A method and system for calculating the accuracy of building mapping based on contour matching

Abstract

This invention provides a method and system for calculating the accuracy of building mapping based on contour matching. The method includes: obtaining an initial set of matching inflection points using a contour inflection point matching algorithm; firstly, removing obvious mismatched inflection points from the initial set; then removing non-unique mismatched inflection points to obtain a preset high-precision matching inflection point set. Unmatched high-precision matching inflection points are selected and added to the preset high-precision matching inflection point set to obtain a final set of matching inflection points. Based on the matching inflection points in the final set, two contours are divided into sets of matching edges with an equal number of edges. The distance between each set of matching edges is calculated using Riemann integrals, and the weighted average distance is calculated based on the average length of the matching edges to obtain the final average distance between the two building contours. This invention calculates the distance between the mapped contour of a building and a reference contour through contour matching and the average distance of matching edges, which is more scientific and accurate than traditional mapping accuracy assessment methods.

Description

Technical Field

[0001] This invention relates to the field of mapping calculation technology, and in particular to a method and system for calculating the accuracy of building mapping based on contour matching. Background Technology

[0002] When surveying buildings, it is usually necessary to measure the error between the obtained surveyed outline and the reference outline in order to determine the quality of the surveyed outline.

[0003] Traditional methods for assessing the accuracy of building contours involve selecting several matching inflection points and calculating the mean square error of the distances between these points as the final assessment result. This method only considers the distances between points, neglecting the positional relationships between line segments. When the distances between endpoints are equal, the distance between two line segments at their intersection is less than the distance between them at their separation, which affects the accuracy of the building contour assessment. Furthermore, since the number of inflection points often differs between automated mapping results and actual reference contour results, the inflection points in the mapped contour and the reference contour are not always in a one-to-one relationship; there may also be "one-to-zero," "one-to-many," or "many-to-one" relationships (e.g., ...). Figure 1 In such cases, traditional methods for assessing the accuracy of building contours struggle to select the optimal corresponding inflection points for calculating contour accuracy. Furthermore, inaccurate inflection point matching frequently occurs in AI-based contour mapping, leading to inaccurate results from traditional methods.

[0004] Existing methods for evaluating the accuracy of AI mapping results are based on the intersection-over-union (IoU) ratio of the regions enclosed by two contours. The problem with this method is that it is an indirect calculation method and cannot directly calculate the distance between the two contours, making IoU an inaccurate evaluation criterion.

[0005] The IoU (Inversion of Union) evaluation method focuses on the overlapping area of ​​two polygons or contours, ignoring the overlap pattern and the intersection of the extracted building vector contour with the true vector contour. Therefore, even with the same IoU value, the average distance between the two contours can differ in many cases. For example, ① when two contours overlap, the IoU evaluation method cannot accurately reflect the accuracy of the building mapping indirectly: from Figure 2It is evident from the data that the average distance between the two contours in the three sets of data are not the same, and d2>d3>d1. In this case, the Intersection over Union (IoU) evaluation method cannot accurately reflect the average distance between the two contours, meaning that the IoU evaluation method cannot accurately express the mapping accuracy of the building. Secondly, when the two contours have no overlap, the IoU evaluation method cannot accurately and indirectly reflect the mapping accuracy of the building: when the two contours are separated, regardless of their distance, the calculated IoU value is the same (IoU value is 0), but the distance between the two contours is different in this case. Boundary IoU basically overcomes this problem and can effectively evaluate the segmentation quality of different segmentation algorithms for object boundaries (effectively evaluating the quality of building edge extraction). However, Boundary IoU can still only perform indirect AI mapping accuracy evaluation; it cannot calculate the mapping error distance or mean square error value. Furthermore, the evaluation result of Boundary IoU is highly dependent on the selection of the threshold d, making it generally difficult to automatically select a suitable threshold during accuracy assessment.

[0006] Therefore, in the above situations, neither the IoU evaluation method nor its improved algorithm can accurately calculate the average distance between the mapped profile and the reference profile, and it is difficult to accurately reflect the mapping accuracy of the building. Summary of the Invention

[0007] This invention provides a method and system for calculating the accuracy of building mapping based on contour matching, in order to solve the deficiency in the prior art of lacking a high accuracy calculation method for building mapping.

[0008] In a first aspect, the present invention provides a method for calculating the accuracy of building mapping based on contour matching, comprising:

[0009] Obtain the mapping outline and reference outline of the building to be measured, and use the inflection point matching algorithm to match the mapping outline and reference outline to generate an initial set of matching inflection points;

[0010] An obvious mismatch inflection point removal algorithm is used to remove obvious mismatch inflection points from the initial set of matching inflection points, resulting in a set of matching inflection points after removing obvious mismatch inflection points.

[0011] A point is considered a clear mismatch inflection point if any one of the following three conditions is met:

[0012] (1) The absolute distance between the same matching inflection point pair exceeds M pixels, where M is generally 30-200 pixels, such as 40, 45, 50, 60, etc., and the preferred value in this embodiment of the invention is 45;

[0013] (2) The average distance between matching inflection point pairs whose distance exceeds N times that of matching inflection point pairs, where N is generally 1.5-5, such as 1.5, 2, 3, 4, etc., and is preferably 3 in this embodiment of the invention;

[0014] (3) The angle between the vectors from the inflection point of the same matching inflection point to the centroid of the closed figure enclosed by the contour where it is located exceeds α, and α is generally 0.5-1.5 (radians). In the embodiment of the present invention, 1 is preferred.

[0015] A non-unique mismatch inflection point removal algorithm is used to remove non-unique mismatch inflection points from the set of matching inflection points after removing obvious mismatch inflection points, thereby obtaining a preset high-precision matching inflection point set.

[0016] A matching inflection point addition algorithm is used to add missing matching inflection points to the preset high-precision matching inflection point set to obtain the final matching inflection point set;

[0017] Based on the final set of matching inflection points, the mapping contour and the reference contour are divided into an equal number of edges. The equal number of edges are matched to obtain a set of matching edges. Based on the Riemann integral method for calculating the distance between matching edges, the average distance between each set of matching edges in the set of matching edges is obtained.

