Method for extracting road network topology correction

Abstract

The application discloses an extraction road network topology correction method, including an extraction road network topology correction method based on direction extension and endpoint matching: extraction road network data preprocessing; screening road endpoint candidate set; traversing the road endpoint candidate set, calculating the matching index of the road endpoint pair; constructing an endpoint matching model; traversing the road endpoint candidate set, calculating the endpoint matching degree, selecting and connecting the endpoint pair with the maximum endpoint matching degree to extract the road network for topology correction; and an extraction topology correction method based on crowd source road network and geometric matching, comprising: OSM road network and extraction road network data preprocessing; screening road candidate set, constructing road network matching model, and obtaining road network matching relationship; according to the road network matching relationship, topologically correcting the topology defects between the extraction roads belonging to the same and different crowd source roads. The application can fit roads with different shapes in the presence or absence of reference data, and solves the problem of extraction road network topology correction in different scenes.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of road network topology correction, and particularly relates to a method for extracting road network topology correction. BACKGROUND

[0002] Quick and accurate acquisition of road data is beneficial to the response of government departments to emergency events and the daily travel of people, and is of great significance to the forefront research such as automatic driving. Although the traditional manual measurement method can obtain relatively accurate road information, the process of collecting data is relatively cumbersome, the workload is large, and it is very time-consuming. With the continuous development of deep learning technology, the remote sensing image road network extraction technology based on deep learning has gradually matured, and this method of acquiring road network has become one of the main methods for global road network data production. However, due to various factors, there is a problem of topology missing in the extraction of road network, which is difficult to be practically applied. Therefore, topology correction needs to be performed on the extracted road network.

[0003] Although the existing road network topology correction method can solve the problem of imperfect topology of the extracted road network to a certain extent, it cannot well repair relatively complex roads due to the lack of reference data. Moreover, when supplementing the topology-missing roads, straight-line-shaped roads are usually used. However, the shape of actual roads is various, and the straight-line-shaped roads obviously cannot well fit the actual roads. SUMMARY

[0004] The purpose of the present application is to solve the problems of the prior art, and provide a method for extracting road network topology correction, which can fit roads of different shapes with or without reference data, and solve the problem of extracting road network topology correction in different scenarios.

[0005] To solve the above technical problems, the present application adopts the following technical solutions:

[0006] A method for extracting road network topology correction based on direction extension and end point matching is suitable for topology correction of the extracted road network without reference data, and includes the following steps:

[0007] Step 1, preprocessing the extracted road network data;

[0008] Step 2, screening a road end point candidate set from the preprocessed extracted road network data;

[0009] Step 3, direction index extraction: traversing the road end point candidate set, calculating the included angle between the road where the road end point is located and the road end point connecting line, i.e. the starting angle and the ending angle ;

[0010] Step 4, distance index extraction: traversing the road end point candidate set, according to the starting angle in step 3, and termination angle Calculate the lateral distance between two road endpoints and longitudinal distance ;

[0011] Step 5: Use the analytic hierarchy process (AHP) to analyze and calculate the weights of the endpoint matching index, and combine the direction index from Step 3 and the distance index from Step 4 to construct an endpoint matching model to calculate the endpoint matching degree.

[0012] Step 6: Traverse the candidate set of road endpoints, calculate the endpoint matching degree, select and connect the endpoint with the highest endpoint matching degree to perform topology correction on the extracted road network.

[0013] Furthermore, the preprocessing in step 1 includes aggregating nodes of unaggregated road intersections in the extracted road network data, and removing road fragments and protrusions at both ends of the road.

[0014] Furthermore, the specific implementation method of step 2 is as follows:

[0015] For each road endpoint, if it is a suspended endpoint, it is continuously extended by a distance of X meters. After each extension, a buffer zone is established with a radius of X meters, and the endpoint candidate set falling into the buffer zone is selected. The above process is repeated until the number of extensions reaches N. If it intersects with other roads during the extension process, it is no longer extended. The endpoint candidate set in each extension process is retained as the final endpoint candidate set.

[0016] Furthermore, the formulas for calculating the starting angle and ending angle in step 3 are as follows:

[0017]

[0018]

[0019] in, As the starting angle, Starting road The direction angle, The direction angle of the line connecting the road endpoints. The termination angle, To terminate the road The direction angle.

[0020] Furthermore, the formulas for calculating the horizontal and vertical distances in step 4 are as follows:

[0021]

[0022]

[0023] in, Let be the Euclidean distance between the road endpoints. is the start angle in step 3, is the end angle in step 3.

