image control points in the image control point database, if there are no image control points in the area where the new stereo mapping satellite image data is located, image control points will be constructed based on the new stereo mapping satellite image data, and the constructed image control points will be added to the image control point database to expand the area covered by the image control point database; at the same time, for the situation where image control points already exist, the present invention improves the ground positioning accuracy of the existing image control points in the image control point database through positioning accuracy comparison.

[0014] Furthermore, when there are image control points in the search area, the update process is as follows:

[0015] a. Perform block adjustment on the new stereo mapping satellite image data, the retrieved image control points, the original image data of the control points, and their positioning parameters to obtain the corresponding image correction parameters;

[0016] B. according to the block adjustment processing result, if new stereo mapping satellite image data positioning accuracy is lower than the positioning accuracy of the image control point that retrieves, then abandon the image control point updating process in the retrieval area, otherwise, using the characteristic point in the new stereo mapping satellite image data regional network control process as new image control point;

[0017] c. Rematch and calculate the elevation coordinates of the original image control points in the retrieval area. On the basis of retaining the original image control points, modify the ground coordinates, original image data and its positioning parameters in the image control point description file, and update the image control points in the retrieval area in the image control point database.

[0018] By comparing the latest stereo mapping satellite image data with the positioning accuracy of existing image control points to update the image control point library, the ground positioning accuracy of the image control point library can be further improved.

[0019] Furthermore, in step c, the data of the original image control points in the search area are deleted during the update.

[0020] By deleting existing image control points with insufficient positioning accuracy in the image control point library during updating, unnecessary image control points in the image control point library can be reduced, making it easier to manage the image control point library.

[0021] Furthermore, the search area is determined based on the geographic coordinate range of the new stereo mapping satellite image data, and a minimum circumscribed circle of the geographic coordinate range is established. The minimum circumscribed circle is expanded according to a set ratio, and the expanded area is the search area of the new stereo mapping satellite image data.

[0022] By expanding the search area, the probability of retrieving image control points in the search area can be increased, avoiding the problem of not being able to retrieve image control points that should be retrieved.

[0023] Further, described step 1) when carrying out block adjustment processing, if there is ground control point, set up the image space affine transformation parameter error equation that ground control point is corresponding, and combine the image space affine transformation parameter error equation that itself and the characteristic point of the same name are corresponding to carry out block adjustment processing; If there is no ground control point, directly utilize the characteristic point of the same name image space affine transformation parameter error equation to carry out block adjustment processing.

[0024] Furthermore, in order to quickly and accurately determine the elevation coordinates of the plane grid points, the multi-view plumb line trajectory method is used in step 2) to obtain the elevation coordinates of the plane grid points.

[0025] Furthermore, in order to ensure that the constructed control image has minimal deformation, in step 3), when constructing the control image, a certain size of image data is intercepted from the original stereo mapping satellite image closest to the plumb line direction with the candidate point as the center, as the control image of the candidate point.

[0026] Furthermore, in order to facilitate the unified management of the image control point database, the step 4) further includes converting the ground coordinates of the image control points in the image control point database into geographic coordinates. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a method for constructing an image control point library based on stereo mapping satellite images according to the present invention;

[0028] Figure 2 This is a flowchart of obtaining image control points based on image feature points and ground grid points in the present invention;

[0029] Figure 3 The present invention is a flow chart for optimizing positioning parameters of non-stereoscopic remote sensing satellite images using an image control point library. DETAILED DESCRIPTION

[0030] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0031] The present invention first performs a large-scale regional network adjustment on the stereo mapping satellite image data to improve the overall ground positioning accuracy of the image data; then extracts feature points from the image, and constructs an image control point database by combining the image control points generated by the intersection of feature points with the image control points generated by matching and intersecting the ground regular grid points. The implementation process is as follows: Figure 1 As shown, the specific implementation process is as follows.

[0032] 1. Obtain multi-temporal stereo mapping satellite images and their positioning parameter data for a specific area, match and obtain feature points with the same name from the image data, and perform joint block adjustment of the satellite images based on an image positioning model with additional image square correction parameters to obtain the image square correction parameters and the overall image positioning accuracy relative to the ground.

