A method for constructing a topological relationship of a street view image based on a road traffic marking line

Abstract

The application relates to a kind of street view image topological relation construction methods based on road traffic marking line, the present application obtains street view track from street view image file;With road traffic marking line class element, the collected street view track is divided into intersection and section;Each section is filtered out a section track line with stop line, and each intersection is selected into intersection entry and exit track point;Based on intersection entry and exit track point, the jump relationship between track points is established;Street view image corresponding to track point is reorganized according to grid;Street view image topological relation table is set, and street view image topological relation at section and intersection is generated.The present application divides street view track into intersection and section for processing respectively with road traffic marking line class element, which is more efficient and reasonable;Through the selection and topological relation construction of track point in intersection, the intersection relationship of different track lines is established, which is consistent with the actual street condition experience when browsing panorama.

Description

Technical Field

[0001] This invention relates to the field of street view image data processing technology, and in particular to a method for constructing the topological relationship of intersection street view images for front-end browsing and display of street view. Background Technology

[0002] With the development of smart cities, urban management is becoming increasingly refined and three-dimensional, requiring more comprehensive and realistic 3D data support. 360° street view browsing and display systems can vividly and realistically represent the natural and cultural phenomena of cities and regions, providing a new model for urban 3D information display and services. Currently, a series of street view maps have been launched both domestically and internationally, and are widely used in public geographic information services for smart cities. However, as the use of vehicle-mounted mobile measurement systems for street view data collection becomes increasingly convenient, the volume of street view data is gradually expanding. How to process the collected street view image data to better facilitate service promotion and online display has become a problem that needs to be solved.

[0003] Street view imagery is collected by mobile survey vehicles following pre-set trajectories in multiple trips. However, research on how to handle the topological connectivity of special locations such as intersections and T-junctions after overlaying street view images from different trips is relatively limited. This results in the inability to capture the intersection relationships of different trajectories after street view data collection, hindering the experience of viewing panoramic views that accurately reflect actual street conditions. Furthermore, how to select the best trajectories from multiple acquisitions for optimal front-end display is also a problem that needs to be considered.

[0004] Chinese patent CN201910877727.0 discloses an automatic map topology construction method. This method uses the orientation of trajectory points to calculate the complexity of driving behavior within a certain area, and then performs intersection detection and recognition based on this complexity. It uses the intersection information and the continuity information of trajectory lines to treat the beginning and end points of the trajectory lines as virtual intersections, and then clusters these virtual intersections to obtain the connectivity between them. However, this intersection detection scheme based on trajectory point orientation requires calculating the orientation of the trajectory points throughout their entire length, resulting in a large computational load. Furthermore, for road sections such as mountain roads where drivers frequently change their orientation, this intersection detection scheme based on trajectory point orientation to calculate the complexity of driving behavior within a certain area is prone to misjudgment.

[0005] Chinese patent application CN201910476143.2 discloses a lane map construction method, which divides the road surface into an intersection surface and a lane surface based on the lane stop line and its extension; determines the center position of a single lane surface, and generates lane reference lines for the single lane surface along the travel direction at the center position; and constructs the topological relationship of the lane reference lines based on the direction of the lane reference lines, the endpoints of the lane reference lines, and the intersection points of the lane reference lines and the intersection surface. This scheme performs well in recognizing regular intersections, but its performance is poor in recognizing slightly more complex irregular intersections (such as intersections located under elevated roads or intersections with large turning radii); furthermore, when constructing the topological relationship, the lane reference lines entering the intersection surface do not consider the actual situation of dedicated left-turn lanes, dedicated right-turn lanes, etc., which will affect the accuracy of the route when vehicles pass through the intersection surface. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a method for constructing the topological relationship of street view images based on road traffic markings.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A method for constructing topological relationships from street view images based on road traffic markings includes the following steps:

[0009] The planar coordinate information of the street view image is obtained from the street view image parameter file and spatialized into trajectory points. The trajectory points are then cleaned and trajectory lines are generated.

[0010] The data of pedestrian crossings and road surfaces are collected and spatial calculations are performed on the generated trajectory lines to divide the trajectory lines into intersection trajectory lines and road segment trajectory lines.