[0024] Further, the specific implementation of step 5 is as follows:

[0025] S5.1, describe the problem to be solved, and establish the corresponding analytic hierarchy model;

[0026] S5.2, analyze the relative importance of each matching index, and construct a judgment matrix;

[0027] S5.3, calculate the matching index weight according to the judgment matrix;

[0028] S5.4, check the consistency of the judgment matrix;

[0029] S5.5, construct an endpoint matching model to calculate the endpoint matching degree, and the calculation formula of the endpoint matching model is as follows:

[0030]

[0031] wherein, and are the weights of the horizontal distance and the vertical distance respectively, is the weight sum of the start angle and the end angle; and are the threshold values of the horizontal distance and the vertical distance respectively, is the threshold sum of the start angle and the end angle.

[0032] A road network topology correction method based on crowd-sourced road network and geometric matching is suitable for extracting road network topology correction under reference data, comprising the following steps:

[0033] Step 1, preprocessing OSM road network and extracted road network data;

[0034] Step 2, obtaining road network matching relationship from the preprocessed data;

[0035] Step 3, according to the road network matching relationship, topologically correcting the topological defects between the extracted roads belonging to the same crowd-sourced road;

[0036] Step 4, according to the road network matching relationship, topologically correcting the topological defects between the extracted roads belonging to different crowd-sourced roads.

[0037] Further, the step 1 includes the following sub-steps:

[0038] S1.1, constructing OSM road network Stroke;

[0039] S1.2, breaking the extracted road network.

[0040] Further, the step S1.1 comprises:

[0041] S1.1.1, sequentially selecting a road in the OSM road network data, if the road does not form a road Stroke with other roads, using it to initialize a new road Stroke;

[0042] S1.1.2, screening roads intersecting with the road endpoint to form a candidate road set;

[0043] S1.1.3, traversing the candidate road set, and eliminating roads different from the road category of the starting road;

[0044] S1.1.4, traversing the candidate road set, calculating the arc segment angle of the candidate road and the starting road, and selecting the candidate road with the largest arc segment angle;

[0045] S1.1.5, judging whether the largest arc segment angle is greater than the arc segment angle threshold, if it is greater than the set arc segment angle threshold, using the road to update the road Stroke;

[0046] S1.1.6, looping S1.1.2 to S1.1.5 until the road Stroke is constructed;

[0047] S1.1.7, looping the above steps until the road network Stroke is constructed.

[0048] Further, the step S1.2 comprises:

[0049] S1.2.1, looping reading the OSM road network data to obtain the OSM road endpoint set in the OSM road network;

[0050] S1.2.2, looping the road endpoint set of the OSM road network, sequentially reading an endpoint, and establishing a buffer area with the endpoint as the center and X meters as the buffer radius, and screening the extracted roads and extracted road endpoints falling into the buffer area from the extracted road network;

[0051] S1.2.3, if there are extracted road endpoints falling into the buffer area, calculating the perpendicular point of the endpoint to the extracted road, and breaking the extracted road at the perpendicular point;

[0052] S1.2.4, if there are no extracted road endpoints falling into the buffer area, selecting the extracted road closest to the endpoint, calculating the perpendicular point of the endpoint to the extracted road, and breaking the extracted road at the perpendicular point.

[0053] Further, the method for obtaining the road network matching relationship in step 2 is:

[0054] S2.1, for each road in the extracted road network dataset, first establish the buffer of the road, and then filter the road candidate set falling into the buffer from the OSM road network;

[0055] S2.2, traverse the road candidate set, encrypt the nodes of each road, and for each road matching pair, select a node from the shorter road, calculate the minimum distance from the node to all nodes of the longer road, and the value is SM HD;

[0056] S2.3, for each road matching pair, respectively establish the minimum area circumscribed rectangle of the road, obtain the diagonal of the minimum area circumscribed rectangle and calculate its direction angle, and the difference between the corresponding direction angles of the two roads is angle dif;

[0057] S2.4, for each road matching pair, respectively establish the road buffer, and the buffer radius is set as SM HD, and the ratio of the intersection area of the two road buffers to the union area is overlap;

[0058] S2.5, use the analytic hierarchy process to analyze and calculate the weight value of the similarity characteristic factor, and construct a road network matching model according to SM HD, angle dif and overlap to calculate the comprehensive matching degree;

[0059] S2.6, traverse the road matching pair, calculate the comprehensive matching degree, and record the road matching pair with the maximum comprehensive matching degree.

