Method, device and equipment for quality inspection of relative geometric accuracy between satellite L1 images
By obtaining the digital elevation model data of satellite L1-level images, performing connection point matching and error calculations, and generating a relative geometric accuracy quality inspection report, solving the problem of the inability to automate and quantify the relative geometric accuracy of quality inspection in the existing technology, achieving high efficiency image quality inspection.
Patent Information
- Application Number
- CN202411026736.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-07-30
AI Technical Summary
The existing satellite L1-level image quality evaluation technology cannot automatically and quantitatively conduct quality inspection of the relative geometric accuracy between satellite original remote sensing images.
By obtaining the digital elevation model data of the L1-level image set of satellites to be inspected, connecting point matching and coarse deviation elimination based on the regional network adjustment method, the pixel coordinate error and relative accuracy error are determined using the rational polynomial coefficient model to generate relative geometry Accuracy quality inspection report.
It realizes automatic and quantitative quality inspection of the relative geometric accuracy between satellite L1-level images, and improves the accuracy and efficiency of quality inspection.
Smart Images

Figure CN118552544B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of automatic precision quality inspection of satellite L1-level images, and in particular to a relative geometric precision quality inspection method, device and equipment between satellite L1-level images. Background Art
[0002] With the development of the economy, optical remote sensing satellite L1-level images are playing a more important role in more and more fields. For many optical satellite L1-level image application fields with high requirements for geometric accuracy, such as surveying and mapping, satellite L1-level images that have undergone reliable and comprehensive qualitative and quantitative geometric quality evaluation can provide a good foundation for subsequent applications. In addition, the geometric quality of satellite L1-level images can also reflect the current status of the satellite and its sensors to a certain extent. The normalized geometric quality evaluation of the same satellite can also provide the possibility of discovering anomalies on the satellite in the first place. However, the existing satellite L1-level image quality evaluation technology has the problem of being unable to automatically and quantitatively inspect the relative geometric accuracy between L1-level satellite original remote sensing images. Summary of the invention
[0003] In view of this, the purpose of the present invention is to provide a method, device and equipment for quality inspection of relative geometric accuracy between satellite L1 level images, which can automatically and quantitatively inspect the relative geometric accuracy between satellite L1 level images.
[0004] In a first aspect, an embodiment of the present invention provides a method for quality inspection of relative geometric accuracy between satellite L1 level images, comprising:
[0005] Obtain a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected;
[0006] The L1-level satellite image set to be inspected is matched with connecting points and gross error points are eliminated based on the regional block adjustment method to obtain the target connecting points contained in the L1-level satellite image set to be inspected;
[0007] For any first L1-level satellite image to be inspected and at least one second L1-level satellite image to be inspected that overlaps with the first L1-level satellite image in the set of L1-level satellite images to be inspected, based on the rational polynomial coefficient model and the digital elevation model, determine the pixel coordinate error between the target connection points respectively included in the first L1-level satellite image to be inspected and the second L1-level satellite image to be inspected, and determine the relative accuracy mean error between the first L1-level satellite image to be inspected and the second L1-level satellite image to be inspected;
[0008] Based on the pixel coordinate error and relative accuracy mean error, a relative geometric accuracy quality inspection report corresponding to the satellite L1 image set to be inspected is generated.
[0009] In one embodiment, based on the digital elevation model, determining the pixel coordinate error between the target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, and determining the relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, comprises:
[0010] The target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected are used as the first target connection point and the second target connection point respectively;
[0011] Determine the estimated pixel coordinates of the first target connection point in the second satellite L1 level image to be inspected based on the rational polynomial coefficient model, the digital elevation model and the first actual pixel coordinates of the first target connection point in the first satellite L1 level image to be inspected;
[0012] Determine the pixel coordinate error according to the estimated pixel coordinate and the second actual pixel coordinate of the second target connection point matching the first target connection point in the second L1 level image of the satellite to be inspected;
[0013] The relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected is determined according to the pixel coordinate error.
[0014] In one embodiment, the relative geometric accuracy quality inspection report includes a scene-by-scene quality inspection report and / or an overall quality inspection report;
[0015] The sub-scene quality inspection report is a relative geometric accuracy quality inspection report between the first satellite L1 level image to be inspected and all the second satellite L1 level images to be inspected;
[0016] The overall quality inspection report is a quality inspection report on the relative geometric accuracy between all satellite L1 images to be inspected.
[0017] In one implementation, the relative geometric accuracy quality inspection report includes a scene-by-scene quality inspection report, and based on the pixel coordinate error and the relative accuracy mean error, a relative geometric accuracy quality inspection report corresponding to the L1-level image set of the satellite to be inspected is generated, including:
[0018] The target connection points contained in the first satellite L1 level image to be inspected and all the second satellite L1 level images to be inspected are taken as the first inspection points;
[0019] Based on the pixel coordinate error corresponding to the first checkpoint, statistical processing is performed on the first checkpoint to obtain a checkpoint statistical result; wherein the checkpoint statistical result includes one or more of the number of first points whose pixel coordinate error is less than a first mean error threshold, the number of second points whose pixel coordinate error is less than a second mean error threshold, the number of third points whose pixel coordinate error is greater than the second mean error threshold, the maximum pixel coordinate error, and the checkpoint gross error rate;
[0020] Generate image quality inspection results based on the gross error rate and relative accuracy error of the inspection points;
[0021] One or more of the pixel coordinate error corresponding to each first checkpoint, the total number of first checkpoints, checkpoint statistical results, and image quality inspection results are used as a scene-by-scene quality inspection report.
[0022] In one embodiment, generating an image quality inspection result according to the gross error rate of the checkpoint and the relative precision error includes:
[0023] If the gross error rate of the checkpoint is less than the gross error rate threshold, and the relative accuracy mean error is less than or equal to the first mean error threshold, the image quality inspection result is determined to be qualified.
[0024] In one embodiment, the method further comprises:
[0025] Based on the comparison results between the pixel coordinate error and the first mean error threshold and the second mean error threshold, different marks are assigned to the first checkpoint to generate a checkpoint thumbnail between the first satellite L1 level image to be inspected and any second satellite L1 level image to be inspected based on the mark.
