Geographic positioning precision test characterization method and device based on microwave remote sensing image

Abstract

The invention discloses a microwave remote sensing image-based geographic positioning precision test characterization method and device, and belongs to the field of remote sensing image analysis, and the method comprises the steps: interacting a microwave imager carried by a meteorological satellite, and obtaining a microwave remote sensing image; in the microwave remote sensing image, extracting a temperature change extreme point, and generating a to-be-detected coastline based on the temperature change extreme point; matching the to-be-detected coastline with a preset coastline to obtain a matched coastline; coastline points are extracted on the preset coastline and the matched coastline respectively, and multiple pairs of coastline points are formed; and the distances between the multiple pairs of coastline points are solved respectively, a positioning error is generated, and a verification characterization result is generated based on the positioning error. According to the method, the technical problem that a microwave remote sensing image geographic positioning precision test and characterization method in the prior art is not accurate enough is solved, and the technical effect of improving the microwave remote sensing image geographic positioning precision test and characterization precision is achieved.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing image analysis, and in particular to a method and a device for verifying and characterizing geographic positioning accuracy based on microwave remote sensing images. Background Art

[0002] At present, the verification and characterization of the geolocation accuracy of microwave remote sensing images is mainly carried out by extracting the coastline with obvious temperature changes in the image, matching it with the preset high-precision coastline, and using the deviation of the matching result as a measure of positioning accuracy. However, the existing coastline extraction and matching methods are not accurate enough, making it difficult to accurately estimate the positioning deviation of microwave remote sensing images. Summary of the invention

[0003] The present invention provides a method and device for detecting and characterizing the geographic positioning accuracy based on microwave remote sensing images, aiming to solve the technical problem that the microwave remote sensing image geographic positioning accuracy detection and characterization methods in the prior art are not accurate enough.

[0004] In view of the above problems, the present invention provides a method and device for verifying the accuracy of geographic positioning based on microwave remote sensing images.

[0005] The first aspect disclosed by the present invention provides a geographic positioning accuracy verification and characterization method based on microwave remote sensing images, the method comprising: using a microwave imager carried by an interactive meteorological satellite to obtain a microwave remote sensing image; extracting temperature change extreme points in the microwave remote sensing image, and generating a coastline to be detected based on the temperature change extreme points; matching the coastline to be detected with a preset coastline to obtain a matching coastline; respectively extracting coastline points on the preset coastline and the matching coastline to form multiple pairs of coastline points; respectively calculating the distances between the multiple pairs of coastline points to generate positioning errors, and generating verification characterization results based on the positioning errors.

[0006] Another aspect disclosed by the present invention provides a geographic positioning accuracy verification and characterization device based on microwave remote sensing images, the device comprising: a remote sensing image acquisition unit, used to interact with a microwave imager carried by a meteorological satellite to obtain a microwave remote sensing image; a coastline generation unit to be detected, used to extract temperature change extreme points in the microwave remote sensing image, and generate a coastline to be detected based on the temperature change extreme points; a matching coastline acquisition unit, used to match the coastline to be detected with a preset coastline to obtain a matching coastline; a coastline point extraction unit, used to extract coastline points on the preset coastline and the matching coastline, respectively, to form multiple pairs of coastline points; a verification characterization result generation unit, used to respectively obtain the distances between multiple pairs of coastline points, generate positioning errors, and generate verification characterization results based on the positioning errors.

[0007] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0008] By adopting the microwave imager carried by the interactive meteorological satellite, microwave remote sensing images are obtained as the data basis for verifying and characterizing the accuracy of geographic positioning; in the microwave remote sensing images, the extreme points of temperature change are extracted, and the coastline to be detected is generated based on the extreme points of temperature change, so as to obtain the position of the coastline more accurately; the coastline to be detected is matched with the preset coastline to obtain the matching coastline; coastline points are respectively extracted on the preset coastline and the matching coastline to form multiple pairs of coastline points; the distances between the multiple pairs of coastline points are respectively calculated to generate positioning errors, and verification characterization results are generated based on the positioning errors, so as to achieve the matching of the coastline to be detected with the preset coastline, and the positioning error is obtained by calculating the distance between the matched coastline point pairs, and then the verification result characterizing the accuracy of geographic positioning is generated. The technical solution solves the technical problem that the method for verifying and characterizing the accuracy of microwave remote sensing images in the prior art is not accurate enough, and achieves the technical effect of improving the accuracy of verifying and characterizing the accuracy of geographic positioning of microwave remote sensing images.

[0009] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic flow chart of a method for characterizing geographic positioning accuracy inspection based on microwave remote sensing images is provided for an embodiment of the present invention;

[0011] Figure 2 A structural schematic diagram of a device for verifying and characterizing geographic positioning accuracy based on microwave remote sensing images is provided for an embodiment of the present invention.

