Map data confidence updating method, device, and storage medium

Abstract

The application provides a map data confidence updating method and device and a storage medium, relates to the technical field of map processing, and comprises the following steps: acquiring map data collected by at least one collection end at least once; associating the map data with road sections of a map according to a driving track of the collection end in the map data, to obtain map data of a plurality of road sections; determining whether the collection end meets an observation condition of the road section; wherein the observation condition is used to determine whether the map data collected by the collection end is reliable; for a target road section meeting the observation condition, using the map data of the target road section to update a confidence of a map element of the target road section in the map and a road topology confidence of the target road section. The application can accurately evaluate the confidence of map data of a high-precision map.

Description

Technical Field

[0001] This application relates to the field of map processing technology, and in particular to a method, device and storage medium for updating map data confidence. Background Technology

[0002] In autonomous driving scenarios, high-definition map data is typically used for positioning and driving mode switching. Therefore, the real-time performance and completeness of map data in high-definition maps are crucial for the safety of autonomous driving.

[0003] High-definition maps (HDMaps) represent real-world physical objects in the digital world. These real-world objects may change over time. However, the collection of map data used to create HDMaps, and the creation of HDMaps themselves, are subject to errors and delays. This means that the updates to HDMaps are also subject to errors and delays, resulting in discrepancies between the map data in HDMaps and the real world. Therefore, accurately evaluating the confidence level of HDMap data is a crucial problem that needs to be solved to ensure the selection of reliable map data for positioning and driving mode switching in autonomous driving scenarios. Summary of the Invention

[0004] This application provides a map data confidence update method, device, and storage medium, which can accurately evaluate the confidence of map data in high-precision maps.

[0005] Firstly, this application provides a method for updating map data confidence, the method comprising:

[0006] Acquire map data from at least one acquisition terminal at least once;

[0007] Based on the driving trajectory of the acquisition terminal in the map data, the map data is associated with road segments in the map to obtain map data of multiple road segments;

[0008] Determine whether the acquisition terminal meets the observation conditions of the road segment; wherein, the observation conditions are used to determine whether the map data acquired by the acquisition terminal is reliable;

[0009] For a target road segment that meets the observation conditions, the confidence level of the map features of the target road segment and the road topology confidence level of the target road segment are updated using the map data of the target road segment.

[0010] Optionally, determining whether the acquisition terminal meets the observation conditions of the road segment includes:

[0011] Extract a first road segment and multiple consecutive second road segments from the multiple road segments. The observation conditions for the first road segment are that it can be located by both the positioning system and visual location. The observation conditions for the second road segment include either visual location or positioning system location.

[0012] Based on the map data of the first road segment and the map elements within the first road segment in the map, determine whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard.

[0013] If it is determined that the camera intrinsic parameter accuracy of the acquisition terminal is not exceeded, then based on the map data of multiple consecutive second road segments, it is determined whether the inertial measurement unit (IMU) of the acquisition terminal is available.

[0014] Obtain the positioning system used by the acquisition terminal;

[0015] Based on the results of whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, whether the IMU is usable, and the positioning system used by the acquisition terminal, it is determined whether the acquisition terminal meets the observation conditions of the road segment.

[0016] Optionally, determining whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard based on the map data of the first road segment and the map features within the first road segment in the map includes:

[0017] The map features within the first road segment are reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features within the first road segment.

[0018] Based on the reprojection error of map features within the first road segment, the camera intrinsic parameters of the acquisition terminal are estimated.

[0019] Based on the estimated camera intrinsic parameters of the acquisition end, the map features in the first road segment are reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features in the first road segment.

[0020] If the proportion of map features whose reprojection error exceeds a preset reprojection error threshold is greater than a preset proportion threshold, then it is determined that the camera intrinsic parameter accuracy of the acquisition end exceeds the standard.

[0021] If the proportion of map features whose reprojection error is greater than a preset reprojection error threshold is less than or equal to the preset proportion threshold, then it is determined that the camera intrinsic parameter accuracy of the acquisition end has not exceeded the standard.

[0022] Optionally, determining whether the inertial measurement unit (IMU) at the acquisition end is available based on map data from multiple consecutive second road segments includes:

[0023] Based on map data of multiple consecutive second road segments, obtain the starting point and ending point of the multiple consecutive second road segments.

[0024] Based on the starting point of the positioning, the endpoint is recursively calculated using the IMU of the acquisition terminal;

[0025] Based on the endpoint of the positioning and the endpoint of the recursion, the recursion accuracy of the IMU at the acquisition end is obtained;

[0026] If the recursive accuracy of the IMU at the acquisition end is greater than or equal to the IMU recursive accuracy threshold, then the IMU at the acquisition end is determined to be available.

[0027] If the recursive accuracy of the IMU at the acquisition end is less than the IMU recursive accuracy threshold, then the IMU at the acquisition end is determined to be unusable.

[0028] Optionally, the observation conditions for the road segment include at least one of the following: visually locatable and locatable by the inertial measurement unit, visually locatable and locating by the positioning system, or visually unlocatable; determining whether the acquisition terminal meets the observation conditions for the road segment based on the results of whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, whether the IMU is usable, and the positioning system used by the acquisition terminal includes:

[0029] For road segments that are both visually and IMU-locatable, if the camera intrinsic parameter accuracy of the acquisition terminal does not exceed the standard and the IMU is usable, then it is determined that the acquisition terminal meets the observation conditions for that road segment.

[0030] For road segments that are visually locatable and can be located by a positioning system, if the camera intrinsic parameter accuracy of the acquisition terminal does not exceed the standard, then based on the positioning system used by the acquisition terminal, it is determined whether the acquisition terminal meets the observation conditions for that road segment.

[0031] For road segments that are visually locatable and IMU-locatable, and for road segments that are visually locatable and located by a positioning system, if the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, then it is determined that the acquisition terminal does not meet the observation conditions for that road segment.

[0032] Optionally, the confidence level includes: confidence level of feature existence, confidence level of attributes of target map features related to autonomous driving behavior; confidence level of updating map features of the target road segment in the map using map data of the target road segment; and road topology confidence level of the target road segment includes:

[0033] Based on the driving trajectory of the data collected from the map data of the target road segment, update the road topology confidence of the target road segment and the attribute confidence of the target map elements;

[0034] Based on the collected images from the map data of the target road segment and the map features within the target road segment, update the confidence level of the existence of the map features of the target road segment.

[0035] Optionally, updating the road topology confidence of the target road segment and the attribute confidence of the target map features based on the driving trajectory of the acquisition terminal in the map data of the target road segment includes:

[0036] The road topology confidence of the target road segment is updated based on the traffic relationship between the driving trajectory of the target road segment and the driving trajectory of its adjacent road segments.

[0037] Based on the driving trajectory of the target road segment, obtain the driving behavior related to the target map elements;

[0038] The attribute confidence of the target map feature is updated based on whether the driving behavior matches the driving behavior constraints represented by the target map feature.

[0039] Optionally, updating the confidence level of the existence of map features in the target road segment based on the collected images from the map data of the target road segment and the map features within the target road segment includes:

[0040] The map features within the target road segment are reprojected onto the acquired image in the map data of the target road segment to obtain the reprojection position of the map feature in the acquired image.

[0041] If, within the set time period, the reprojection location of the map feature is unobstructed in all images collected, the confidence level of the map feature's existence is updated based on whether the map feature exists within a preset range of its reprojection location.

[0042] Optionally, the method further includes:

[0043] The roads in the map are divided into segments to obtain multiple road segments;

[0044] Based on the driving trajectories in the historical map data used when constructing the map, the historical map data is associated with the road segments of the map to obtain historical map data for multiple road segments;

[0045] Based on historical map data of the road segments, obtain the observation conditions for each road segment;

[0046] Determine whether the historical data acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment;

[0047] Based on whether the historical data acquisition terminal meets the observation conditions of the road segment, initialize the confidence level of the map features of the road segment, and the road topology confidence level of the road segment.

[0048] Optionally, obtaining the observation conditions for each road segment based on historical map data includes:

[0049] Based on the location data of at least one positioning system in the historical map data of the road segment, determine whether the road segment can be located by the positioning system.

[0050] If the road segment positioning system can locate it, then obtain the group of positioning systems that can locate the road segment;

[0051] Based on the distribution and changes of map elements within the visual range of the road segment in the map, determine whether the road segment is visually locatable;

[0052] If the road segment positioning system cannot locate it, and the distribution makes it visually unlocatable, then the road segment IMU is determined to be locatable, and the IMU recursive accuracy threshold is obtained based on the continuous road segments where the IMU can locate it.

[0053] Optionally, determining whether a road segment is locatable by a positioning system based on positioning data from at least one positioning system in the historical map data of the road segment includes:

[0054] Based on the positioning data of at least one positioning system in the historical map data of the road section, obtain the satellite elevation angles of the at least one positioning system in at least four directions of the road section;

[0055] Based on the satellite elevation angles of the at least one positioning system in at least four directions of the road segment, determine whether there is a positioning system in the at least one positioning system that can locate the road segment;

[0056] If it exists, then the observation conditions of the road segment are determined to be that the positioning system can locate it.

[0057] Optionally, determining whether a road segment is visually locatable based on the distribution and changes of map elements within the visual range of the road segment in the map includes:

[0058] Based on the distribution of map elements within the visual range of the road segment in the map, the visual relocation level of the road segment is determined;

[0059] The road segment is divided into grids;

[0060] The map features in the collected images from the historical map data of the road section are associated with the grid to obtain the number of times the map features of the grid change within a certain period of time.

[0061] If the visual relocation level of the road segment meets the requirements for visual localization and the proportion of the target grid is less than or equal to a preset proportion, then the road segment is determined to be visually localizable; the target grid is a grid whose map feature changes more than a first preset number of times within a certain time period.

[0062] If the visual relocation level of the road segment does not meet the requirements for visual localization, and / or the proportion of the target grid is greater than a preset proportion, then the road segment is determined to be visually unlocalizable.