[0060] Further, the step S2.5 is specifically implemented as follows:

[0061] S2.5.1, describe the problem to be solved, and establish the corresponding analytic hierarchy model;

[0062] S2.5.2, analyze the relative importance of each similarity characteristic index, and construct a judgment matrix;

[0063] S2.5.3, calculate the matching index weight value according to the judgment matrix;

[0064] S2.5.4, check the consistency of the judgment matrix;

[0065] S2.5.5, construct a road network matching model to calculate the comprehensive matching degree, wherein the road network matching model is:

[0066]

[0067] In the formula, is the weight value of the distance similarity characteristic factor, is the weight value of the direction similarity characteristic factor, The weights of the overlap similarity feature factors are... and These are the thresholds for the distance similarity feature factor and the direction similarity feature factor, respectively.

[0068] Furthermore, the specific method for step 3 is as follows:

[0069] S3.1 Calculate the road matching segments and obtain the missing topology segments;

[0070] S3.2 Calculate the starting point of the missing road segment respectively. Distance to OSM road and the finish line Distance to OSM road ;

[0071] S3.3, Set the starting point The feet of the perpendiculars to the OSM road correspond to points respectively. , This will lead OSM down the path of All nodes of the road segment are added to the node set, let the node set be . ;

[0072] S3.4 Calculate the distance With distance The difference ;

[0073] S3.5, the difference The scaling distance is distributed across all nodes in the node set according to length. For each node... Calculate the scaling distance of the node. The specific formula is as follows:

[0074]

[0075] in, For road section Length, For road section Length;

[0076] S3.6. Trace the coordinates of the corresponding nodes on the road. The specific calculation formula is as follows:

[0077]

[0078] Where, vector sum vector Each is a road segment and road sections unit vector, The included angle, The nodes of the original road. To copy the nodes of the road behind.

[0079] Further, step 4 specifically comprises the following sub-steps:

[0080] S4.1, extract the OSM road intersection, record the road information constituting the road intersection;

[0081] S4.2, sequentially select an OSM road intersection, obtain the corresponding extracted road intersection constituting road information;

[0082] S4.3, obtain the extracted road endpoint from the extracted road, if the extracted road endpoint is the same point, it is not processed, otherwise it goes to the next step;

[0083] S4.4, calculate the distance between the extracted road endpoint and the OSM road intersection point one by one, filter the extracted road endpoints with a distance less than the threshold value and calculate the center point coordinates;

[0084] S4.5, for the road where the extracted road endpoint is less than the threshold value, the node is aggregated; for the road where the extracted road endpoint is greater than the threshold value, the road is copied with the road endpoint as the starting point and the center point as the end point.

[0085] Compared with the prior art, the beneficial effects of the present application are:

[0086] (1) The extracted road network topology correction method provided by the present application can fit roads of different shapes with or without reference data, solve the problem of extracting road network topology correction in different scenes, and greatly improve the topological integrity of automatic extraction of road network from remote sensing images;

[0087] (2) The extracted road network topology correction method based on direction extension and endpoint matching improves the efficiency of the algorithm by using the direction extension algorithm, and to some extent, avoids the influence of artificial experience value by using the analytic hierarchy process;

[0088] (3) The topology correction method based on crowd source road network and geometric matching of the present application greatly improves the topology correction effect by introducing OSM road network, increases the accuracy of geometric matching by using a multi-factor geometric matching model and the analytic hierarchy process to calculate the weight to obtain the road network matching relationship, and makes the topology correction road fit the actual road, which is more consistent with human visual perception. BRIEF DESCRIPTION OF DRAWINGS

[0089] Figure 1 The extracted road network topology correction flowchart based on direction extension and endpoint matching of the embodiment of the present application;

[0090] Figure 2This is a schematic diagram of using the directional extension algorithm to filter the endpoint candidate set in an embodiment of the present invention, where the extension distance is set to X meters each time;

[0091] Figure 3 This is a schematic diagram of the endpoint matching index in an embodiment of the present invention; wherein, Line_1 and Line_2 are endpoints. and endpoints The road where it is located; As endpoints and endpoints Midpoint of the line; and Midpoints to the perpendicular point of the two lines; and Horizontal distance; and Vertical distance; and These are the angles between the lines connecting road Line_1 and road Line_2 and the endpoints, respectively.

[0092] Figure 4 This is a schematic diagram of topology correction based on directional extension and endpoint matching in an embodiment of the present invention; wherein, (a) is the original extracted road network, and (b) is the extracted road network after topology correction;

[0093] Figure 5 This is a flowchart illustrating the extraction and correction of road network topology based on crowdsourced road networks and geometric matching, as described in an embodiment of the present invention.

