Map data conversion method and device, equipment and medium

Abstract

The embodiment of the invention discloses a map data conversion method and device, equipment and a medium, and the method comprises the steps: obtaining the position information of a to-be-converted region of map data, the to-be-converted region of map data being a region in which the position information of road elements in a high-precision map does not meet the training data requirement, and the to-be-converted region of map data being a region in which the position information does not meet the training data requirement; the training data requirement is a position requirement which sample data for training the perception model meets, and the trained perception model is used for creating a local map; the position information of road elements in the map data to-be-converted area is adjusted, so that the adjusted position information meets the training data requirement, and the road elements comprise one or more of different types of lane lines. By adopting the technical scheme, the high-precision map data can be reused in the mapping process of the local map, and the problem that time and labor are wasted when the map data are collected again is avoided.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of autonomous driving technology, and in particular, to a method, device, equipment and medium for converting map data. Background Art

[0002] HD map is a high-precision map used for autonomous driving. It contains rich map elements such as road shapes, road markings, traffic signs and obstacles. When creating a local map of a certain area, if the local map data is collected again, the collection process will take up a lot of manpower and material resources and be inefficient. In order to avoid the above problems, the map data of the high-precision map that has been built can be reused in the creation process of the local map.

[0003] Since the mapping rules of high-precision maps and local maps are inconsistent, if the high-precision map data is directly reused in the creation process of the local map, it is impossible to obtain map data that conforms to the mapping rules of the local map. Summary of the invention

[0004] The embodiments of the present invention provide a method, device, equipment and medium for converting map data, so as to reuse high-precision map data in the process of building a local map, thereby avoiding the time-consuming and labor-intensive problem of re-collecting map data.

[0005] The specific technical solutions are as follows:

[0006] In a first aspect, an embodiment of the present invention provides a method for converting map data, comprising:

[0007] Obtaining location information of the area to be converted of the map data, wherein the area to be converted of the map data is an area where the location information of the road elements in the high-precision map does not meet the training data requirements, the training data requirements are location requirements that the sample data used to train the perception model meet, and the trained perception model is used to create a local map;

[0008] The position information of the road elements in the area to be converted of the map data is adjusted so that the adjusted position information meets the training data requirements, wherein the road elements include one or more of different types of lane lines.

[0009] Through the above scheme, it can be known that after determining the map data to be converted area where the position information of the road elements in the high-precision map does not meet the requirements of the training data, by obtaining the position information of the map data to be converted area, the position information of the road elements in the map data to be converted area can be adjusted so that the adjusted position information meets the requirements of the training data, and the adjusted high-precision map data that meets the requirements of the training data can be used in the training process of the perception model to create a local map through the trained perception model. By adopting the above technical scheme, the high-precision map data is reused in the creation process of the local map, avoiding the time-consuming and labor-intensive problem of re-collecting map data when creating the local map, saving a lot of manpower and material costs, and effectively improving the efficiency of creating the local map.

[0010] Optionally, the method provided by the embodiment of the present invention further includes:

[0011] If there is training data that does not belong to the local map in the high-precision map data, the training data that does not belong to the local map will be deleted;

[0012] Accordingly, the step of obtaining the location information of the area to be converted of the map data includes:

[0013] From the high-precision map data after the deletion operation is completed, the location information of the area to be converted of the map data is obtained.

[0014] Optionally, the training data requirements include that the distance information is within a preset distance range, the angle information is within a preset angle range, and the elevation information is within a preset elevation range, and the lane line is in a connected state;

[0015] Accordingly, the position information of the road elements in the area to be converted of the map data is adjusted, including:

[0016] For lane lines of the same type, the starting point and the end point of the lane lines of the same type that are in a disconnected state in the area to be converted in the map data are clustered respectively to obtain a starting point group and an end point group, wherein the types of lane lines include lane boundary lines and road edge lines;

[0017] Determine a first position center of each starting position in the starting point group, and a second position center of each ending position in the ending point group;

[0018] Searching with the first position center and the second position center respectively, determining that the target first position center meets the target second position center that meets the preset matching rule, wherein the preset matching rule is that the distance between the first position center and the second position center is within a first set distance threshold, and there are no other matching objects for both the first position center and the second position center;

[0019] If the number of starting points in the target starting point group corresponding to the target first center position is the same as the number of endpoints in the target endpoint group corresponding to the target second center position, then each starting point in the target starting point group is correspondingly connected to each endpoint in the target endpoint group so that the sum of the distances of the topologically connected lane lines obtained after the connection is minimized.