[0063] Optionally, determining whether the historical data acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment includes:

[0064] If the observation conditions of the road segment include visual non-localization, then it is determined that the historical data acquisition terminal does not meet the observation conditions of the road segment.

[0065] If the observation conditions of the road segment include visual location and location system location, then determine whether the location system used by the historical data acquisition terminal includes a group of location systems that can locate the road segment; if it includes a group of location systems that can locate the road segment, then determine that the historical data acquisition terminal meets the observation conditions of the road segment.

[0066] If the observation conditions of the road segment include visual localization and IMU localization, then determine whether the recursive accuracy of the IMU of the historical acquisition terminal is greater than or equal to the IMU recursive accuracy threshold; if it is greater than or equal to the IMU recursive accuracy threshold, then determine that the historical acquisition terminal meets the observation conditions of the road segment.

[0067] Optionally, the step of initializing the confidence level of the map features of the road segment and the road topology confidence level of the road segment based on whether the historical data acquisition terminal meets the observation conditions of the road segment includes:

[0068] For road segments that do not meet the observation conditions, the confidence level of the existence of map elements in that road segment is set to low confidence; for road segments that meet the observation conditions, the confidence level of the existence of map elements in that road segment is set to high confidence.

[0069] For road segments that are visually unlocatable and visually locatable, the attribute confidence of map elements corresponding to locations in the road segment where the number of element changes within a certain time period exceeds the second preset threshold is set to low confidence; the attribute confidence of map elements corresponding to locations where the number of element changes within a certain time period is less than the second preset threshold is set to high confidence.

[0070] The road topology confidence of the road segment is initialized based on the number of times the driving trajectory of the historical acquisition terminal in the historical map data of the adjacent road segment represents whether the adjacent road segment is passable.

[0071] Secondly, this application provides a map data confidence update device, the device comprising:

[0072] The acquisition module is used to acquire map data collected at least once by at least one acquisition terminal;

[0073] The processing module is used to associate the map data with road segments in the map data based on the driving trajectory of the acquisition terminal in the map data data, so as to obtain map data data of multiple road segments;

[0074] A determination module is used to determine whether the acquisition terminal meets the observation conditions of the road segment; wherein, the observation conditions are used to determine whether the map data acquired by the acquisition terminal is reliable;

[0075] The update module is used to update the confidence level of map features of the target road segment and the road topology confidence level of the target road segment in the map, using map data of the target road segment, for target road segments that meet the observation conditions.

[0076] Thirdly, this application provides an electronic device, including: a processor and a memory; the processor is communicatively connected to the memory;

[0077] The memory stores computer instructions;

[0078] The processor executes computer instructions stored in the memory to implement the method as described in any one of the first aspects.

[0079] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the method as described in any one of the first aspects.

[0080] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the first aspects.

[0081] The map data confidence update method, device, and storage medium provided in this application differentiate the observation conditions of different road segments. This allows for the determination of the acquisition reliability of the acquisition terminal based on whether its hardware meets the observation conditions for the corresponding road segment. In other words, it determines the reliability of the map data acquired by the acquisition terminal. Furthermore, the confidence level of the map data for that road segment can be updated using map data from road segments that meet the observation conditions. This approach avoids completely rejecting or accepting the acquisition results entirely. Instead, it selects map data from acquisition terminals that meet the observation conditions for that road segment to evaluate its confidence level. This fully utilizes the acquisition results from acquisition terminals with different capabilities, improving the accuracy and efficiency of evaluating map data confidence, increasing the utilization rate of map data, and reducing the possibility of contaminated confidence levels. Additionally, because the method in this application can utilize the acquisition results from acquisition terminals with different capabilities, it is not limited to acquisition terminals with high requirements. Therefore, it can also reduce the cost of acquiring map data and fully utilize acquisition terminals with various capabilities. Attached Figure Description

[0082] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0083] Figure 1 A flowchart illustrating a map data confidence initialization method provided in this application embodiment;

[0084] Figure 2 A schematic flowchart illustrating a method for obtaining observation conditions of a road segment, provided in an embodiment of this application;

[0085] Figure 3 A flowchart illustrating a map data confidence update method provided in an embodiment of this application;

[0086] Figure 4 A flowchart illustrating a process for determining whether a data acquisition terminal meets the observation conditions of a road segment, as provided in an embodiment of this application.

[0087] Figure 5 A schematic diagram illustrating a process for updating map data confidence levels, provided as an embodiment of this application;

[0088] Figure 6 This is a schematic diagram of the structure of a map data confidence update device provided in an embodiment of this application;

[0089] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0090] The accompanying drawings have illustrated specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to specific embodiments. Detailed Implementation

[0091] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0092] The following is a brief explanation of some of the terms and concepts used in this application:

[0093] 1. Map data: This is the basic component of a map and represents physical objects from the real world in the digital world. Examples include lane lines, speed limit signs, stop lines, and left-turn arrows.

[0094] 2. Confidence level of map data: This describes whether map data is reliable, that is, it represents the probability that the map data exists in the real world. The higher the confidence level of map data, the greater the likelihood that the map data actually exists in the real world. The lower the confidence level of map data (i.e., the lower the probability), the smaller the likelihood that the map data actually exists in the real world.

[0095] 3. Road topology: Used to describe the traffic relationships between lane lines in a road.

[0096] 4. Map Data: This includes map data collected through consumer devices or through professional sensors. Consumer devices can be mobile devices deployed in vehicles traveling on the road. These mobile devices can be, for example, smartphones, dashcams, or customized lightweight devices deployed in taxis, private cars, etc., equipped with cameras, positioning systems, and inertial measurement units (IMUs). The cameras can capture images of the road conditions while the vehicle is traveling.

[0097] In autonomous driving scenarios, high-definition maps are typically used for localization and driving mode switching. Therefore, high-definition maps have very high requirements for freshness; that is, the speed at which high-definition maps are updated directly affects the safety of autonomous driving.

[0098] Currently, high-precision map creators primarily collect map data using their own specialized data collection vehicles (equipped with professional sensors), and then create and update high-precision maps based on this data. However, this method suffers from poor autonomy and effectiveness, failing to meet actual usage needs.

[0099] In recent years, with the development of basic capabilities such as data acquisition sensors, on-device computing, and network communication, using consumer devices to collect map data has become a trend. However, due to the large quantity and diverse types of consumer devices, and the difficulty in guaranteeing the reliability of online processing results from these devices, evaluating the confidence level of high-precision map data using map data collected by consumer devices, in order to select reliable map data for use in autonomous driving scenarios, is an urgent problem to be solved.

[0100] Currently, some existing solutions directly use map data collected by consumer devices to evaluate the confidence level of high-precision map data. However, this approach places high demands on the consumer devices themselves (hardware requirements, data acquisition process requirements, etc.) to ensure the reliability of the acquired map data. This makes it unsuitable for all types of consumer devices, resulting in low utilization rates. Furthermore, if the consumer devices used for the map data do not meet the requirements, it can easily contaminate the confidence level of the high-precision map data.

[0101] In view of this, this application provides a map data confidence update method. By distinguishing the observation conditions of different road segments, the reliability of the map data collected by the acquisition terminal can be determined based on whether the hardware of the acquisition terminal meets the observation conditions of the corresponding road segment. Then, the confidence of the map data of the road segment can be updated using the map data of the road segment that meets the observation conditions.

[0102] This approach neither completely rejects nor accepts the data collected by the acquisition end. Instead, it selects map data from acquisition ends that meet the observation conditions for that road segment to evaluate the confidence level of that segment. This allows for full utilization of data collected from different acquisition ends, improving the accuracy and efficiency of evaluating the confidence level of map data, increasing the utilization rate of map data, and reducing the possibility of contaminating the confidence level.

[0103] It should be understood that the method provided in the embodiments of this application can be applied to updating the confidence level of map data in high-precision maps, and can also be used to update the confidence level of map data in maps of other precision (e.g., standard precision maps).

[0104] The road segment divisions in the above map, as well as the observation conditions of the road segments, can be obtained based on the map data used when the map was made (hereinafter referred to as historical map data).

[0105] Furthermore, if the map is a newly created map, the confidence level of the map data recorded in the map can be obtained by initializing the confidence level based on the map data used when creating the map and the observation conditions of each road segment. If the map is not a newly created map, the confidence level of the map data recorded in the map can be obtained by updating the confidence level based on the map data collected by the acquisition terminal (e.g., a consumer device) in the previous iteration.

[0106] In other words, this application consists of two parts:

[0107] Part 1 (i.e., the pre-calibration stage): Using historical map data used in map making, obtain the observation conditions of each road segment in the map, and initialize the confidence level of the map data in the map.

[0108] Part 2 (i.e., the actual use phase): By using the observation conditions of different road segments, we screen whether the hardware of the acquisition terminal meets the observation conditions of the road segments to determine the acquisition reliability of the acquisition terminal, and use the map data of the road segments that meet the observation conditions to update the confidence level of the map data of the road segments.

[0109] Part 1 and Part 2 may be executed by the same entity or by different entities; this application does not limit this. For example, the executing entities of Part 1 and Part 2 may both be map data processing platforms, or any electronic device with processing capabilities (such as a server).

[0110] The technical solutions of this application will be described in detail below with reference to specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0111] To facilitate understanding, the following explains how to obtain the observation conditions for each road segment in the map using historical map data used in map creation, and how to initialize the confidence level of the map data. Part 1 will be explained first.

[0112] Figure 1 This is a flowchart illustrating a map data confidence initialization method provided in an embodiment of this application. Figure 1 As shown, the method may include the following steps:

[0113] S101. Divide the roads in the map into segments to obtain multiple road segments.

[0114] For example, roads in a map can be segmented based on their road topology to obtain multiple roads and intersections. A road can be a segment between two intersections.