[0094] Figure 6 This is a schematic diagram of using the buffer growth method to extract a road from the road network in an embodiment of the present invention, and filtering the road candidate set from the OSM road network; wherein, A1 is a road to be extracted from the road network, and B1 and B2 are roads in the OSM road network that intersect with the buffer established with road A1, which are the road candidate set;

[0095] Figure 7 This is a schematic diagram showing the calculation results of road distance before and after node encryption in an embodiment of the present invention. The distance between the two roads before node encryption is 28.3m, and the distance between the two roads after node encryption is 20m.

[0096] Figure 8 This is a schematic diagram illustrating the calculation of the diagonal angular difference between the minimum area circumscribed rectangles of two roads in an embodiment of the present invention.

[0097] Figure 9 This is a schematic diagram illustrating the calculation of the ratio of the intersection area to the union area of ​​two road buffer zones in an embodiment of the present invention, wherein the black area represents the intersection of the buffer zones and the gray area represents the union of the buffer zones;

[0098] Figure 10This is a schematic diagram of topology correction based on crowdsourced road network and geometric matching in an embodiment of the present invention; wherein, (a) is the OSM road network, (b) is the original extracted road network, and (c) is the topology-corrected extracted road network. Detailed Implementation

[0099] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0100] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0101] The present invention will be further described below with reference to specific embodiments, but these are not intended to limit the scope of the invention.

[0102] The road network topology correction method provided in this invention includes a road network number extraction topology correction method based on directional extension and endpoint matching, and a road network topology correction method based on crowdsourced road networks and geometric matching. The road network number extraction topology correction method based on directional extension and endpoint matching is suitable for extracting road network topology correction without reference data; while the road network topology correction method based on crowdsourced road networks and geometric matching is suitable for extracting road network topology correction with reference data.

[0103] Example 1

[0104] like Figure 1 As shown, this embodiment provides a method for extracting road network number topology correction based on directional extension and endpoint matching, including the following steps:

[0105] Step 1: Preprocess the extracted road network data to eliminate the influence of irrelevant factors on the results, so as to facilitate the subsequent process; specifically, the data preprocessing process in this example includes: aggregating nodes of unaggregated road intersections, removing road fragments, and removing protrusions at both ends of the road.

[0106] Step 2 involves filtering the candidate set of road endpoints from the preprocessed road network data obtained in Step 1 using a directional extension algorithm. This step specifically includes the following sub-steps:

[0107] S2.1 Extract all road endpoints from the extracted road network;

[0108] S2.2, Traverse the road endpoints and determine whether they are suspended endpoints. If they are, proceed to the next step.

[0109] S2.3. Let n be the number of times the road is extended, and let X be the distance of each extension.

[0110] S2.4. Starting from the road endpoint, extend the road by n*X meters to obtain the new road endpoint. Establish a buffer zone with the new road endpoint as the center and X meters as the buffer radius, and filter out road endpoints that fall within the buffer zone.

[0111] S2.5, Road extension times n+1;

[0112] S2.6 Repeat steps S2.3 to S2.5 until the extension is performed N times. Retain the endpoint candidate set from each extension process to obtain the final endpoint candidate set. If the extension intersects with other roads, the process must be terminated.

[0113] In this example, the single-direction extension distance and buffer radius of the road endpoint are X = 10 meters, and the maximum number of extensions is N = 5. The values ​​of X and N can be selected as needed. The extension method is as follows: Figure 2 As shown, this figure is a schematic diagram of the endpoint extension;

[0114] Step 3, Directional Indicator Extraction: Traverse the candidate set of road endpoints and calculate the angle between the road containing the road endpoint and the line connecting the road endpoint, i.e., the starting angle. and termination angle Among them, the starting angle and termination angle like Figure 3 As shown, the calculation formula is as follows:

[0115] ；

[0116] ；

[0117] in, Starting road The direction angle, The direction angle of the line connecting the road endpoints. To terminate the road The direction angle.

[0118] Step 4, Distance index extraction: Traverse the candidate set of road endpoints and calculate the lateral distance between two road endpoints. and longitudinal distance In this embodiment, the lateral distance and longitudinal distance like Figure 3 As shown, the calculation formula is as follows:

[0119] ；

[0120] ；

[0121] wherein, is the Euclidean distance of the road endpoints, and are the start angle and end angle in step C, respectively.