[0011] For all the defined road segment trajectory lines, use the stop line element to filter the trajectory lines collected multiple times for the same road segment in the same direction, and retain one of them;

[0012] Based on the defined intersection trajectory lines, the entry and exit trajectory points of all intersections are selected;

[0013] Based on the selected entry and exit trajectory points at intersections, determine the trajectory points that each trajectory point can jump to, and establish the jump relationship between trajectory points;

[0014] After establishing the jump relationship, the street view images corresponding to the trajectory points are reorganized and renamed according to the grid to obtain a unique ID for each street view image.

[0015] Design a street view image topology table, including three fields: current image ID, associated image ID, and angle;

[0016] Based on the street view image topology table, generate and store the street view image topology relationships at road segments; and based on the street view image topology table and the trajectory point jump relationships within intersections, generate and store the street view image topology relationships at intersections.

[0017] Furthermore, the specific process of generating the trajectory line is as follows:

[0018] Traverse all street view image trajectory points in reverse order of acquisition; set a distance threshold, and if a trajectory point is within the threshold range of the previous acquired trajectory point, then discard that trajectory point.

[0019] After cleaning, the trajectory points are connected into trajectory lines according to the image exposure order during a single acquisition, with one trajectory line corresponding to each acquisition vehicle.

[0020] Furthermore, the specific process of dividing the trajectory line into intersection trajectory lines and road segment trajectory lines is as follows:

[0021] Collect the closed line formed by the boundary of the pedestrian crossing as the pedestrian crossing line element, generate the bounding rectangle for each pedestrian crossing line element, and extract the center line of the long side of the bounding rectangle in the same direction as the center line of the pedestrian crossing line.

[0022] Use the extracted pedestrian crossing centerline to break the trajectory lines that intersect with the centerline;

[0023] Filter the road intersection surfaces in the road surface data, perform spatial calculations on the lines and intersection surfaces after breaking the trajectory lines, and take the trajectory lines located in the road intersection surfaces as the intersection trajectory lines;

[0024] Trajectory lines other than those at intersections are considered as road segment trajectory lines.

[0025] Furthermore, the specific process of using stop line elements to filter trajectory lines collected multiple times for the same lane in the same direction on the same road segment is as follows:

[0026] Perform line intersection operation between stop line elements and trajectory line elements, filter out all road segment trajectory lines that intersect with a certain stop line, and use the filtered trajectory lines as the trajectory lines collected multiple times for the same road segment in the same direction.

[0027] For lanes traveling in the same direction on the same road segment, the latest trajectory line is retained based on the data collection time, and the remaining trajectory lines are no longer retained.

[0028] Furthermore, the specific process for selecting the entry and exit trajectory points of all intersections is as follows:

[0029] For the intersection currently awaiting processing, identify all related road segments;

[0030] The trajectory line of a vehicle entering the intersection to be processed on the associated road segment is defined as the entry trajectory line, and the trajectory line of a vehicle exiting the intersection to be processed on the associated road segment is defined as the exit trajectory line.

[0031] Traverse each entry trajectory line of the intersection to be processed, select the intersection trajectory line that connects with the entry trajectory line, and take the first trajectory point after entering the intersection area on the intersection trajectory line as the entry trajectory point of the associated road segment within the intersection.

[0032] Iterate through each exit trajectory line of the intersection to be processed, select the intersection trajectory line that connects with the exit trajectory line, and take the last trajectory point of the exit area on the intersection trajectory line as the exit trajectory point of the associated road segment within the intersection.

[0033] Apart from the designated entry and exit trajectory points, no other trajectory points are retained within the intersection.

[0034] Furthermore, the specific process for establishing the jump relationship between trajectory points is as follows:

[0035] For each trajectory point within the intersection, a reference point is determined as the endpoint of the road segment trajectory line corresponding to that trajectory point within the pedestrian crossing line;

[0036] For a given reference point, the direction it can jump to is the one that is closest to the other reference points in the related road segments;

[0037] The jump relationships between reference points are mapped onto their corresponding trajectory points to form the jump relationships between trajectory points.

[0038] Furthermore, the specific process for obtaining the IDs of each street view image is as follows:

[0039] The street view trajectory point set is formed by the trajectory points retained at the intersection and all the trajectory points that make up the trajectory lines of the road segment;

[0040] Spatial calculations are performed on street view trajectory points and grid surfaces, and street view images falling into a certain grid surface are arranged in the original numbering order;

[0041] Each street view image is re-encoded with a normalized sequential code, and the re-encoded sequential code is merged with the grid code to form the ID of the street view image.