[0018] The average distance between each set of matching edges is calculated by weighting the average distance based on the average length of the matching edges, and the weighted average distance is used as the distance between the mapped contour and the reference contour.

[0019] According to the present invention, a method for calculating the accuracy of building mapping based on contour matching is provided, which involves obtaining the mapping contour and reference contour of the building to be measured, and using an inflection point matching algorithm to match the mapping contour and reference contour to generate an initial set of matching inflection points, including:

[0020] The mapping contour is extracted using artificial intelligence algorithms or other automated mapping methods, and the reference contour is obtained using traditional mapping methods or artificial intelligence sample annotation methods.

[0021] Determine the first set of inflection points in the mapped contour and the second set of inflection points in the reference contour;

[0022] Calculate the Euclidean distance between each inflection point in the first inflection point set and each inflection point in the second inflection point set in sequence;

[0023] A distance matrix is ​​constructed based on the Euclidean distances between all inflection points. This distance matrix is ​​used as the initial matching inflection point set. The number of rows in the distance matrix is ​​equal to the number of inflection points in the first inflection point set, and the number of columns in the distance matrix is ​​equal to the number of inflection points in the second inflection point set.

[0024] According to the present invention, a method for calculating the accuracy of building mapping based on contour matching employs an obvious mismatch inflection point removal algorithm to remove obvious mismatch inflection points from the initial set of matching inflection points, resulting in a set of matching inflection points after removing obvious mismatch inflection points, including:

[0025] If it is determined that the absolute distance between the same pair of matching inflection points in the initial set of matching inflection points exceeds a preset number of pixels, then the same pair of matching inflection points is removed.

[0026] Alternatively, if it is determined that the distance between the same pair of matching inflection points in the initial set of matching inflection points exceeds a preset multiple of the average distance between the pairs of matching inflection points, then the same pair of matching inflection points is removed.

[0027] Alternatively, if it is determined that the angle between the vectors of the inflection points of the same matching inflection point pair in the initial matching inflection point set and the centroid of the closed shape enclosed by the contours where the inflection points of the same matching inflection point pair are located exceeds a preset angle, then the same matching inflection point pair is discarded.

[0028] According to the present invention, a method for calculating the accuracy of building mapping based on contour matching employs a non-unique mismatch inflection point removal algorithm to remove non-unique mismatch inflection points from the set of matching inflection points after removing obvious mismatch inflection points, thereby obtaining a preset high-precision matching inflection point set, including:

[0029] Initialize the high-precision matching inflection point set, and set the preset high-precision matching inflection point set to an empty set;

[0030] The matching inflection points in the set of matching inflection points after removing obvious mismatches are extracted in descending order of the distance between the inflection points. If it is determined that the extracted inflection point is different from all the matching inflection points in the set of matching inflection points after removing obvious mismatches, the extracted inflection point is added to the preset high-precision matching inflection point set.

[0031] According to the present invention, a method for calculating the accuracy of building mapping based on contour matching is provided. This method employs a matching inflection point addition algorithm to add missed matching inflection points to the accuracy matching inflection point set, resulting in a final matching inflection point set, including:

[0032] The inflection points in the preset high-precision matching inflection point set divide the mapped contour and the reference contour into one-to-one corresponding matching edges.

[0033] If there are no matching inflection points among the matching edges on the reference contour, they are taken as inflection points to be filtered. If it is determined that the ratio of the length of any two inflection points to the corresponding vertex of their respective edges to the length of their respective edges is less than a preset ratio, then the two inflection points are added to the set of matching inflection points to be filtered.

[0034] Add the shortest matching inflection point in the set of matching inflection points to be filtered to the set of matching inflection points to be added;

[0035] The set of matching inflection points to be added is filtered based on the non-unique matching inflection point elimination algorithm to obtain the final set of matching inflection points to be added.

[0036] The final set of added matching inflection points is merged with the preset high-precision matching inflection point set to obtain the final set of matching inflection points.

[0037] According to the present invention, a method for calculating the accuracy of building mapping based on contour matching is provided. Based on the final set of matching inflection points, the mapping contour and the reference contour are divided into an equal number of edges. The equal number of edges are matched to obtain a set of matching edges. A method for calculating the distance between matching edges based on Riemann integrals is used to obtain the average distance between each group of matching edges in the set of matching edges, including:

[0038] In the survey contour and the reference contour, determine any two corresponding sides, obtain the two endpoints of the first corresponding side and the two endpoints of the second corresponding side, determine a first arbitrary point between the two endpoints of the first corresponding side, and determine a second arbitrary point between the two endpoints of the second corresponding side.

[0039] The line connecting the two endpoints of the first corresponding edge and the two endpoints of the second corresponding edge, as well as the line connecting the two endpoints of the first corresponding edge to the first arbitrary point and the line connecting the two endpoints of the second corresponding edge to the second arbitrary point, are used to obtain the line ratio between 0 and 1.

[0040] Determine the coordinates of the first arbitrary point and the second arbitrary point, and establish a distance calculation function from the first arbitrary point to the second arbitrary point based on the coordinates of the first arbitrary point, the coordinates of the second arbitrary point, and the ratio of the connecting lines.

[0041] The average of the distance calculation function with respect to the ratio of the connecting lines is taken as the distance between any two corresponding sides;

[0042] If any two corresponding edges are line segments, then the distance calculation function is used;

[0043] If any two corresponding edges are a polyline and a line segment, or a polyline and a polyline, then multiple matching inflection points are added between the two endpoints of the first corresponding edge and the two endpoints of the second corresponding edge, and multiple segment distance calculation functions and multiple segment distance weights are obtained. The multiple segment distance calculation functions and the multiple segment distance weights are used to form a comprehensive distance calculation function.

[0044] According to the present invention, a method for calculating the accuracy of building mapping based on contour matching is provided, wherein a weighted average distance is calculated based on the average length of the matching edges for the average distance between each set of matching edges, and the weighted average distance is used as the distance between the mapped contour and the reference contour, comprising:

[0045] The matched inflection points in the measured contour and the reference contour are divided into several groups of matching edges, and the distance between each group of matching edges and the weight corresponding to each group of matching edges are obtained.