[0012] Explanation of reference numerals: remote sensing image acquisition unit 11, to-be-detected coastline generation unit 12, matched coastline acquisition unit 13, coastline point extraction unit 14, verification characterization result generation unit 15. DETAILED DESCRIPTION

[0013] The overall idea of ​​the technical solution provided by the present invention is as follows:

[0014] The embodiment of the present invention provides a method and device for verifying the accuracy of geographic positioning based on microwave remote sensing images. First, a microwave remote sensing image is obtained from a microwave imager carried by a meteorological satellite as a data basis. Then, by extracting the extreme points of temperature change in the microwave remote sensing image and generating a coastline to be detected based on the extreme points, the temperature difference at the boundary between land and sea is used to accurately extract the coastline position. Next, the coastline to be detected is matched with the preset coastline to obtain a matching coastline, and multiple pairs of coastline points are extracted on the preset coastline and the matching coastline. The positioning error is obtained by calculating the distance between the coastline point pairs, and a verification characterization result is generated.

[0015] After introducing the basic principles of the present invention, various non-limiting implementation methods of the present invention will be specifically described below in conjunction with the accompanying drawings.

[0016] Embodiment 1

[0017] like Figure 1 As shown, an embodiment of the present invention provides a method for verifying and characterizing geographic positioning accuracy based on microwave remote sensing images, the method comprising:

[0018] S1: A microwave imager carried by an interactive meteorological satellite to obtain microwave remote sensing images;

[0019] Specifically, microwave remote sensing images are acquired by microwave imagers carried by meteorological satellites. Microwave imagers are active remote sensors for earth observation, which form microwave images reflecting the characteristics of the earth's surface by receiving microwave signals scattered and radiated by the target. When acquiring microwave remote sensing images, the microwave imager data transmitted is received and parsed through interaction with the meteorological satellite ground station to obtain digital microwave image data. Among them, the area covered by the acquired microwave remote sensing image should contain a sufficiently long coastline to support the subsequent coastline extraction and matching steps. At the same time, the acquired microwave remote sensing images should also have appropriate spatial resolution and radiation resolution to ensure the accuracy of coastline extraction and positioning error calculation, laying a data foundation for the implementation of subsequent steps.

[0020] S2: extracting temperature change extreme points in the microwave remote sensing image, and generating a coastline to be detected based on the temperature change extreme points;

[0021] Specifically, due to the difference in specific heat capacity between the ocean and land, there is usually an obvious temperature gradient change at the sea-land boundary in microwave remote sensing images. By taking advantage of this physical property, the location of the coastline can be accurately obtained by extracting the extreme points in the area with large temperature changes.

[0022] First, the microwave remote sensing image is preprocessed, including denoising and smoothing operations to reduce noise and interference in the image and highlight the temperature difference characteristics at the sea-land boundary. Then, by calculating the temperature gradient field of the image, the local maximum points in the area with large gradient values ​​are extracted as candidate coastline points.

[0023] In order to further improve the accuracy of coastline extraction, some post-processing measures can also be introduced, such as removing isolated noise points through morphological filtering, optimizing the shape of the coastline through curve fitting, and removing pseudo coastline points with small temperature gradients through threshold segmentation.

[0024] After the above processing, a set of temperature change extreme points reflecting the coastline position is obtained. Based on these extreme points, a continuous coastline to be detected can be generated. The coastline to be detected has a high degree of consistency with the actual coastline position, providing a reliable reference for subsequent coastline matching.

[0025] The method for extracting the extreme points of temperature change is not limited to the gradient method described above, and other edge detection algorithms may also be used, such as the Canny algorithm, the Sobel algorithm, etc. At the same time, the method for generating the coastline to be detected may also be adjusted according to actual needs, such as using curve fitting, line segment fitting, etc., to obtain a smoother and continuous coastline.

[0026] By utilizing the physical characteristics of microwave remote sensing images, the precise extraction of coastline positions is achieved, laying the foundation for subsequent coastline matching and positioning error calculation. Compared with traditional coastline extraction methods, the extraction accuracy is effectively improved and the interference of false coastlines is reduced.

[0027] S3: Matching the coastline to be detected with a preset coastline to obtain a matching coastline;

[0028] Specifically, the geometric features between the coastline to be detected and the preset coastline are used to optimize the matching algorithm to solve the optimal transformation relationship between the two coastlines and obtain the matched coastline.

[0029] First, obtain a preset high-precision coastline as a reference. The preset coastline comes from high-resolution geographic information data products, such as the global coastline dataset GSHHS. The spatial position and shape accuracy of the preset coastline are high, and can be used as a benchmark for evaluating the positioning accuracy of the coastline to be detected. Then, the iterative closest point (ICP) algorithm is used to achieve the matching between the coastline to be detected and the preset coastline. The ICP algorithm finds the optimal rigid body transformation relationship between the two coastlines through iterative optimization, so that the shape and position of the transformed coastline to be detected are as consistent as possible with the preset coastline. After optimization by the ICP algorithm, the transformed coastline to be detected is obtained, that is, the matching coastline. The matching coastline has the best geometric fit with the preset coastline, and its position deviation reflects the positioning accuracy of the coastline to be detected.