[0042] Furthermore, the street view image topology table is specifically represented as follows:

[0043] If the current image is in the middle of a road segment, the associated image ID is the ID of the previous and next images; if the current image is at an intersection, the associated image is determined and stored according to the jump relationship at the intersection; the angle is the angle between the line connecting the associated image and the current image and the north direction.

[0044] Furthermore, the method also includes the following specific process for calculating the angle between street view images for each record in the street view image topology table:

[0045] If the two street view image trajectory points corresponding to this record are not the entry trajectory point and the exit trajectory point, calculate the angle between the two street view image trajectory points;

[0046] If the current street view image trajectory point corresponding to this record is an entry trajectory point within the intersection, and the associated image is an entry or exit trajectory point within the intersection to be jumped to, then an illustrative direction calculation is performed:

[0047] Select the trajectory line within a certain distance of the associated road segment as the driving trend line and calculate the illustrative direction; obtain the road segment trajectory line corresponding to the trajectory point to be jumped to, cut the line segment of the trajectory line at a set distance in front of it, and generate the outer rectangle of the line segment.

[0048] Extract the centerline of the circumscribed rectangular skeleton, calculate the angle between the starting and ending coordinates of the skeleton centerline, obtain the schematic direction, and store it.

[0049] Furthermore, the formula for calculating the angle between the coordinates is:

[0050] y = sin(x) p2 -x p1 )×cosy p2

[0051] x = cosy p1 ×siny p2 -siny p1× cosy p2 ×cos(x p2 -x p1 )

[0052] A = tan -1 (y÷x)

[0053] In the formula, (x p1 y p1 ) and (x p2 y p2 ) are the coordinates of the two points to be calculated.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1) This invention utilizes more refined road traffic marking elements in road holographic mapping to divide the street scene trajectory collected by the 3D scanning vehicle into two parts, intersection and road segment, for separate processing, which is more efficient and reasonable;

[0056] 2) By selecting trajectory points within intersections and constructing topological relationships, the intersection relationships of different trajectory lines are established to achieve an experience consistent with actual street conditions when browsing the panorama;

[0057] 3) Use illustrative directions when calculating the driving direction between jump trajectory points within an intersection to avoid dizziness caused by large differences in the driving direction of the trajectory points themselves after the jump, thus providing a better visual experience for users. Attached Figure Description

[0058] Figure 1 This is a general flowchart of an embodiment of the present invention;

[0059] Figure 2 This is a schematic diagram of the intersection and road segment trajectory line division according to an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of trajectory line selection at road sections according to an embodiment of the present invention;

[0061] Figure 4 This is a schematic diagram illustrating the selection of trajectory points and the construction of jump relationships at intersections according to an embodiment of the present invention. Detailed Implementation

[0062] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0063] Example 1

[0064] This invention fully utilizes the results of road holographic mapping to extract road traffic markings such as pedestrian crossings, stop lines, and lane lines. Through spatial calculations between road traffic markings and street scene image trajectory lines, it divides road segment street scenes and intersection street scenes. It then designs corresponding processing methods for the trajectory characteristics of road segment street scenes and intersection street scenes respectively, selects repeated road segment street scenes collected multiple times, and establishes association jump relationships between key trajectory points at intersections, forming a method for constructing street scene image data topology relationships.

[0065] The specific process of the street view image topology relationship construction method based on road traffic markings provided by this invention is as follows: Figure 1 As shown, it includes the following steps:

[0066] Step 1: Obtain the planar coordinate information of each street view image from the street view image parameter file, spatialize it into trajectory points, and perform trajectory point cleaning and trajectory line generation.

[0067] Step 1.1: Traverse all street view image trajectory points in reverse order of acquisition. Set a distance threshold (take 1 / 5 of the exposure distance when the mobile measurement vehicle is collecting data). If a trajectory point is within the threshold range from the previous trajectory point, remove the trajectory point from the dataset displayed on the front end. This step aims to eliminate the dense exposure of the acquisition vehicle within a short distance due to reasons such as waiting for traffic lights at intersections.

[0068] Step 1.2: Connect the cleaned trajectory points into trajectory lines according to the image exposure order during a single acquisition, with one trajectory line corresponding to each acquisition vehicle.

[0069] Step 2, as follows Figure 2 As shown, pedestrian crossing lines and road surface data collected by intelligent holographic mapping are spatially calculated with the trajectory lines to divide the trajectory lines into intersection trajectory lines and road segment trajectory lines.