[0026] In one embodiment, the relative geometric accuracy quality inspection report includes an overall quality inspection report, and the method further includes:
[0027] All target connection points included in the satellite L1 level images to be inspected are used as the second inspection points;
[0028] And, statistical processing is performed on the image quality inspection report to obtain report statistical results; wherein the report statistical results include the number of scenes that pass the quality inspection, the number of scenes that fail the quality inspection, and the quality inspection pass rate;
[0029] The pixel coordinate error corresponding to each second checkpoint, the total number of second checkpoints, the relative accuracy mean error corresponding to each satellite L1 image to be inspected, the image quality inspection report, and one or more of the report statistical results are used as the overall quality inspection report.
[0030] In a second aspect, an embodiment of the present invention further provides a relative geometric accuracy quality inspection device between satellite L1 level images, comprising:
[0031] The data acquisition module is used to acquire a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected;
[0032] The connection point determination module is used to match the connection points of the L1-level satellite image set to be inspected and eliminate the gross error points based on the regional block adjustment method to obtain the target connection points contained in the L1-level satellite image set to be inspected;
[0033] The error calculation module is used to determine, for any first L1-level satellite image to be inspected and at least one second L1-level satellite image to be inspected that overlaps with the first L1-level satellite image in the set of L1-level satellite images to be inspected, pixel coordinate errors between target connection points respectively included in the first L1-level satellite image to be inspected and the second L1-level satellite image to be inspected, and to determine a mean error in relative accuracy between the first L1-level satellite image to be inspected and the second L1-level satellite image to be inspected based on a rational polynomial coefficient model and a digital elevation model;
[0034] The quality inspection module is used to generate a relative geometric accuracy quality inspection report corresponding to the satellite L1 image set to be inspected based on the pixel coordinate error and the relative accuracy mean error.
[0035] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising a processor and a memory, wherein the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement any one of the methods provided in the first aspect.
[0036] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement any one of the methods provided in the first aspect.
[0037] The embodiment of the present invention provides a relative geometric accuracy quality inspection method, device and equipment between satellite L1 level images. First, a set of satellite L1 level images to be inspected and digital elevation model data covering each satellite L1 level image to be inspected in the set of satellite L1 level images to be inspected are obtained; then, the set of satellite L1 level images to be inspected is matched with connection points and gross error points are eliminated based on a regional block adjustment method to obtain target connection points included in the set of satellite L1 level images to be inspected; for any first satellite L1 level image to be inspected in the set of satellite L1 level images to be inspected and at least one second satellite L1 level image to be inspected that overlaps with the first satellite L1 level image to be inspected in the set of satellite L1 level images to be inspected, based on a rational polynomial coefficient model and a digital elevation model, the pixel coordinate error between the target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected is determined, and the relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected is determined; finally, based on the pixel coordinate error and the relative accuracy mean error, a relative geometric accuracy quality inspection report corresponding to the set of satellite L1 level images to be inspected is generated. After automatically matching the connection points between the satellite L1 images to be inspected, the above method automatically eliminates erroneous connection points in combination with the digital elevation model data of the region, and on this basis, determines the pixel coordinate errors between the connection points and the relative accuracy mean error between the images, and then automatically generates a relative geometric accuracy quality inspection report. The embodiment of the present invention realizes automated and quantitative quality inspection of the relative geometric accuracy between satellite L1 images.
[0038] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.
[0039] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0041] Figure 1 A schematic diagram of a flow chart of a relative geometric accuracy quality inspection method between satellite L1-level images provided in an embodiment of the present invention;
[0042] Figure 2A schematic diagram of a scene-by-scene quality inspection report provided by an embodiment of the present invention;
[0043] Figure 3 A schematic diagram of another scene-by-scene quality inspection report provided by an embodiment of the present invention;
[0044] Figure 4 A schematic diagram of an overall quality inspection report provided by an embodiment of the present invention
[0045] Figure 5 An overall flow chart of a relative geometric accuracy quality inspection method between satellite L1 level images provided by an embodiment of the present invention;
[0046] Figure 6 A schematic diagram of the structure of a relative geometric accuracy quality inspection device between satellite L1-level images provided by an embodiment of the present invention;
[0047] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described in combination with the embodiments below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0049] At present, relevant technologies provide the following solutions for automated precision quality inspection of remote sensing images:
[0050] (1) A method, device, equipment and storage medium for evaluating the geometric quality of optical satellite images: The invention discloses a method, device, equipment and storage medium for evaluating the geometric quality of optical satellite images. The method comprises: obtaining an image to be evaluated and a parameter file, an image metafile and an image file header for geometric quality evaluation generated based on the image to be evaluated; obtaining the path, band and evaluation mode of the band of the image to be evaluated from the parameter file for geometric quality evaluation; reading the parameters of the image to be evaluated from the image metafile or the image file header; selecting the parameters of control points that meet preset conditions and the image blocks corresponding to the control points from a preset control point library according to the parameters of the image to be evaluated; matching the image to be evaluated and the image blocks corresponding to the control points to obtain a matching result; and calculating the geometric quality evaluation index of the image to be evaluated according to the matching result.
[0051] (2) A method for automatically checking the geometric accuracy of remote sensing satellite images: A method for automatically checking the geometric accuracy of remote sensing satellite images is provided, the method comprising: inputting an image to be tested, and automatically obtaining a corresponding reference image based on the image to be tested; obtaining a set of same-name point pairs based on the image to be tested and the reference image by a method based on phase congruency image matching, wherein the set of same-name point pairs includes at least twenty same-name point pairs, which are key points evenly distributed in the image to be tested and the corresponding matching points of the key points in the reference image; calculating the positioning error of the same-name point pairs based on the positions of the key points and matching points in the set of same-name point pairs and drawing an error direction map, thereby realizing automatic inspection of the geometric accuracy of remote sensing images. This method of obtaining a set of same-name point pairs by means of phase congruency matching reduces time and material consumption compared to traditional manual point puncturing; it improves matching accuracy compared to the classic SIFT matching method, and effectively improves the demand for geometric accuracy inspection of large-scale remote sensing satellite image products.
