Method for evaluating the reasonableness of road marking setting based on crowd-sourced vehicle-mounted perception information

Abstract

The present application relates to a kind of road marking setting rationality evaluation method based on crowd-sourcing vehicle-mounted sensing information, comprising the following steps: S1, the basic attribute information of each road section unit is extracted, and scene classification is completed;S2, vehicle-mounted terminal is perceived lane line by camera, and positioning module obtains vehicle positioning information;S3, line-pressing event is recorded and attribute is marked, and line-pressing timestamp and longitude and latitude are uploaded to cloud platform;S4, cloud platform matches line-pressing event and space-time information to corresponding road section unit;S5, line-pressing event is aggregated according to different time periods, and the event space-time feature table of road section unit by unit is formed;S6, group vehicle line-pressing event is marked to form list according to preset rule determination group vehicle line-pressing event;S7, output road marking setting rationality evaluation conclusion and grade according to preset rule.The present application provides quantifiable basis for highway marking maintenance and optimization, reduces marking unreasonable, and improves road traffic safety.

Description

Technical Field

[0001] This invention relates to the fields of traffic management and intelligent transportation, and more specifically, to a method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information. Background Technology

[0002] Road markings, as crucial control facilities for road traffic organization, are used to regulate vehicle lane-changing behavior and ensure the operational order of key areas such as merging and diverging. Due to factors such as mismatches between road marking placement and traffic organization, violations such as crossing the lines are increasingly common on some road sections. Therefore, accurately identifying improperly placed road markings is of great significance for optimizing road facilities and managing violations. However, existing technologies have the following limitations in assessing the rationality of road marking placement: Insufficient aggregation and judgment of spatiotemporal features. Existing methods mostly focus on identifying and warning of events at fixed locations or times, lacking aggregated statistics of line-crossing events across multiple spatiotemporal dimensions, making it difficult to identify high-incidence road marking locations for violations. Insufficient collaborative utilization of crowdsourced vehicle-mounted sensing information. Existing technologies have low utilization of multi-source crowdsourced vehicle-mounted sensing information, lacking collaborative processing methods for vehicle-side collection and cloud-based aggregation, making it difficult to fully utilize large-scale sensing data from multiple vehicles and time periods. Insufficient evaluation of the rationality of road marking placement. Traditional road marking detection methods are mostly focused on marking location or damage identification, primarily concerned with the visual performance of the markings. The test results were not considered for further transformation into indicators and criteria for evaluating the rationality of the benchmark settings. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information, which can provide quantifiable basis for the maintenance and optimization of road markings, thereby reducing unreasonable marking layout and improving road traffic safety.

[0004] The technical solution adopted by this invention to solve its technical problem is: to construct a method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information, comprising the following steps: S1. Divide the roads in the area to be evaluated into road segment units according to the preset segmentation rules and assign road segment labels. Extract the basic attribute information of each road segment unit and complete the scene classification based on lane function and lane line type elements. S2. The connected vehicle terminal senses lane lines through a camera and obtains vehicle location information through a positioning module. S3. The vehicle terminal processing module identifies line crossing events of its own vehicle and other vehicles, records and labels the line crossing events, and uploads the line crossing timestamp and latitude and longitude to the cloud platform. S4. The cloud platform will match the line crossing event and its spatiotemporal information to the corresponding road segment unit. S5. Aggregate and statistically analyze multi-vehicle lane crossing events according to different time periods to form a spatiotemporal characteristic table of events for each unit of the road segment; S6. Based on the aggregation characteristics of road segment unit line crossing events in time and space, determine group vehicle line crossing events according to preset rules, and mark the group vehicle line crossing events to form a list. S7. Based on the road segment unit event list, output the rationality assessment conclusion and level of road marking settings according to preset rules.

[0005] According to the above scheme, in step S1, the segmentation rules include fixed-length segmentation rules, feature boundary segmentation rules, and combined segmentation rules; the basic attribute information of the road segment unit includes road segment identifier, road segment length, number of lanes, lane function, and lane line type; lane function includes mainline section, divergence section, merging section, entrance / exit related section, and guide zone related section; scene classification classifies the road segment unit according to lane function and lane line type and generates scene category identifier.

[0006] According to the above scheme, in step S2, the lane line perception information includes a timestamp and vehicle location information; the vehicle positioning information includes latitude and longitude and mileage markers, the latitude and longitude are used to upload to the cloud platform, and the mileage markers are used for spatial positioning and road segment unit matching of subsequent lane crossing events.