[0030] By accurately matching the coastline to be detected with the preset coastline, a matching coastline reflecting the positioning accuracy is obtained. Compared with the traditional single-point matching method, the overall geometric characteristics of the coastline are utilized, the matching relationship is optimized through the ICP algorithm, the influence of local deformation and noise interference is effectively overcome, and the matching accuracy and reliability are improved.

[0031] S4: extracting coastline points on the preset coastline and the matching coastline respectively to form multiple pairs of coastline points;

[0032] Specifically, after obtaining the matching coastline, the characteristic points on the preset coastline and the matching coastline are further extracted to construct a point pair relationship in preparation for the subsequent calculation of the positioning error.

[0033] First, the preset coastline and the matching coastline are resampled so that they have the same number of points and uniform point spacing. Resampling can use methods such as linear interpolation and spline interpolation to simplify subsequent processing steps while maintaining the shape of the coastline. Then, a certain number of feature points are equally spaced on the resampled preset coastline and the matching coastline as coastline points. The selection interval of the coastline points can be adaptively adjusted according to the length and complexity of the coastline to balance computational efficiency and representation accuracy. Next, based on the correspondence between the coastline points on the preset coastline and the matching coastline, a coastline point pair is constructed. Specifically, for each coastline point on the matching coastline, a one-to-one point pair relationship is established by calculating its nearest point on the preset coastline.

[0034] A set of paired coastline points is obtained, each pair of points comes from the preset coastline and the matching coastline. These point pairs reflect the correspondence between the two coastlines in the local area, providing a basis for the subsequent calculation of positioning errors.

[0035] The extraction method of coastline points is not limited to equally spaced sampling, but can also be adaptively selected based on the curvature, inflection points and other features of the coastline. In addition, in order to improve the stability and reliability of point pairs, some optimization measures can be introduced, such as filtering abnormal point pairs by setting a distance threshold, optimizing the distribution of point pairs by curvature constraints, and optimizing the correspondence of point pairs by matching local shape features.

[0036] By extracting multiple pairs of corresponding coastline points on the preset coastline and the matching coastline, the foundation for the calculation of positioning error is laid. Compared with directly using the original coastline for error calculation, resampling and point pair construction effectively simplifies the calculation complexity and improves the calculation efficiency. At the same time, through point pair optimization measures, the accuracy and reliability of positioning error calculation can be further improved.

[0037] S5: respectively calculating the distances between multiple pairs of coastline points, generating positioning errors, and generating verification characterization results based on the positioning errors.

[0038] Specifically, the obtained coastline point pairs are used to quantitatively evaluate the positioning error of the matching coastline relative to the preset coastline by calculating the distance between each pair of coastline points, and further generate a verification characterization result that characterizes the overall positioning accuracy.

[0039] First, traverse all the coastline point pairs, and for each pair of coastline points, calculate their Euclidean distance in the longitude and latitude coordinate system as the positioning error of the point pair, which reflects the offset of the matching coastline at different positions relative to the preset coastline. In order to further quantify the overall positioning accuracy, some statistical indicators such as mean, median, root mean square error, etc. can also be calculated based on this set of positioning error values ​​as verification characterization results.

[0040] Through quantitative evaluation and characterization of the positioning accuracy of the matched coastline, verification characterization results reflecting the overall positioning error level were obtained.

[0041] Furthermore, the embodiment of the present invention also includes:

[0042] Linearly upsampling the longitude and latitude of the microwave remote sensing image to increase the resolution to 1 km×1 km;

[0043] Surface fitting of the observed brightness temperatures was performed to increase the resolution to the same as longitude and latitude;

[0044] The area with larger gradient is extracted from the fitted brightness temperature surface as the coastline to generate the coastline to be detected.

[0045] In a feasible implementation, since the original microwave remote sensing image has a low resolution, directly extracting the coastline on it may affect the accuracy. Therefore, the spatial resolution of the image is increased to 1km×1km by linearly upsampling the longitude and latitude data. The upsampling process is implemented through interpolation algorithms, such as bilinear interpolation, bicubic interpolation, etc. Through upsampling, the grid density of the microwave remote sensing image is increased, which helps to capture the position changes of the coastline more finely in the future.

[0046] Unlike longitude and latitude data, brightness temperature data observed by microwave remote sensing images are often affected by noise and interference, and direct upsampling may introduce large errors. In order to overcome this problem, a surface fitting method is used to construct a smooth surface model on the brightness temperature data, and resample according to the fitted surface to obtain brightness temperature data with the same resolution as longitude and latitude. Among them, surface fitting can use mathematical models such as polynomial functions and spline functions, and solve model parameters through optimization algorithms such as least squares method. Through surface fitting, the noise of brightness temperature data is effectively suppressed, while maintaining the continuity and smoothness of the data, providing high-quality input data for subsequent coastline extraction.

[0047] After obtaining a high-resolution, smooth and continuous brightness temperature surface, the gradient field of the surface is calculated to extract areas with larger gradient values ​​as candidate coastline locations. The gradient field can be calculated using numerical differentiation methods, such as central difference, Sobel operator, etc. Areas with larger gradient values ​​correspond to areas with drastic changes in brightness temperature, which usually indicate the location of the land-sea boundary. In order to further optimize the coastline extraction results, post-processing operations such as threshold segmentation and morphological filtering can be performed on the gradient field to remove noise points and connect breakpoints to obtain a more continuous and smooth coastline to be detected.