[0020] Optionally, the training data requirements provided by the embodiment of the present invention also include:

[0021] At the vehicle divergence start position and the vehicle merging end position, there is no virtual lane line whose distance from the lane separation line meets the second set distance threshold, wherein the lane separation line is a solid lane line used to separate two lanes in the lane after divergence and the lane before merging, and the virtual lane line is a dotted lane line added in the high-precision map for lane grouping of the lane after divergence or the lane before merging;

[0022] Accordingly, the position information of the road elements in the area to be converted of the map data is adjusted, including:

[0023] At the vehicle divergence start position and the vehicle merging end position, the virtual lane line whose distance to the separation lane line meets the second set distance threshold is deleted.

[0024] Optionally, the training data requirements provided by the embodiment of the present invention also include:

[0025] The lane centerline of any lane after lane divergence and the lane centerline of any lane before lane merging have a vehicle continuing driving attribute, which is a parameter of training data and is used to indicate that the vehicle continues to drive along the current road after divergence or before merging;

[0026] Accordingly, the position information of the road elements in the area to be converted of the map data is adjusted, including:

[0027] Vehicle continued driving attribute information is added to the lane center line of any lane after lane divergence and the lane center line of any lane before lane merging in the area to be converted of the map data.

[0028] Optionally, the method provided by the embodiment of the present invention further includes:

[0029] The position information of the road elements before and after the adjustment are visualized respectively, and the visualization results before the position information adjustment are compared with the visualization results after the position information adjustment to test the effect of the position adjustment.

[0030] In a second aspect, an embodiment of the present invention further provides a map data conversion device, comprising:

[0031] The module for acquiring the area to be adjusted is configured to acquire the position information of the area to be converted in the map data, wherein the area to be converted in the map data is an area where the position information of the road elements in the high-precision map does not meet the training data requirements, and the training data requirements are the position requirements that the sample data used to train the perception model meet, and the trained perception model is used to create a local map;

[0032] The position adjustment module is configured to adjust the position information of the road elements in the area to be converted of the map data so that the adjusted position information meets the training data requirements, wherein the road elements include one or more of different types of lane lines.

[0033] Optionally, the device provided by the embodiment of the present invention further includes:

[0034] The data deletion module is configured to: if there is training data that does not belong to the local map in the high-precision map data, delete the training data that does not belong to the local map;

[0035] Correspondingly, the module for acquiring the area to be adjusted is specifically configured as follows:

[0036] From the high-precision map data after the deletion operation is completed, the location information of the area to be converted of the map data is obtained.

[0037] Optionally, the training data requirements include that the distance information is within a preset distance range, the angle information is within a preset angle range, and the elevation information is within a preset elevation range, and the lane line is in a connected state;

[0038] Accordingly, the position adjustment module is specifically configured as follows:

[0039] For lane lines of the same type, the starting point and the end point of the lane lines of the same type that are in a disconnected state in the area to be converted in the map data are clustered respectively to obtain a starting point group and an end point group, wherein the types of lane lines include lane boundary lines and road edge lines;

[0040] Determine a first position center of each starting position in the starting point group, and a second position center of each ending position in the ending point group;

[0041] Searching with the first position center and the second position center respectively, determining that the target first position center meets the target second position center that meets the preset matching rule, wherein the preset matching rule is that the distance between the first position center and the second position center is within a first set distance threshold, and there are no other matching objects for both the first position center and the second position center;

[0042] If the number of starting points in the target starting point group corresponding to the target first center position is the same as the number of endpoints in the target endpoint group corresponding to the target second center position, then each starting point in the target starting point group is correspondingly connected to each endpoint in the target endpoint group so that the sum of the distances of the topologically connected lane lines obtained after the connection is minimized.

[0043] Optionally, training data requirements also include:

[0044] At the vehicle divergence start position and the vehicle merging end position, there is no virtual lane line whose distance from the lane separation line meets the second set distance threshold, wherein the lane separation line is a solid lane line used to separate two lanes in the lane after divergence and the lane before merging, and the virtual lane line is a dotted lane line added in the high-precision map for lane grouping of the lane after divergence or the lane before merging;

[0045] Accordingly, the position adjustment module is specifically configured as follows:

[0046] At the vehicle divergence start position and the vehicle merging end position, the virtual lane line whose distance to the separation lane line meets the second set distance threshold is deleted.