[0046] The distance between the mapped contour and the reference contour is obtained by weighted summation of the distance and the weight.

[0047] Secondly, the present invention also provides a building mapping accuracy calculation system based on contour matching, comprising:

[0048] The initial matching module is used to obtain the mapping outline and reference outline of the building to be measured, and to use the inflection point matching algorithm to match the mapping outline and reference outline to generate an initial matching inflection point set;

[0049] The mismatch removal module is used to remove obvious mismatch inflection points from the initial set of matching inflection points using an obvious mismatch inflection point removal algorithm, to obtain a set of matching inflection points after removing obvious mismatch inflection points; and to remove non-unique mismatch inflection points from the set of matching inflection points after removing obvious mismatch inflection points using a non-unique mismatch inflection point removal algorithm, to obtain a preset high-precision set of matching inflection points.

[0050] The matching addition module is used to add missing matching inflection points to the preset high-precision matching inflection point set by employing a matching inflection point addition algorithm, thereby obtaining the final matching inflection point set.

[0051] The edge matching module is used to divide the mapping contour and the reference contour into an equal number of edges according to the final matching inflection point set, match the equal number of edges to obtain a matching edge set, and obtain the average distance between each group of matching edges in the matching edge set based on the matching edge distance calculation method of Riemann integral.

[0052] The distance calculation module is used to calculate the weighted average distance between each set of matching edges according to the average length of the matching edges, and use the weighted average distance as the distance between the mapped contour and the reference contour.

[0053] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the building mapping accuracy calculation method based on contour matching as described above.

[0054] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the building mapping accuracy calculation method based on contour matching as described above.

[0055] The method and system for calculating building mapping accuracy based on contour matching provided by this invention are mainly used to objectively and accurately evaluate the mapping accuracy of existing or future automated mapping methods and software. That is, by using one or more standard datasets (containing reference contours), the extraction results of existing automated mapping methods are first tested, and then the mapping accuracy is objectively and accurately evaluated using the method described in this invention. Compared with traditional mapping accuracy evaluation methods and IOU and its derivative accuracy evaluation methods in the field of artificial intelligence, this method is more scientific and accurate. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0057] Figure 1 This is a schematic diagram illustrating the inconsistencies in the number of inflection points, non-parallelism of line segments, and intersection of line segments in the building reference outlines and their automated mapping results (automated outline extraction) provided by existing technologies.

[0058] Figure 2 This is a schematic diagram showing the inconsistencies in the number of inflection points, non-parallelism of line segments, and intersection of line segments between the complex building outlines (reference outlines) provided by traditional mapping methods and the building outlines extracted by automated mapping methods (extracted outlines).

[0059] Figure 3 This is one of the flowcharts illustrating the method for calculating the accuracy of building mapping based on contour matching provided by the present invention;

[0060] Figure 4 This is the second flowchart illustrating the method for calculating the accuracy of building mapping based on contour matching provided by the present invention.

[0061] Figure 5 This is a flowchart of a method for calculating mapping accuracy based on the results obtained from an automated roof extraction method for buildings, provided by the present invention.

[0062] Figure 6 This is a flowchart of the automatic calculation method for mapping accuracy between two vector contours provided by the present invention;

[0063] Figure 7 This invention provides a graph of matching inflection point pairs and matching edge pairs.

[0064] Figure 8 This is a schematic diagram of the elimination of non-unique mismatched inflection points provided by the present invention (when performing contour inflection point matching between the measured contour and the reference contour, there may be "one-to-many" or "many-to-one" situations in the inflection point matching results).

[0065] Figure 9 This is a schematic diagram of the matching between the building mapping outline AB and the reference outline CD provided by the present invention;

[0066] Figure 10 This is a schematic diagram of the matching between line segment AB and polyline CF (CDEF) provided by the present invention;

[0067] Figure 11 This is a schematic diagram of the matching between the polyline AC (ABC) and the polyline DG (DEFG) provided by the present invention;

[0068] Figure 12 This is a schematic diagram of the building mapping accuracy calculation system based on contour matching provided by the present invention;

[0069] Figure 13 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0071] Figure 3 This is one of the flowcharts illustrating the method for calculating building mapping accuracy based on contour matching provided in this embodiment of the invention. Figure 3 As shown, it includes:

[0072] Step 100: Obtain the mapping outline and reference outline of the building to be measured, and use the inflection point matching algorithm to match the mapping outline and reference outline to generate an initial set of matching inflection points;

[0073] Step 200: Use the obvious mismatch inflection point removal algorithm to remove obvious mismatch inflection points from the initial matching inflection point set to obtain the set of inflection points without mismatches;

[0074] Step 300: Use a non-unique mismatch inflection point removal algorithm to remove non-unique matching inflection points from the set of mismatch inflection points to obtain a preset high-precision matching inflection point set;

[0075] Step 400: Using the matching inflection point addition algorithm, add the missing matching inflection points to the preset high-precision matching inflection point set to obtain the final matching inflection point set;

[0076] Step 500: Based on the final set of matching inflection points, divide the mapping contour and the reference contour into an equal number of edges, match the equal number of edges to obtain a set of matching edges, and use the Riemann integral-based method for calculating the distance between matching edges to obtain the average distance between each set of matching edges in the set of matching edges.

[0077] Step 600: Calculate the weighted average distance between the average distances of each set of matching edges according to the average length of the matching edges, and use the weighted average distance as the distance between the mapped contour and the reference contour.