[0052] (3) Method, device and electronic equipment for quality inspection of stereo satellite images: This invention provides a method, device and electronic equipment for quality inspection of stereo satellite images, which relates to the field of photogrammetry and remote sensing technology. The method includes: obtaining stereo image pairs in the same survey area of stereo satellite images, and constructing an image pyramid corresponding to the stereo image pairs; performing step-by-step matching in the image pyramid based on a preset search radius to determine the same-name pixel points in the stereo image pairs corresponding to the reference image; performing spatial forward intersection processing on the same-name pixel points to obtain the object coordinates corresponding to the stereo image pairs; determining the accuracy data between stereo satellite images based on the object coordinates, and performing accuracy quality inspection based on the accuracy data. This application can automatically remove erroneous connection points, and introduce multi-degree control points to synchronously check absolute accuracy.
[0053] Existing remote sensing image automated accuracy quality inspection technologies, such as (1) and (2), all automatically match the same-name points between the image to be inspected and the reference image, and then use the same-name points as the basic unit to calculate the error of each point, and then automatically calculate the absolute geometric accuracy error of the image to be inspected, and automatically form an absolute accuracy quality inspection report. (3) proposed an automated quality inspection method for the relative accuracy and absolute accuracy of stereo satellite L1-level images, but it is not suitable for the automated quality inspection of the relative geometric accuracy between non-stereo satellite original images. In summary, none of the current methods can automatically and quantitatively solve the problem of quality inspection of the relative geometric accuracy between L1-level satellite original remote sensing images.
[0054] Based on this, the present invention provides a method, device and equipment for quality inspection of relative geometric accuracy between satellite L1 level images, which can automatically and quantitatively inspect the relative geometric accuracy between satellite L1 level images.
[0055] To facilitate understanding of this embodiment, firstly, a method for quality inspection of relative geometric accuracy between satellite L1-level images disclosed in an embodiment of the present invention is described in detail. Figure 1 The flowchart of a relative geometric accuracy quality inspection method between satellite L1 level images is shown, and the method mainly includes the following steps S102 to S108:
[0056] Step S102 , obtaining a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected.
[0057] The set of L1-level satellite images to be inspected includes at least two L1-level satellite images to be inspected that have overlapping relationships and corresponding RPC (Rational Polynomial Coefficient) files, and the digital elevation model data needs to cover all L1-level satellite images to be inspected.
[0058] Step S104, performing tie point matching on the L1-level satellite image set to be inspected and eliminating gross error points based on a block adjustment method to obtain target tie points included in the L1-level satellite image set to be inspected.
[0059] Among them, the connection points refer to the same-name points with the same ground features between the L1-level satellite images to be inspected, which are the basis of quantitative geometric accuracy inspection. In one example, the connection points of the L1-level satellite images to be inspected that have overlapping relationships in the L1-level satellite images to be inspected can be matched using the fully automatic matching technology to extract the initial connection points from the L1-level satellite images to be inspected, and then the initial connection points are removed by the regional block adjustment algorithm to ensure that the matching target connection points are correct as much as possible to ensure the accuracy of the final relative accuracy inspection.
[0060] Step S106, for any first satellite L1 level image to be inspected in the set of satellite L1 level images to be inspected and at least one second satellite L1 level image to be inspected that overlaps with the first satellite L1 level image to be inspected, determine the pixel coordinate error between the target connection points respectively contained in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected based on the rational polynomial coefficient model and the digital elevation model, and determine the relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected.
[0061] In one example, for at least two L1-level satellite images to be inspected that have an overlapping relationship, one of the L1-level satellite images to be inspected can be used as the first L1-level satellite image to be inspected, and the other L1-level satellite images to be inspected can be used as the second L1-level satellite images to be inspected. The target connection points included in the first L1-level satellite image to be inspected are recorded as the first target connection points, and the target connection points included in the second L1-level satellite image to be inspected are recorded as the second target connection points. Each first target connection point is traversed, and firstly, the estimated pixel coordinates of the first target connection point in the second L1-level satellite image to be inspected are determined according to the rational polynomial coefficient model, the digital elevation model and the first actual pixel coordinates of the first target connection point in the first L1-level satellite image to be inspected; then, the pixel coordinate errors between the target connection points are determined in combination with the second actual pixel coordinates of the second target connection point matched with the first target connection point in the second L1-level satellite image to be inspected; finally, the relative accuracy mean error between the first L1-level satellite image to be inspected and the second L1-level satellite image to be inspected is calculated based on each pixel coordinate error.
[0062] Step S108, generating a relative geometric accuracy quality inspection report corresponding to the L1 level satellite image set to be inspected based on the pixel coordinate error and the relative accuracy mean error.
[0063] The relative geometric accuracy quality inspection report includes a scene-by-scene quality inspection report and / or an overall quality inspection report. The scene-by-scene quality inspection report is a relative geometric accuracy quality inspection report between the first L1-level image of the satellite to be inspected and all the second L1-level images of the satellite to be inspected; the overall quality inspection report is a relative geometric accuracy quality inspection report between all L1-level images of the satellite to be inspected. In one example, the pixel coordinate error and the relative accuracy mean error can be counted separately to obtain the scene-by-scene quality inspection report, and the scene-by-scene quality inspection report can be further counted to obtain the overall quality inspection report.
[0064] The above-mentioned relative geometric accuracy quality inspection method between satellite L1 level images provided by the embodiment of the present invention automatically matches the connection points between the satellite L1 level images to be inspected, and then automatically eliminates the erroneous connection points in combination with the digital elevation model data of the area. On this basis, the pixel coordinate errors between the connection points and the relative accuracy mean error between the images are determined respectively, and then a relative geometric accuracy quality inspection report is automatically generated. The embodiment of the present invention realizes automated and quantitative quality inspection of the relative geometric accuracy between satellite L1 level images.
[0065] For ease of understanding, an embodiment of the present invention provides a specific implementation method of a relative geometric accuracy quality inspection method between satellite L1 level images.