[0070] Step 2.1: Pedestrian crossing elements are collected. The boundaries of the pedestrian crossings form closed lines. For each pedestrian crossing element, an outer rectangle is generated, and the center line of the outer rectangle in the same direction as its long side is extracted.

[0071] Step 2.2: Use the extracted center line of the pedestrian crossing to break the trajectory lines that intersect with it;

[0072] Step 2.3: Filter out the road intersection surfaces in the road surface data, and perform spatial calculations on the broken trajectory lines and intersection surfaces to determine if the lines are within the surface. The trajectory lines located within the road intersection surfaces are the intersection trajectory lines.

[0073] Step 2.4, the trajectory lines other than the intersection trajectory lines are the road segment trajectory lines;

[0074] Step 3: For all the road segment trajectory lines divided in Step 2, use the stop line elements (stop line is the parking positioning line for vehicles waiting for the traffic signal) collected by intelligent holographic mapping to filter out the trajectory lines collected multiple times in the same direction of the same road segment, and retain the one with the latest collection time.

[0075] Step 3.1: Perform a "line intersection" operation on the stop line elements and trajectory line elements to filter out all road segment trajectory lines that intersect with a certain stop line. These trajectory lines are the trajectory lines collected multiple times for the same road segment in the same direction.

[0076] Step 3.2, as follows Figure 3 As shown, for lanes traveling in the same direction on the same road segment, the latest trajectory line is retained based on the collection time, and the remaining trajectory lines are no longer retained;

[0077] Step 4: For the intersection trajectory lines divided in Step 2, select any intersection as the current processing intersection and select intersection trajectory points until all intersections have been processed.

[0078] Step 4.1: For the current intersection to be processed, determine all its associated road segments. For example, a crossroads has four associated road segments, and a T-junction has three associated road segments.

[0079] Step 4.2: Define the trajectory lines of vehicles entering the intersection to be processed on the associated road segment as entry trajectory lines, and define the trajectory lines of vehicles exiting the intersection to be processed on the associated road segment as exit trajectory lines. After step 3, each associated road segment has one entry trajectory line and one exit trajectory line;

[0080] Step 4.3: Traverse each entry trajectory line at the intersection, select the intersection trajectory line that connects with the entry trajectory line, and take the first trajectory point on the intersection trajectory line after entering the intersection area as the entry trajectory point of the associated road segment within the intersection.

[0081] Step 4.4: Traverse each exit trajectory line at the intersection, select the intersection trajectory line that connects with the exit trajectory line, and take the last trajectory point of the exit area on the intersection trajectory line as the exit trajectory point of the associated road segment within the intersection.

[0082] In step 4.5, apart from the entry and exit trajectory points determined in steps 4.3 and 4.4, no other trajectory points are retained within the intersection;

[0083] Step 5: For the remaining trajectory points within the intersection, iterate through each trajectory point, determine other trajectory points it can jump to, and establish the jump relationships between trajectory points, such as... Figure 4 As shown;

[0084] Step 5.1: First, determine a reference point for each trajectory point within the intersection, which is the endpoint of the road segment trajectory line corresponding to that trajectory point within the pedestrian crossing line;

[0085] Step 5.2: For a given reference point, the possible jump direction is the one that is closer to the other two reference points in each road segment;

[0086] Step 5.3: Map the jump relationships between reference points to their corresponding trajectory points to form the jump relationships between trajectory points;

[0087] Step 6: Reorganize and rename the street view images corresponding to the processed trajectory points according to the grid to ensure that the ID of the street view image is unique.

[0088] Step 6.1: After the aforementioned processing, the trajectory points retained at the intersection and all trajectory points constituting the road segment trajectory lines together form the set of street view trajectory points displayed at the front end.

[0089] Step 6.2: Perform spatial calculations on the street view trajectory points and grid surfaces, and arrange the street view images that fall into a certain grid surface according to their original numbering order;

[0090] Step 6.3: Re-add a sequence number to each street view image and normalize it into a seven-digit sequence code;

[0091] Step 6.4: Combine the newly compiled seven-digit sequence code with the grid code to form "grid number_seven-digit sequence code", which will serve as the new ID for the street view image;

[0092] Step 7: Design the street view image topology table, as shown in Table 1, which includes three fields: current image ID, associated image ID, and angle. If the current image is in the middle of a road segment, its associated image ID is the ID of the previous and next images. If the current image is at an intersection, it may have multiple associated images, which need to be determined and stored according to the construction of the jump relationship at the intersection. The angle is the angle between the line connecting the associated image and the current image and the north direction.