[0033] 1-A: Based on the stereo mapping satellite image positioning parameters (rational function model parameters), a satellite image rational function positioning model is constructed. By calculating the ground coordinates of the four corner points of each image, the overlapping relationship between the given images is analyzed to determine the overlapping range between the images.

[0034] 1-B: Within the overlapping range of the images, a scale-invariant feature transform (SIFT) feature extraction and matching method is used to obtain SIFT feature points with the same name. As another implementation method, other existing methods for extracting feature points with the same name may also be used.

[0035] 1-C: Select the general model of affine transformation and use the initial values without affine transformation to correct the image-space coordinates of the satellite image rational function positioning model. That is, combine the general model of affine transformation and the rational function model to calculate the initial ground coordinates of the SIFT feature points with the same name, and establish the image-space affine transformation parameter error equation of the SIFT feature points with the same name.

[0036] 1-D: When there are ground control points, the image space affine transformation parameter error equations corresponding to the ground control points (including plane control points, elevation control points, and horizontal-height control points) are established and combined with the image space affine transformation parameter error equations corresponding to the SIFT feature points with the same name to perform regional block adjustment processing to obtain the overall ground positioning accuracy of the adjusted satellite image, and obtain the optimized image space affine transformation parameters of each satellite image and the ground coordinates of the SIFT feature points with the same name; when there are no ground control points or there are few ground control points, the image space affine transformation parameter error equations of the feature points with the same name can be directly used for regional block adjustment processing.

[0037] 2. Calculate the ground area of the image data, divide the area into regular grid points, and use the "Multi-view Vertical Line Locus" (MVLL) model to match and obtain the elevation coordinates of the grid points.

[0038] 2-A: Set the spacing of regular plane grid points based on the imaging resolution of the stereo mapping satellite imagery. Based on the number of rational function model parameters in the satellite imagery and their solution requirements, after the grid spacing is set, the number of plane grid points on each satellite imagery should not be less than the number of rational function model parameters.

[0039] 2-B: For each plane grid point, given its plane coordinates, the MVLL model is used to match the elevation coordinates of that plane grid point and its corresponding image points in multiple overlapping images. The MVLL model is based on a given plane grid point elevation. The corresponding image point coordinates are then calculated using the plane grid point coordinates and the given elevation. These coordinates are then matched against the actual image point coordinates. If a match is unsuccessful, the elevation of the plane grid point is adjusted until the calculated image point coordinates successfully match the actual image point coordinates.

[0040] 3. Select the feature points matched in step 1 and the image points with the same name corresponding to the plane grid points obtained in step 2 as candidate points; select the local image centered on the candidate point as the control image, and write the image coordinates, ground coordinates and coordinate system, positioning accuracy, control image and original data information of the candidate point into a description file to form the image control point. The process is as follows: Figure 2 As shown, the specific process is as follows.

[0041] 3-A: In order to ensure that the control image is slightly deformed, based on the known imaging angle of the stereo mapping satellite, the present invention selects the original image closest to the plumb line direction, and with the alternative point as the center, intercepts a certain size of image data (usually the image length and width are both about 1000 pixels) from the original image closest to the plumb line direction as the control image of the alternative point.

[0042] 3-B: Perform local image capture based on the image point position of the candidate point on the original image, try to ensure that the candidate point is in the center of the local image, and determine the image point position of the candidate point on the local image.

[0043] 3-C: The overall ground positioning accuracy after regional adjustment of the satellite image is used as the positioning accuracy of the candidate point, and the ground coordinates of the candidate point and its corresponding coordinate system, control image name, positioning accuracy, and image point location information on the control image are written into an XML file to form an image control point description file.

[0044] 3-D: Combine the candidate point control images and their description files as components of image control points to form image control points.

[0045] 4. The image control points, the original image data of the stereo mapping satellite and its corrected positioning parameters are stored in the database at the same time to form an image control point database, thereby achieving a quantitative expression of the positioning capability of the stereo mapping satellite.