[0115] Optionally, a road can be considered as a single road segment, or it can be further subdivided at certain intervals to obtain multiple road segments. Similarly, an intersection can be considered as a single road segment, or it can be further subdivided at certain intervals, etc., and this application does not impose any limitations. The distance mentioned here may be related to the precision of the observation conditions for the road segments to be determined, for example, it may be 5 meters, etc.

[0116] S102. Based on the driving trajectory in the historical map data used when constructing the map, associate the historical map data with the road segments of the map to obtain historical map data of multiple road segments.

[0117] For example, based on the trajectory points on the driving trajectory, the historical map data corresponding to the driving trajectory where the trajectory points are located on the road segment is used as the historical map data of that road segment.

[0118] It should be understood that the historical map data mentioned here can be, for example, map data collected by our own professional data collection vehicle as mentioned above, or map data collected using consumer devices, depending on the map data used in map production. That is, the historical data collection terminal can be our own professional data collection vehicle or a consumer device.

[0119] Historical map data can include: captured images, the vehicle's trajectory during the historical data acquisition, and positioning data from the positioning system. The captured images can be images of the road the vehicle traveled on. The trajectory includes the poses of multiple trajectory points. The positioning data can be observation data from the positioning system or positioning results obtained after processing the observation data.

[0120] It should be noted that the collected images, driving trajectories, and location data all have timestamps, so the three can be correlated using the timestamps.

[0121] S103. Obtain the observation conditions for each road segment based on the historical map data of the road segment.

[0122] For example, the observation conditions of a road segment are related to the hardware conditions of the data acquisition terminal to be evaluated. Taking the data acquisition terminal as including a positioning system, an IMU, and a camera sensor as an example, the observation conditions of a road segment can be categorized as follows: the positioning system can locate the segment, and the visual system can also locate it; or, the visual system can locate the segment, and the IMU can also locate it.

[0123] Taking the observation conditions of a road segment as an example, it can be categorized as either location-system locatorable and visually locatorable, or visually locatorable and IMU-locatorable. For instance, based on the location data of one or more positioning systems in the historical map data of the road segment, it can be determined whether the road segment is location-system locatorable; if it is, the location-system locatorable positioning system is obtained; based on the collected images in the historical map data of the road segment, it can be determined whether the road segment is visually locatorable; road segments that are not location-system locatorable can be designated as road segments that are location-system locatorable, etc.

[0124] The positioning system mentioned here can be any system capable of positioning. A data acquisition terminal can use one or more positioning systems. For example, a data acquisition terminal can simultaneously use the Global Positioning System (GPS) and the BeiDou Navigation Satellite System. Furthermore, for a road segment, the positioning system capable of locating that road segment can be one or a combination of multiple positioning systems (referred to as a positioning system group).

[0125] S104. Determine whether the historical data acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment.

[0126] Taking the observation conditions of a road segment as an example, such as: being able to be located by a positioning system and being able to be located visually; or being able to be located visually and being able to be located by an IMU, the determination of whether the historical acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment can be as follows:

[0127] If the observation conditions for the road segment include visual non-location, then it is determined that the historical data acquisition terminal does not meet the observation conditions for that road segment.

[0128] If the observation conditions for the road segment include visual location capability and location system capability, then it is determined whether the location system used by the historical data acquisition terminal includes a group of location systems capable of locating the road segment. If it includes a group of location systems capable of locating the road segment, then it is determined that the historical data acquisition terminal meets the observation conditions for that road segment. If it does not include one or more of the group of location systems capable of locating the road segment, then it is determined that the historical data acquisition terminal does not meet the observation conditions for that road segment.

[0129] For example, if the positioning system group that can locate the road segment is GPS and BeiDou positioning system, and the positioning system used by the historical data acquisition terminal is GPS, then it is determined that the historical data acquisition terminal does not meet the observation conditions for the road segment. If the positioning system used by the historical data acquisition terminal includes GPS and BeiDou positioning system, then it is determined that the historical data acquisition terminal meets the observation conditions for the road segment.

[0130] If the observation conditions for the road segment include both visual localization and IMU localization, then it is determined whether the recursive accuracy of the IMU at the historical acquisition end is greater than or equal to the IMU recursive accuracy threshold. If it is greater than or equal to the IMU recursive accuracy threshold, then it is determined that the historical acquisition end meets the observation conditions for the road segment. For example, positioning data from a positioning system in historical map data can be used for localization, and / or visual localization can be used for acquisition images to obtain the start and end points of a road composed of multiple consecutive road segments. This information is then compared with the IMU recursive endpoint to obtain the IMU recursive accuracy.

[0131] S105. Based on whether the historical data acquisition terminal meets the observation conditions of the road segment, initialize the confidence level of the map elements of the road segment and the road topology confidence level of the road segment.

[0132] The confidence level of map features for a road segment depends on the specific application scenario. Taking this application to an autonomous driving scenario as an example, the confidence level of map features for a road segment may include, for example, the confidence level of feature existence. For target map features in the road segment that are related to autonomous driving behavior, the confidence level may also include, for example, attribute confidence. The target map features mentioned here may be, for example, lane lines (dashed lines, solid lines), speed limit signs, left turn arrows, etc.

[0133] Among them, road topology confidence is used to characterize the reliability of communication relationships with adjacent road segments. Feature existence confidence is used to characterize the probability that the map feature actually exists in the real world. Attribute confidence is used to characterize the reliability of the driving behavior constraints represented by the map feature.

[0134] Furthermore, this application does not limit the representation of the aforementioned confidence level in the map. For example, the confidence level of map elements can be set by assigning high and low values, or by assigning specific probability values. Additionally, this application does not limit the granularity of the confidence level assignment for map elements. Taking high and low as an example, the confidence level can be divided into three values: high, medium, and low, or it can be divided into high 1, high 2, high 3; medium 1, medium 2, medium 3; low 1, low 2, low 3, etc.

[0135] The following example uses confidence level assignment, assuming that confidence levels include: confidence level of feature existence, road topology confidence level, and attribute confidence level. For example, the following method can be used to initialize the confidence level of map features of the road segment and the road topology confidence level of the road segment based on whether the historical data acquisition terminal meets the observation conditions of the road segment:

[0136] For example, if a road segment does not meet the observation conditions at the historical data collection point, it indicates that the reliability of the historical data collection point is low, and the confidence level of the existence of map elements in that road segment is set to low confidence. If a road segment meets the observation conditions, it indicates that the reliability of the historical data collection point is high, and the confidence level of the existence of map elements in that road segment is set to high confidence.

[0137] For road segments that are visually unlocatable and visually locatable, the confidence level of map features can be assigned based on the number of changes to map features at the same location in the historical map data of that road segment. For example, the confidence level of map features corresponding to locations in that road segment where the number of changes within a certain time period is greater than a second preset threshold is set to low confidence; the confidence level of map features corresponding to locations where the number of changes within a certain time period is less than the second preset threshold is set to high confidence.

[0138] Additionally, based on the historical map data of adjacent road segments, the number of driving trajectories from historical data collection terminals indicating whether adjacent road segments are passable is used to initialize the road topology confidence of the road segment. For example, if a driving trajectory indicates a passable relationship between two road segments (i.e., the driving trajectory passes through the first road segment and then the second road segment), and the number of driving trajectories indicating a passable relationship between the two road segments exceeds the corresponding set value, then the confidence of the topology indicating a topological relationship between the two road segments is set to high confidence. Otherwise, it is set to low confidence.

[0139] The method in this application embodiment can utilize historical map data used in map making to obtain the observation conditions of each road segment in the map, so as to accurately determine the hardware requirements of the acquisition terminal for different road segments. Furthermore, it can use the observation conditions of each road segment to evaluate whether the historical acquisition terminal corresponding to the historical map data meets the observation conditions. Based on this, it can accurately initialize the confidence of existing map elements in the map and the road topology confidence of the road segments.

[0140] In other words, by utilizing historical map data used in map creation, the observation conditions of each road segment in the map are obtained to accurately determine the hardware requirements of the acquisition terminal for different road segments, providing accurate prior information for the selection of map data collected by subsequent acquisition terminals. Furthermore, by using the observation conditions of each road segment, the historical acquisition terminals corresponding to the historical map data are evaluated to determine whether they meet the observation conditions. Based on this, the confidence levels of existing map elements and the road topology confidence levels of road segments can be accurately initialized, providing an accurate benchmark for subsequent confidence level updates.

[0141] Taking high-precision maps as an example, currently, in scenarios where map data collected by consumer devices is used to update the confidence level of map data, it is assumed that all existing map data is reliable. That is, it is assumed that the map data in high-precision maps made using map data collected by proprietary professional data collection vehicles is of high confidence. However, the map data collected by proprietary professional data collection vehicles is not filtered according to the observation conditions of each road segment, nor is the confidence level of the map data in the high-precision map accurately initialized based on whether the hardware of the data collection vehicle meets the observation conditions. This can easily lead to a situation where the initialized confidence level does not match the actual level.

[0142] The method of this application embodiment, based on the high-precision map created using map data collected by its own professional data collection vehicle, can evaluate the reliability of the map data collected by the professional data collection vehicle through the observation conditions of road sections. For data collection that meets the requirements, the map data created using it is set to a high confidence level. This allows for accurate initialization of the confidence level of the high-precision map data, avoiding discrepancies between the confidence level and reality, and providing an effective and accurate reference for the mapping process.

[0143] The following section uses the data acquisition end as including a positioning system, an IMU, and camera sensors, and the observation conditions of a road segment are, for example, divided into: positioning system capable of location and visual location; or visual location and IMU capable of location. It then provides a detailed explanation of how to obtain the observation conditions for each road segment based on historical map data:

[0144] Figure 2 This is a schematic flowchart illustrating a method for obtaining observation conditions of a road segment, provided as an embodiment of this application. Figure 2 As shown, S103 may include the following steps:

[0145] S201. Based on the positioning data of at least one positioning system in the historical map data of the road segment, determine whether the road segment is locatable by a positioning system. That is, determine the satellite observability of the road segment. If the positioning system is locatable, proceed to step S202; if the positioning system is not locatable, proceed to step S204.