[0093] Table 1. Storage table structure of street view image topology in this embodiment of the invention.

[0094] 1 Current Image ID character varying(50) Image ID 2 Associated Image ID character varying(50) Image ID associated with the current image 3 angle numeric The angle between the line connecting the associated image and the current image and the north direction.

[0095] Step 8: Generate and store the street view image topology relationship table designed in Step 7 for each road segment;

[0096] Step 9: Based on the street view image topology table designed in Step 7 and the trajectory point jump relationship established in Step 5, generate and store the street view image topology relationship at the intersection.

[0097] Step 10: Calculate the angle between the associated street view images for each record in the street view image topology table;

[0098] Step 10.1: If the two street view image trajectory points corresponding to this record are not the entry trajectory point and the exit trajectory point, the angle calculation formula is as follows:

[0099] y = sin(x) p2 -x p1 )×cosy p2 Equation 10-1

[0100] x = cosy p1 ×siny p2 -siny p1× cosy p2 ×cos(x p2 -x p1 Equation 10-2

[0101] A = tan -1(y÷x) equation 10-3

[0102] In the formula (x p1 y p1 (x) represents the coordinates of the trajectory point in the current image. p2 y p2 ) represents the coordinates of the trajectory points associated with the street view image.

[0103] Step 10.2: If the current street view image trajectory point corresponding to the record is an entry trajectory point within the intersection, and the associated image is an entry or exit trajectory point within the intersection to be jumped to, then the illustrative direction is calculated.

[0104] Step 11: Select the trajectory line within a certain distance of the associated road segment as the driving trend line and calculate the illustrative direction;

[0105] Step 11.1: Obtain the road segment trajectory line corresponding to the trajectory point to be jumped to, extract the first 50 meters of the trajectory line, and generate the bounding rectangle of this line segment.

[0106] Step 11.2: Extract the center line of the circumscribed rectangular skeleton, substitute the starting coordinates and ending coordinates of the center line into the angle calculation formula in step 10.1, calculate the schematic direction and store it;

[0107] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for constructing topological relationships from street view images based on road traffic markings, characterized in that, Includes the following steps: The planar coordinate information of the street view image is obtained from the street view image parameter file and spatialized into trajectory points. The trajectory points are then cleaned and trajectory lines are generated. The collected pedestrian crossing and road surface data are then used to perform spatial calculations with the generated trajectory lines, dividing the trajectory lines into intersection trajectory lines and road segment trajectory lines. The specific process is as follows: Collect the closed line formed by the boundary of the pedestrian crossing as the pedestrian crossing line element, generate the bounding rectangle for each pedestrian crossing line element, and extract the center line of the long side of the bounding rectangle in the same direction as the center line of the pedestrian crossing line. Use the extracted pedestrian crossing centerline to break the trajectory lines that intersect with the centerline; Filter the road intersection surfaces in the road surface data, perform spatial calculations on the lines and intersection surfaces after breaking the trajectory lines, and take the trajectory lines located in the road intersection surfaces as the intersection trajectory lines; Trajectory lines other than those at intersections are considered as road segment trajectory lines; For all the defined road segment trajectory lines, use the stop line element to filter the trajectory lines collected multiple times for the same road segment in the same direction, and retain one of them; Based on the defined intersection trajectory lines, the entry and exit trajectory points of all intersections are selected; Based on the selected entry and exit trajectory points at intersections, determine the trajectory points that each trajectory point can jump to, and establish the jump relationship between trajectory points; After establishing the jump relationship, the street view images corresponding to the trajectory points are reorganized and renamed according to the grid to obtain a unique ID for each street view image. Design a street view image topology table, including three fields: current image ID, associated image ID, and angle; Based on the street view image topology table, generate and store the street view image topology relationships at road segments; and based on the street view image topology table and the trajectory point jump relationships within intersections, generate and store the street view image topology relationships at intersections.

2. The method for constructing street view image topology based on road traffic markings according to claim 1, characterized in that, The specific process for generating the trajectory line is as follows: Traverse all street view image trajectory points in reverse order of acquisition; set a distance threshold, and if a trajectory point is within the threshold range of the previous acquired trajectory point, then discard that trajectory point. After cleaning, the trajectory points are connected into trajectory lines according to the image exposure order during a single acquisition, with one trajectory line corresponding to each acquisition vehicle.