[0066] For the aforementioned step S102, the embodiment of the present invention provides an implementation method of obtaining a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected, including:
[0067] The L1-level satellite images to be inspected, i.e., the original satellite images to be inspected, have certain geometric positioning errors due to in-orbit geometric calibration errors and other reasons, and it is necessary to automatically inspect the relative geometric accuracy between the original satellite images; in addition, for the original satellite images that have been triangulated in the air, it is also necessary to automatically quantify the relative geometric accuracy of the original satellite images after the triangulation in the air. The original satellite images in the embodiments of the present invention refer to non-stereoscopic original satellite images with RPC files, such as original satellite images of Gaofen-1 and Gaofen-2.
[0068] Specifically, the data that needs to be prepared are as follows: (a) at least two overlapping satellite original images and the corresponding RPC files; (b) digital elevation model (DEM) data that can cover all satellite original images to be inspected.
[0069] For the aforementioned step S104, the embodiment of the present invention provides an implementation method of matching connection points and eliminating gross difference points on the set of L1 level satellite images to be inspected to obtain target connection points included in the set of L1 level satellite images to be inspected, including:
[0070] (1) Tie point matching: Tie points refer to the same-name points with the same ground features between images, which are the basis for quantitative geometric accuracy quality inspection. Automation mainly relies on full-automatic matching technology. The full-automatic matching algorithm used in the embodiment of the present invention is as follows:
[0071] (1.1) The feature point extraction algorithm is the Harris operator: the point with the largest grayscale change (i.e. gradient) in the local window is selected as the feature point in a block-by-block manner, so as to find the same-name points on adjacent images, ensuring the uniformity and characteristic of the feature point selection;
[0072] (1.2) The automatic matching strategy adopted is the pyramid level-by-level matching strategy: The pyramid level-by-level matching strategy refers to starting the matching from a certain level of the image pyramid (usually set to a level smaller than 1024), and the matching result of each point is passed to the next level to narrow the search range until the original level of the image. The purpose is to speed up the matching efficiency and automatically match the image with a large initial positioning difference with a smaller search radius. For details, please refer to the patent application with application number CN202110934043.7 and application name "A Matching Method and Device for Image Pyramid", and the embodiments of the present invention will not be repeated here.
[0073] (1.3) The main matching measure used is the correlation coefficient method.
[0074] After each scene image is matched, a connection point file is generated, which records the points with the same name between the scene image and all adjacent images with overlapping relationships. The format of the connection points is shown in Table 1 below:
[0075] Table 1
[0076]
[0077] Where, ID: the number of the connection point; : Column coordinates of the L1-level satellite image to be inspected; : Row coordinates of the L1-level satellite image to be inspected; : The image path of the nth scene with overlapping area images; : The column coordinates of the image with overlapping area of the nth scene; : The row coordinates of the image with overlapping area in the nth scene; : The geographic X coordinate of the intersection of the connection points, obtained by the regional block adjustment solution; The geographic Y coordinate of the intersection of the connecting points is obtained by the regional network adjustment solution; Z: elevation value, which is read from the DEM data based on the X and Y of the intersection of the connecting points.
[0078] (2) Elimination of gross error points, that is, elimination of erroneous points by regional block adjustment:
[0079] The purpose of eliminating gross error points is to automatically find and eliminate the points with large errors obtained by automatic matching through some technical means. The purpose is to ensure that the matched connection points are correct as much as possible to ensure the accuracy of the final relative precision quality inspection.
[0080] The basic model of satellite L1-level raw images is the RPC model, which is essentially a high-precision mathematical fit to a strict physical model and can achieve a geometric accuracy that is almost consistent with that of a strict model.
[0081] The RPC parameters are calculated and fitted based on the strict imaging model at the object space virtual control points. It describes the relationship between the image point coordinates and the object space coordinates. The model has a total of 90 parameters, which are defined as follows:
[0082] ;
[0083] ;
[0084] in , , , are all polynomials of the following form, for example:
[0085] ;
[0086] Where (U, V, W) is the standardized ground coordinates, and (r, c) is the standardized image point coordinates:
[0087] ;
[0088] ;
[0089] Where ( , , , , ) is the standardized translation parameter, ( , , , , ) are standardized scale parameters, and they are related to ~ , ~ , ~ , ~ A total of 90 parameters are saved in the RPC file.
[0090] The free network adjustment method is used to iteratively calculate the image square error value of each point, set the error threshold, and delete the points that are greater than the error threshold. The method used in the free network adjustment is the bundle method regional network adjustment in digital photogrammetry. The free network adjustment is an iterative operation process, and each iteration requires the following steps:
[0091] (2.1) Calculate the geographic coordinates X, Y, and Z of each connection point based on the RPC model forward intersection of each image. Z is read from the digital elevation model DEM based on X and Y.
[0092] (2.2) The X, Y, and Z coordinates of each tie point on each image are calculated using the RPC model to obtain the pixel coordinates Col and Row. The pixel coordinates are then subtracted from the pixel coordinates of each image obtained by automatic matching to calculate the image square residual RMS of the tie point in each image. Tie points with a mean error greater than three times the mean error are deleted.
[0093] (2.3) The improved affine transformation parameters of the RPC of each image are calculated based on the deleted tie points, which are used to correct the RPC model of each image, and then enter the first iterative processing step.
[0094] Affine transformation model (image space), A and B are the six parameters of RPC image space affine correction of each image solved by free network adjustment:
[0095] ;
[0096] in, , It is the image coordinate compensation function, which is used to correct the RPC model by converting geographic coordinates into pixel coordinates. are the pixel coordinates of the tie points on the image.
[0097] The iteration termination condition is that the error in the image cube of all connection points is less than the set error threshold or the number of iterations exceeds the set maximum number of iterations. After the free network adjustment is solved, the RPC improved affine transformation parameters finally solved need to be cleared. This can not only maintain the RPC model of the original satellite image itself, but also delete the connection points with large gross errors, ensuring the objectivity of subsequent quantitative quality inspections.