[0007] According to the above scheme, in step S3, the vehicle terminal processing module identifies lane line crossing events of its own vehicle and other vehicles within the perception range of the vehicle terminal based on the relative positional relationship information between the vehicle and the lane line, and uploads the lane line crossing timestamp and latitude and longitude to the cloud platform; the relative positional relationship information between the vehicle and the lane line includes: the lateral distance between the vehicle boundary and the left lane line boundary. Lateral distance between the vehicle boundary and the right lane line boundary and lane line width .

[0008] According to the above scheme, the method for determining the line-pressing status based on relative positional relationships, identifying line-pressing events and recording attributes through a preset time threshold includes the following steps: S301, if the lateral distance between the vehicle's boundary and the boundaries of the left and right lane lines... The vehicle is determined to be in a normal driving condition; if If so, the vehicle is determined to be in a state of crossing the line; S302. Based on the pressure line state sequence obtained in step S301, the vehicle-mounted terminal processing module extracts continuous state segments, with the timestamp recorded at the beginning of each segment being... The timestamp recorded at the end of the segment is , Duration of the segment ,like Event time threshold greater than the preset threshold If so, the continuous segment is determined to be a line-crossing event.

[0009] According to the above scheme, the recorded attributes of the lane crossing event include the event type, start and end timestamps, event duration, event location, lane line type, and lane line side.

[0010] According to the above scheme, in step S4, the cloud platform performs map matching between the line crossing event and spatiotemporal information based on the vehicle positioning information, and maps it to the corresponding road segment unit.

[0011] According to the above scheme, in step S5, multiple vehicle crossing events within the same road segment unit are aggregated and statistically analyzed by time period, and a spatiotemporal feature table of events for each unit of the road segment is output. The spatiotemporal feature table of events for each unit of the road segment includes the road segment identifier, the number of crossings, the cumulative observation time, and the road segment length.

[0012] According to the above scheme, in step S6, based on the aggregation characteristics of the road segment unit line crossing events in time and space obtained in step S5, the group vehicle line crossing events are determined according to the threshold rule, marked and formed into a road segment unit event list; adjacent merging and other methods are used to spatially aggregate continuous group vehicle line crossings and output the road segment unit event list.

[0013] According to the above scheme, in step S7, based on the road segment unit event list and the road segment unit event spatiotemporal characteristic table, the rationality of road marking settings is comprehensively evaluated from two dimensions: event quantity and event intensity, and the rationality evaluation conclusion and level of road marking settings are output according to the preset grading rules.

[0014] The road marking setting rationality evaluation method based on crowdsourced vehicle-mounted perception information of the present invention has the following beneficial effects: This invention enables the evaluation and classification of the rationality of road marking settings using low-cost crowdsourced vehicle-mounted sensing data without the need to deploy additional road testing equipment. It can provide quantifiable basis for the maintenance and optimization of highway markings, thereby reducing unreasonable marking layouts and improving road traffic safety. Attached Figure Description

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a flowchart of the road marking setting rationality evaluation method based on crowdsourced vehicle-mounted perception information of the present invention; Figure 2 This is a schematic diagram of the first embodiment of the road marking setting rationality evaluation method based on crowdsourced vehicle-mounted perception information of the present invention; Figure 3 This is a schematic diagram of a line-crossing event from the first embodiment of the road marking setting rationality evaluation method based on crowdsourced vehicle-mounted perception information of the present invention. Detailed Implementation

[0016] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0017] like Figure 1-3 As shown, the road marking setting rationality evaluation method based on crowdsourced vehicle-mounted perception information of the present invention includes the following steps: S1. Divide the roads in the evaluation area into road segment units according to the preset segmentation rules and assign road segment labels. Extract the basic attribute information of each road segment unit and complete the scene classification based on elements such as lane function and lane line type.

[0018] Segmentation rules include one or more of the following: fixed-length segmentation rules, feature-boundary segmentation rules, and combined segmentation rules. The basic attribute information of a road segment unit includes road segment identifier, road segment length, number of lanes, lane function, and lane line type; lane functions include mainline segments, divergence segments, merging segments, entrance / exit related segments, and guide zone related segments; scene classification categorizes road segment units based on lane function and lane line type and generates scene category identifiers.