[0122] Step 5, using the analytic hierarchy process to analyze and calculate the weight of the endpoint matching index, combining the direction index of step 3 and the distance index of step 4 to construct an endpoint matching model to calculate the endpoint matching degree; wherein, this step specifically includes:

[0123] S5.1, describe the problem to be solved, in this embodiment, the problem to be solved is how to provide suitable weights for the endpoint matching model, and establish a corresponding analytic hierarchy model according to the problem;

[0124] S5.2, according to the analytic hierarchy model, analyze the relative importance of each matching index, and construct a judgment matrix; the relative importance of any two matching indexes is between 1-9, the larger the value, the higher the importance, which is generally preset according to the needs; in this embodiment, the relative importance of two matching indexes is used to construct the judgment matrix;

[0125] S5.3, calculate the matching index weight according to the judgment matrix;

[0126] S5.4, check the consistency of the judgment matrix, in this embodiment, the consistency check coefficient CR is calculated according to the judgment matrix, if it is less than 0.01, the consistency check is passed;

[0127] S5.5, determine the matching index weight that can be used according to the consistency result of the judgment matrix, that is, if the consistency check is passed, it can be determined as the matching index that can be used, and according to the determined matching index weight that can be used, construct an endpoint matching model to calculate the endpoint matching degree, the endpoint matching model is:

[0128]

[0129] wherein, , correspond to the weight of the horizontal distance and the vertical distance, respectively, is the weight sum of the start angle and the end angle; , correspond to the threshold of the horizontal distance and the vertical distance, respectively, is the threshold sum of the start angle and the end angle;

[0130] In this example, = 0.1210, = 0.4020, = 0.4770, = 50, = 10, = 0.5 The above parameters are not limited to the present application, and the corresponding values can be selected according to actual needs.

[0131] Step 6, traversing the road endpoint candidate set, calculating the endpoint matching degree according to the above endpoint matching model, selecting and connecting the endpoint pair with the largest endpoint matching degree to extract the topological correction of the road network, and the topological correction result is as shown in Figure 4 .

[0132] Embodiment 2

[0133] As shown in Figure 5 , the present embodiment provides a method for extracting road network topology correction based on crowd-sourced road network and geometric matching, which is suitable for topology correction of extracted road network under reference data, comprising:

[0134] Step 1, preprocessing the data of OSM road network and extracted road network; as shown in Figure 6 , this step includes:

[0135] S1.1, OSM road network Stroke construction; the specific implementation mode is as follows,

[0136] First, read the OSM road network data to obtain geometric information and semantic information, wherein the geometric information includes coordinates of each node of the road, and the semantic information includes road categories;

[0137] Second, sequentially select a road in the OSM road network data, if the road does not form a road Stroke with other roads, use it to initialize a new road Stroke;

[0138] Third, filter the roads intersecting with the road endpoints to form a candidate road set;

[0139] Fourth, traverse the candidate road set to remove roads with different road categories from the starting road;

[0140] Fifth, traverse the candidate road set to calculate the arc segment angle of the candidate road and the starting road, and select the candidate road with the largest arc segment angle;

[0141] Sixth, determine whether the largest arc segment angle is greater than the arc segment angle threshold, if it is greater than the set arc segment angle threshold, use the road to update the road Stroke; in the present embodiment, the arc segment angle threshold can be 120 degrees, of course, in other embodiments, other angles can also be selected according to actual needs;

[0142] Step 7, repeat Step 3 to Step 4 until the construction of the road Stroke is completed.

[0143] Step 8, repeat Step 2 until the construction of the road network Stroke is completed.

[0144] S1.2, break the extracted road network, which is implemented as follows:

[0145] Step 1, obtain the set of road endpoints of the OSM road network;

[0146] Step 2, cycle through the set of road endpoints of the OSM road network, sequentially read one endpoint. Take the endpoint as the center of the buffer, and establish a buffer with a radius of X meters (in this embodiment, the buffer radius X is 10 meters), and select the extracted roads and extracted road endpoints that fall within the buffer from the extracted road network;

[0147] Step 3, if there are no extracted road endpoints falling within the buffer, return to Step 2, otherwise go to Step 4;

[0148] Step 4, cycle through the extracted roads falling within the buffer, calculate the distance from the endpoint to the extracted road. Select the closest extracted road, calculate the perpendicular point of the endpoint to the extracted road, and break the extracted road at the perpendicular point.

[0149] Step 2, obtain the road network matching relationship; this step specifically includes the following sub-steps:

[0150] S2-1, for each road in the extracted road network dataset, first establish a buffer Buffer for the road, and then select a set of road candidates that fall within the buffer Buffer from the OSM road network;

[0151] S2-2, traverse the road candidate set, and perform node matching for each road, as shown in Figure 7 ; In this embodiment, the encrypted road node encryption distance is set to 5 meters. For each road matching pair, select a node from the shorter road, and calculate the minimum distance from this node to all nodes of the longer road, where the value is the SM_HD (the Short-line Median Hausdorff Distance), SM_HD The calculation formula is as follows:

[0152] ；

[0153] Where a is a node of road A, and b is a node of road B, and the length of road A is less than the length of road B;