[0098] For the aforementioned step S106, the embodiment of the present invention provides an implementation method for determining the pixel coordinate error between the target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected based on the rational polynomial coefficient model and the digital elevation model, and determining the relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, see the following steps 1 to 4:
[0099] Step 1: The target connection points contained in the first L1-level satellite image to be inspected (recorded as image 1) and the second L1-level satellite image to be inspected (recorded as image 2) are respectively used as the first target connection points and the second target connection points.
[0100] Step 2: Determine the estimated pixel coordinates of the first target connection point in the second satellite L1 image to be inspected based on the rational polynomial coefficient model, the digital elevation model and the first actual pixel coordinates of the first target connection point in the first satellite L1 image to be inspected.
[0101] After free network adjustment, each connection point has a geographic coordinate , , ,in , It is a geographic coordinate obtained by intersection during the adjustment process. This elevation value Z is also obtained through , Read from DEM data, but the geographic coordinates obtained by this adjustment , , It is just a process product and cannot be directly used to calculate relative error, but The value is relatively accurate and can be used as the initial elevation value for converting pixel coordinates to geographic coordinates.
[0102] Loop and count the connection points of each image and all the images that overlap with it. For example, there are N connection points between image 1 and image 2. Record the connection points as the first target connection points. Calculate the first actual pixel coordinates of these N first target connection points on image 1. , Use the RPC model of image 1 to convert it into geographic coordinates, convert the Z value obtained by adjustment as the initial elevation value, and obtain a new geographic coordinate value ( , , ), using the new planimetric geographic coordinate values ( , ) Read new elevation values from DEM data , further improving the accuracy of the elevation value, and then using and pixel coordinates , The RPC model of image 1 is used to convert to geographic coordinates (X, Y, Z), which represents the geographic coordinate value of the first target connection point calculated using the RPC model of image 1. The geographic coordinates (X, Y, Z) are converted to pixel coordinates ( , ), the pixel coordinates are the estimated pixel coordinates of the first target connection point in the second satellite L1 image to be inspected.
[0103] Step 3: Determine the pixel coordinate error based on the estimated pixel coordinates and the second actual pixel coordinates of the second target connection point matching the first target connection point in the second L1 level image of the satellite to be inspected.
[0104] In one example, the first target connection point is based on image 1, and the pixel coordinate error between image 1 and image 2 is As follows: , ) is the second actual pixel coordinate of the second target connection point in image 2 obtained by matching the first target connection point.
[0105] .
[0106] Step 4: Determine the relative accuracy mean error between the first L1-level satellite image to be inspected and the second L1-level satellite image to be inspected according to the pixel coordinate error.
[0107] In one example, the pixel error of all connection points between image 1 and image 2 is After the calculation is completed, a relative accuracy error is calculated. , represents the relative accuracy value between image 1 and image 2.
[0108]
[0109] The same method is used to calculate the relative accuracy mean error of image 1 and other overlapping images. … The relative accuracy error of image 1 can also be calculated using the above formula If you want to calculate , N represents the number of all connection points in image 1.
[0110] Calculate the relative accuracy error of image 1 Then, the relative accuracy mean error of image 2...image N can be calculated by the same method. … .
[0111] For the aforementioned step S108, an embodiment of the present invention provides an implementation method for generating a relative geometric accuracy quality inspection report corresponding to the satellite L1 image set to be inspected based on the pixel coordinate error and the relative accuracy mean error. Two relative accuracy quality inspection reports need to be exported, one is the relative accuracy quality inspection report of the scene, referred to as the scene quality inspection report, and the other is the overall relative accuracy quality inspection report, referred to as the overall quality inspection report.
[0112] The embodiment of the present invention first provides a method for generating a scene-by-scene quality inspection report, including:
[0113] Step a1: The target connection points contained in the first L1-level satellite image to be inspected and all the second L1-level satellite images to be inspected are all taken as first inspection points. The inspection points are the matching connection points mentioned above, which are called inspection points in the exported quality inspection report.
[0114] Assume that there is an overlap between image 1 and image 2 and image 3, that is, image 1 is the first satellite L1 image to be inspected, and image 2 and image 3 are the second satellite L1 images to be inspected. If a scene-by-scene quality inspection report needs to be generated, all target connection points contained in image 1, image 2, and image 3 need to be used as the first inspection point.
[0115] Step a2, based on the pixel coordinate error corresponding to the first checkpoint, statistically process the first checkpoint to obtain the checkpoint statistical results; wherein the checkpoint statistical results include one or more of the number of first points whose pixel coordinate error is less than a first mean error threshold, the number of second points whose pixel coordinate error is less than a second mean error threshold, the number of third points whose pixel coordinate error is greater than the second mean error threshold, the maximum pixel coordinate error, and the checkpoint gross error rate.
[0116] In one example, the pixel coordinate error can be compared with a first mean error threshold, and the number of first checkpoints whose pixel coordinate error is less than the first mean error threshold (referred to as the first point number) is counted; in one example, the pixel coordinate error is compared with a second mean error threshold, where the second mean error threshold is a specified number of digits of the first mean error threshold, and the number of first checkpoints whose pixel coordinate error is less than the second mean error threshold (referred to as the second point number) is counted, and the number of first checkpoints whose pixel coordinate error is greater than the second mean error threshold (referred to as the third point number) is counted; in one example, the maximum pixel coordinate error is determined from the pixel coordinate error; in one example, the ratio of the third point number to the total number of first checkpoints is used as the checkpoint gross error rate.
[0117] Step a3, generating an image quality inspection result according to the checkpoint gross error rate and the relative precision mean error. In one example, if the checkpoint gross error rate is less than the gross error rate threshold and the relative precision mean error is less than or equal to the first mean error threshold, the image quality inspection result is determined to be qualified.
[0118] Step a4: Use one or more of the pixel coordinate error corresponding to each first checkpoint, the total number of first checkpoints, checkpoint statistics, and image quality inspection results as a scene-by-scene quality inspection report.