[0047] Optionally, training data requirements also include:

[0048] The lane centerline of any lane after lane divergence and the lane centerline of any lane before lane merging have a vehicle continuing driving attribute, which is a parameter of training data and is used to indicate that the vehicle continues to drive along the current road after divergence or before merging;

[0049] Accordingly, the position adjustment module is specifically configured as follows:

[0050] Vehicle continued driving attribute information is added to the lane center line of any lane after lane divergence and the lane center line of any lane before lane merging in the area to be converted of the map data.

[0051] Optionally, the device provided by the embodiment of the present invention further includes:

[0052] The visualization module is configured to visualize the position information of the road elements before and after the adjustment, and compare the visualization results before the position information adjustment with the visualization results after the position information adjustment to test the effect of the position adjustment.

[0053] In a third aspect, an embodiment of the present invention provides a computer device, the computer device comprising:

[0054] At least one processor is coupled to a memory, and the memory stores a program or instruction running on the processor. When the program or instruction is executed by the processor, the map data conversion method provided by any embodiment of the present invention is implemented.

[0055] In a fourth aspect, an embodiment of the present invention provides a readable storage medium having a program or instruction stored thereon, and when the program or instruction is executed by a processor, a method for converting map data as provided in any embodiment of the present invention is implemented.

[0056] In a fifth aspect, an embodiment of the present invention provides a vehicle, which includes a map data conversion device provided by any embodiment of the present invention, or includes a computer device provided by any embodiment of the present invention.

[0057] In a sixth aspect, an embodiment of the present invention provides a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a computer, a method for converting map data as provided in any embodiment of the present invention is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1a A flowchart of a method for converting map data provided in the first embodiment of the present invention;

[0060] Figure 1b A flowchart for adjusting the position of road elements in a high-precision map provided in the first embodiment of the present invention;

[0061] Figure 1c A schematic diagram of a breakpoint connection provided in the first embodiment of the present invention;

[0062] Figure 2 A structural block diagram of a map data conversion device provided in Embodiment 2 of the present invention;

[0063] Figure 3 A structural block diagram of a computer device provided in Embodiment 3 of the present invention;

[0064] Figure 4 A schematic diagram of a vehicle provided in accordance with a fourth embodiment of the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0067] The embodiments of the present invention disclose a method, device, equipment and medium for converting map data, which are described in detail below.

[0068] Embodiment 1

[0069] Figure 1a This is a flowchart of a method for converting map data provided in the first embodiment of the present invention. The method can be applied to vehicle terminals such as vehicle computers and vehicle industrial personal computers (IPC), and can also be applied to servers, and the present embodiment does not limit this. The method provided in this embodiment can be applied to the scenario of reusing high-precision map data in the mapping process of a local map. The method provided in this embodiment can be executed by a map data conversion device, which can be implemented by software and / or hardware. Figure 1a As shown, the method provided in this embodiment specifically includes:

[0070] S110, obtaining location information of the area to be converted in the map data.

[0071] The map data to be converted area is the area where the position information of the road elements in the high-precision map does not meet the training data requirements. The road elements include one or more of different types of lane lines, such as lane boundary lines, road edge lines, road center lines, and lane center lines. The training data requirements are the position requirements that the sample data used to train the perception model meets, which are set according to the mapping rules of the local map. For example, the same type of lane lines within a preset distance range, angle information within a preset angle range, and elevation information within a preset elevation range are in a connected state, or the solid lane lines in the same lane can only be connected to the solid lane lines of the same type in the lane, and the dashed lane lines in the same lane can only be connected to the dashed lane lines of the same type in the lane.

[0072] In this embodiment, the perception model is a deep learning model for creating local maps. During the training of the perception model, map data in the high-precision map that meets the requirements of the local map training data can be used as training sample data for the perception model. After the perception model training is completed, the image data collected by the vehicle camera and the vehicle posture and other data can be used as input. The output of the model is the location information of various types of lane lines, road signs and other road environment data. The road environment data can be used to create local maps in real time, that is, without using high-precision maps, a high-precision local map corresponding to the actual road conditions can be obtained through the trained perception model.