[0146] One possible implementation is to obtain the satellite obstruction status (i.e., obstruction by objects (e.g., buildings) of at least one positioning system) based on the positioning data of at least one positioning system in the historical map data of the road segment, thereby determining whether the road segment is locatable by the positioning system based on the satellite obstruction status of the positioning system.

[0147] For example, satellite elevation angles of at least one positioning system in at least four azimuths of the road segment can be obtained based on positioning data from at least one positioning system in historical map data of the road segment. For instance, the horizontal 360 degrees can be divided into multiple azimuths at 15 degrees each, and then, based on the positioning data, the satellite elevation angle of each positioning system at a certain point (e.g., start point, end point, midpoint, etc.) of the road segment can be obtained in each azimuth. For details on how to obtain satellite elevation angles based on positioning data, please refer to the description of existing technologies, which will not be repeated here.

[0148] Then, based on the satellite elevation angles of the at least one positioning system in at least four directions of the road segment, it can be determined whether there is a positioning system capable of locating the road segment. That is, based on the satellite elevation angles of the at least one positioning system in at least four directions of the road segment, the obstruction of the positioning system's satellites in that road segment is determined, thereby determining whether there is a positioning system capable of locating the road segment. If so, the observation conditions for the road segment are determined to be suitable for positioning by the positioning system. If not, the observation conditions for the road segment are determined to be unsuitable for positioning by the positioning system.

[0149] For example, if the satellite elevation angle is higher than a preset elevation angle threshold, then the azimuth is considered unobstructed. If the number of unobstructed azimuths for a single positioning system is greater than a preset number, or if the number of unobstructed azimuths for multiple positioning systems combined is greater than a preset number, then the positioning system for that road segment is considered capable of positioning. Conversely, if the elevation angle is lower than a preset threshold, then the positioning system for that road segment is considered uncapable of positioning. The specific satellite elevation angle threshold mentioned above is related to positioning accuracy; for example, the satellite elevation angle threshold could be 45 degrees.

[0150] For example, if there are two positioning systems in the historical map data, namely GPS and Beidou positioning system, the horizontal 360 degrees can be divided into multiple directions at 15 degrees each. At the midpoint of the road segment, the satellite elevation angle of GPS satellites in each direction can be obtained through GPS positioning data, and the satellite elevation angle of Beidou positioning system satellites in each direction can be obtained through Beidou positioning system positioning data.

[0151] Then, based on the satellite elevation angles of GPS satellites in each direction and the satellite elevation angles of the BeiDou positioning system satellites in each direction, it can be determined whether the road segment is locatable.

[0152] If the number of GPS satellites with elevation angles greater than a preset elevation angle threshold is greater than a corresponding preset number, and / or the number of BeiDou satellites with elevation angles greater than a preset elevation angle threshold is greater than a corresponding preset number, then the positioning system for that road segment is determined to be capable of positioning. The corresponding preset number can be set according to actual needs.

[0153] If the combined elevation angles of GPS satellites and BeiDou satellites are greater than the preset elevation angle threshold, and the number of such locations exceeds the set value, then the positioning system for that road segment is deemed capable of providing positioning.

[0154] S202. Obtain the group of positioning systems that can locate the road segment.

[0155] For example, if a single positioning system can be used to locate a road segment, then the positioning system group for that road segment includes that single positioning system. If multiple positioning systems work together to achieve positioning for that road segment, that is, if the positioning requirements described in step S201 are met, then the positioning system group for that road segment includes those multiple positioning systems.

[0156] S203. Determine whether the road segment is visually locatable based on the distribution and changes of map elements within the visual range of the road segment in the map. That is, determine whether the road segment is visually locatable based on the visual relocation observability of the road segment and the stability of the scene itself. If visually locatable, the process ends; if changes make it visually unlocatable, the process ends; if distribution makes it visually unlocatable, proceed to S204.

[0157] For example, the visual relocation level of a road segment can be determined based on the distribution of map features within the visual range of that road segment in the map. The visual range referred to here can be the camera's field of vision.

[0158] For example, at a certain point on the road segment, the distribution of map elements within the visual range of the road segment is statistically analyzed at 15-degree azimuth and 15-degree elevation angles. Then, based on the statistical results and the mapping relationship between the distribution statistics and the visual relocation level, the visual relocation level of the road segment is obtained. Alternatively, the distribution statistics results can be input into a pre-trained model to obtain the output visual relocation level.

[0159] Additionally, the road segment can be divided into grids. For example, each grid can be a 1m x 1m area. Then, map features in the collected images from the historical map data of the road segment are associated with the grids to obtain the number of times the map features of each grid change within a certain time period. This "certain time period" can be the length of the historical map data collection period, or a value shorter than that collection period.

[0160] That is, among the map features associated with the grid in multiple acquired images within a certain time period, if the map features associated with the grid in two adjacent acquired images are different, it is considered that a change has occurred, thereby obtaining the number of times the map features of the grid have changed within a certain time period. This application does not limit the above-described method of associating map features in the acquired images with the grid. For example, the grid can be reprojected onto the acquired images, and the map features within the range of the reprojected grid can be used as the map features associated with that grid.

[0161] If the visual relocation level of the road segment meets the requirements for visual localization, and the proportion of the target grid is less than or equal to a preset proportion, it indicates that the visual relocation observability of the road segment meets the requirements, and the scene itself of the road segment has high stability, which also meets the requirements for visual localization. Therefore, the road segment is determined to be visually localizable. Here, the target grid refers to a grid whose map feature changes more than a first preset threshold number within a certain time period.

[0162] If the visual relocation level of the road segment does not meet the requirements for visual localization, and / or the proportion of the target grid is greater than the preset proportion, it indicates that the visual relocation observability of the road segment does not meet the requirements, and / or the scene stability of the road segment itself is poor and does not meet the requirements for visual localization, then the road segment is determined to be visually unlocalizable.

[0163] It should be noted that the execution of steps S201 and S203 above is not in any particular order.

[0164] S204. If the road segment positioning system cannot locate, and visual positioning is not possible due to the distribution, then the IMU of the road segment is determined to be locatable, and the IMU recursive accuracy threshold is obtained based on the continuous road segments where the IMU can be located.

[0165] For example, based on the length of the continuous road segment, the IMU recursion accuracy that meets that length can be obtained and used as the IMU recursion accuracy threshold for subsequent evaluation of whether the hardware at the acquisition end meets the observation conditions. The length mentioned here could be, for example, the length of the maximum continuous road segment.

[0166] The method in this application accurately distinguishes the minimum equipment required to collect map data for each road segment from the perspectives of satellite observation capability, visual relocation observation capability, and the stability of the scene itself. This allows for accurate determination of the observation conditions for each road segment and accurate evaluation of the reliability of the map data collected by the subsequent acquisition end.

[0167] The above explains how to use historical map data used in map making to obtain the observation conditions of each road segment in the map, and how to initialize the confidence level of the map data.

[0168] The following section explains how to use the observation conditions of different road segments to screen whether the hardware of the acquisition terminal meets the observation conditions of the road segment, determine the reliability of the map data collected by the acquisition terminal, and update the confidence level of the map data of the road segment using the map data of the road segment that meets the observation conditions. That is, it explains Part 2.

[0169] Figure 3 This is a flowchart illustrating a map data confidence update method provided in an embodiment of this application. Figure 3 As shown, the method may include the following steps:

[0170] S301. Obtain map data collected at least once by at least one acquisition terminal.

[0171] The data acquisition device mentioned here can be a consumer device or a dedicated data acquisition vehicle. The map data collected by the acquisition device can include: captured images, the vehicle's trajectory during data acquisition, and positioning data from the positioning system. The captured images can be images of the road along which the vehicle is traveling. The trajectory includes the poses of multiple trajectory points. The positioning data can be observation data from the positioning system or positioning results obtained after processing the observation data.

[0172] It should be noted that the collected images, driving trajectories, and location data all have timestamps, so the three can be correlated using the timestamps.

[0173] For example, map data collected at least once by at least one data acquisition terminal can be received via an application programming interface (API) or a graphical user interface (GUI). Alternatively, the data acquisition terminal can send the map data it has collected. Alternatively, the map data collected at least once by the aforementioned at least one data acquisition terminal can also be pre-acquired and stored locally.

[0174] S302. Based on the driving trajectory of the acquisition terminal in the map data, associate the map data with the road segments of the map to obtain map data for multiple road segments.

[0175] For example, based on the trajectory points on the driving trajectory, the map data corresponding to the driving trajectory where the trajectory points are located on the road segment is used as the map data for that road segment.

[0176] S303. Determine whether the acquisition terminal meets the observation conditions of the road section.

[0177] The observation conditions are used to determine whether the map data collected by the acquisition terminal is reliable.

[0178] For example, the data acquisition device can be evaluated based on the observation conditions of each road segment to determine if it meets those conditions. That is, the entire set of road segments can be used to evaluate the data acquisition device. Alternatively, multiple road segments can be identified to determine the various observation conditions they may have. For each observation condition, one or more road segments can be selected to evaluate the data acquisition device. That is, a subset of road segments can be used to evaluate the data acquisition device to obtain results showing whether it meets the observation conditions of all road segments. This method of evaluating only a subset of road segments can improve processing efficiency.

[0179] Furthermore, the specific dimensions used to determine whether the data acquisition terminal meets the observation conditions of a road segment depend on those conditions. For example, if the observation conditions are categorized as either location-capable by the positioning system and visually identifiable, or visually identifiable and IMU-identifiable, then the suitability of the data acquisition terminal for the road segment's observation conditions can be determined by considering the camera parameters of the acquisition terminal, the positioning system used by the acquisition terminal, and the recursive accuracy of the acquisition terminal's IMU.