3. The method for constructing street view image topology based on road traffic markings according to claim 1, characterized in that, The specific process of using stop line features to filter trajectory lines collected multiple times for the same lane in the same direction on the same road segment is as follows: Perform line intersection operation between stop line elements and trajectory line elements, filter out all road segment trajectory lines that intersect with a certain stop line, and use the filtered trajectory lines as the trajectory lines collected multiple times for the same road segment in the same direction. For lanes traveling in the same direction on the same road segment, the latest trajectory line is retained based on the data collection time, and the remaining trajectory lines are no longer retained.

4. The method for constructing topological relationships of street view images based on road traffic markings according to claim 1, characterized in that, The specific process for selecting the entry and exit trajectory points of all intersections is as follows: For the intersection currently awaiting processing, identify all related road segments; The trajectory line of a vehicle entering the intersection to be processed on the associated road segment is defined as the entry trajectory line, and the trajectory line of a vehicle exiting the intersection to be processed on the associated road segment is defined as the exit trajectory line. Traverse each entry trajectory line of the intersection to be processed, select the intersection trajectory line that connects with the entry trajectory line, and take the first trajectory point after entering the intersection area on the intersection trajectory line as the entry trajectory point of the associated road segment within the intersection. Iterate through each exit trajectory line of the intersection to be processed, select the intersection trajectory line that connects with the exit trajectory line, and take the last trajectory point of the exit area on the intersection trajectory line as the exit trajectory point of the associated road segment within the intersection. Retain the identified entry and exit trajectory points, and delete other trajectory points within the intersection.

5. The method for constructing street view image topology based on road traffic markings according to claim 4, characterized in that, The specific process for establishing the jump relationship between trajectory points is as follows: For each trajectory point within the intersection, a reference point is determined as the endpoint of the road segment trajectory line corresponding to that trajectory point within the pedestrian crossing line; For a given reference point, the direction it can jump to is the one that is closest to the other reference points in the related road segments; The jump relationships between reference points are mapped onto their corresponding trajectory points to form the jump relationships between trajectory points.

6. The method for constructing street view image topology based on road traffic markings according to claim 1, characterized in that, The specific process for obtaining the IDs of each street view image is as follows: The street view trajectory point set is formed by the trajectory points retained at the intersection and all the trajectory points that make up the trajectory lines of the road segment; Spatial calculations are performed on street view trajectory points and grid surfaces, and street view images falling into a certain grid surface are arranged in the original numbering order; Each street view image is re-encoded with a normalized sequential code, and the re-encoded sequential code is merged with the grid code to form the ID of the street view image.

7. The method for constructing topological relationships of street view images based on road traffic markings according to claim 1, characterized in that, The street view image topology table is specifically represented as follows: If the current image is in the middle of a road segment, the associated image ID is the ID of the previous and next images; if the current image is at an intersection, the associated image is determined and stored according to the jump relationship at the intersection; the angle is the angle between the line connecting the associated image and the current image and the north direction.

8. The method for constructing street view image topology based on road traffic markings according to claim 1, characterized in that, The method also involves calculating the angle between street view images for each record in the street view image topology table, as follows: If the two street view image trajectory points corresponding to this record are not the entry trajectory point and the exit trajectory point, calculate the angle between the two street view image trajectory points; If the current street view image trajectory point corresponding to this record is an entry trajectory point within the intersection, and the associated image is an entry or exit trajectory point within the intersection to be jumped to, then an illustrative direction calculation is performed: Select the trajectory line within a certain distance of the associated road segment as the driving trend line and calculate the illustrative direction; Obtain the road segment trajectory line corresponding to the trajectory point to be jumped to, extract the line segment at a set distance from the beginning of the trajectory line, and generate the bounding rectangle of the line segment; Extract the centerline of the circumscribed rectangular skeleton, calculate the angle between the starting and ending coordinates of the skeleton centerline, obtain the schematic direction, and store it.

9. A method for constructing topological relationships of street view images based on road traffic markings according to claim 8, characterized in that, The formula for calculating the angle between the coordinates is: In the formula, (x p1 y p1 ) and (x p2 y p2 ) are the coordinates of the two points to be calculated.

Images