[0046] 4-A: To facilitate the subsequent unified management of global image control points, convert the ground coordinates of the image control points to geographic coordinates (latitude and longitude) based on the coordinate system of the image control points. The control point data is then stored and retrieved based on geographic coordinates. If the coordinate system of the image control points is already the geographic coordinate system, no conversion is required. This approach provides favorable support for subsequent retrieval based on geographic location and facilitates the unified visualization of image control points.

[0047] 4-B: In the image control point database, data field editing is used to implement functions such as image control point data modification, addition, and deletion to meet the needs of image control point update and editing.

[0048] 5. When new stereo mapping satellite image data is added, the geographic coordinate range corresponding to the image data is calculated and a search area is constructed. The constructed search area is used to search in the image control point database. If an image control point exists in the search area, it is necessary to update the image control point according to the positioning accuracy of the new stereo mapping satellite image data; if not, the new stereo mapping satellite image data is directly used to form a new image control point and added to the image control point database.

[0049] 5-A: Calculate the minimum circumscribed circle of the geographic coordinate range of the new stereo mapping satellite image data and obtain the center and radius of the minimum circumscribed circle; to avoid failing to find the image control point within the minimum circumscribed circle, the search area needs to be expanded. The present invention maintains the center position unchanged and constructs a new circular search range with twice the radius as the search area of the image control point database. As another embodiment,

[0050] 5-B: When there are no image control points in the search area, directly use steps 1 to 4 to process the newly added satellite image data, generate new image control points and store them in the database for management.

[0051] 5-C: When there are image control points in the search area, image control point update processing is performed, which specifically includes the following steps.

[0052] 5-C-1: Perform the block adjustment processing corresponding to step 1 on the newly added satellite image data, the retrieved image control points, the original image data of the control points, and their positioning parameters to obtain the new image square correction parameters of the satellite image.

[0053] 5-C-2: Analyze the results of the regional block adjustment processing. If the positioning accuracy of the newly added satellite image data is lower than the average positioning accuracy of the retrieval control points, abandon the image control point update process in the retrieval area and end step 5-C; otherwise, continue with the subsequent processing steps.

[0054] 5-C-3: For the newly added satellite image data, add the feature points in the regional block adjustment process in step 5-C-1 as new image control points and manage the image control points in the database.

[0055] 5-C-4: Use the MVLL model to re-match and recalculate the elevation coordinates of the original image control points in the search area. While retaining the original version of the image control points, modify the ground coordinates, original image data, and positioning parameters in the image control point description file to update the image control point database.

[0056] As another embodiment, during the update, the original image control point data in the search area in the image control point library may be deleted.

[0057] Through the above process, it is possible to use stereo mapping remote sensing satellite images to build an image control point database. Based on the image control point database, it is possible to quantitatively analyze the ground positioning capability of stereo mapping remote sensing satellite images and carry out special applications to improve the ground positioning accuracy of non-stereo observation remote sensing satellite data, such as Figure 3 shown.

[0058] Given a given regional scope, control point data is retrieved from the image control point database. Based on the control point ground positioning accuracy information, the stereo mapping satellite's ground positioning accuracy within the given regional scope is analyzed to achieve regional expression of the stereo mapping satellite's ground positioning capability. When non-stereo observation remote sensing satellite data is available, the geographic coordinate range corresponding to this data is calculated. When the positioning accuracy of the image data is lower than the positioning accuracy of the stereo mapping satellite within the regional scope, image control points are retrieved from the image control point database and matched to achieve control information-assisted optimization of non-stereo observation remote sensing satellite image positioning parameters. Based on the optimization results of the non-stereo observation remote sensing satellite image positioning parameters, the parameters of the rational function model of the non-stereo observation remote sensing satellite image are regenerated to improve the ground positioning accuracy of the non-stereo observation remote sensing satellite data.