[0073] In this embodiment, the location information of the area to be converted of the map data obtained includes the area number (ID number) of the map area, the longitude, latitude and elevation information of the area. Exemplarily, the area to be converted of the map data can be obtained by manual calibration or automatic identification, and this embodiment does not specifically limit the identification method.

[0074] Furthermore, after obtaining the position information of the area to be converted in the map data, the present embodiment can visualize the position information of the road elements before adjustment, and after adjusting the position information of the area to be converted in the map data, can also visualize the position information of the road elements after adjustment, and then the visualization result before the position information is adjusted can be compared with the visualization result after the position information is adjusted to verify the position adjustment effect, and the road element position adjustment algorithm can also be updated or optimized through the comparison results.

[0075] S120, adjusting the position information of the road elements in the area to be converted in the map data so that the adjusted position information meets the training data requirements.

[0076] In this embodiment, when adjusting the position information of road elements in the area to be converted of the map data, the adjustment can be made according to the pre-set training data requirements, so that the adjusted position information can meet the training data requirements, and the perception model can be trained using high-precision map data after the position information is adjusted.

[0077] Specifically, since different road elements may have different corresponding adjustment methods, in this embodiment, the adjustment algorithm for the position information of road elements in different map data to be converted areas can be managed in a modular manner, and each module can use the same entry function. When adjusting the position information of road elements in the map data to be converted area, the entry function can be called to adjust the position information of road elements in each data to be converted area in turn.

[0078] As an optional implementation, in high-precision map data, for two endpoints whose distance information is within a preset distance range, whose angle information is within a preset angle range, and whose elevation information is within a preset elevation range, even if the two endpoints are not in a connected state, the two endpoints that meet the above conditions will be processed as points at the same position during the processing of the high-precision map data, that is, the fact that the points at the same position in the high-precision map are in a non-connected state does not affect the processing of their positions. Unlike high-precision maps, in local maps, if the two endpoints of the lane lines that meet the above conditions are in a connected state. That is, in this embodiment, the training requirements for the data used to train the perception model include that the distance information is within a preset distance range, the angle information is within a preset angle range, and the elevation information is within a preset elevation range, and the lane lines are in a connected state. Among them, the preset distance range can be set to 0-0.1 meters, the preset angle range can be set to 0-10°, and the preset elevation range can be set to 0-0.1 meters.

[0079] like Figure 1b As shown, in this embodiment, for road elements in the high-precision map that do not meet the above training data requirements, the positions of the road elements can be adjusted according to the following steps S121 to S124 to make them meet the training data requirements of the perception model:

[0080] S121. For lane lines of the same type, the start and end points of the lane lines of this type that are in a disconnected state in the area to be converted in the map data are clustered to obtain a start point group and an end point group.

[0081] The types of lane lines include lane boundary lines and road edge lines. There are many clustering algorithms, for example, a hierarchical clustering algorithm can be used, which is not specifically limited in this embodiment.

[0082] S122: Determine a first position center of each starting point position in the starting point group and a second position center of each ending point position in the ending point group.

[0083] S123, searching with the first position center and the second position center respectively to determine the target first position center and the target second position center that meet the preset matching rule.

[0084] The preset matching rule is that the distance between the first position center and the second position center is within a first set distance threshold, and there are no other matching objects between the first position center and the second position center. The first set distance threshold may be 0.5 to 1 meter.

[0085] S124. If the number of starting points in the target starting point group corresponding to the target first center position is the same as the number of endpoints in the target endpoint group corresponding to the target second center position, then each starting point in the target starting point group is correspondingly connected to each endpoint in the target endpoint group so that the sum of the distances of the topologically connected lane lines obtained after the connection is minimized.

[0086] Furthermore, after connecting each starting point in the starting point group with each end point in the end point group to obtain a topologically connected lane line, the lane line may be smoothed by using a smoothing algorithm to make the lane line smoother.

[0087] In this embodiment, for lane line endpoints that do not meet the above training data requirements, by adopting the above position adjustment method, different types of lane lines with distance information within a preset distance range, angle information within a preset angle range, and elevation information within a preset elevation range can be connected. The above position adjustment method can be applied to the scenario where two disconnected endpoints that do not meet the above training requirements belong to the same mapping area (cell).