[0019] Taking fixed-length segments as an example, the rules are as follows: using the road reference line as a baseline, according to the preset segment length... The road is divided into road segment units; This is a configurable parameter, which can be 50m or 100m; each road segment unit includes a start point and an end point, and each road segment unit is assigned a unique value. The actual length of the road segment unit is calculated based on its start and end positions. Among them, the non-boundary road segment units satisfy The boundary segment unit satisfies .

[0020] The extracted basic information includes road segment unit ID, start and end point locations, road segment length, number of lanes, lane function, and lane line type. This information is used for scene classification, providing a basis for subsequent scene recognition and lane crossing event determination. For example, in functional areas such as merging and diverging zones, vehicles frequently cross lane lines. Using the same determination method as for mainline sections could easily lead to misjudgments. Therefore, it is necessary to design differentiated determination rules based on different scenarios.

[0021] S2. The connected vehicle terminal senses lane lines through cameras, and the positioning module obtains vehicle positioning information.

[0022] The structured data transmitted by the connected vehicle terminal includes timestamps and vehicle location; the vehicle information acquired by the positioning module includes latitude and longitude information and mileage marker information synchronized with the structured data. The latitude and longitude information uses... It means that among them Longitude Latitude. Mileage markers can be used. The location is represented in a format indicating that the vehicle is currently 15.1 km away in the given direction on the road, thus mapping the event to a road segment unit. Latitude and longitude coordinates are used to upload data to the cloud platform, while the mileage marker is used for subsequent spatial positioning and matching of line-crossing events with road segment units. S3, the vehicle terminal processing module identifies line crossing events of its own vehicle and other vehicles, records and labels the line crossing events, and uploads the line crossing timestamp and latitude and longitude to the cloud platform.

[0023] The vehicle-mounted terminal processing module identifies lane line crossing information, including lane line information and the relative positional relationship between the vehicle and the lane lines. Based on the relative positional relationship and a preset time threshold, it identifies lane line crossing events of the vehicle itself and other vehicles within the perception range of the vehicle-mounted terminal, and uploads the lane line crossing timestamp and latitude and longitude to the cloud platform. The relative positional relationship information between the vehicle and the lane lines includes: the lateral distance between the vehicle boundary and the left lane line boundary. Lateral distance between the vehicle boundary and the right lane line boundary Lane width To standardize the determination of the left and right sides of lane lines, the lateral distance between the vehicle boundary and the boundaries of the left and right lane lines is defined. A positive value indicates a gap between the vehicle boundary and the lane line boundary, while a negative value indicates overlap or intrusion between the vehicle and the lane line. For each frame of video data, based on... With lane line width Determine vehicle status:

[0024] Based on the vehicle state sequence, continuous lane-crossing segments are extracted. The timestamp recorded at the beginning of each segment is... The timestamp recorded at the end of the segment is The duration of the segment is:

[0025] like Event time threshold greater than the preset threshold If so, the continuous segment is determined to be a line-crossing event. A timeout of 0.3-1.0 seconds is recommended to avoid false detections or missed wire-pressing events.

[0026] For continuous segments identified as lane-crossing events, event attributes are labeled. Event attributes include: event type, start timestamp, end timestamp, event duration, event location, lane line type, and lane line side.

[0027] S4, the cloud platform will match the line-crossing event and its spatiotemporal information to the corresponding road segment unit. The cloud platform matches the lane-crossing event with spatiotemporal information based on vehicle location information on the map and maps it to the corresponding road segment unit.

[0028] Vehicle location information, using mileage markers as an example, is represented as follows: Convert to mileage location :

[0029] With preset segment length For example, the road segment marker corresponding to the location where the line crossing incident occurred is:

[0030] but The corresponding road segment sign is This allows the line-crossing event to be matched to the corresponding road segment unit. The timestamp information of the line-crossing event is then matched synchronously, as follows:

[0031] in, The timestamp of the time when the incident occurred. The timestamp at which the statistics began. This is the preset statistics window. For example, if the timestamp at the start of the statistics is 12:00:00, and a line crossing event occurs at 14:00:01 on the same day, then the time ID of the event is recorded as 3, and thus matched to the corresponding road segment unit.

[0032] S5. Aggregate and statistically analyze multi-vehicle lane crossing events according to different time periods to form a spatiotemporal characteristic table of events for each unit of the road segment.