[0078] Specifically, such as Figure 4 As shown, this embodiment of the invention first uses an inflection point matching algorithm to establish a matching relationship between the inflection points of the building's mapped outline and the reference outline, generating an initial set of matched inflection points; then, it employs an obvious mismatched inflection point removal algorithm based on the distance between matched inflection points and the rotation angle difference between matched inflection points to remove obvious mismatched inflection points from the initial set of matched inflection points, obtaining a set of matched inflection points after removing obvious mismatched inflection points; finally, it employs a "one-to-many" and "many-to-one" matched inflection point removal algorithm based on the distance between matched inflection points to remove "one-to-many" and "many-to-one" matched inflection points from the set of matched inflection points after removing obvious mismatched inflection points, generating a high-precision "one-to-one" matched inflection point set; and then employs... The matching inflection point addition algorithm based on high-precision matching inflection points adds correct matching inflection points that were not in the initial matching inflection point set to obtain the final matching inflection point set. Based on the matching inflection points in the final matching inflection point set, the building mapping contour and the reference contour are divided into an equal number of edges, and the edges are matched one by one according to the correspondence to obtain the matching edge set. The matching edge distance calculation method based on Riemann integral is used to obtain the average distance between each group of matching edges in the matching edge set. The distance between the matching edges is calculated into a weighted average distance according to the average length of the matching edges, and the calculated weighted average distance is used as the distance between the mapping contour and the reference contour.

[0079] The purpose and function of the method described in this invention is mainly to objectively and accurately evaluate the mapping accuracy of existing or future automated mapping methods and software. That is, by using one or more standard datasets (containing reference contours), the extraction results of existing automated mapping methods are first tested, and then the mapping accuracy is objectively and accurately evaluated using the method described in this invention. Compared with traditional mapping accuracy evaluation methods and IOU and its derivative accuracy evaluation methods in the field of artificial intelligence, this method is more scientific and accurate.

[0080] It should be noted that assessing the mapping accuracy of results obtained from automated roof extraction methods typically requires two steps: data preprocessing and mapping accuracy evaluation. Figure 5 As shown:

[0081] When performing data preprocessing, the first step for building rooftop patches extracted by artificial intelligence algorithms generally requires binarization, vectorization, and regularization (e.g., Douglas-Plück DP algorithm and its improved algorithms), as well as data preprocessing work such as extracting building outline inflection points; the second step for building rooftop vector outlines extracted by traditional methods generally requires regularization and data preprocessing work such as extracting building outline inflection points.

[0082] When evaluating mapping accuracy, the automatic calculation method for building mapping accuracy based on contour matching and Riemann integral in this embodiment of the invention generally requires the following steps when automatically calculating the mapping accuracy between two input vectorized contours (automatically extracted and preprocessed vectorized contour of the building roof and reference vector contour) (e.g. Figure 6 and Figure 7 As shown, except for the graph corresponding to the matching edge, the blue outline in the graph is the outline of the real building, the red outline is the outline of the building in the projection, the two points connected by the green dashed line are a pair of matching inflection points, and the matching edges in the graph corresponding to the matching edge have the same color.

[0083] Based on the above embodiments, step 100 includes:

[0084] The mapping contour is extracted using artificial intelligence algorithms or other automated mapping methods, and the reference contour is obtained using traditional mapping methods or artificial intelligence sample annotation methods.

[0085] Determine the first set of inflection points in the mapped contour and the second set of inflection points in the reference contour;

[0086] Calculate the Euclidean distance between each inflection point in the first inflection point set and each inflection point in the second inflection point set in sequence;

[0087] A distance matrix is ​​constructed based on the Euclidean distances between all inflection points. This distance matrix is ​​used as the initial matching inflection point set. The number of rows in the distance matrix is ​​equal to the number of inflection points in the first inflection point set, and the number of columns in the distance matrix is ​​equal to the number of inflection points in the second inflection point set.

[0088] Specifically, in this embodiment of the invention, the measured contour and the reference contour are matched based on an inflection point matching algorithm to generate an initial set of matching inflection points:

[0089] The inflection points of the extracted contour S and the reference contour T are s and s respectively. i (i = 1, 2, ..., m) and t i (i = 1, 2, ..., n), s i t i Euclidean distance d(s) i ,t i That is, the distance of all corresponding inflection points can be represented as an m×n distance matrix, where the rows of the matrix represent the inflection points of the extracted contour S and the columns represent the inflection points of the reference contour T.

[0090] 1) This algorithm utilizes the basic framework of approximate dynamic programming, relying on a dynamic programming (DP) table. A path in the DP table can be represented as {(m,1),(m-1,2),...,(1,n)}, with its starting point at the bottom edge. It moves in a certain direction (left or right) towards the cell with the smallest value in the DP table, moving only one cell at a time until the vertex is reached and the row ends.

[0091] 2) The row and column number (i,j) of each cell traversed by this route corresponds to the matching of two contours and one inflection point (s i ,t i The optimal matching of two contour inflection points involves finding a path from the bottom row to the top row in a dynamic programming table that minimizes the sum of the cell values ​​it passes through. This path can extend horizontally or vertically, representing a one-to-many or many-to-one relationship. Multiple values ​​in the same row indicate the extraction of an inflection point s of contour S. i For multiple inflection points of the reference contour T; multiple values ​​in the same column indicate that multiple inflection points of the extracted contour S correspond to one inflection point t of the reference contour T. i .

[0092] 3) Based on the final dynamic programming table, convert it into a set of matching inflection point pairs {(s i ,t j )}, and use it as the initial match.

[0093] Based on the above embodiments, step 200 includes:

[0094] If it is determined that the absolute distance between the same pair of matching inflection points in the initial set of matching inflection points exceeds a preset number of pixels, then the same pair of matching inflection points is removed.

[0095] Alternatively, if it is determined that the distance between the same pair of matching inflection points in the initial set of matching inflection points exceeds a preset multiple of the average distance between the pairs of matching inflection points, then the same pair of matching inflection points is removed.

[0096] Alternatively, if it is determined that the angle between the vectors of the inflection points of the same matching inflection point pair in the initial matching inflection point set and the centroid of the closed shape enclosed by the contours where the inflection points of the same matching inflection point pair are located exceeds a preset angle, then the same matching inflection point pair is discarded.