[0088] Similarly, in the scenario of two adjacent cells, for lane line endpoints that do not meet the above training data requirements, the above position adjustment method can also be used to make the lane boundary lines, road edge lines, road center lines and lane center lines whose distance information is within the preset distance range, angle information is within the preset angle range and elevation information is within the preset elevation range are respectively in a connected state. It should be noted that when performing position adjustment, in the matching process of the starting group and the end point group, the corresponding preset matching rule in the adjacent cell scenario is that the distance between the first position center and the second position center is within the third set distance threshold, and there are no other matching objects in the first position center and the second position center. Among them, the third set distance threshold is greater than the first set distance threshold, for example, it can be 3 to 5 meters.

[0089] For example, Figure 1c A schematic diagram of a breakpoint connection provided in the first embodiment of the present invention is shown in FIG. Figure 1b As shown, starting points a1 to a4 form a starting point group, and end points b1 to b4 form an end point group. The starting point group and the end point group belong to different cells, and the starting point group and the end point group are uniquely matched and the number of starting points and the number of end points are the same. Figure 1b As shown, according to Figure 1b The connection method of each starting point and each end point shown can minimize the sum of distances of the topologically connected lane lines obtained after connection.

[0090] As another optional implementation, in the high-precision map, in the vehicle divergence and merging scenarios, there is a solid lane line at the divergence start position for separating the lanes after the divergence, and there is a solid lane line at the merging end position for separating the lanes before the merging. In the process of processing high-precision map data, different lanes need to be divided into lane groups according to the lane lines therein. For example, after a main road is divided into two lanes, there are two lane groups. In the process of dividing lane groups according to lane lines, a lane line separating two lanes at the divergence start position or the merging end position can only belong to one lane group, which will lead to the problem of inaccurate lane group division. In order to ensure the accuracy of lane group division, a virtual dashed lane line is usually added in the high-precision map within the range of the second set distance threshold (for example, 20 to 30 cm) between the divergence start position or the merging end position and the separation lane line. Through this virtual lane line, the lanes after the divergence and the lanes before the merging can be accurately divided into lane line groups. For example, the solid lane line at the original divergence start position in the two lane groups after the divergence belongs to one lane group, and the added virtual lane line belongs to another lane group. Unlike the high-precision map, the mapping rules of the local map do not need to divide the lane groups, that is, the virtual lane line added in the high-precision map cannot be directly reused in the local map. If it is reused directly, the virtual lane line added in the high-precision map will be used as the dashed lane line actually existing in the road during the processing of the local map data. This will cause the two lanes that could not merge after the divergence or before the merging to merge at the dashed lane line position. This is inconsistent with the actual situation of the actual road surface, and the obtained local map cannot reflect the real road surface information. In this embodiment, in order to avoid the above problems, the data requirements for training data for the perception model also include:

[0091] At the vehicle divergence start position and the vehicle merging end position, there is no virtual lane line whose lane separation distance meets the second set distance threshold, wherein the lane separation line is a solid lane line used to separate two lanes in the lane after divergence and the lane before merging, and the virtual lane line is a dotted lane line added in the high-precision map for grouping the lane after divergence and the lane before merging;

[0092] Accordingly, in this embodiment, the position information of the road elements in the area to be converted in the map data is adjusted, which may be specifically:

[0093] Processing is performed according to the mapping area (cell). For each cell, at the vehicle divergence start position and the vehicle merging end position, the virtual lane line whose distance to the separation lane line meets the second set distance threshold is deleted.

[0094] As another optional implementation, in the process of processing high-precision map data, there may be a situation where the divergence and converging attributes of the lane centerline are lost. Among them, the divergence and converging attributes include divergence attributes, converging attributes and continue driving attributes, wherein the divergence attribute is used to indicate that the current lane centerline is in a diverging position, the converging attribute is used to indicate that the current lane centerline belongs to a converging position, and the continue driving attribute indicates that the vehicle is driving along the current driving mode on the current road. For local maps, the lane centerline is usually mapped in real time only where the lane is formed. However, the mapping specifications of high-precision maps do not require that the lane centerline must have the continue driving attribute. Therefore, the mapping specifications of local maps must have the vehicle continue driving attribute for the lane centerline after lane divergence or the lane centerline before lane merging, that is, in this embodiment, the training data requirements for the map data used for perception model training also include:

[0095] The lane centerline of any lane after lane divergence and the lane centerline of any lane before lane merging both have a vehicle continuing driving attribute, which is a parameter of training data and is used to indicate that the vehicle continues to drive along the current road after divergence or before merging. Accordingly, in this embodiment, the position information of the road elements in the area to be converted in the map data is adjusted, which can be specifically:

[0096] Vehicle continued driving attribute information is added to the lane center line of any lane after lane divergence and the lane center line of any lane before lane merging in the area to be repaired.