[0119] In a specific embodiment, such as Figure 2 The following is a schematic diagram of a scene quality inspection report. The scene quality inspection report needs to include the following indicators:
[0120] 1) Checkpoint error ;
[0121] 2) Total number of checkpoints (N);
[0122] 3) The number of checkpoints (N1) that are less than the median error threshold M0;
[0123] 4) Number of checkpoints less than 2 times the median error threshold (N2)--(M0 <RMS<2*M0);
[0124] 5) The number of checkpoints greater than 2 times the mean error threshold (C) -- (RMS>2*M0);
[0125] 6) Maximum error (Max(Di)) (Definition: the error in the checkpoint with the largest error);
[0126] 7) Gross error rate (Cp) (Definition: the number of gross error points divided by the total number of points (C / N*100%);
[0127] 8) Quality inspection results (Explanation: The qualified condition is that the gross error rate (Cp) is less than 5% and the mean error ( )<= median error threshold (M0).
[0128] In a specific implementation, a relative accuracy index of the cyclic statistics image 1 and other overlapping images is also required, such as Figure 3 Another schematic diagram of a quality inspection report for different scenes is shown below:
[0129] 1) Checkpoint error , represents the relative accuracy mean error between the Mth image and the Nth image;
[0130] 2) The total number of checkpoints (N); represents the number of checkpoints of the Mth image and the Nth image;
[0131] 3) The number of checkpoints (N1) that are less than the median error threshold M0;
[0132] 4) Number of checkpoints less than 2 times the median error threshold (N2)--(M0 <RMS<2*M0);
[0133] 5) The number of checkpoints greater than 2 times the mean error threshold (C) -- (RMS>2*M0);
[0134] 6) Maximum error (Max(Di)) (Definition: the error in the checkpoint with the largest error);
[0135] 7) Gross error rate (Cp) (Definition: the number of gross error points divided by the total number of points (C / N*100%);
[0136] 8) Quality inspection results (Explanation: The qualified condition is that the gross error rate (Cp) is less than 5% and the mean error ( )<= median error threshold (M0).
[0137] Further report content should also include checkpoint thumbnails. In one embodiment, different marks are assigned to the first checkpoint based on the comparison results between the pixel coordinate error and the first mean error threshold and the second mean error threshold, so as to generate checkpoint thumbnails between the first satellite L1 level image to be inspected and any second satellite L1 level image to be inspected based on the marks. For example, in the checkpoint thumbnails between every two images, checkpoints with a mean error threshold of M0 are displayed in green, checkpoints with a mean error threshold of 2 times are displayed in yellow, and checkpoints with a mean error threshold of 2 times are displayed in red. This not only can quantitatively evaluate the quality inspection results, but also can visually evaluate which positions of the image have larger errors.
[0138] The image internal accuracy quality inspection report automatically generated in the embodiment of the present invention can visually clearly and quickly locate where in the image there are large relative accuracy errors, and can quantitatively accurately give the error in the internal geometric distortion of the image and whether it is qualified.
[0139] The embodiment of the present invention first provides a method for generating an overall quality inspection report, including:
[0140] Step b1, taking all target connection points included in the L1-level satellite images to be inspected as the second inspection points;
[0141] Step b2, performing statistical processing on the image quality inspection report to obtain report statistical results; wherein the report statistical results include the number of scenes that pass the quality inspection, the number of scenes that fail the quality inspection, and the quality inspection pass rate;
[0142] Step b3, taking one or more of the pixel coordinate error corresponding to each second checkpoint, the total number of second checkpoints, the relative accuracy mean error corresponding to each satellite L1 image to be inspected, the image quality inspection report, and the report statistical results as the overall quality inspection report.
[0143] In a specific implementation, all L1-level satellite images to be inspected will output an overall accuracy inspection report, such as Figure 4 The following is a schematic diagram of an overall quality inspection report. The main statistical indicators are as follows:
[0144] 1) Number of images to be examined (ALL_IMG_NUM);
[0145] 2) Total number of checkpoints (ALL_N);
[0146] 3) Mean error of all checkpoints (ALL_M);
[0147] 4) Overall quality inspection results (mean error (ALL_M) <= mean error threshold (M0);
[0148] 5) Number of scenes that passed the quality inspection (Qualified_NUM);
[0149] 6) Number of scenes that failed quality inspection (Qualified_NUM);
[0150] 7) Quality inspection qualified rate (Qualified_Ratio): Qualified_NUM / ALL_IMG_NUM;
[0151] 8) List the relative mean error value of each image, adjacent images, whether it is qualified, single-image quality inspection report path, etc.
[0152] In summary, the overall flow chart of the relative geometric accuracy quality inspection method between satellite L1-level images provided by the embodiment of the present invention is as follows: Figure 5As shown, the process includes data preparation, connection point matching, regional network adjustment to eliminate erroneous points, relative error calculation, and relative accuracy quality inspection report. In view of the problem of automatic quality inspection of relative geometric accuracy of satellite L1 remote sensing images, after automatically matching the same-name points between the images to be inspected, the embodiment of the present invention combines the regional digital elevation model data (DEM) of the region and uses the regional network adjustment method in the digital photogrammetry discipline to automatically eliminate erroneous same-name points, and automatically generates thumbnails and quality inspection reports for each remote sensing image. The thumbnails use different color levels to characterize the relative error degree of each same-name point according to the set relative geometric accuracy error threshold, and introduce the percentage parameter exceeding the threshold to assist in determining whether the relative geometric accuracy of the image to be inspected meets the standard. Through the method proposed in the embodiment of the present invention, not only can the relative geometric accuracy of the image to be inspected be accurately captured visually, but also the relative accuracy value of each image to be inspected and all adjacent images can be quantitatively described to determine whether it has passed the quality inspection.
[0153] Based on the above-mentioned embodiment, the present invention provides a relative geometric accuracy inspection device between satellite L1 level images, see Figure 6 The schematic diagram of the structure of a relative geometric accuracy quality inspection device between satellite L1-level images is shown, and the device mainly includes the following parts:
[0154] The data acquisition module 602 is used to acquire a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected;
[0155] The connection point determination module 604 is used to match the connection points of the L1-level satellite image set to be inspected and remove the gross error points based on the regional block adjustment method to obtain the target connection points included in the L1-level satellite image set to be inspected;
[0156] The error calculation module 606 is used to determine, for any first satellite L1 level image to be inspected and at least one second satellite L1 level image to be inspected that overlaps with the first satellite L1 level image in the set of satellite L1 level images to be inspected, pixel coordinate errors between target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, and to determine the relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected based on the rational polynomial coefficient model and the digital elevation model;
[0157] The quality inspection module 608 is used to generate a relative geometric accuracy quality inspection report corresponding to the satellite L1 image set to be inspected based on the pixel coordinate error and the relative accuracy mean error.