[0154] S2-3, for each road matching pair, the minimum area circumscribed rectangle of the road is established respectively, the diagonal of the minimum area circumscribed rectangle is obtained and the direction angle thereof is calculated, see Figure 8 ; the difference between the corresponding direction angles of the two roads is the angle_dif , The calculation formula is as follows,

[0155]

[0156] wherein, and are the direction angles of the diagonals of the minimum area circumscribed rectangles of the roads a and b respectively;

[0157] S2-4, for each road matching pair, the road buffer zone is established respectively, the buffer zone radius is set as SM_HD , the ratio of the intersection area of the two road buffer zones to the union area is the overlap , see Figure 9 , The calculation formula is as follows,

[0158]

[0159] wherein, and are the buffer zones of the roads a and b respectively;

[0160] S2.5, the analytic hierarchy process is used to analyze and calculate the weight values of the similarity characteristic factors, and the road network matching model is constructed according to the SM_HD, angle_dif and overlap to calculate the comprehensive matching degree; this step is similar to example one, including the following sub-steps:

[0161] S5.1, the problem to be solved is described, in this embodiment, the problem to be solved is how to provide suitable weight values for the end point matching model, and the corresponding analytic hierarchy model is established according to the problem;

[0162] S5.2, the relative importance of each matching index is analyzed according to the analytic hierarchy model, and the judgment matrix is constructed; the relative importance of any two matching indexes is between 1-9, the larger the value is, the higher the importance degree is, which is generally preset according to the needs; in this embodiment, the judgment matrix is constructed using the relative importance of the two-by-two matching indexes;

[0163] S5.3, the matching index weight value is calculated according to the judgment matrix;

[0164] S5.4, the consistency of the judgment matrix is checked, in this embodiment, the test coefficient CR is calculated according to the judgment matrix, if it is less than 0.01, the consistency test is passed;

[0165] S5.5, determining the matching index weight that can be used according to the consistency result of the judgment matrix, that is, if the consistency test is passed, it can be determined as the matching index that can be used, and a road network matching model is constructed according to the matching index weight that can be used to calculate the comprehensive matching degree, and the calculation formula of the road network matching model is as follows:

[0166]

[0167] In the formula, is the weight of the distance similarity characteristic factor, is the weight of the direction similarity characteristic factor, is the weight of the overlap degree similarity characteristic factor, and are the threshold values of the distance similarity characteristic factor and the direction similarity characteristic factor respectively.

[0168] Step 3, according to the road network matching relationship, topological correction is performed on the topological missing between the extracted roads belonging to the same crowd source road; wherein, it includes:

[0169] S3-1, calculating the road matching section to obtain the topological missing section;

[0170] S3-2, using the road sketching algorithm to supplement the topological missing section; specifically:

[0171] Firstly, the distances from the start point and the end point of the missing section road to the OSM road are calculated respectively , . If the end point is located on the left side of the OSM road, the distance is set as a positive value, otherwise it is set as a negative value;

[0172] Secondly, the perpendicular feet of the start point to the OSM road are set as points , , all the nodes on the OSM road falling into the road section are added to the node set, and the node set is set as ;

[0173] Thirdly, the difference between the distance and the distance is calculated;

[0174] Fourthly, the difference is distributed to the scaled distance of all nodes in the node set according to the length. For each node , the scaled distance of the node is calculated using the following formula ;

[0175] ​

[0176] wherein, is the length of the road segment , is the length of the road segment ;

[0177] In the fifth step, the coordinates of the corresponding nodes of the imitated road can be obtained using the following formula:

[0178]

[0179] wherein, the vector , the vector are the unit vectors of the road segment , the road segment , respectively, is the included angle thereof, is the node of the original road, is the node of the imitated road.

[0180] In step 4, the topological correction of the topological defects between the extracted roads belonging to different crowdsourced roads is performed according to the road network matching relationship. This step includes:

[0181] S4-1, extract the OSM road intersection and record the road information constituting the road intersection;

[0182] S4-2, take the OSM road intersection as a reference and perform topological correction on the corresponding road intersection in the extracted road network according to the road network matching relationship. Specifically,

[0183] In the first step, an OSM road intersection is sequentially selected and the constituent road information of the corresponding extracted road intersection can be easily obtained according to the road network matching relationship;

[0184] In the second step, the extracted road endpoint is obtained from the extracted road. If the extracted road endpoint is the same point, it is not processed, otherwise the next step is entered;

[0185] In the third step, the distance between the extracted road endpoint and the OSM road intersection point is calculated one by one, the extracted road endpoint with a distance less than a threshold value is screened and the center point coordinates are calculated;

[0186] In the fourth step, the nodes of the road with the extracted road endpoint with a distance less than a threshold value are aggregated. For the road with the extracted road endpoint with a distance greater than a threshold value, the road imitation is performed with the road endpoint as the starting point and the center point as the ending point. The threshold value can be set to 10 meters. In this embodiment, the topological correction result of the extracted road network is shown in FIG. 8. Figure 10

[0187] ​The above merely preferred embodiments of the present application, and not therefore limit the embodiments and protection scope of the present application, for those skilled in the art, it should be realized that the equivalent replacement and obvious changes made by the application description, the resulting scheme should be included in the protection scope of the present application.