[0033] Multiple vehicle lane-crossing events within the same road segment unit are aggregated and statistically analyzed by time period, and a spatiotemporal characteristic table of events for each unit of the road segment is output. The spatiotemporal characteristic table of events for each unit of the road segment includes the road segment identifier, the number of lane-crossing events, the cumulative observation duration, and the road segment length.

[0034] Aggregate and statistically analyze multiple vehicle lane-crossing events at different times that are matched to the same road segment unit. The aggregation and statistical results are output in the form of a spatiotemporal feature table of events per unit of the road segment. It can be divided by hour, day, or by peak and off-peak hours. Number of times the line is crossed. For statistics window The total number of line-crossing events that occurred within this road section, and the cumulative observation duration. For statistics window The sum of the detection times of all vehicles passing through the unit within this road segment is calculated as follows.

[0035]

[0036] in In the statistics window The number of vehicles that have been inspected in this section of the road. Let be the detection time for the i-th vehicle within the road segment unit.

[0037] Table 1. Spatiotemporal characteristics of events per unit along a road segment

[0038] S6. Based on the aggregation characteristics of road segment unit line crossing events in time and space, determine group vehicle line crossing events according to preset rules, and mark the group vehicle line crossing events to form a list.

[0039] Based on the aggregation characteristics of the road segment unit line crossing events in time and space obtained in step S5, the group vehicle line crossing events are determined according to the threshold rule, marked and formed into a road segment unit event list; and further, methods such as adjacent merging can be used to spatially aggregate continuous group vehicle line crossings and output the road segment unit event list.

[0040] The determination of lane-crossing events by group vehicles is based on records in the spatiotemporal characteristic table of events per unit of road segment. Statistical windows are divided by hour. For example, to avoid misjudging incidents of group vehicles crossing the line due to insufficient cumulative observation time, if the cumulative observation time is longer... Less than a given threshold If the detection data for that road segment is insufficient to determine a group of vehicles crossing the line, then it is determined that the detection data for that segment is insufficient to determine a line-crossing event. The time interval can be set to 120-300 seconds depending on the number of observation vehicles or the length of the road segment unit to ensure sufficient data for determining group vehicle crossing events.

[0041] In satisfying Under the premise of [specific conditions], threshold rules are used to determine the line-crossing events of each road segment unit. When the threshold rule conditions are met, the road segment unit is determined to have a group of vehicles crossing the line. The threshold rule can be an event frequency threshold rule, a unit time event rate threshold rule, etc., where the threshold rule is set differently according to the scenario category.

[0042] Event frequency threshold rules:

[0043] in Given a pre-defined threshold for the number of events, if a certain road segment unit is in the same statistical window If the number of lane-crossing events occurring within a given event count threshold is greater than or equal to the number of events, then it is determined that there are group vehicle lane-crossing events in that road segment unit.

[0044] Event rate per unit time The following data was calculated from the spatiotemporal characteristic table of each unit event of the road segment:

[0045] Event rate threshold rule per unit time:

[0046] in Given a pre-defined threshold for the event rate per unit time, if a certain road segment unit falls within the statistical window... The event rate per unit time within the specified time is greater than or equal to a given event rate per unit time threshold. If so, it is determined that there is a group of vehicles crossing the line in that road segment.

[0047] If there exists a pair of adjacent road segment units In the same statistics window If all adjacent road segments are determined to have experienced a group of vehicles crossing the lane line, then adjacent segments will be spatially aggregated using a neighbor merging method. The spatial range can be determined using... It is indicated that the road segment markings can use the section center stake number. This is indicated. An example of a list of events for a road segment unit is shown in Table 2.

[0048] Table 2 List of events for road segment units

[0049] S7. Based on the road segment unit event list, output the rationality assessment conclusion and level of road marking settings according to preset rules.

[0050] Based on the list of events per road segment unit and the spatiotemporal characteristics table of events per road segment unit, the rationality of road marking settings is comprehensively evaluated from two dimensions: the number of events and the intensity of events. The evaluation conclusion and level of the rationality of road marking settings are output according to the preset grading rules.

[0051] Number of events and the intensity of the event A comprehensive assessment of the rationality of road marking placement is conducted. The intensity of the incident is determined by the coverage area of ​​the road markings. The maximum event rate per unit time for all group vehicle lane-crossing events. The road marking rationality grading rules are as follows: Reasonable: and ; Things to note: or ; unreasonable: or ; in , , , All are preset thresholds, with statistical windows divided by hours. For example, among which The value can be between 1 and 2. The value can be between 2 and 4; The range of values ​​can be: , The range of values ​​can be: By number of events With event intensity The assigned classification rule outputs the rationality level of the road marking settings.