[0048] The spatial resolution of the data was improved by upsampling, the noise interference of the brightness temperature data was eliminated by surface fitting, and the coastline extraction accuracy was optimized through gradient field analysis and post-processing. Finally, a high-quality coastline to be detected was obtained, laying a solid foundation for the subsequent coastline matching and positioning error calculation.

[0049] Furthermore, the embodiment of the present invention also includes:

[0050] Extract the coastline points to be detected from the coastline to be detected, and obtain a coastline point set to be detected;

[0051] Find reference coastline points corresponding to the coastline points to be detected from the preset coastline to obtain a reference coastline point set;

[0052] The ICP algorithm is used to match the detected coastline point set with the reference coastline point set, solve the optimal rigid transformation parameters, and obtain the matching coastline.

[0053] In a feasible implementation, first, a series of discrete points are selected from the generated coastline to be detected as the coastline points to be detected to form a set of coastline points to be detected. The selection of the coastline points to be detected can adopt methods such as equidistant sampling and curvature sampling to evenly cover the entire coastline. At the same time, using the preset coastline, the corresponding reference point is found for each coastline point to be detected to form a reference coastline point set. The search for the reference point can be based on criteria such as Euclidean distance and curvature similarity to search for the position that best matches the coastline point to be detected on the preset coastline. In order to improve the robustness of the matching, some constraints can also be introduced, such as setting a search range threshold, filtering outliers using topological relationships, etc. By constructing a reference coastline point set, a corresponding relationship between the coastline points to be detected and the preset coastline is established, providing an initial match for subsequent iterative optimization. After obtaining the coastline point set to be detected and the reference coastline point set, the iterative ICP algorithm is used to achieve a fine match between the coastline point set to be detected and the reference coastline point set. The ICP algorithm continuously updates the posture of the coastline point set to be detected through iterative optimization, so that it overlaps with the reference coastline point set as much as possible, and the transformed coastline point set to be detected is the matching coastline.

[0054] By finely matching the coastline to be detected with the preset coastline, a high-quality matching coastline is obtained. Compared with the existing technology, the complexity of the matching problem is reduced by extracting a discrete coastline point set; a reliable initial match is provided by constructing a reference coastline point set; and the precise matching of the overall geometric features of the coastline is achieved through iterative optimization of the ICP algorithm. The final matching coastline is highly consistent with the preset coastline, laying the foundation for the accurate calculation of the positioning error.

[0055] Furthermore, the embodiment of the present invention also includes:

[0056] Calculate the mean Euclidean distance between each point in the detected coastline point set and the nearest point in the reference coastline point set as the distance error of the current iteration;

[0057] Calculate the geometric center points of the coastline point set to be detected and the reference coastline point set, and calculate the coordinate difference between the coastline point set to be detected and the reference coastline point set relative to their respective geometric center points to obtain decentralized coordinates;

[0058] Construct a matrix based on the decentralized coordinates and perform singular value decomposition to obtain the matrix and vector representing the rotation and translation relationship, and obtain the current rigid transformation parameters;

[0059] The current rigid transformation parameters are used to transform the coastline point set to be detected to obtain an updated coastline point set to be detected, and the nearest point in the reference coastline point set is re-searched and a new distance error is calculated to obtain an updated distance error;

[0060] The iteration is performed, and if the absolute value of the difference between the distance error and the updated distance error is less than a preset threshold or the number of iterations reaches a preset upper limit, the current rigid transformation parameter is used as the optimal rigid transformation parameter.

[0061] In a feasible implementation, for each point in the coastline point set to be detected, by calculating its Euclidean distance to all points in the reference coastline point set, the nearest point is found as the corresponding point, and the distances of all point pairs are averaged to obtain the distance error of the current iteration. The distance error reflects the overall distance difference between the current coastline point set to be detected and the reference coastline point set, and is an important indicator for evaluating the goodness of matching. Then, the coordinates of the geometric center points of the coastline point set to be detected and the reference coastline point set are calculated respectively, and the coordinates of each point are subtracted from the coordinates of the center point of the corresponding point set to obtain the coordinate difference relative to the center point, that is, the decentralized coordinates. Through the decentralized processing, the rigid transformation problem is transformed into a rotation and translation problem around the origin, reducing the complexity of the parameter solution. Subsequently, a 3×N matrix is ​​constructed using the coordinates of the coastline point set to be detected and the coordinates of the reference coastline point set after decentralization, which contains the rotation and translation relationship between the two point sets. Among them, N is the number of points in the coastline point set to be detected and the reference coastline point set. Each column of the matrix consists of a pair of decentralized point coordinate differences, that is, the i-th column of the matrix is ​​the difference between the decentralized coordinates of the i-th point in the coastline point set to be detected and the decentralized coordinates of the i-th point in the reference coastline point set. Then, the matrix is ​​subjected to singular value decomposition to obtain an orthogonal matrix representing the rotation relationship and a vector representing the translation relationship. These matrices and vectors together constitute the rigid transformation parameters of the current iteration.