[0180] S304. For a target road segment that meets the observation conditions, use the map data of the target road segment to update the confidence level of the map elements of the target road segment in the map, and the road topology confidence level of the target road segment.

[0181] As mentioned earlier, the confidence level of map features for a road segment is specifically related to the application scenario. Taking the application of this application to an autonomous driving scenario as an example, the confidence level of map features for a road segment may include, for example, the confidence level of feature existence. For target map features in the road segment that are related to autonomous driving behavior, their confidence level may also include, for example, attribute confidence level.

[0182] Taking autonomous driving scenarios as an example, if the target road segment includes target map elements related to autonomous driving behavior, the road topology confidence and attribute confidence of the target road segment can be updated based on the driving trajectory of the acquisition terminal in the map data of the target road segment; the existence confidence of the map elements of the target road segment can be updated based on the acquired images in the map data of the target road segment and the map elements within the target road segment.

[0183] If the target road segment does not include target map features related to autonomous driving behavior, the road topology confidence of the target road segment can be updated based on the driving trajectory of the acquisition terminal in the map data of the target road segment; the feature existence confidence of the map features of the target road segment can be updated based on the acquired images in the map data of the target road segment and the map features within the target road segment.

[0184] It should be understood that the confidence scores of other road segments that do not meet the observation conditions will not be updated this time. This avoids updating the confidence scores of road segments using unreliable map data, thus preventing the contamination of the confidence scores.

[0185] The method in this application distinguishes the observation conditions of different road segments, thereby determining the reliability of the map data collected by the acquisition terminal based on whether the hardware of the acquisition terminal meets the observation conditions of the corresponding road segment. Then, the confidence level of the map data for that road segment can be updated using the map data of the road segment that meets the observation conditions. This processing method does not completely reject or accept the acquisition results of the acquisition terminal. Instead, it selects map data from acquisition terminals that meet the observation conditions of the road segment to evaluate the confidence level of the map data for that road segment. This fully utilizes the acquisition results of acquisition terminals with different acquisition capabilities, improving the accuracy and efficiency of evaluating the confidence level of map data, increasing the utilization rate of map data, and reducing the possibility of contaminating the confidence level. Furthermore, because the method in this application can utilize the acquisition results of acquisition terminals with different acquisition capabilities, it is not limited to acquisition terminals with high requirements. Therefore, it can also reduce the cost of collecting map data and fully utilize acquisition terminals with various acquisition capabilities.

[0186] The following section uses a data acquisition terminal that includes a positioning system, an IMU, and camera sensors. The observation conditions for a road segment are categorized as follows: either the positioning system can locate the segment, and visual location is also possible; or, visual location is possible, and IMU location is also possible. This section will explain in detail how to determine whether the data acquisition terminal meets the observation conditions for the road segment.

[0187] Figure 4 This is a flowchart illustrating a method for determining whether a data acquisition terminal meets the observation conditions of a road segment, as provided in an embodiment of this application. Figure 4 As shown, S303 may include the following steps:

[0188] S401. Extract a first road segment and multiple consecutive second road segments from the plurality of road segments. The observation conditions of the first road segment are that it can be located by both the positioning system and visual location. The observation conditions of the second road segment include either visual location or positioning system location.

[0189] The observation conditions for the above-mentioned road sections can be obtained by adopting the aforementioned methods. Figure 1 and Figure 2 The method shown is used to obtain it.

[0190] The map described above stores the observation conditions for each road segment, or a dedicated database is used to store these conditions. Therefore, observation conditions for multiple road segments can be obtained from the map, and a first road segment and multiple consecutive second road segments can be selected from them. It should be understood that the multiple consecutive second road segments mentioned here refer to road segments with a traffic topology relationship.

[0191] S402. Based on the map data of the first road segment and the map elements within the first road segment in the map, determine whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard.

[0192] For example, the accuracy of the camera intrinsics at the acquisition end can be determined by reprojecting map features onto the acquired images in the map data. Alternatively, other methods for evaluating camera intrinsics accuracy can be used to determine whether the accuracy of the camera intrinsics at the acquisition end exceeds the standard based on the map data and map features. This application does not limit the methods.

[0193] Taking reprojection as an example, for instance, map features within the first road segment can be reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features within the first road segment. That is, the reprojection error of each reprojected map feature. For how to reproject three-dimensional map features onto a two-dimensional acquired image, please refer to existing reprojection methods, which will not be elaborated upon here.

[0194] Then, the camera intrinsic parameters of the acquisition terminal can be estimated based on the reprojection error of the map features within the first road segment. Based on the estimated camera intrinsic parameters of the acquisition terminal, the map features within the first road segment are reprojected again onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features within the first road segment.

[0195] If the proportion of map features with a reprojection error greater than a preset reprojection error threshold is greater than a preset proportion threshold, it indicates that the camera at the acquisition end does not meet the requirements for visual positioning. Therefore, it is determined that the intrinsic parameter accuracy of the camera at the acquisition end exceeds the standard, and the acquired map data is unusable. Step S405 is then executed. If the proportion of map features with a reprojection error greater than a preset reprojection error threshold is less than or equal to the preset proportion threshold, it indicates that the camera at the acquisition end meets the requirements for visual positioning. Therefore, it is determined that the intrinsic parameter accuracy of the camera at the acquisition end does not exceed the standard, and step S403 is executed.

[0196] This application does not limit the aforementioned preset reprojection error threshold or preset percentage threshold. For example, it can be set according to the required visual positioning accuracy. For example, the preset percentage threshold can be 95%. That is, if the reprojection error of 95% of the map features reprojected in the target road segment is greater than the preset reprojection error threshold, then the camera intrinsic parameter accuracy at the acquisition end is considered to be out of tolerance.

[0197] S403. Based on the map data of multiple consecutive second road segments, determine whether the IMU of the acquisition terminal is available.

[0198] That is, the IMU is determined using continuous road segments that can be located visually or by a positioning system. In other words, by using continuous road segments whose location can be obtained through positioning or repositioning, the accuracy of the IMU is inferred from the repositioning or positioning locations of these road segments to determine whether the IMU is usable.

[0199] It should be understood that the multiple consecutive second road segments can all be visually locatable road segments, all be locatable road segments by a positioning system, or be partially visually locatable road segments and partially locatable road segments by a positioning system.

[0200] For example, based on map data of multiple consecutive second road segments, the starting and ending points of these segments can be obtained; that is, the starting and ending points of a road formed by these multiple consecutive second road segments. In other words, the starting point of the first second road segment and the ending point of the last second road segment can be obtained. It should be understood that the starting and ending points mentioned here are determined relative to the mode of travel.

[0201] If the first second road segment is a segment that can be located by the positioning system, then the starting point of the second road segment located by the positioning system is obtained. If the first second road segment is a visually locatable segment, then the starting point of the second road segment is obtained by using the acquired image of the second road segment and employing visual relocalization.

[0202] Accordingly, if the last second road segment is a segment that can be located by the positioning system, then the endpoint of the second road segment located by the positioning system is obtained. If the last second road segment is a visually locatable segment, then the endpoint of the second road segment is obtained by using the acquired image of the second road segment and employing visual relocalization.

[0203] The aforementioned map data includes information related to the IMU used by the acquisition terminal, such as the IMU's identifier and / or configuration. Alternatively, in addition to the map data, the acquisition terminal may also specifically report the identifier and / or configuration of the IMU it uses. Therefore, the IMU used by the acquisition terminal can be determined, and then, based on the aforementioned starting point of positioning, the endpoint can be recursively calculated using the acquisition terminal's IMU. In this way, the recursive accuracy of the acquisition terminal's IMU can be obtained based on the positioning endpoint and the recursively calculated endpoint.

[0204] For details on how to use the IMU to predict the endpoint and how to obtain its prediction accuracy, please refer to existing technologies, which will not be elaborated here. For example, the prediction accuracy per second of the IMU can be obtained by analyzing the mapping relationship between the difference between the positioning endpoint and the prediction endpoint and the linear increase of the IMU error over time.

[0205] If the recursive accuracy of the IMU at the acquisition end is greater than or equal to the IMU recursive accuracy threshold, then the IMU at the acquisition end is determined to be usable; if the recursive accuracy of the IMU at the acquisition end is less than the IMU recursive accuracy threshold, then the IMU at the acquisition end is determined to be unusable. The IMU recursive accuracy threshold used here can be, for example, obtained in the pre-calibration stage as described above.

[0206] S404. Obtain the positioning system used by the acquisition terminal.

[0207] The aforementioned map data includes information related to the positioning system used by the data acquisition terminal, such as the positioning system identifier and / or the positioning data (which can identify the positioning system used). Therefore, the positioning system used by the data acquisition terminal can be obtained through the map data. Alternatively, in addition to the map data, the data acquisition terminal may also specifically report the identifier of the positioning system it uses. Therefore, the positioning system used by the data acquisition terminal can be determined through its report.

[0208] S405. Based on the results of whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, whether the IMU is available, and the positioning system used by the acquisition terminal, determine whether the acquisition terminal meets the observation conditions of the road segment.

[0209] For road sections that are visually and IMU-locatable, if the camera intrinsic parameter accuracy at the acquisition end is not exceeded and the IMU is available, then it is determined that the acquisition end meets the observation conditions for that road section.

[0210] For road segments that are visually and systematically locatable, if the camera intrinsic parameter accuracy at the data acquisition terminal is within acceptable limits, then the data acquisition terminal's suitability for the observation conditions of that road segment is determined based on the positioning system used by the acquisition terminal. For example, if the data acquisition terminal includes a group of positioning systems capable of locating the road segment, then it is determined that the acquisition terminal meets the observation conditions for that road segment. If it does not include one or more of the positioning systems capable of locating the road segment, then it is determined that the acquisition terminal does not meet the observation conditions for that road segment.

[0211] For road segments that are visually locatable and IMU-locatable, and road segments that are visually locatable and location system-locatable, if the camera intrinsic parameter accuracy at the acquisition end exceeds the standard, it is determined that the acquisition end does not meet the observation conditions for that road segment.