[0052] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information, characterized in that, Includes the following steps: S1. Divide the roads in the area to be evaluated into road segment units according to the preset segmentation rules and assign road segment labels. Extract the basic attribute information of each road segment unit and complete the scene classification based on lane function and lane line type elements. S2. The connected vehicle terminal senses lane lines through a camera and obtains vehicle location information through a positioning module. S3. The vehicle terminal processing module identifies line crossing events of its own vehicle and other vehicles, records and labels the line crossing events, and uploads the line crossing timestamp and latitude and longitude to the cloud platform. S4. The cloud platform will match the line crossing event and its spatiotemporal information to the corresponding road segment unit. S5. Aggregate and statistically analyze multi-vehicle lane crossing events according to different time periods to form a spatiotemporal characteristic table of events for each unit of the road segment; S6. Based on the aggregation characteristics of road segment unit line crossing events in time and space, determine group vehicle line crossing events according to preset rules, and mark the group vehicle line crossing events to form a list. S7. Based on the road segment unit event list, output the rationality assessment conclusion and level of road marking settings according to preset rules.

2. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S1, the segmentation rules include fixed-length segmentation rules, feature boundary segmentation rules, and combined segmentation rules; the basic attribute information of the road segment unit includes road segment identifier, road segment length, number of lanes, lane function, and lane line type; lane function includes mainline section, divergence section, merging section, entrance / exit related section, and guide zone related section; scene classification classifies the road segment unit according to lane function and lane line type and generates scene category identifier.

3. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S2, the lane line perception information includes a timestamp and vehicle location information; the vehicle positioning information includes latitude and longitude and mileage markers, the latitude and longitude are used to upload to the cloud platform, and the mileage markers are used for spatial positioning and road segment unit matching of subsequent lane crossing events.

4. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S3, the vehicle terminal processing module identifies lane crossing events of its own vehicle and other vehicles within the perception range of the vehicle terminal based on the relative positional relationship information between the vehicle and the lane line, and uploads the lane crossing timestamp and latitude and longitude to the cloud platform. The relative positional relationship information between the vehicle and the lane lines includes: the lateral distance between the vehicle boundary and the left lane line boundary. Lateral distance between the vehicle boundary and the right lane line boundary and lane line width .

5. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 4, characterized in that, The method for determining the line-pressing status based on relative positional relationships, and for identifying line-pressing events and recording attributes through a preset time threshold, includes the following steps: S301, if the lateral distance between the vehicle's boundary and the boundaries of the left and right lane lines... The vehicle is determined to be in a normal driving condition; if If so, the vehicle is determined to be in a state of crossing the line; S302. Based on the pressure line state sequence obtained in step S301, the vehicle-mounted terminal processing module extracts continuous state segments, with the timestamp recorded at the beginning of each segment being... The timestamp recorded at the end of the segment is , Duration of the segment ,like Event time threshold greater than the preset threshold If so, the continuous segment is determined to be a line-crossing event.

6. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 5, characterized in that, The recorded attributes of the lane crossing event include the event type, start and end timestamps, event duration, event location, lane line type, and lane line side.

7. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S4, the cloud platform performs map matching between the line crossing event and spatiotemporal information based on vehicle positioning information, and maps it to the corresponding road segment unit.

8. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S5, multiple vehicle lane-crossing events within the same road segment unit are aggregated and statistically analyzed by time period, and a spatiotemporal feature table of events for each unit of the road segment is output. The spatiotemporal feature table of events for each unit of the road segment includes the road segment identifier, the number of lane-crossing events, the cumulative observation duration, and the road segment length.

9. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S6, based on the aggregation characteristics of the road segment unit line crossing events in time and space obtained in step S5, the group vehicle line crossing events are determined according to the threshold rule, marked and formed into a road segment unit event list; adjacent merging and other methods are used to spatially aggregate continuous group vehicle line crossings and output the road segment unit event list.

10. The method for evaluating the rationality of road marking settings based on crowdsourced vehicle-mounted perception information according to claim 1, characterized in that, In step S7, based on the road segment unit event list and the road segment unit event spatiotemporal characteristic table, the rationality of road marking settings is comprehensively evaluated from two dimensions: event quantity and event intensity. The rationality evaluation conclusion and level of road marking settings are then output according to the preset grading rules.

Images