[0097] As another optional implementation, during the construction of the local map, for lane lines in the same lane in the same direction, solid lane lines of the same type are only connected to solid lane lines of the same type, and dashed lane lines of the same type are only connected to dashed lane lines of the same type. In this embodiment, if lane lines of the same type or dashed-solid line connection errors occur in the acquired data to be converted area, the incorrectly connected lane lines can be adjusted according to the positional relationship between the lane lines in the lane to meet the lane line connection requirements in the local map.

[0098] It should be noted that the present embodiment processes the high-precision map data so that the processed data can meet the training requirements of the perception model, thereby creating a local map. That is, by adopting the data conversion method provided in the present embodiment, the high-precision map data can be reused in the local map construction process. The application scenario of the present embodiment is not limited to the above-mentioned high-precision map reuse scenario. For the conversion method of map data provided by the present embodiment, the training data obtained after the high-precision map data is converted can also be converted twice using the method provided in the present embodiment to make the data in the training data that does not meet the preset training requirements meet the preset training requirements. Alternatively, for the newly collected sample data for training the perception model, the method provided in the present embodiment can also be used to convert the data to obtain map data that meets the training data requirements, so as to create a local map that conforms to the real road surface.

[0099] The technical solution provided in this embodiment can adjust the position information of the road elements in the map data to be converted area by obtaining the position information of the map data to be converted area after determining that the position information of the road elements in the high-precision map does not meet the requirements of the training data, so that the adjusted position information meets the requirements of the training data, and the adjusted high-precision map data that meets the requirements of the training data can be used in the training process of the perception model to create a local map through the trained perception model. By adopting the above technical solution, the high-precision map data is reused in the creation process of the local map, avoiding the time-consuming and labor-intensive problem of re-collecting map data when creating the local map, saving a lot of manpower and material costs, and effectively improving the efficiency of creating the local map.

[0100] Embodiment 2

[0101] Figure 2 A structural block diagram of a map data conversion device provided in the second embodiment of the present invention, such as Figure 2 As shown, the device includes: a module 210 for acquiring an area to be adjusted and a module 220 for adjusting a position, wherein:

[0102] The module 210 for acquiring the area to be adjusted is configured to acquire the position information of the area to be converted in the map data, wherein the area to be converted in the map data is an area where the position information of the road elements in the high-precision map does not meet the training data requirements, and the training data requirements are the position requirements that the sample data used to train the perception model meet, and the trained perception model is used to create a local map;

[0103] The position adjustment module 220 is configured to adjust the position information of the road elements in the area to be converted of the map data so that the adjusted position information meets the training data requirements, wherein the road elements include one or more of different types of lane lines.

[0104] Optionally, the device provided by the embodiment of the present invention further includes:

[0105] The data deletion module is configured to: if there is training data that does not belong to the local map in the high-precision map data, delete the training data that does not belong to the local map;

[0106] Correspondingly, the module for acquiring the area to be adjusted is specifically configured as follows:

[0107] From the high-precision map data after the deletion operation is completed, the location information of the area to be converted of the map data is obtained.

[0108] Optionally, the training data requirements include that the distance information is within a preset distance range, the angle information is within a preset angle range, and the elevation information is within a preset elevation range, and the lane line is in a connected state;

[0109] Accordingly, the position adjustment module is specifically configured as follows:

[0110] For lane lines of the same type, the starting point and the end point of the lane lines of the same type that are in a disconnected state in the area to be converted in the map data are clustered respectively to obtain a starting point group and an end point group, wherein the types of lane lines include lane boundary lines and road edge lines;

[0111] Determine a first position center of each starting position in the starting point group, and a second position center of each ending position in the ending point group;

[0112] Searching with the first position center and the second position center respectively, determining that the target first position center meets the target second position center that meets the preset matching rule, wherein the preset matching rule is that the distance between the first position center and the second position center is within a first set distance threshold, and there are no other matching objects for both the first position center and the second position center;

[0113] If the number of starting points in the target starting point group corresponding to the target first center position is the same as the number of endpoints in the target endpoint group corresponding to the target second center position, then each starting point in the target starting point group is correspondingly connected to each endpoint in the target endpoint group so that the sum of the distances of the topologically connected lane lines obtained after the connection is minimized.