[0212] The method in this application embodiment can select one or more road segments to judge the acquisition end for each observation condition. That is, it uses a subset of road segments to judge the acquisition end to obtain the result of whether the acquisition end meets the observation conditions of all road segments. By using a subset of road segments for judgment, processing efficiency can be improved.

[0213] In addition, by assessing the visual performance, IMU performance, and satellite observation capabilities of the aforementioned acquisition terminals, it is determined whether the acquisition terminals meet the observation conditions for the road segment. This allows for the screening of map data subsequently used for confidence evaluation, ensuring the accuracy of the map data and thus guaranteeing the accuracy of the map data used in the subsequent confidence evaluation.

[0214] The following example uses confidence levels, including the confidence level of feature existence and the confidence level of target map features related to autonomous driving behavior, to illustrate how to use map data of the target road segment to update the confidence level of map features and the road topology confidence level of the target road segment. The following examples illustrate this by assigning high to low confidence levels.

[0215] Figure 5 This is a schematic diagram illustrating a process for updating map data confidence levels, provided as an embodiment of this application. Figure 5 As shown, S304 may include the following steps:

[0216] S501. Based on the driving trajectory of the acquisition terminal in the map data of the target road segment, update the road topology confidence of the target road segment and the attribute confidence of the target map element.

[0217] For example, the road topology confidence of a target road segment can be determined based on whether the travel trajectory passes through road segments (usually adjacent road segments) that have a topological relationship with the target road segment. For instance, the communication road topology confidence of the target road segment can be updated based on the traffic relationship between the travel trajectory of the target road segment and the travel trajectories of its adjacent road segments. Optionally, if a certain number of trajectories pass through the target road segment, and there are road segments that do not have topological connections with the target road segment, the road topology confidence of the target road segment is updated to low confidence. If a certain number of trajectories pass through the target road segment, and there are road segments that have topological connections with the target road segment, the road topology confidence of the target road segment is updated to high confidence.

[0218] For target map features, driving behaviors related to the target map feature can be obtained based on the driving trajectory of the target road segment. The attribute confidence of the target map feature is then updated based on whether the driving behavior matches the driving constraints represented by the target map feature. That is, the attribute confidence is determined by analyzing a large number of trajectories based on the influence of different target map features on driving behavior. For example, if most driving trajectories match the driving constraints represented by the target map feature, its attribute confidence is considered high; otherwise, it is considered low. In this way, the confidence of a single map feature can be evaluated.

[0219] Taking a solid lane line as an example of a map feature, if a certain number of driving trajectories cross the solid line to change lanes, it indicates that the probability of the solid line being a dashed line in the real world is relatively high, and the confidence level of the solid line's attribute is updated to low confidence. If a certain number of trajectories change lanes near the solid line and its adjacent dashed line, it indicates that the probability of the solid line being a solid line in the real world is relatively high, and the confidence level of the solid line's attribute is updated to high confidence.

[0220] Taking a dashed line adjacent to a solid line as an example of a map feature, if a certain number of driving trajectories change lanes near the point of transition between the dashed line and its adjacent solid line, and the change in their heading angle exceeds a set threshold, it indicates that the probability of the dashed line being a solid line in the real world is relatively high, and the attribute confidence of the dashed line is updated to low confidence. If a certain number of driving trajectories are evenly distributed at the locations crossing the dashed line, it indicates that the probability of the dashed line being a dashed line in the real world is relatively high, and the attribute confidence of the dashed line is updated to high confidence.

[0221] The determination of the solid and dashed lines can be made based on their positions recorded on the map, thus determining their location on the driving trajectory, and subsequently, based on the driving behavior corresponding to that trajectory. The heading angle used in this determination process can be obtained from the pose of the trajectory points.

[0222] Taking a ground arrow representing a lane as a map feature as an example, if a certain number of driving trajectories within that lane do not match the driving direction indicated by the ground arrow (e.g., a left or right turn), but the driving constraint represented by the ground arrow is straight-ahead, then there is a mismatch. When there is a mismatch, it indicates that the probability of the ground arrow's attribute existing in the real world is low, and the attribute confidence level of the ground arrow is updated to low confidence. Conversely, if a certain number of driving trajectories within that lane match the driving constraint represented by the ground arrow, it indicates that the probability of the ground arrow's attribute existing in the real world is high, and the attribute confidence level of the ground arrow is updated to high confidence.

[0223] Taking a speed limit sign as a map element as an example, if a certain number of driving trajectories on a road segment indicate that the driving speed exceeds the preset percentage of the maximum speed limit set by the sign, then the two do not match. For example, a certain number of trajectories show a speed exceeding the speed limit sign by 10%. When there is a mismatch, it means that the probability of the speed limit sign's attribute existing in the real world is low, and the confidence level of the speed limit sign's attribute is updated to low confidence. Conversely, it means that the probability of the ground arrow's attribute existing in the real world is high, and the confidence level of the ground arrow's attribute is updated to high confidence.

[0224] S502. Based on the collected images in the map data of the target road segment and the map elements within the target road segment, update the confidence level of the existence of the map elements of the target road segment.

[0225] For example, reprojection can be used to determine whether the map feature exists in the acquired image, thereby updating the confidence level of the map feature's existence.

[0226] For example, map features within the target road segment can be reprojected onto the acquired images in the map data of the target road segment to obtain the reprojection position of the map feature in the acquired images. If, within a set time period, the reprojection position of the map feature is unobstructed by any occlusion in the acquired images, it indicates that the quality of these acquired images is high and can be used to assess the existence of map features. By using multiple acquisition results for the target road segment and determining whether each map feature is obstructed, it is possible to ensure full utilization of the acquired images, reduce the coverage requirements of a single acquired image for map features, and achieve the evaluation of the confidence level of individual map features.

[0227] Thus, when no occlusion is confirmed, the confidence level of a map element's existence can be updated based on whether other map elements exist within a preset range of its reprojection location. For example, if a map element exists within the preset range of its reprojection location, it indicates a high probability that the map element exists in the real world, and the confidence level of its existence is updated to high confidence. Conversely, if it does not exist, it is updated to low confidence.

[0228] Since the reprojection in this step is based on the premise that the acquisition end meets the observation requirements for that road segment—that is, the acquisition end's positioning system is capable of localization, or visual localization, or the IMU recursive accuracy meets the requirements—reliable position and orientation results can be obtained based on the map data acquired by the acquisition end. Therefore, by using the above reprojection method to transform the map feature's position to the pixel coordinate system on the acquired image, and based on the image where its position is unobstructed, it is possible to accurately determine whether the map feature exists in the real world, thereby accurately updating the confidence level of its existence.

[0229] The confidence update method provided in this application distinguishes the observation conditions of different road segments, thereby determining the reliability of the map data collected by the acquisition terminal based on whether the hardware of the acquisition terminal meets the observation conditions of the corresponding road segment. Then, the confidence of the map data of the road segment can be updated one by one using the map data of the road segment that meets the observation conditions. This can make full use of the acquisition results of acquisition terminals with different acquisition capabilities, improve the accuracy and efficiency of evaluating the confidence of map data, increase the utilization rate of map data, and reduce the possibility of contaminating the confidence.

[0230] Furthermore, confidence levels can be updated for individual map data. Therefore, there's no need to pre-set the confidence update cycle; confidence updates can be automatically triggered based on whether the map data at the granular level has reached a sufficient number of confidence levels to be updated. This improves the automation and efficiency of confidence updates. Simultaneously, because it updates individual map data, the requirements for coverage and acquisition technology during a single acquisition are lower, allowing for full utilization of acquisition terminals with various acquisition capabilities.

[0231] Figure 6 This is a schematic diagram of a map data confidence update device provided in an embodiment of this application. Figure 6 As shown, the device includes: an acquisition module 11, a processing module 12, a determination module 13, and an update module 14. Optionally, the device may further include: an initialization module 15.

[0232] The acquisition module 11 is used to acquire map data collected at least once by at least one acquisition terminal;

[0233] Processing module 12 is used to associate the map data with road segments in the map data based on the driving trajectory of the acquisition terminal in the map data data, so as to obtain map data data of multiple road segments;

[0234] The determining module 13 is used to determine whether the acquisition terminal meets the observation conditions of the road segment; wherein, the observation conditions are used to determine whether the map data acquired by the acquisition terminal is reliable;

[0235] The update module 14 is used to update the confidence level of map features of the target road segment and the road topology confidence level of the target road segment in the map, using the map data of the target road segment, for target road segments that meet the observation conditions.

[0236] One possible implementation involves determining module 13, specifically for:

[0237] Extract a first road segment and multiple consecutive second road segments from the multiple road segments. The observation conditions for the first road segment are that it can be located by both the positioning system and visual location. The observation conditions for the second road segment include either visual location or positioning system location.

[0238] Based on the map data of the first road segment and the map elements within the first road segment in the map, determine whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard.

[0239] If it is determined that the camera intrinsic parameter accuracy of the acquisition terminal is not exceeded, then based on the map data of multiple consecutive second road segments, it is determined whether the inertial measurement unit (IMU) of the acquisition terminal is available.

[0240] Obtain the positioning system used by the acquisition terminal;

[0241] Based on the results of whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, whether the IMU is usable, and the positioning system used by the acquisition terminal, it is determined whether the acquisition terminal meets the observation conditions of the road segment.

[0242] For example, module 13 is specifically used for:

[0243] The map features within the first road segment are reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features within the first road segment.

[0244] Based on the reprojection error of map features within the first road segment, the camera intrinsic parameters of the acquisition terminal are estimated.

[0245] Based on the estimated camera intrinsic parameters of the acquisition end, the map features in the first road segment are reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features in the first road segment.

[0246] If the proportion of map features whose reprojection error exceeds a preset reprojection error threshold is greater than a preset proportion threshold, then it is determined that the camera intrinsic parameter accuracy of the acquisition end exceeds the standard.