[0097] Specifically, in this embodiment of the invention, if a matching inflection point satisfies any one of the following three conditions, the match is considered a clear mismatch and is removed from the initial matching set, finally obtaining the matching inflection point set after removing clear mismatch inflection points:

[0098] (1) The absolute distance between the same matching inflection point pair exceeds M pixels, where M is generally 30-200 pixels, such as 40, 45, 50, 60, etc., and the preferred value in this embodiment of the invention is 45;

[0099] (2) The average distance between matching inflection point pairs whose distance exceeds N times that of matching inflection point pairs, where N is generally 1.5-5, such as 1.5, 2, 3, 4, etc., and is preferably 3 in this embodiment of the invention;

[0100] (3) The angle between the vectors from the inflection point of the same matching inflection point to the centroid of the closed figure enclosed by the contour where it is located exceeds α, and α is generally 0.5-1.5 (radians). In the embodiment of the present invention, 1 is preferred.

[0101] Based on the above embodiments, step 300 includes:

[0102] Initialize a preset high-precision matching inflection point set, and set the preset high-precision matching inflection point set to an empty set;

[0103] The matching inflection points in the set of matching inflection points after removing obvious mismatches are extracted in descending order of the distance between the inflection points. If it is determined that the extracted inflection point is different from all the matching inflection points in the set of matching inflection points after removing obvious mismatches, the extracted inflection point is added to the preset high-precision matching inflection point set.

[0104] Specifically, in this embodiment of the invention, the matching inflection point elimination algorithm based on the distance between matching inflection points first emptys the high-precision matching inflection point set, then takes out the matching inflection points in the mismatched inflection point set in descending order of the distance between the inflection points, and then determines whether the inflection points in the taken matching inflection points have the same inflection points as the inflection points of all matching inflection points in the mismatched inflection point set. If there are no same inflection points, the taken matching inflection points are added to the high-precision matching inflection point set.

[0105] Eliminate according to the above approach Figure 8 The problem involves "one-to-many" and "many-to-one" matching inflection point pairs. Since only ac, bc, and bd share a common inflection point, all other matches besides these three are returned to set A after being extracted, regardless of the extraction order. Therefore, we only need to discuss the case of these three matches. d has the largest distance and is extracted first. It shares an inflection point b with bc, so d is removed. Then bc is extracted, and it shares an inflection point c with ac, so bc is removed. Finally, ac is extracted, and it has no common inflection point with any match in set A, so ac is returned to set A. In the end, ac is retained out of the three matches ac, bc, and bd.

[0106] Based on the above embodiments, step 400 includes:

[0107] The inflection points in the preset high-precision matching inflection point set divide the mapped contour and the reference contour into one-to-one corresponding matching edges.

[0108] If there are no matching inflection points among the matching edges on the reference contour, they are taken as inflection points to be filtered. If it is determined that the ratio of the length of any two inflection points to the corresponding vertex of their respective edges to the length of their respective edges is less than a preset ratio, then the two inflection points are added to the set of matching inflection points to be filtered.

[0109] Add the shortest matching inflection point in the set of matching inflection points to be filtered to the set of matching inflection points to be added;

[0110] The set of matching inflection points to be added is filtered based on the non-unique matching inflection point elimination algorithm to obtain the final set of matching inflection points to be added.

[0111] The final set of added matching inflection points is merged with the preset high-precision matching inflection point set to obtain the final set of matching inflection points.

[0112] Specifically, in this embodiment of the invention, two contours are divided into one-to-one matching edges based on the high-precision matching inflection points in the high-precision matching inflection point set.

[0113] For inflection points on the reference contour that do not match, perform the following operation:

[0114] 1) Traverse the unmatched inflection points on the matching edges of the inflection points as inflection points to be filtered. If the ratio of the length of the two inflection points to the corresponding vertex of their respective edges to the length of the two edges is less than P, add the matching inflection point to the set of matching inflection points to be filtered. P is generally 1%-5%, preferably 3%.

[0115] 2) Add the shortest matching inflection point in the set of matching inflection points to be filtered to the set of matching inflection points to be added;

[0116] The matching inflection point set to be added is removed using "one-to-many" and "many-to-one" matching inflection point removal algorithms based on the distance between matching inflection points, resulting in the final set of added matching inflection points.

[0117] The final set of matching inflection points is merged with the high-precision set of matching inflection points to obtain the final set of matching inflection points.

[0118] Based on the above embodiments, step 500 includes:

[0119] In the survey contour and the reference contour, determine any two corresponding sides, obtain the two endpoints of the first corresponding side and the two endpoints of the second corresponding side, determine a first arbitrary point between the two endpoints of the first corresponding side, and determine a second arbitrary point between the two endpoints of the second corresponding side.

[0120] The line connecting the two endpoints of the first corresponding edge and the two endpoints of the second corresponding edge, as well as the line connecting the two endpoints of the first corresponding edge to the first arbitrary point and the line connecting the two endpoints of the second corresponding edge to the second arbitrary point, are used to obtain the line ratio between 0 and 1.

[0121] Determine the coordinates of the first arbitrary point and the second arbitrary point, and establish a distance calculation function from the first arbitrary point to the second arbitrary point based on the coordinates of the first arbitrary point, the coordinates of the second arbitrary point, and the ratio of the connecting lines.

[0122] The average of the distance calculation function with respect to the ratio of the connecting lines is taken as the distance between any two corresponding sides;

[0123] If any two corresponding edges are line segments, then the distance calculation function is used;

[0124] If any two corresponding edges are a polyline and a line segment, or a polyline and a polyline, then multiple matching inflection points are added between the two endpoints of the first corresponding edge and the two endpoints of the second corresponding edge, and multiple segment distance calculation functions and multiple segment distance weights are obtained. The multiple segment distance calculation functions and the multiple segment distance weights are used to form a comprehensive distance calculation function.

[0125] Specifically, in this embodiment of the invention, matching edges are obtained by dividing two vector contours based on matching inflection point pairs, such as... Figure 9 As shown, in the matching inflection point pairs, inflection point A corresponds to inflection point C, and inflection point B corresponds to inflection point D. Therefore, edges AB and CD are corresponding matching edges. Based on the matching inflection point pairs, the survey contour and the reference contour are divided into multiple sets of corresponding edges, and then the distance between each set of matching edges is calculated. Matching edges can be divided into two categories—line segments and polyline segments—based on whether they contain unmatched inflection points. There are three matching relationships: line segment matching with line segment, line segment matching with polyline segment, and polyline segment matching with polyline segment.