[0114] Optionally, training data requirements also include:

[0115] At the vehicle divergence start position and the vehicle merging end position, there is no virtual lane line whose distance from the lane separation line meets the second set distance threshold, wherein the lane separation line is a solid lane line used to separate two lanes in the lane after divergence and the lane before merging, and the virtual lane line is a dotted lane line added in the high-precision map for lane grouping of the lane after divergence or the lane before merging;

[0116] Accordingly, the position adjustment module is specifically configured as follows:

[0117] At the vehicle divergence start position and the vehicle merging end position, the virtual lane line whose distance to the separation lane line meets the second set distance threshold is deleted.

[0118] Optionally, training data requirements also include:

[0119] The lane centerline of any lane after lane divergence and the lane centerline of any lane before lane merging have a vehicle continuing driving attribute, which is a parameter of training data and is used to indicate that the vehicle continues to drive along the current road after divergence or before merging;

[0120] Accordingly, the position adjustment module is specifically configured as follows:

[0121] Vehicle continued driving attribute information is added to the lane center line of any lane after lane divergence and the lane center line of any lane before lane merging in the area to be converted of the map data.

[0122] Optionally, the device provided by the embodiment of the present invention further includes:

[0123] The visualization module is configured to visualize the position information of the road elements before and after the adjustment, and compare the visualization results before the position information adjustment with the visualization results after the position information adjustment to test the effect of the position adjustment.

[0124] The map data conversion device provided in the embodiment of the present invention can execute the map data conversion method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not described in detail in the above embodiment, please refer to the map data conversion method provided in any embodiment of the present invention.

[0125] Embodiment 3

[0126] Figure 3 A structural block diagram of a computer device provided in Embodiment 3 of the present invention is shown in FIG. Figure 3 As shown, the computer device includes:

[0127] At least one processor ( Figure 3 A processor 520 is shown,

[0128] The processor 520 is coupled to the memory 510 , and the memory 510 stores a program or instruction running on the processor 520 . When the program or instruction is executed by the processor 520 , the map data conversion method provided by any embodiment of the present invention is implemented.

[0129] Based on the above embodiments, another embodiment of the present invention provides a vehicle, wherein the vehicle includes the apparatus as described in any of the above embodiments, or includes the computer device as described above.

[0130] Embodiment 4

[0131] Figure 4 This is a schematic diagram of a vehicle provided in Embodiment 4 of the present invention. Figure 4 As shown, the vehicle includes a speed sensor 61, an ECU (Electronic Control Unit) 62, a GPS (Global Positioning System) positioning device 63, and a T-Box (Telematics Box) 64. The speed sensor 61 is used to measure the vehicle speed and use the vehicle speed as an empirical speed for model training; the GPS positioning device 63 is used to obtain the current geographical location of the vehicle; the T-Box 64 can communicate with the server as a gateway; and the ECU 62 can execute the above-mentioned map data conversion method.

[0132] In addition, the vehicle may also include: a V2X (Vehicle-to-Everything) module 65, a radar 66 and a camera 67. The V2X module 65 is used to communicate with other vehicles, roadside equipment, etc.; the radar 66 or the camera 67 is used to sense the road environment information in the front and / or other directions to obtain original point cloud data; the radar 66 and / or the camera 67 can be configured at the front and / or rear of the vehicle body.

[0133] Based on the above method embodiments, another embodiment of the present invention provides a readable storage medium having a program or instruction stored thereon, which, when executed by a processor, enables the processor to implement the map data conversion method as described in any of the above embodiments.

[0134] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0135] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for converting map data, characterized in that: include: Obtaining location information of a region to be converted in map data, wherein the region to be converted in map data is a region where location information of road elements in a high-precision map does not meet training data requirements, the training data requirements are location requirements that sample data used to train a perception model meet, and the trained perception model is used to create a local map; The position information of the road elements in the area to be converted of the map data is adjusted so that the adjusted position information meets the training data requirements, wherein the road elements include one or more of different types of lane lines.