[0062] After that, the obtained rigid transformation parameters are applied to the coastline point set to be detected, and the coastline point set to be detected is rotated and translated to obtain the updated coastline point set to be detected. Then, the operation is repeated to find the nearest point in the reference coastline point set for each point in the updated coastline point set to be detected, and the new distance error is calculated to obtain the updated distance error. The updated distance error reflects the degree of matching between the coastline point set to be detected and the reference coastline point set after the current rigid transformation, and is a key indicator for evaluating the effect of iterative optimization. Through iterative execution, the rigid transformation parameters are continuously updated to gradually reduce the matching error between the coastline point set to be detected and the reference coastline point set. During the iterative process, the absolute value of the distance error difference between the two iterations before and after is compared with the preset threshold to determine whether the algorithm has reached the convergence condition. When the error difference is less than the threshold, it means that the iterative optimization effect is not obvious and the iteration can be terminated. At the same time, in order to prevent the algorithm from oscillating near the local optimal solution, the maximum number of iterations can also be set as another termination condition. When any of the above termination conditions is met, the rigid transformation parameters obtained in the current iteration are used as the optimal rigid transformation parameters, and the matching process of the coastline point set is completed.

[0063] Furthermore, the embodiment of the present invention also includes:

[0064] uniformly selecting a plurality of sampling points on the matching coastline to obtain a matching coastline sampling point set;

[0065] The nearest point from each sampling point in the matching coastline sampling point set to the preset coastline is calculated to form multiple pairs of coastline points, each pair of coastline points including a matching coastline sampling point and a nearest point on the preset coastline.

[0066] In a feasible implementation, multiple sampling points are uniformly selected at certain intervals on the obtained matching coastline to form a matching coastline sampling point set. The sampling points are selected using the equal arc length sampling method, that is, the matching coastline is divided into several arcs of equal length, and then a sampling point is selected at the midpoint of each arc. The number of sampling points is adaptively determined according to the length and complexity of the coastline, ensuring that the sampling points can fully characterize the shape characteristics of the coastline, while avoiding excessive computational burden caused by too dense sampling points. Through sampling, a uniformly distributed and representative matching coastline sampling point set is obtained, which is ready for the subsequent construction of coastline point pairs.

[0067] Subsequently, the preset coastline data is used to find the nearest point on the preset coastline for each sampling point in the matching coastline sampling point set, and a coastline point pair is constructed. For each sampling point, the Euclidean distance to all points on the preset coastline is calculated, and the point with the smallest distance is found as the corresponding nearest point. The search for the nearest point can utilize efficient data structures and algorithms, such as KD trees, R trees, etc., to speed up the calculation process. In order to improve the robustness of the nearest point search, additional constraints can be introduced. For example, a distance threshold is set. When the distance from the sampling point to the nearest point on the preset coastline exceeds the threshold, it is considered that the sampling point has not found a suitable corresponding point and is marked as an outlier. In addition, the topological structure of the coastline can be used to constrain the selection of the nearest point to ensure that adjacent sampling points also have adjacent nearest points on the preset coastline.

[0068] By establishing a one-to-one correspondence between the matched coastline sampling point set and the preset coastline, multiple pairs of coastline points are obtained. Each pair of coastline points consists of a sampling point on the matched coastline and the closest point on the preset coastline, reflecting the correspondence between the two coastlines in the local area. These coastline point pairs are the basis for the subsequent calculation of positioning errors, and their quality directly affects the reliability of positioning accuracy assessment.

[0069] Furthermore, the embodiment of the present invention also includes:

[0070] Decomposing the positioning error into an along-track component and a cross-track component;

[0071] The mean values ​​of the along-track component and the cross-track component are calculated respectively to obtain the mean value of the along-track positioning error and the mean value of the cross-track positioning error;

[0072] The mean along-track positioning error and the mean across-track positioning error are taken as the verification characterization results.

[0073] In a feasible implementation, the positioning error is characterized from the intuitive aspect of the object space image, and its image positioning error is directly decomposed into two components, along the track and across the track, and used as the output result. The along-track component represents the position deviation of the coastline point pair in the direction of satellite flight, and the across-track component represents the position deviation of the coastline point pair perpendicular to the direction of satellite flight. The decomposition of the positioning error can utilize the orbital parameters and imaging geometry model of the satellite. First, according to the orbital parameters of the satellite, such as orbital altitude, inclination, etc., the projection of the satellite flight direction on the ground is calculated to obtain the unit vector along the track. Then, according to the geographic coordinates of the matching coastline point and the preset coastline point, their connecting vector on the ground is calculated. Finally, the connecting vector is projected onto the along-track unit vector and the direction perpendicular to the along-track unit vector, respectively, to obtain the components of the positioning error along the track and across the track. Then, the along-track error components and across-track error components of all the obtained coastline point pairs are statistically analyzed, and their arithmetic means are calculated to obtain the mean along-track positioning error and the mean across-track positioning error, which reflects the average level of the positioning error along the track and across the track on the entire coastline. Next, the obtained along-track positioning error mean and cross-track positioning error mean are used as the final verification characterization results to evaluate and represent the geolocation accuracy of microwave remote sensing images. These two error means quantitatively describe the positioning accuracy level from the along-track direction and cross-track direction respectively.