Claims

1. A method for constructing an image control point library based on stereo mapping satellite images, characterized in that: The method comprises the following steps: 1) Obtain multi-temporal stereo mapping satellite images and their positioning parameter data for a certain area, obtain the same-name feature points from the multi-temporal stereo mapping satellite images, and obtain the image square correction parameters and the overall image positioning accuracy of the satellite images through regional block adjustment processing; 2) Gridding the area, matching and obtaining the elevation coordinates of the plane grid points, and obtaining the same-name image points of the plane grid points on multiple overlapping images; 3) The obtained feature points and image points with the same name are used as candidate points, and a local image centered on the candidate points is selected as a control image. The image coordinates, ground coordinates, ground positioning accuracy, and control image of the candidate points are used as a description file to form image control points. The description file is used to update the image control points. 4) The image control points, the stereo mapping satellite images obtained in step 1), and the corresponding image correction parameters are stored in a set database to form an image control point database; when non-stereo observation remote sensing satellite image data is available, and the positioning accuracy of the image data is lower than the positioning accuracy of the stereo mapping satellite within the area, the image control points are retrieved from the image control point database, and image control point matching is performed to achieve optimization of the non-stereo observation remote sensing satellite image positioning parameters with the assistance of control information.

2. The method for constructing an image control point library based on stereo mapping satellite images according to claim 1, characterized in that: When new stereo mapping satellite image data is added, the geographic coordinate range corresponding to the image data is calculated and a search area is constructed. The constructed search area is used to search in the image control point database. If an image control point exists in the search area, the image control point is updated according to the positioning accuracy of the new stereo mapping satellite image data; if not, the new stereo mapping satellite image data is directly used to form a new image control point and added to the image control point database.

3. The method for constructing an image control point library based on stereo mapping satellite images according to claim 2, characterized in that: When there are image control points in the search area, the update process is as follows: a. Perform block adjustment on the new stereo mapping satellite image data, the retrieved image control points, the original image data of the control points, and their positioning parameters to obtain the corresponding image correction parameters; B. according to the block adjustment processing result, if new stereo mapping satellite image data positioning accuracy is lower than the positioning accuracy of the image control point that retrieves, then abandon the image control point updating process in the retrieval area, otherwise, using the characteristic point in the new stereo mapping satellite image data regional network control process as new image control point; c. Rematch and calculate the elevation coordinates of the original image control points in the retrieval area. On the basis of retaining the original image control points, modify the ground coordinates, original image data and its positioning parameters in the image control point description file, and update the image control points in the retrieval area in the image control point database.

4. The method for constructing an image control point library based on stereo mapping satellite images according to claim 3, characterized in that: In step c, the data of the original image control points in the search area are deleted during the update.

5. The method for constructing an image control point library based on stereo mapping satellite images according to claim 2, characterized in that: The search area is determined based on the geographic coordinate range of the new stereo mapping satellite image data. The minimum circumscribed circle of the geographic coordinate range is established, and the minimum circumscribed circle is expanded according to a set ratio. The expanded area is the search area of the new stereo mapping satellite image data.

6. The method for constructing an image control point library based on stereo mapping satellite images according to claim 1 or 2, characterized in that: Described step 1) when carrying out block adjustment processing, if there is ground control point, set up the image square affine transformation parameter error equation that ground control point is corresponding, and itself and the image square affine transformation parameter error equation corresponding to the characteristic point of the same name are united to carry out block adjustment processing; If there is no ground control point, directly utilize the characteristic point of the same name image square affine transformation parameter error equation to carry out block adjustment processing.

7. The method for constructing an image control point library based on stereo mapping satellite images according to claim 1 or 2, characterized in that: In the step 2), the elevation coordinates of the plane grid points are obtained by using a multi-view plumb line trajectory method.

8. The method for constructing an image control point library based on stereo mapping satellite images according to claim 1 or 2, characterized in that: In step 3), when constructing the control image, with the candidate point as the center, image data of a certain size is intercepted from the original stereo mapping satellite image closest to the plumb line direction as the control image of the candidate point.

9. The method for constructing an image control point library based on stereo mapping satellite images according to claim 1 or 2, characterized in that: The step 4) further includes converting the ground coordinates of the image control points in the image control point database into geographic coordinates.

Images