2. The method according to claim 1, characterized in that The method further comprises: If there is training data that does not belong to the local map in the high-precision map data, the training data that does not belong to the local map is deleted; Accordingly, the step of obtaining the location information of the area to be converted of the map data includes: From the high-precision map data after the deletion operation is completed, the location information of the area to be converted of the map data is obtained.

3. The method according to claim 1 or 2, characterized in that: The training data requirements include: the same type of lane lines whose distance information is within a preset distance range, whose angle information is within a preset angle range, and whose elevation information is within a preset elevation range are in a connected state; Accordingly, the adjusting the position information of the road elements in the area to be converted of the map data includes: For lane lines of the same type, the starting point and the end point of the lane lines of the same type that are in a disconnected state in the area to be converted of the map data are clustered respectively to obtain a starting point group and an end point group, wherein the types of the lane lines include lane boundary lines and road edge lines; Determine a first position center of each starting position in the starting point group, and a second position center of each ending position in the ending point group; Searching with the first position center and the second position center respectively, to determine the target first position center and the target second position center that meet the preset matching rule, wherein the preset matching rule is that the distance between the first position center and the second position center is within a first set distance threshold, and there are no other matching objects for both the first position center and the second position center; If the number of starting points in the target starting point group corresponding to the first center position of the target is the same as the number of end points in the target end point group corresponding to the center of the second target position, then each starting point in the target starting point group is correspondingly connected to each end point in the target end point group so that the sum of the distances of the topologically connected lane lines obtained after the connection is minimized.

4. The method according to claim 3, characterized in that The training data requirements also include: At the vehicle divergence start position and the vehicle merging end position, there is no virtual lane line whose distance from the lane separation line meets the second set distance threshold, wherein the lane separation line is a solid lane line used to separate two lanes in the lane after divergence and the lane before merging, and the virtual lane line is a dotted lane line added in the high-precision map for lane grouping of the lane after divergence or the lane before merging; Accordingly, the adjusting the position information of the road elements in the area to be converted of the map data includes: At the vehicle divergence start position and the vehicle merging end position, the virtual lane line whose distance to the lane separation line meets the second set distance threshold is deleted.

5. The method according to claim 3, characterized in that: The training data requirements also include: A lane centerline of any lane after lane divergence and a lane centerline of any lane before lane merging have a vehicle continuing driving attribute, where the continuing driving attribute is a parameter of training data, and is used to indicate that the vehicle continues to drive along the current road after divergence or before merging; Accordingly, the adjusting the position information of the road elements in the area to be converted of the map data includes: Vehicle continued driving attribute information is added to the lane center line of any lane after lane divergence and the lane center line of any lane before lane merging in the area to be converted of the map data.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: The position information of the road elements before and after the adjustment are visualized respectively, and the visualization results before the position information adjustment are compared with the visualization results after the position information adjustment to test the effect of the position adjustment.

7. A map data conversion device, characterized in that: include: The module for acquiring the area to be adjusted is configured to acquire the position information of the area to be converted in the map data, wherein the area to be converted in the map data is an area where the position information of the road elements in the high-precision map does not meet the training data requirements, and the training data requirements are the position requirements that the sample data used to train the perception model meet, and the trained perception model is used to create a local map; The position adjustment module is configured to adjust the position information of the road elements in the area to be converted of the map data so that the adjusted position information meets the training data requirements, wherein the road elements include one or more of different types of lane lines.

8. The device according to claim 7, characterized in that The device also includes: The data deletion module is configured to: if there is training data that does not belong to the local map in the high-precision map data, delete the training data that does not belong to the local map; Correspondingly, the module for acquiring the area to be adjusted is specifically configured as follows: From the high-precision map data after the deletion operation is completed, the location information of the area to be converted of the map data is obtained.

9. A computer device, characterized in that: The method comprises at least one processor, wherein the processor is coupled to a memory, wherein the memory stores a program or instruction running on the processor, and when the program or instruction is executed by the processor, the steps of the method for converting map data as claimed in any one of claims 1 to 6 are implemented.

10. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or instruction is executed by a processor, the steps of the map data conversion method according to any one of claims 1 to 6 are implemented.

Images