[0074] In another feasible implementation of the present invention, from the perspective of the image coordinate system, the positioning error can be decomposed into the rotation error of the instrument around the XYZ axis of its own coordinate system. Among them, the rotation error of the instrument coordinate system around the Y axis is reflected in the overall offset in the X direction of flight, which is manifested as an along-track error; the rotation error of the instrument coordinate system around the X axis is reflected in the overall offset in the Y direction across the track, which is manifested as a cross-track error; the rotation error of the instrument coordinate system around the Z axis has different effects on microwave instruments with different scanning modes. For cross-track scanning microwave instruments, its essence is an installation error, which presents a rotation effect when there is no XY direction error. The along-track and cross-track errors on both sides of the field of view show a symmetrical effect of equal size and opposite sign. The rotation error angle can be determined by the ratio of the field of view edge positioning deviation to the half field of view length; for conical scanning microwave instruments, the error around the Z axis can be equivalent to the instrument scanning timing error, which is an error that the instrument is inconsistent with the laboratory calibration timing under the on-track working condition. In the image footprint, along-track and cross-track deviations coexist, and the size of these two deviation components changes with the scanning angle of the pixel. If the pixel deviation is directly decomposed into the along-track-cross-track coordinate system, the size of its components will change. In order to obtain the true along-track and cross-track deviations, the following transformations need to be made:

[0075] Assume that the pixel deviation number obtained by image matching is (R N ,C N ), and the actual spatial resolution of IFOV along the track and in the scanning direction is expressed as (R L ,C L ). Establish a cross-track-along-track coordinate system, set the scanning angle to β, and write the along-track deviation as a vector A = [R L 0], the scanning direction deviation is written as vector s = [C L sin(β) C L cos(β)], so the actual along-track and across-track deviation vectors are d L =A+S=[R L +C L sin(β) C L cos(β)]. After the above transformation, the along-track and across-track positioning deviations of conical scanning instruments can be expressed uniformly.

[0076] Furthermore, the embodiment of the present invention also includes:

[0077] Acquire microwave remote sensing images at multiple times and generate positioning errors at multiple times;

[0078] Decompose the positioning error at each time into an along-track component and a cross-track component, and obtain the along-track positioning error and cross-track positioning error at each time;

[0079] Based on the along-track positioning error and cross-track positioning error at multiple times, the functional relationship between the along-track positioning error, the cross-track positioning error and the scanning angle is established to obtain the scanning timing error;

[0080] Calculate the mean and standard deviation of along-track positioning error, cross-track positioning error and scanning timing error at multiple times to generate verification characterization results.

[0081] In a preferred embodiment, first, multiple microwave remote sensing images collected at different time points are obtained, and corresponding positioning errors are generated for each image. By processing data from multiple time points, error samples reflecting the temporal variation characteristics of positioning performance are obtained, providing a data basis for subsequent statistical analysis. Then, the positioning error obtained at each time point is decomposed into an along-track component and a cross-track component. By decomposition, the positioning errors in the along-track direction and cross-track direction of each time point are obtained, reflecting the distribution characteristics of the positioning errors in two directions at different time points.

[0082] Afterwards, the multi-time samples of along-track positioning errors and cross-track positioning errors are used to analyze their functional relationship with the scanning angle, and the scanning timing error reflecting the scanning timing characteristics is obtained. The scanning timing error is caused by the difference in observation characteristics at different scanning angles, which reflects the systematic change of positioning error with the scanning angle. After that, for each time, the mean and standard deviation of the along-track positioning error and cross-track positioning error are calculated to obtain the statistics reflecting the positioning accuracy level of each time, generate the verification characterization results, and comprehensively reflect the overall positioning accuracy level of the microwave remote sensing image.

[0083] In summary, the method for verifying the accuracy of geographic positioning based on microwave remote sensing images provided by the embodiment of the present invention has the following technical effects:

[0084] The microwave imager carried by the interactive meteorological satellite obtains microwave remote sensing images as the data basis for subsequent geographic positioning accuracy verification and characterization. In the microwave remote sensing image, the extreme points of temperature change are extracted, and the coastline to be detected is generated based on the extreme points of temperature change, so as to achieve accurate extraction of the coastline and lay the foundation for subsequent coastline matching. The coastline to be detected is matched with the preset coastline to obtain the matching coastline, which provides a basis for the calculation of subsequent positioning errors. Coastline points are extracted on the preset coastline and the matching coastline respectively to form multiple pairs of coastline points for the calculation of subsequent positioning errors. The distances between multiple pairs of coastline points are calculated respectively to generate positioning errors, and verification and characterization results are generated based on the positioning errors, which realizes the quantitative evaluation and characterization of positioning errors and improves the accuracy of geographic positioning accuracy verification and characterization of microwave remote sensing images.