[0247] If the proportion of map features whose reprojection error is greater than a preset reprojection error threshold is less than or equal to the preset proportion threshold, then it is determined that the camera intrinsic parameter accuracy of the acquisition end has not exceeded the standard.

[0248] For example, module 13 is specifically used for:

[0249] Based on map data of multiple consecutive second road segments, obtain the starting point and ending point of the multiple consecutive second road segments.

[0250] Based on the starting point of the positioning, the endpoint is recursively calculated using the IMU of the acquisition terminal;

[0251] Based on the endpoint of the positioning and the endpoint of the recursion, the recursion accuracy of the IMU at the acquisition end is obtained;

[0252] If the recursive accuracy of the IMU at the acquisition end is greater than or equal to the IMU recursive accuracy threshold, then the IMU at the acquisition end is determined to be available.

[0253] If the recursive accuracy of the IMU at the acquisition end is less than the IMU recursive accuracy threshold, then the IMU at the acquisition end is determined to be unusable.

[0254] For example, the observation conditions of the road segment include at least one of the following: visually locatable and locatable by the inertial measurement unit, visually locatable and locating by the positioning system, or visually unlocatable; the determination module 13 is specifically used for:

[0255] For road segments that are both visually and IMU-locatable, if the camera intrinsic parameter accuracy of the acquisition terminal does not exceed the standard and the IMU is usable, then it is determined that the acquisition terminal meets the observation conditions for that road segment.

[0256] For road segments that are visually locatable and can be located by a positioning system, if the camera intrinsic parameter accuracy of the acquisition terminal does not exceed the standard, then based on the positioning system used by the acquisition terminal, it is determined whether the acquisition terminal meets the observation conditions for that road segment.

[0257] For road segments that are visually locatable and IMU-locatable, and for road segments that are visually locatable and located by a positioning system, if the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, then it is determined that the acquisition terminal does not meet the observation conditions for that road segment.

[0258] One possible implementation is that the confidence level includes: confidence level of feature existence and confidence level of attributes of target map features related to autonomous driving behavior; update module 14 is specifically used for:

[0259] Based on the driving trajectory of the data collected from the map data of the target road segment, update the road topology confidence of the target road segment and the attribute confidence of the target map elements;

[0260] Based on the collected images from the map data of the target road segment and the map features within the target road segment, update the confidence level of the existence of the map features of the target road segment.

[0261] For example, update module 14 is specifically used to update the road topology confidence of the target road segment based on the traffic relationship between the driving trajectory of the target road segment and the driving trajectory of its adjacent road segments; to obtain driving behavior related to the target map element based on the driving trajectory of the target road segment; and to update the attribute confidence of the target map element based on whether the driving behavior matches the driving behavior constraints represented by the target map element.

[0262] For example, update module 14 is specifically used to reproject map features within the target road segment onto the acquired images in the map data of the target road segment, and obtain the reprojection position of the map feature in the acquired images; if there are no obstructions at the reprojection position of the map feature in the acquired images within a set time period, then update the confidence level of the existence of the map feature based on whether there are map features within a preset range of the reprojection position of the map feature.

[0263] One possible implementation is to initialize module 15, which is used for:

[0264] The roads in the map are divided into segments to obtain multiple road segments;

[0265] Based on the driving trajectories in the historical map data used when constructing the map, the historical map data is associated with the road segments of the map to obtain historical map data for multiple road segments;

[0266] Based on historical map data of the road segments, obtain the observation conditions for each road segment;

[0267] Determine whether the historical data acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment;

[0268] Based on whether the historical data acquisition terminal meets the observation conditions of the road segment, initialize the confidence level of the map features of the road segment, and the road topology confidence level of the road segment.

[0269] For example, initialization module 15 is specifically used for:

[0270] Based on the location data of at least one positioning system in the historical map data of the road segment, determine whether the road segment can be located by the positioning system.

[0271] If the road segment positioning system can locate it, then obtain the group of positioning systems that can locate the road segment;

[0272] Based on the distribution and changes of map elements within the visual range of the road segment in the map, determine whether the road segment is visually locatable;

[0273] If the road segment positioning system cannot locate it, and the distribution makes it visually unlocatable, then the road segment IMU is determined to be locatable, and the IMU recursive accuracy threshold is obtained based on the continuous road segments where the IMU can locate it.

[0274] For example, initialization module 15 is specifically used for:

[0275] Based on the positioning data of at least one positioning system in the historical map data of the road section, obtain the satellite elevation angles of the at least one positioning system in at least four directions of the road section;

[0276] Based on the satellite elevation angles of the at least one positioning system in at least four directions of the road segment, determine whether there is a positioning system in the at least one positioning system that can locate the road segment;

[0277] If it exists, then the observation conditions of the road segment are determined to be that the positioning system can locate it.

[0278] For example, initialization module 15 is specifically used for:

[0279] Based on the distribution of map elements within the visual range of the road segment in the map, the visual relocation level of the road segment is determined;

[0280] The road segment is divided into grids;

[0281] The map features in the collected images from the historical map data of the road section are associated with the grid to obtain the number of times the map features of the grid change within a certain period of time.

[0282] If the visual relocation level of the road segment meets the requirements for visual localization and the proportion of the target grid is less than or equal to a preset proportion, then the road segment is determined to be visually localizable; the target grid is a grid whose map feature changes more than a first preset number of times within a certain time period.

[0283] If the visual relocation level of the road segment does not meet the requirements for visual localization, and / or the proportion of the target grid is greater than a preset proportion, then the road segment is determined to be visually unlocalizable.

[0284] For example, initialization module 15 is specifically used for:

[0285] If the observation conditions of the road segment include visual non-localization, then it is determined that the historical data acquisition terminal does not meet the observation conditions of the road segment.

[0286] If the observation conditions of the road segment include visual location and location system location, then determine whether the location system used by the historical data acquisition terminal includes a group of location systems that can locate the road segment; if it includes a group of location systems that can locate the road segment, then determine that the historical data acquisition terminal meets the observation conditions of the road segment.

[0287] If the observation conditions of the road segment include visual localization and IMU localization, then determine whether the recursive accuracy of the IMU of the historical acquisition terminal is greater than or equal to the IMU recursive accuracy threshold; if it is greater than or equal to the IMU recursive accuracy threshold, then determine that the historical acquisition terminal meets the observation conditions of the road segment.

[0288] For example, initialization module 15 is specifically used for:

[0289] For road segments that do not meet the observation conditions, the confidence level of the existence of map elements in that road segment is set to low confidence; for road segments that meet the observation conditions, the confidence level of the existence of map elements in that road segment is set to high confidence.

[0290] For road segments that are visually unlocatable and visually locatable, the attribute confidence of map elements corresponding to locations in the road segment where the number of element changes within a certain time period exceeds the second preset threshold is set to low confidence; the attribute confidence of map elements corresponding to locations where the number of element changes within a certain time period is less than the second preset threshold is set to high confidence.

[0291] The road topology confidence of the road segment is initialized based on the number of times the driving trajectory of the historical acquisition terminal in the historical map data of the adjacent road segment represents whether the adjacent road segment is passable.

[0292] The map data confidence update device provided in this application is used to execute the aforementioned map data confidence update method embodiment. Its implementation principle and technical effect are similar, and will not be described again.

[0293] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 7The illustrated electronic device 20 includes a memory 21 and a processor 22. Optionally, in some embodiments, the electronic device 20 may further include a communication interface 23. The memory 21, processor 22, and communication interface 23 are communicatively connected to each other. For example, the memory 21, processor 22, and communication interface 23 may be connected via a network. Alternatively, the electronic device 20 may further include a bus 24. The memory 21, processor 22, and communication interface 23 are communicatively connected to each other via the bus 24. Figure 7 This is a schematic diagram of an electronic device 20 in which the memory 21, processor 22, and communication interface 23 communicate with each other through a bus 24.

[0294] The memory 21 may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 21 may store a program, which, when executed by the processor 22, performs the method described in any of the foregoing embodiments. The memory may also store data required to perform the method.

[0295] The processor 22 may be a general-purpose CPU, microprocessor, application-specific integrated circuit (ASIC), graphics processing unit (GPU), or one or more integrated circuits.

[0296] Processor 22 can also be an integrated circuit chip with signal processing capabilities. In implementation, the method embodiments of this application can be completed through integrated logic circuits in the hardware of processor 22 or through software instructions. The aforementioned processor 22 can also be a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 21, and processor 22 reads information from memory 21 and, in conjunction with its hardware, completes the method embodiments of this application.

[0297] Communication interface 23 uses transceiver modules, such as, but not limited to, transceivers, to enable communication between electronic device 20 and other devices or communication networks. For example, map data collected at least once by at least one acquisition end can be acquired through communication interface 23.

[0298] When the aforementioned electronic device 20 includes a bus 24, the bus 24 may include a path for transmitting information between various components of the electronic device 20 (e.g., memory 21, processor 22, communication interface 23).

[0299] This application also provides a computer-readable storage medium, which may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. Specifically, the computer-readable storage medium stores program instructions, which are used in the methods described in the above embodiments.

[0300] This application also provides a program product including executable instructions stored in a readable storage medium. At least one processor of an electronic device can read the executable instructions from the readable storage medium, and the at least one processor executes the executable instructions to cause the electronic device to perform the methods provided in the various embodiments described above.

[0301] The term "multiple" in this document refers to two or more. The term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Furthermore, the character " / " in this document generally indicates an "or" relationship between the preceding and following related objects; in formulas, " / " indicates a "division" relationship. Additionally, it should be understood that in the description of this application, words such as "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or order.

[0302] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application.