[0126] First, the first calculation method is proposed, such as Figure 9 As shown, inflection point A corresponds to inflection point C, and inflection point B corresponds to inflection point D; both sides AB and CD are line segments or broken lines; PQ are the corresponding points on sides AB and CD.

[0127] (1) Take two corresponding points P and Q on the two sides AB and CD, satisfying the equation:

[0128]

[0129] (2) Simplify the expression to get

[0130]

[0131] (3) Let the coordinates of points P and Q be P(x) and Q(q) respectively. p ,y p ), Q(x) Q ,y Q If we can find the parametric equations of points P and Q with respect to t, then the coordinates of points P and Q are both continuous functions of t.

[0132] (4) According to the distance calculation formula, we can obtain:

[0133]

[0134] It satisfies the condition for the existence of a definite integral.

[0135] (5) Find the average of the function d(t) with respect to t, and use it as the distance between the two sides, i.e.:

[0136]

[0137] When the matching edges are line segments matching line segments, the distance function has a simple form:

[0138]

[0139] However, when the matching edge is a line segment and a polyline or a polyline and a polyline, the distance function in Method 1 is a piecewise function, which results in a large error when performing numerical calculations using a computer. To address this issue, this embodiment of the invention proposes Method 2 for the case where the matching edge is a line segment and a polyline or a polyline and a polyline:

[0140] (1) Add matching inflection points to the matching edge of the edge for all unmatched inflection points, and the added matching inflection points should satisfy the ratio of the distances along the edge from the matching inflection point to the matching endpoint of the matching edge is equal to the ratio of the lengths of the edge.

[0141] Figure 10 Conditions for matching the midline segment with the broken line

[0142] Figure 11 Conditions for matching a broken line with a broken line

[0143] (2) By adding inflection points, the two matching edges can be divided into multiple sets of matching segments. Then, the distance between each set of matching segments can be calculated using Method 1.

[0144] Figure 10 There are three sets of corresponding line segments: AD'-CD, D'E'-DE, and E'B-EF. The distance d between the three sets of line segments is calculated using the average distance between line segments. AD’-CD d D’E’-DE and d E’B-EF .

[0145] Figure 11 There are four sets of corresponding line segments: AE'-DE, E′B-EB′, B′F-BF′, and F′C-FG. The distance d between the four sets of line segments is calculated using the average distance between line segments. AE’-DE d E′B-EB′ d B′F-BF′ and d F′C-FG .

[0146] (3) Take the average length of the two line segments in each set of matching line segments as the weight of that set of matching line segments. Take the average distance between each set of line segments calculated in step (2) and calculate the weighted average distance according to the corresponding weight. Take the final average distance as the distance between the two matching edges.

[0147] Figure 10 Weights and results in:

[0148] Let l1=|AB|, l2=|CD|+|DE|+|EF|

[0149] Weight:

[0150] Distance between matching edges: d AB-CDEF

[0151] =f AD’-CD d AD’-CD +f D’E’-DE d D’E’-DE +f E’B-EF d E’B-EF

[0152] Figure 11 Weights and results in:

[0153]

[0154]

[0155]

[0156]

[0157] Result: d ABC-DEFG

[0158] =f AE’-DE d AE’-DE +f E′B-EB′ d E′B-EB′ +f B′F-BF′ d B′F-BF′

[0159] +f F′C-FG d F′C-FG

[0160] It can be seen that the two methods for calculating the distance between matching edges proposed in the embodiments of the present invention are theoretically equivalent. However, when there are too many segments in the piecewise integral in Method 1, the numerical calculation result has a large error or even cannot be calculated. Method 2 realizes the transformation from "matching line segments (polylines) with polylines" to "matching line segments with line segments", effectively solving the problem of inaccurate numerical integral results of piecewise functions.

[0161] Based on the above embodiments, step 600 includes:

[0162] The matched inflection points in the measured contour and the reference contour are divided into several groups of matching edges, and the distance between each group of matching edges and the weight corresponding to each group of matching edges are obtained.

[0163] The distance between the mapped contour and the reference contour is obtained by weighted summation of the distance and the weight.

[0164] Specifically, assume that the two contours are divided into n groups of matching edges by the matching inflection points, and the distance between each group of matching edges is d. i(i = 1, 2, ..., n), the weight corresponding to each pair of matching edges (the ratio of the average length of the matching edge to the average perimeter of the two contours, or the ratio of the sum of the lengths of the matching edges to the sum of the perimeters of the two contours) is f. i (i = 1, 2, ..., n), then the distance between the vector contours is:

[0165]

[0166] Furthermore, the above algorithm is applied to all building outlines within a mapping area to calculate the error for each building outline. This yields an error statistical distribution histogram, maximum error, mean error (MAD), and mean square error (RMSE), allowing for a precise assessment of the accuracy of the automated building mapping algorithm. Assume there are n mapping outlines in total, and the error of each building outline is ε. i (i = 1, 2, ..., n), the calculation formula is as follows:

[0167] Maximum error = maxε i

[0168]

[0169]

[0170] This invention obtains the theoretical distance error between the vector contours of two buildings, and further calculates the theoretical mean error of a set of data, thus realizing the theoretical accuracy calculation of an automated building mapping method.

[0171] The following describes the building mapping accuracy calculation system based on contour matching provided by the present invention. The building mapping accuracy calculation system based on contour matching described below and the building mapping accuracy calculation method based on contour matching described above can be referred to in correspondence.