[0085] Embodiment 2

[0086] Based on the same inventive concept as the method for testing and characterizing geographic positioning accuracy based on microwave remote sensing images in the aforementioned embodiment, Figure 2 As shown, an embodiment of the present invention provides a device for testing and characterizing geographic positioning accuracy based on microwave remote sensing images, the device comprising:

[0087] A remote sensing image acquisition unit 11 is used for interacting with a microwave imager carried by a meteorological satellite to acquire a microwave remote sensing image;

[0088] A coastline generation unit 12 to be detected is used to extract temperature extreme value points in the microwave remote sensing image, and generate a coastline to be detected based on the temperature extreme value points;

[0089] A matched coastline acquisition unit 13 is used to match the coastline to be detected with a preset coastline to acquire a matched coastline;

[0090] A coastline point extraction unit 14 is used to extract coastline points on the preset coastline and the matching coastline respectively to form a plurality of pairs of coastline points;

[0091] The verification characterization result generating unit 15 is used to respectively obtain the distances between multiple pairs of coastline points, generate positioning errors, and generate verification characterization results based on the positioning errors.

[0092] Furthermore, the to-be-detected coastline generation unit 12 includes the following execution steps:

[0093] Linearly upsampling the longitude and latitude of the microwave remote sensing image to increase the resolution to 1 km×1 km;

[0094] Surface fitting of the observed brightness temperatures was performed to increase the resolution to the same as longitude and latitude;

[0095] The area with larger gradient is extracted from the fitted brightness temperature surface as the coastline to generate the coastline to be detected.

[0096] Furthermore, the matching coastline acquisition unit 13 includes the following execution steps:

[0097] Extract the coastline points to be detected from the coastline to be detected, and obtain a coastline point set to be detected;

[0098] Find reference coastline points corresponding to the coastline points to be detected from the preset coastline to obtain a reference coastline point set;

[0099] The ICP algorithm is used to match the detected coastline point set with the reference coastline point set, solve the optimal rigid transformation parameters, and obtain the matching coastline.

[0100] Furthermore, the matching coastline acquisition unit 13 further includes the following execution steps:

[0101] Calculate the mean Euclidean distance between each point in the detected coastline point set and the nearest point in the reference coastline point set as the distance error of the current iteration;

[0102] Calculate the geometric center points of the coastline point set to be detected and the reference coastline point set, and calculate the coordinate difference between the coastline point set to be detected and the reference coastline point set relative to their respective geometric center points to obtain decentralized coordinates;

[0103] Construct a matrix based on the decentralized coordinates and perform singular value decomposition to obtain the matrix and vector representing the rotation and translation relationship, and obtain the current rigid transformation parameters;

[0104] The current rigid transformation parameters are used to transform the coastline point set to be detected to obtain an updated coastline point set to be detected, and the nearest point in the reference coastline point set is re-searched and a new distance error is calculated to obtain an updated distance error;

[0105] The iteration is performed, and if the absolute value of the difference between the distance error and the updated distance error is less than a preset threshold or the number of iterations reaches a preset upper limit, the current rigid transformation parameter is used as the optimal rigid transformation parameter.

[0106] Furthermore, the coastline point extraction unit 14 includes the following execution steps:

[0107] uniformly selecting a plurality of sampling points on the matching coastline to obtain a matching coastline sampling point set;

[0108] The nearest point from each sampling point in the matching coastline sampling point set to the preset coastline is calculated to form multiple pairs of coastline points, each pair of coastline points including a matching coastline sampling point and a nearest point on the preset coastline.

[0109] Furthermore, the verification characterization result generating unit 15 includes the following execution steps:

[0110] Decomposing the positioning error into an along-track component and a cross-track component;

[0111] The mean values ​​of the along-track component and the cross-track component are calculated respectively to obtain the mean value of the along-track positioning error and the mean value of the cross-track positioning error;

[0112] The mean along-track positioning error and the mean across-track positioning error are taken as the verification characterization results.

[0113] Furthermore, the verification characterization result generating unit 15 further includes the following execution steps:

[0114] Acquire microwave remote sensing images at multiple times and generate positioning errors at multiple times;

[0115] Decompose the positioning error at each time into an along-track component and a cross-track component, and obtain the along-track positioning error and cross-track positioning error at each time;

[0116] Based on the along-track positioning error and cross-track positioning error at multiple times, the functional relationship between the along-track positioning error, the cross-track positioning error and the scanning angle is established to obtain the scanning timing error;

[0117] Calculate the mean and standard deviation of along-track positioning error, cross-track positioning error and scanning timing error at multiple times to generate verification characterization results.

[0118] Any steps of the method described above may be stored as computer instructions or programs in an unlimited computer memory, and may be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present invention, without further limitation.

[0119] Furthermore, the first or second mentioned above may not only represent an order relationship, but may also represent a specific concept, and / or refer to multiple elements that can be selected individually or in full. Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is intended to include these modifications and variations.