Claims

1. A method for extracting road network topology correction based on direction extension and endpoint matching, characterized in that, The method comprises the following steps: Step 1, preprocessing the extracted road network data; Step 2, screening a road endpoint candidate set from the preprocessed extracted road network data; Step 3, direction indicator extraction: traverse the road endpoint candidate set, calculate the angle between the road where the road endpoint is located and the line connecting the road endpoints, i.e. the start angle and the end angle ; Step 4, distance indicator extraction: iterate through the road endpoint candidate set, and calculate the lateral distance and longitudinal distance between two road endpoints according to the start angle and end angle in step 3; Step 5, using the analytic hierarchy process to analyze and calculate the weight of the endpoint matching index, and combining the direction index of step 3 and the distance index of step 4 to construct an endpoint matching model to calculate the endpoint matching degree; Step 6, traversing the road endpoint candidate set, calculating the endpoint matching degree, selecting and connecting the endpoint pair with the maximum endpoint matching degree to extract the topological correction of the road network; In step 3, the calculation formula of the start angle and the end angle is as follows: wherein is a start angle, is a direction angle of the start road is a direction angle of the start road is a direction angle of the road end line, is a finish angle, is a direction angle of the finish road is a direction angle of the finish road In step 4, the calculation formula of the horizontal distance and the vertical distance is as follows: wherein is the Euclidean distance to the road endpoint, is the start angle in step 3, is the end angle in step 3; The specific implementation mode of step 5 is as follows: S5.1, describe the problem to be solved, and establish the corresponding analytic hierarchy model; S5.2, analyze the relative importance of each matching index, and construct a judgment matrix; S5.3, calculate the matching index weight according to the judgment matrix; S5.4, check the consistency of the judgment matrix; S5.5, construct an endpoint matching model to calculate the endpoint matching degree, and the calculation formula of the endpoint matching model is as follows: wherein, and are weights for lateral and longitudinal distances, respectively, is a sum of weights for start and end angles; and are threshold values for lateral and longitudinal distances, respectively, is a sum of threshold values for start and end angles.

2. The extraction road network topology revision method based on directional extension and end point matching according to claim 1, characterized in that, The preprocessing in step 1 includes node aggregation on the un-aggregated road intersections in the extracted road network data, and removal of road fragments and road protrusions at both ends.

3. The method of claim 1, wherein the method is characterized by, The specific implementation mode of step 2 is as follows: For each road endpoint, if it is a hanging endpoint, it is continuously extended by X meters; after each extension, a buffer area with a radius of X meters is established, and a candidate set of endpoints falling within the buffer area is screened; Repeat the above process until the number of extensions reaches N, and if the extension process intersects with other roads, stop extending; Keep the endpoint candidate set in each extension process, which is the final endpoint candidate set.

4. A method for extracting road network topology correction based on crowd-sourced road network and geometric matching, characterized in that, The method comprises the following steps: Step 1, preprocessing the OSM road network and the extracted road network data; Step 2, obtaining the road network matching relationship from the preprocessed data; Step 3, according to the road network matching relationship, topologically correcting the topological defects between the extracted roads belonging to the same crowd source road; Step 4, according to the road network matching relationship, topologically correcting the topological defects between the extracted roads belonging to different crowd source roads; In step 2, the method for obtaining the road network matching relationship is as follows: S2.1, for each road in the extracted road network data set, first establish a buffer area Buffer for the road, and then screen a road candidate set falling into the buffer area Buffer from the OSM road network; S2.2, traverse the road candidate set, encrypt each road, and for each road matching pair, select a node from the shorter road, calculate the minimum distance from the node to all nodes of the longer road, and the value is SM_HD; S2.3, for each road matching pair, respectively establish the minimum area circumscribed rectangle of the road, obtain the diagonal of the minimum area circumscribed rectangle and calculate the direction angle, and the difference between the corresponding direction angles of the two roads is angle_dif; S2.4, for each road matching pair, respectively establish the road buffer area, and the buffer area radius is SM_HD, and the ratio of the intersection area of the two road buffer areas to the union area is overlap; S2.5, using the analytic hierarchy process to analyze and calculate the weight of the similarity characteristic factor, and constructing a road network matching model according to SM HD, angle dif and overlap to calculate the comprehensive matching degree; S2.6, traversing the road matching pair, calculating the comprehensive matching degree, and recording the road matching pair with the maximum comprehensive matching degree; The specific implementation method of step S2.5 is as follows: S2.5.1, describe the problem to be solved, and establish the corresponding analytic hierarchy model; S2.5.2, analyze the relative importance of each similarity characteristic index, and construct a judgment matrix; S2.5.3, calculate the matching index weight according to the judgment matrix; S2.5.4, check the consistency of the judgment matrix; S2.5.5, construct a road network matching model to calculate the comprehensive matching degree, wherein the road network matching model is: wherein is a weight of the distance similarity feature factor, is a weight of the direction similarity feature factor, is a weight of the overlap similarity feature factor, and are threshold values of the distance similarity feature factor and the direction similarity feature factor, respectively.