[0172] Figure 12 This is a schematic diagram of the structure of the building mapping accuracy calculation system based on contour matching provided in an embodiment of the present invention, as shown below. Figure 12 As shown, it includes: an initial matching module 1201, a mismatch removal module 1202, a matching addition module 1203, an edge matching module 1204, and a distance calculation module 1205, wherein:

[0173] The initial matching module 1201 is used to acquire the surveyed outline and reference outline of the building to be measured, and to match the surveyed outline and reference outline using an inflection point matching algorithm to generate an initial matching inflection point set; the mismatch removal module 1202 is used to remove obvious mismatch inflection points from the initial matching inflection point set using an obvious mismatch inflection point removal algorithm to obtain a matching inflection point set after removing obvious mismatch inflection points; and to remove non-unique mismatch inflection points from the matching inflection point set after removing obvious mismatch inflection points using a non-unique mismatch inflection point removal algorithm to obtain a preset high-precision matching inflection point set; the matching addition module 1203 is used to add matching inflection points using a matching inflection point addition algorithm. The method involves adding missing matching inflection points to the preset high-precision matching inflection point set to obtain the final matching inflection point set; the edge matching module 1204 is used to divide the mapping contour and the reference contour into an equal number of edges according to the final matching inflection point set, match the equal number of edges to obtain a matching edge set, and obtain the average distance between each group of matching edges in the matching edge set based on the matching edge distance calculation method of Riemann integral; the distance calculation module 1205 is used to calculate the weighted average distance of the average distance between each group of matching edges according to the average length of the matching edges, and use the weighted average distance as the distance between the mapping contour and the reference contour.

[0174] Figure 13 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 13As shown, the electronic device may include a processor 1310, a communication interface 1320, a memory 1330, and a communication bus 1340. The processor 1310, communication interface 1320, and memory 1330 communicate with each other via the communication bus 1340. The processor 1310 can call logical instructions in the memory 1330 to execute a method for calculating the accuracy of building mapping based on contour matching. This method includes: acquiring the mapping contour and reference contour of the building to be measured; matching the mapping contour and reference contour using an inflection point matching algorithm to generate an initial set of matching inflection points; using an obvious mismatch inflection point removal algorithm to remove obvious mismatch inflection points from the initial set of matching inflection points, obtaining a set of matching inflection points after removing obvious mismatch inflection points; and using a non-unique mismatch inflection point removal algorithm to remove non-unique mismatch inflection points from the set of matching inflection points after removing obvious mismatch inflection points, obtaining a preset set of matching inflection points. A high-precision matching inflection point set is obtained. A matching inflection point addition algorithm is used to add missing matching inflection points to the preset high-precision matching inflection point set, resulting in a final matching inflection point set. Based on the final matching inflection point set, the mapping contour and the reference contour are divided into an equal number of edges. These equal number of edges are matched to obtain a matching edge set. A matching edge distance calculation method based on Riemann integrals is used to obtain the average distance between each group of matching edges in the matching edge set. A weighted average distance is calculated for the average distance between each group of matching edges according to the average length of the matching edges. This weighted average distance is used as the distance between the mapping contour and the reference contour.

[0175] Furthermore, the logical instructions in the aforementioned memory 1330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0176] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program implements the method for calculating the accuracy of building mapping based on contour matching provided by the above methods. The method includes: acquiring the mapping contour and reference contour of the building to be measured; matching the mapping contour and reference contour using an inflection point matching algorithm to generate an initial set of matching inflection points; using an obvious mismatching inflection point removal algorithm to remove obvious mismatching inflection points from the initial set of matching inflection points to obtain a set of matching inflection points after removing obvious mismatching inflection points; and using a non-unique mismatching inflection point removal algorithm to remove the matching inflection points after removing obvious mismatching inflection points. Non-unique mismatched inflection points in the point set are identified to obtain a preset high-precision matching inflection point set. A matching inflection point addition algorithm is used to add missed matching inflection points to the preset high-precision matching inflection point set, resulting in a final matching inflection point set. Based on the final matching inflection point set, the mapping contour and the reference contour are divided into an equal number of edges. These equal number of edges are matched to obtain a matching edge set. A matching edge distance calculation method based on Riemann integrals is used to obtain the average distance between each group of matching edges in the matching edge set. A weighted average distance is calculated for the average distance between each group of matching edges according to the average length of the matching edges. This weighted average distance is used as the distance between the mapping contour and the reference contour.

[0177] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0178] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0179] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for calculating the accuracy of building mapping based on contour matching, characterized in that, include: Obtain the mapping outline and reference outline of the building to be measured, and use the inflection point matching algorithm to match the mapping outline and reference outline to generate an initial set of matching inflection points; An obvious mismatch inflection point removal algorithm is used to remove obvious mismatch inflection points from the initial set of matching inflection points, resulting in a set of matching inflection points after removing obvious mismatch inflection points. A non-unique mismatch inflection point elimination algorithm is used to eliminate non-unique mismatch inflection points in the matching inflection point set after removing obvious mismatch inflection points, thereby obtaining a preset high-precision matching inflection point set. A matching inflection point addition algorithm is used to add missing matching inflection points to the preset high-precision matching inflection point set to obtain the final matching inflection point set; Based on the final set of matching inflection points, the mapped contour and the reference contour are divided into an equal number of edges. Matching these equal number of edges yields a set of matching edges. Using a Riemann integral-based method for calculating the matching edge distance, the average distance between each group of matching edges in the set is obtained, including: In the survey contour and the reference contour, determine any two corresponding sides, obtain the two endpoints of the first corresponding side and the two endpoints of the second corresponding side, determine a first arbitrary point between the two endpoints of the first corresponding side, and determine a second arbitrary point between the two endpoints of the second corresponding side. The line connecting the two endpoints of the first corresponding edge and the two endpoints of the second corresponding edge, as well as the line connecting the two endpoints of the first corresponding edge to the first arbitrary point and the line connecting the two endpoints of the second corresponding edge to the second arbitrary point, are used to obtain the line ratio between 0 and 1. Determine the coordinates of the first arbitrary point and the second arbitrary point, and establish a distance calculation function from the first arbitrary point to the second arbitrary point based on the coordinates of the first arbitrary point, the coordinates of the second arbitrary point, and the ratio of the connecting lines. The average of the distance calculation function with respect to the ratio of the connecting lines is taken as the distance between any two corresponding sides; If any two corresponding edges are line segments, then the distance calculation function is used; If any two corresponding edges are a polyline and a line segment, or a polyline and a polyline, then multiple matching inflection points are added between the two endpoints of the first corresponding edge and the two endpoints of the second corresponding edge, and multiple segment distance calculation functions and multiple segment distance weights are obtained. The multiple segment distance calculation functions and the multiple segment distance weights are used to form a comprehensive distance calculation function. The average distance between each set of matching edges is calculated by weighting the average distance based on the average length of the matching edges, and the weighted average distance is used as the distance between the mapped contour and the reference contour.