[0158] The above-mentioned relative geometric accuracy quality inspection device between satellite L1 level images provided by the embodiment of the present invention automatically matches the connection points between the satellite L1 level images to be inspected, and then automatically eliminates erroneous connection points in combination with the digital elevation model data of the area, and on this basis, determines the pixel coordinate errors between the connection points and the relative accuracy mean error between the images, and then automatically generates a relative geometric accuracy quality inspection report. The embodiment of the present invention realizes automated and quantitative quality inspection of the relative geometric accuracy between satellite L1 level images.
[0159] In one implementation, the error calculation module 606 is specifically used to:
[0160] The target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected are used as the first target connection point and the second target connection point respectively;
[0161] Determine the estimated pixel coordinates of the first target connection point in the second satellite L1 level image to be inspected based on the rational polynomial coefficient model, the digital elevation model and the first actual pixel coordinates of the first target connection point in the first satellite L1 level image to be inspected;
[0162] Determine the pixel coordinate error according to the estimated pixel coordinate and the second actual pixel coordinate of the second target connection point matching the first target connection point in the second L1 level image of the satellite to be inspected;
[0163] The relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected is determined according to the pixel coordinate error.
[0164] In one embodiment, the relative geometric accuracy quality inspection report includes a scene-by-scene quality inspection report and / or an overall quality inspection report;
[0165] The scene-by-scene quality inspection report is a quality inspection report on the relative geometric accuracy between the first L1-level satellite image to be inspected and all the second L1-level satellite images to be inspected;
[0166] The overall quality inspection report is a quality inspection report on the relative geometric accuracy between all satellite L1 images to be inspected.
[0167] In one implementation, the relative geometric accuracy quality inspection report includes a scene segmentation quality inspection report, and the quality inspection module 608 is specifically used to:
[0168] The target connection points contained in the first satellite L1 level image to be inspected and all the second satellite L1 level images to be inspected are taken as the first inspection points;
[0169] Based on the pixel coordinate error corresponding to the first checkpoint, statistical processing is performed on the first checkpoint to obtain a checkpoint statistical result; wherein the checkpoint statistical result includes one or more of the number of first points whose pixel coordinate error is less than a first mean error threshold, the number of second points whose pixel coordinate error is less than a second mean error threshold, the number of third points whose pixel coordinate error is greater than the second mean error threshold, the maximum pixel coordinate error, and the checkpoint gross error rate;
[0170] Generate image quality inspection results based on the gross error rate and relative accuracy error of the inspection points;
[0171] One or more of the pixel coordinate error corresponding to each first checkpoint, the total number of first checkpoints, checkpoint statistical results, and image quality inspection results are used as a scene-by-scene quality inspection report.
[0172] In one implementation, the quality inspection module 608 is specifically used to:
[0173] If the gross error rate of the checkpoint is less than the gross error rate threshold, and the relative accuracy mean error is less than or equal to the first mean error threshold, the image quality inspection result is determined to be qualified.
[0174] In one implementation, the quality inspection module 608 is specifically used to:
[0175] Based on the comparison results between the pixel coordinate error and the first mean error threshold and the second mean error threshold, different marks are assigned to the first checkpoint to generate a checkpoint thumbnail between the first satellite L1 level image to be inspected and any second satellite L1 level image to be inspected based on the mark.
[0176] In one implementation, the relative geometric accuracy quality inspection report includes an overall quality inspection report, and the quality inspection module 608 is specifically used to:
[0177] All target connection points included in the satellite L1 level images to be inspected are used as the second inspection points;
[0178] And, statistical processing is performed on the image quality inspection report to obtain report statistical results; wherein the report statistical results include the number of scenes that pass the quality inspection, the number of scenes that fail the quality inspection, and the quality inspection pass rate;
[0179] The pixel coordinate error corresponding to each second checkpoint, the total number of second checkpoints, the relative accuracy mean error corresponding to each satellite L1 image to be inspected, the image quality inspection report, and one or more of the report statistical results are used as the overall quality inspection report.
[0180] The device provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.
[0181] An embodiment of the present invention provides an electronic device. Specifically, the electronic device includes a processor and a storage device. The storage device stores a computer program, and when the computer program is executed by the processor, it executes the method described in any one of the above-mentioned implementation methods.
[0182] Figure 7 A structural diagram of an electronic device provided in an embodiment of the present invention, the electronic device 100 includes: a processor 70, a memory 71, a bus 72 and a communication interface 73, wherein the processor 70, the communication interface 73 and the memory 71 are connected via the bus 72; the processor 70 is used to execute an executable module stored in the memory 71, such as a computer program.
[0183] The computer program product of the readable storage medium provided in the embodiment of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be referred to the previous method embodiments, which will not be repeated here.