Claims

1. A method for testing the accuracy of geographic positioning based on microwave remote sensing images, characterized in that: The method comprises: The microwave imager carried by the interactive meteorological satellite obtains microwave remote sensing images; In the microwave remote sensing image, extreme temperature change points are extracted, and a coastline to be detected is generated based on the extreme temperature change points; Matching the coastline to be detected with a preset coastline to obtain a matching coastline; Extracting coastline points on the preset coastline and the matching coastline respectively to form multiple pairs of coastline points; The distances between multiple pairs of coastline points are respectively calculated to generate positioning errors, and a verification characterization result is generated based on the positioning errors.

2. The method according to claim 1, characterized in that Extracting temperature extreme value points in the microwave remote sensing image, and generating a coastline to be detected based on the temperature extreme value points, including: Linearly upsampling the longitude and latitude of the microwave remote sensing image to increase the resolution to 1 km×1 km; Surface fitting of the observed brightness temperatures was performed to increase the resolution to the same as longitude and latitude; The area with larger gradient is extracted from the fitted brightness temperature surface as the coastline to generate the coastline to be detected.

3. The method according to claim 1, characterized in that Matching the coastline to be detected with a preset coastline to obtain a matching coastline, including: Extract the coastline points to be detected from the coastline to be detected, and obtain a coastline point set to be detected; Find reference coastline points corresponding to the coastline points to be detected from the preset coastline to obtain a reference coastline point set; The ICP algorithm is used to match the detected coastline point set with the reference coastline point set, solve the optimal rigid transformation parameters, and obtain the matching coastline.

4. The method according to claim 3, characterized in that The ICP algorithm is used to match the detected coastline point set with the reference coastline point set, solve the optimal rigid transformation parameters, and obtain the matching coastline, including: Calculate the mean Euclidean distance between each point in the detected coastline point set and the nearest point in the reference coastline point set as the distance error of the current iteration; Calculate the geometric center points of the coastline point set to be detected and the reference coastline point set, and calculate the coordinate difference between the coastline point set to be detected and the reference coastline point set relative to their respective geometric center points to obtain decentralized coordinates; Construct a matrix based on the decentralized coordinates and perform singular value decomposition to obtain the matrix and vector representing the rotation and translation relationship, and obtain the current rigid transformation parameters; The current rigid transformation parameters are used to transform the coastline point set to be detected to obtain an updated coastline point set to be detected, and the nearest point in the reference coastline point set is re-searched and a new distance error is calculated to obtain an updated distance error; The iteration is performed, and if the absolute value of the difference between the distance error and the updated distance error is less than a preset threshold or the number of iterations reaches a preset upper limit, the current rigid transformation parameter is used as the optimal rigid transformation parameter.

5. The method according to claim 1, characterized in that Coastline points are extracted respectively on the preset coastline and the matching coastline to form multiple pairs of coastline points, including: uniformly selecting a plurality of sampling points on the matching coastline to obtain a matching coastline sampling point set; The nearest point from each sampling point in the matching coastline sampling point set to the preset coastline is calculated to form multiple pairs of coastline points, each pair of coastline points including a matching coastline sampling point and a nearest point on the preset coastline.

6. The method according to claim 1, characterized in that Generating a verification characterization result based on the positioning error includes: Decomposing the positioning error into an along-track component and a cross-track component; The mean values ​​of the along-track component and the cross-track component are calculated respectively to obtain the mean value of the along-track positioning error and the mean value of the cross-track positioning error; The mean along-track positioning error and the mean across-track positioning error are taken as the verification characterization results.

7. The method according to claim 1, characterized in that The method further comprises: Acquire microwave remote sensing images at multiple times and generate positioning errors at multiple times; Decompose the positioning error at each time into an along-track component and a cross-track component, and obtain the along-track positioning error and cross-track positioning error at each time; Based on the along-track positioning error and cross-track positioning error at multiple times, the functional relationship between the along-track positioning error, the cross-track positioning error and the scanning angle is established to obtain the scanning timing error; Calculate the mean and standard deviation of along-track positioning error, cross-track positioning error and scanning timing error at multiple times to generate verification characterization results.

8. A device for testing and characterizing geographic positioning accuracy based on microwave remote sensing images, characterized in that: The device comprises: A remote sensing image acquisition unit, wherein the remote sensing image acquisition unit is used to interact with a microwave imager carried by a meteorological satellite to acquire a microwave remote sensing image; A coastline generation unit to be detected, the coastline generation unit to be detected is used to extract temperature change extreme value points in the microwave remote sensing image, and generate the coastline to be detected based on the temperature change extreme value points; A matching coastline acquisition unit, the matching coastline acquisition unit is used to match the coastline to be detected with a preset coastline to acquire a matching coastline; A coastline point extraction unit, the coastline point extraction unit is used to extract coastline points on the preset coastline and the matching coastline respectively, to form a plurality of pairs of coastline points; A verification characterization result generating unit is used to respectively obtain the distances between a plurality of pairs of coastline points, generate positioning errors, and generate a verification characterization result based on the positioning errors.

Images