2. The method for calculating the accuracy of building mapping based on contour matching according to claim 1, characterized in that, Obtain the mapped outline and reference outline of the building to be measured. Use an inflection point matching algorithm to match the mapped outline and reference outline to generate an initial set of matched inflection points, including: The mapping contour is extracted using artificial intelligence algorithms or other automated mapping methods, and the reference contour is obtained using traditional mapping methods or artificial intelligence sample annotation methods. Determine the first set of inflection points in the mapped contour and the second set of inflection points in the reference contour; Calculate the Euclidean distance between each inflection point in the first inflection point set and each inflection point in the second inflection point set in sequence; A distance matrix is ​​constructed based on the Euclidean distances between all inflection points. This distance matrix is ​​used as the initial matching inflection point set. The number of rows in the distance matrix is ​​equal to the number of inflection points in the first inflection point set, and the number of columns in the distance matrix is ​​equal to the number of inflection points in the second inflection point set.

3. The method for calculating the accuracy of building mapping based on contour matching according to claim 1, characterized in that, An obvious mismatch inflection point removal algorithm is used to remove obvious mismatch inflection points from the initial set of matching inflection points, resulting in a set of matching inflection points after removing obvious mismatch inflection points, including: If it is determined that the absolute distance between the same pair of matching inflection points in the initial set of matching inflection points exceeds a preset number of pixels, then the same pair of matching inflection points is removed. Alternatively, if it is determined that the distance between the same pair of matching inflection points in the initial set of matching inflection points exceeds a preset multiple of the average distance between the pairs of matching inflection points, then the same pair of matching inflection points is removed. Alternatively, if it is determined that the angle between the vectors of the inflection points of the same matching inflection point pair in the initial matching inflection point set and the centroid of the closed shape enclosed by the contours where the inflection points of the same matching inflection point pair are located exceeds a preset angle, then the same matching inflection point pair is discarded.

4. The method for calculating the accuracy of building mapping based on contour matching according to claim 1, characterized in that, A non-unique mismatch inflection point removal algorithm is used to remove non-unique mismatch inflection points from the set of matching inflection points after removing obvious mismatch inflection points, resulting in a preset high-precision matching inflection point set, including: Initialize a preset high-precision matching inflection point set, and set the preset high-precision matching inflection point set to an empty set; The matching inflection points in the set of matching inflection points after removing obvious mismatches are extracted in descending order of the distance between the inflection points. If it is determined that the extracted inflection point is different from all the matching inflection points in the set of matching inflection points after removing obvious mismatches, the extracted inflection point is added to the preset high-precision matching inflection point set.

5. The method for calculating the accuracy of building mapping based on contour matching according to claim 1, characterized in that, A matching inflection point addition algorithm is used to add missed matching inflection points to the preset high-precision matching inflection point set, resulting in a final matching inflection point set, including: The inflection points in the preset high-precision matching inflection point set divide the mapped contour and the reference contour into one-to-one corresponding matching edges. If there are no matching inflection points among the matching edges on the reference contour, they are taken as inflection points to be filtered. If it is determined that the ratio of the length of any two inflection points to the corresponding vertex of their respective edges to the length of their respective edges is less than a preset ratio, then the two inflection points are added to the set of matching inflection points to be filtered. Add the shortest matching inflection point in the set of matching inflection points to be filtered to the set of matching inflection points to be added; The set of matching inflection points to be added is filtered based on the non-unique matching inflection point elimination algorithm to obtain the final set of matching inflection points to be added. The final set of added matching inflection points is merged with the preset high-precision matching inflection point set to obtain the final set of matching inflection points.

6. The method for calculating the accuracy of building mapping based on contour matching according to claim 1, characterized in that, The average distance between each set of matching edges is calculated using a weighted average distance based on the average length of the matching edges. This weighted average distance is then used as the distance between the mapped contour and the reference contour, including: The matched inflection points in the measured contour and the reference contour are divided into several groups of matching edges, and the distance between each group of matching edges and the weight corresponding to each group of matching edges are obtained. The distance between the mapped contour and the reference contour is obtained by weighted summation of the distance and the weight.

7. A system for calculating the accuracy of building mapping based on contour matching, comprising the method for calculating the accuracy of building mapping based on contour matching as described in any one of claims 1 to 6, characterized in that, include: The initial matching module is used to obtain the mapping outline and reference outline of the building to be measured, and to match the mapping outline and reference outline using an inflection point matching algorithm to generate an initial matching inflection point set; The mismatch removal module is used to remove obvious mismatch inflection points from the initial set of matching inflection points using an obvious mismatch inflection point removal algorithm, to obtain a set of matching inflection points after removing obvious mismatch inflection points; and to remove non-unique mismatch inflection points from the set of matching inflection points after removing obvious mismatch inflection points using a non-unique mismatch inflection point removal algorithm, to obtain a preset high-precision set of matching inflection points. The matching addition module is used to add missing matching inflection points to the preset high-precision matching inflection point set by employing a matching inflection point addition algorithm, thereby obtaining the final matching inflection point set. The edge matching module is used to divide the mapping contour and the reference contour into an equal number of edges according to the final matching inflection point set, match the equal number of edges to obtain a matching edge set, and obtain the average distance between each group of matching edges in the matching edge set based on the matching edge distance calculation method of Riemann integral. The distance calculation module is used to calculate the weighted average distance between each set of matching edges according to the average length of the matching edges, and use the weighted average distance as the distance between the mapped contour and the reference contour.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for calculating the accuracy of building mapping based on contour matching as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for calculating the accuracy of building mapping based on contour matching as described in any one of claims 1 to 6.

Images