[0303] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for updating map data confidence, characterized in that, The method includes: Acquire map data from at least one acquisition terminal at least once; Based on the driving trajectory of the acquisition terminal in the map data, the map data is associated with road segments in the map to obtain map data for multiple road segments; Extract a first road segment and multiple consecutive second road segments from the multiple road segments. The observation conditions for the first road segment are that it can be located by the positioning system and visually. The observation conditions for the second road segment include either visually locatable or location system locatable. Based on the map data of the first road segment and the map elements within the first road segment in the map, determine whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard. If it is determined that the camera intrinsic parameter accuracy of the acquisition terminal is not exceeded, then the availability of the inertial measurement unit of the acquisition terminal is determined based on the map data of multiple consecutive second road segments. Obtain the positioning system used by the acquisition terminal; Based on the results of whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, whether the inertial measurement unit is available, and the positioning system used by the acquisition terminal, it is determined whether the acquisition terminal meets the observation conditions of the road segment, wherein the observation conditions are used to determine whether the map data collected by the acquisition terminal is reliable. For a target road segment that meets the observation conditions, the confidence level of the map features of the target road segment and the road topology confidence level of the target road segment are updated using the map data of the target road segment.

2. The method according to claim 1, characterized in that, The step of determining whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard based on the map data of the first road segment and the map features within the first road segment in the map includes: The map features within the first road segment are reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features within the first road segment. Based on the reprojection error of map features within the first road segment, the camera intrinsic parameters of the acquisition terminal are estimated. Based on the estimated camera intrinsic parameters of the acquisition end, the map features in the first road segment are reprojected onto the acquired image in the map data of the first road segment to obtain the reprojection error of the map features in the first road segment. If the proportion of map features whose reprojection error exceeds a preset reprojection error threshold is greater than a preset proportion threshold, then it is determined that the camera intrinsic parameter accuracy of the acquisition end exceeds the standard. If the proportion of map features whose reprojection error is greater than a preset reprojection error threshold is less than or equal to the preset proportion threshold, then it is determined that the camera intrinsic parameter accuracy of the acquisition end has not exceeded the standard.

3. The method according to claim 1, characterized in that, The step of determining whether the inertial measurement unit at the acquisition end is available based on map data from multiple consecutive second road segments includes: Based on map data of multiple consecutive second road segments, obtain the starting point and ending point of the multiple consecutive second road segments. Based on the starting point of the positioning, the endpoint is recursively calculated using the inertial measurement unit of the acquisition end; Based on the endpoint of the positioning and the endpoint of the recursion, the recursion accuracy of the inertial measurement unit at the acquisition end is obtained. If the recursive accuracy of the inertial measurement unit at the acquisition end is greater than or equal to the recursive accuracy threshold of the inertial measurement unit, then the inertial measurement unit at the acquisition end is determined to be usable. If the recursive accuracy of the inertial measurement unit at the acquisition end is less than the recursive accuracy threshold of the inertial measurement unit, then the inertial measurement unit at the acquisition end is determined to be unusable.

4. The method according to claim 1, characterized in that, The observation conditions include the ability of the positioning system to locate and visual positioning, or visual positioning and inertial measurement unit (IMU) positioning; determining whether the acquisition terminal meets the observation conditions of the road segment based on the results of whether the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, whether the inertial measurement unit is available, and the positioning system used by the acquisition terminal includes: For road segments that are visually locatable and can be located by inertial measurement units, if the camera intrinsic parameter accuracy of the acquisition terminal does not exceed the standard and the inertial measurement unit is available, then it is determined that the acquisition terminal meets the observation conditions for that road segment. For road segments that are visually locatable and can be located by a positioning system, if the camera intrinsic parameter accuracy of the acquisition terminal does not exceed the standard, then based on the positioning system used by the acquisition terminal, it is determined whether the acquisition terminal meets the observation conditions for that road segment. For road segments that are visually locatable and locatable by both inertial measurement units and visually locatable and locating systems, if the camera intrinsic parameter accuracy of the acquisition terminal exceeds the standard, then it is determined that the acquisition terminal does not meet the observation conditions for that road segment.

5. The method according to any one of claims 1-4, characterized in that, The confidence level includes: the confidence level of the existence of the feature, and the confidence level of the attributes of the target map feature related to autonomous driving behavior; the confidence level of updating the map features of the target road segment in the map using the map data of the target road segment, and the road topology confidence level of the target road segment includes: Based on the driving trajectory of the data collected from the map data of the target road segment, update the road topology confidence of the target road segment and the attribute confidence of the target map elements; Based on the collected images from the map data of the target road segment and the map features within the target road segment, update the confidence level of the existence of the map features of the target road segment.

6. The method according to claim 5, characterized in that, The step of updating the road topology confidence of the target road segment and the attribute confidence of the target map features based on the driving trajectory of the data collected from the map data of the target road segment includes: The road topology confidence of the target road segment is updated based on the traffic relationship between the driving trajectory of the target road segment and the driving trajectory of its adjacent road segments. Based on the driving trajectory of the target road segment, obtain the driving behavior related to the target map elements; The attribute confidence of the target map feature is updated based on whether the driving behavior matches the driving behavior constraints represented by the target map feature.

7. The method according to claim 5, characterized in that, The step of updating the confidence level of the existence of map features in the target road segment based on the collected images from the map data of the target road segment and the map features within the target road segment includes: The map features within the target road segment are reprojected onto the acquired image in the map data of the target road segment to obtain the reprojection position of the map feature in the acquired image. If, within the set time period, the reprojection location of the map feature is unobstructed in all images collected, the confidence level of the map feature's existence is updated based on whether the map feature exists within a preset range of its reprojection location.

8. The method according to any one of claims 1-4, characterized in that, The method further includes: The roads in the map are divided into segments to obtain multiple road segments; Based on the driving trajectories in the historical map data used when constructing the map, the historical map data is associated with the road segments of the map to obtain historical map data for multiple road segments; Based on historical map data of the road segments, obtain the observation conditions for each road segment; Determine whether the historical data acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment; Based on whether the historical data acquisition terminal meets the observation conditions of the road segment, initialize the confidence level of the map features of the road segment, and the road topology confidence level of the road segment.

9. The method according to claim 8, characterized in that, The step of obtaining the observation conditions for each road segment based on historical map data includes: Based on the location data of at least one positioning system in the historical map data of the road segment, determine whether the road segment can be located by the positioning system. If the road segment positioning system can locate it, then obtain the group of positioning systems that can locate the road segment; Based on the distribution and changes of map elements within the visual range of the road segment in the map, determine whether the road segment is visually locatable; If the road segment positioning system cannot locate the road segment, and the visual positioning is also unavailable due to the distribution, then the inertial measurement unit (IMU) of the road segment is determined to be locatable, and the recursive accuracy threshold of the IMU is obtained based on the continuous road segments where the IMU can be located.

10. The method according to claim 9, characterized in that, The step of determining whether a road segment is locatable by a positioning system based on positioning data from at least one positioning system in historical map data of the road segment includes: Based on the positioning data of at least one positioning system in the historical map data of the road section, obtain the satellite elevation angles of the at least one positioning system in at least four directions of the road section; Based on the satellite elevation angles of the at least one positioning system in at least four directions of the road segment, determine whether there is a positioning system in the at least one positioning system that can locate the road segment; If it exists, then the observation conditions of the road segment are determined to be that the positioning system can locate it.

11. The method according to claim 9, characterized in that, Determining whether a road segment is visually locatable based on the distribution and changes of map elements within the visual range of the road segment in the map includes: Based on the distribution of map elements within the visual range of the road segment in the map, the visual relocation level of the road segment is determined; The road segment is divided into grids; The map features in the collected images from the historical map data of the road section are associated with the grid to obtain the number of times the map features of the grid change within a certain period of time. If the visual relocation level of the road segment meets the requirements for visual localization and the proportion of the target grid is less than or equal to a preset proportion, then the road segment is determined to be visually localizable; the target grid is a grid whose map feature changes more than a first preset number of times within a certain time period. If the visual relocation level of the road segment does not meet the requirements for visual localization, and / or the proportion of the target grid is greater than a preset proportion, then the road segment is determined to be visually unlocalizable.

12. The method according to claim 11, characterized in that, The process of determining whether the historical data acquisition terminal corresponding to the historical map data meets the observation conditions of the road segment includes: If the road segment cannot be located visually, then it is determined that the historical data acquisition terminal does not meet the observation conditions for that road segment; If the road segment is visually locatable and the positioning system is also locatable, then determine whether the positioning system used by the historical data acquisition terminal includes a group of positioning systems that can locate the road segment; if it includes a group of positioning systems that can locate the road segment, then determine that the historical data acquisition terminal meets the observation conditions for that road segment. If the road segment is visually locatable and the inertial measurement unit (IMU) is locatable, then determine whether the recursive accuracy of the IMU at the historical acquisition end is greater than or equal to the IMU recursive accuracy threshold; if it is greater than or equal to the IMU recursive accuracy threshold, then determine that the historical acquisition end meets the observation conditions for the road segment.

13. The method according to claim 12, characterized in that, The process of initializing the confidence level of map features for the road segment based on whether the historical data acquisition terminal meets the observation conditions for the road segment, and the road topology confidence level for the road segment, includes: For road segments that do not meet the observation conditions, the confidence level of the existence of map elements in that road segment is set to low confidence; for road segments that meet the observation conditions, the confidence level of the existence of map elements in that road segment is set to high confidence. For road segments that are visually unlocatable and visually locatable, the attribute confidence of map elements corresponding to locations in the road segment where the number of element changes within a certain time period exceeds the second preset threshold is set to low confidence; the attribute confidence of map elements corresponding to locations where the number of element changes within a certain time period is less than the second preset threshold is set to high confidence. The road topology confidence of the road segment is initialized based on the number of times the driving trajectory of the historical acquisition terminal in the historical map data of the adjacent road segment represents whether the adjacent road segment is passable.

14. An electronic device, characterized in that, include: Processor and memory; The processor is communicatively connected to the memory; The memory stores computer instructions; The processor executes computer instructions stored in the memory to implement the method as described in any one of claims 1-13.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed by a processor, implement the method as described in any one of claims 1-13.

Images