5. The method of claim 4, wherein, The step 1 includes the following sub-steps: S1.1, constructing an OSM road network Stroke; S1.2, breaking the extracted road network.

6. The method of claim 5, wherein the method further comprises: Step S1.1 specifically includes: S1.1.1, sequentially selecting a road in the OSM road network data, if the road does not form a road Stroke with other roads, using it to initialize a new road Stroke; S1.1.2, screening roads intersecting with the road endpoint to form a candidate road set; S1.1.3, traversing the candidate road set, and removing roads different from the starting road in road category; S1.1.4, traversing the candidate road set, calculating the arc segment angle of the candidate road and the starting road, and selecting the candidate road with the maximum arc segment angle; S1.1.5, judging whether the maximum arc segment angle is greater than the arc segment angle threshold, if it is greater than the set arc segment angle threshold, using the road to update the road Stroke; S1.1.6, looping S1.1.2 to S1.1.5 until the road Stroke is constructed; S1.1.7, looping the above steps until the road network Stroke is constructed.

7. The method of claim 5, wherein the method further comprises: Step S1.2 includes: S1.2.1, looping to read the OSM road network data to obtain the OSM road endpoint set in the OSM road network; S1.2.2, looping the road endpoint set of the OSM road network, sequentially reading an endpoint, and establishing a buffer area with the endpoint as the center and X meters as the buffer radius, and screening the extracted roads and extracted road endpoints falling into the buffer area from the extracted road network; S1.2.3, if there are extracted road endpoints falling into the buffer area, calculating the vertical point of the endpoint to the extracted road, and breaking the extracted road at the vertical point; S1.2.4, if there is no extracted road endpoint falling into the buffer area, selecting the extracted road closest to the endpoint, calculating the vertical point of the endpoint to the extracted road, and breaking the extracted road at the vertical point.

8. The method of claim 4, wherein the method further comprises: The specific method of step 3 is: S3.1, calculating the road matching segment to obtain the topological missing segment; S3.2, calculate the start point of the missing segment road respectively distance to the OSM road and the end point distance to the OSM road ; S3.3, Set the starting point The feet of the perpendiculars to the OSM road correspond to points respectively. , This will lead OSM down the path of All nodes of the road segment are added to the node set, let the node set be . ; S3.4, calculating distance with distance difference ; S3.5, the difference value For each node in the set of nodes, the scaled distance assigned to the node according to the length , the scaled distance of the node is calculated The specific formula is: wherein is the length of the road segment is the length of the road segment is the length of the road segment is the length of the road segment S3.6, copying the coordinates of the corresponding nodes on the road, and the specific calculation formula is: wherein the vector and the vector are unit vectors of the road segments and the road segments respectively, is the included angle therebetween, is a node of the original road, is a node of the traced road.

9. The method of claim 4, wherein the method further comprises: Step 4 specifically includes the following sub-steps: S4.1, extracting the OSM road intersection, and recording the road information constituting the road intersection; S4.2, sequentially select an OSM road intersection, and obtain the corresponding extracted road intersection component road information; S4.3, obtain the extracted road endpoint from the extracted road, if the extracted road endpoint is the same point, do not process it, otherwise go to the next step; S4.4, calculate the distance between the extracted road endpoint and the OSM road intersection point one by one, screen the extracted road endpoints with a distance less than the threshold value, and calculate the center point coordinates; S4.5, for the road where the extracted road endpoint with a distance less than the threshold value is located, node aggregation is performed; for the road where the extracted road endpoint with a distance greater than the threshold value is located, the road endpoint is taken as the starting point, and the center point is taken as the ending point, and road copying is performed.

Images