[0184] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims
1. A method for quality inspection of relative geometric accuracy between satellite L1 level images, characterized in that: include: Acquire a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected; Performing tie point matching and gross error point elimination based on a regional block adjustment method on the satellite L1 level image set to be inspected, so as to obtain target tie points included in the satellite L1 level image set to be inspected; For any first satellite L1 level image to be inspected in the set of satellite L1 level images to be inspected and at least one second satellite L1 level image to be inspected that overlaps with the first satellite L1 level image to be inspected, based on a rational polynomial coefficient model and the digital elevation model, determine a pixel coordinate error between the target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, and determine a relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected; Based on the pixel coordinate error and the relative accuracy mean error, a relative geometric accuracy quality inspection report corresponding to the set of L1-level satellite images to be inspected is generated; wherein the relative geometric accuracy quality inspection report includes a sub-scene quality inspection report and an overall quality inspection report; the sub-scene quality inspection report is a relative geometric accuracy quality inspection report between the first L1-level satellite image to be inspected and all the second L1-level satellite images to be inspected; the overall quality inspection report is a relative geometric accuracy quality inspection report between all the L1-level satellite images to be inspected; The relative geometric accuracy quality inspection report includes a scene-by-scene quality inspection report. Based on the pixel coordinate error and the relative accuracy mean error, a relative geometric accuracy quality inspection report corresponding to the L1-level image set of the to-be-inspected satellite is generated, including: The target connection points respectively included in the first satellite L1 level image to be inspected and all the second satellite L1 level images to be inspected are used as first inspection points; Based on the pixel coordinate error corresponding to the first checkpoint, statistical processing is performed on the first checkpoint to obtain a checkpoint statistical result; wherein the checkpoint statistical result includes the number of first points whose pixel coordinate error is less than a first mean error threshold, the number of second points whose pixel coordinate error is less than a second mean error threshold, the number of third points whose pixel coordinate error is greater than the second mean error threshold, the maximum pixel coordinate error, and the checkpoint gross error rate; Generate an image quality inspection result according to the gross error rate of the inspection point and the mean error of the relative accuracy; One or more of the pixel coordinate error corresponding to each of the first checkpoints, the total number of the first checkpoints, the checkpoint statistical results, and the image quality inspection results are used as the scene-by-scene quality inspection report.
2. The relative geometric accuracy quality inspection method between satellite L1 level images according to claim 1 is characterized in that: Based on the rational polynomial coefficient model and the digital elevation model, determining the pixel coordinate error between the target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, and determining the relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, including: The target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected are used as the first target connection point and the second target connection point; Determine the estimated pixel coordinates of the first target connection point in the second satellite L1 level image to be inspected based on the rational polynomial coefficient model, the digital elevation model and the first actual pixel coordinates of the first target connection point in the first satellite L1 level image to be inspected; Determine a pixel coordinate error according to the estimated pixel coordinates and a second actual pixel coordinate of the second target connection point matched with the first target connection point in the second L1-level image of the satellite to be inspected; The relative accuracy mean square error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected is determined according to the pixel coordinate error.
3. The relative geometric accuracy quality inspection method between satellite L1 level images according to claim 1 is characterized in that: Generating an image quality inspection result according to the gross error rate of the inspection point and the relative accuracy mean error, including: If the gross error rate of the checkpoint is less than the gross error rate threshold, and the relative accuracy mean error is less than or equal to the first mean error threshold, the image quality inspection result is determined to be qualified.
4. The relative geometric accuracy quality inspection method between satellite L1 level images according to claim 1 is characterized in that: The method further comprises: Based on the comparison results between the pixel coordinate error and the first mean error threshold and the second mean error threshold, different marks are assigned to the first checkpoints to generate checkpoint thumbnails between the first satellite L1 level image to be inspected and all the second satellite L1 level images to be inspected based on the marks.
5. The relative geometric accuracy quality inspection method between satellite L1 level images according to claim 1 is characterized in that: The relative geometric accuracy quality inspection report includes an overall quality inspection report, and the method further includes: The target connection points included in all the satellite L1 level images to be inspected are used as second inspection points; And, statistical processing is performed on the image quality inspection results to obtain report statistical results; wherein the report statistical results include the number of scenes that pass the quality inspection, the number of scenes that fail the quality inspection, and the quality inspection pass rate; One or more of the pixel coordinate error corresponding to each of the second checkpoints, the total number of the second checkpoints, the relative accuracy mean error corresponding to each of the L1-level satellite images to be inspected, the image quality inspection results, and the report statistical results are used as the overall quality inspection report.
6. A relative geometric accuracy inspection device between satellite L1 level images, characterized in that: include: A data acquisition module is used to acquire a set of L1-level satellite images to be inspected, and digital elevation model data covering each L1-level satellite image to be inspected in the set of L1-level satellite images to be inspected; A connection point determination module is used to perform connection point matching on the satellite L1 level image set to be inspected and eliminate gross error points based on a regional block adjustment method to obtain target connection points included in the satellite L1 level image set to be inspected; an error calculation module, for determining, for any first satellite L1 level image to be inspected and at least one second satellite L1 level image to be inspected that overlaps with the first satellite L1 level image in the set of satellite L1 level images to be inspected, a pixel coordinate error between the target connection points respectively included in the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected, and determining a relative accuracy mean error between the first satellite L1 level image to be inspected and the second satellite L1 level image to be inspected based on a rational polynomial coefficient model and the digital elevation model; A quality inspection module is used to generate a relative geometric accuracy quality inspection report corresponding to the set of L1-level satellite images to be inspected based on the pixel coordinate error and the relative accuracy mean error; wherein the relative geometric accuracy quality inspection report includes a sub-scene quality inspection report and an overall quality inspection report; the sub-scene quality inspection report is a relative geometric accuracy quality inspection report between the first L1-level satellite image to be inspected and all the second L1-level satellite images to be inspected; the overall quality inspection report is a relative geometric accuracy quality inspection report between all the L1-level satellite images to be inspected; The relative geometric accuracy quality inspection report includes a scene-by-scene quality inspection report, and the quality inspection module is specifically used for: The target connection points included in the first satellite L1 level image to be inspected and all the second satellite L1 level images to be inspected are taken as first inspection points; Based on the pixel coordinate error corresponding to the first checkpoint, statistical processing is performed on the first checkpoint to obtain a checkpoint statistical result; wherein the checkpoint statistical result includes the number of first points whose pixel coordinate error is less than a first mean error threshold, the number of second points whose pixel coordinate error is less than a second mean error threshold, the number of third points whose pixel coordinate error is greater than the second mean error threshold, the maximum pixel coordinate error, and the checkpoint gross error rate; Generate an image quality inspection result according to the gross error rate of the inspection point and the mean error of the relative accuracy; One or more of the pixel coordinate error corresponding to each of the first checkpoints, the total number of the first checkpoints, the checkpoint statistical results, and the image quality inspection results are used as the scene-by-scene quality inspection report.
7. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 5.
8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method according to any one of claims 1 to 5.
