A method, device, equipment and readable medium for generating vector semantic data in a set format

Abstract

The embodiments of this specification disclose a method, apparatus, device and readable medium for generating vector semantic data in a set format, obtaining vector semantic data in a first format generated based on raw data collected by a sensor; processing dimensional data of each dimension in the vector semantic data in the first format to obtain dimensional data of each dimension in the vector semantic data in a second format, wherein the vector semantic data in the set format includes at least geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information; and generating vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format.

Description

Technical Field

[0001] The present application relates to the field of map technology, and in particular to a method, device, equipment and readable medium for generating vector semantic data in a set format. Background Art

[0002] High-precision maps serve the autonomous driving system of smart cars, providing information on lanes, slopes, curvatures, headings, etc., and predicting roads and surrounding environments in advance. The high reliability of high-precision maps is mainly reflected in map freshness, granularity, and map accuracy. With the deepening of high-precision map applications, higher requirements and challenges are put forward for map quality and freshness. The production and update of high-precision maps based on traditional self-collected equipment requires a large amount of data collection and processing. The data production investment is large, the cost is high, and the output efficiency is low and the cycle is long. It can no longer meet the demand of autonomous driving for nationwide wide-area high-precision maps. Most map service providers have begun to rely on multi-source crowdsourcing data to achieve timely update technology research and application. Among them, semantic data based on vehicle-side sensors is a crucial data source in the crowdsourcing update of high-precision maps.

[0003] However, data from different OEMs and data provided by different sensor data processors are different because they are used for different assisted driving or autonomous driving applications, such as GPS / Camera, RTK / Radar / Lidar, etc. Different autonomous driving applications also have different subjective perceptions of the objective world. The richness of the attributes of the identified elements and the granularity of their expression are also inconsistent. Therefore, data from different sources do not have semantic consistency, especially for attributes with semantic information of different granularities, which are difficult to maintain uniformity in the final data model. For example, the perception of sign shapes, such as quadrilaterals, rectangles, oblongs, and squares, is difficult to unify the expressions of different sources. Summary of the invention

[0004] The embodiments of the present specification provide a method, apparatus, device and readable medium for generating vector semantic data in a set format to solve the problem that data from different data sources are difficult to unify.

[0005] To solve the above technical problems, the embodiments of this specification are implemented as follows:

[0006] An embodiment of the present specification provides a method for generating vector semantic data in a set format, including:

[0007] Acquire vector semantic data in a first format generated based on raw data collected by the sensor;

[0008] Processing the dimensional data of each dimension in the vector semantic data in the first format to obtain the dimensional data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information;

[0009] The vector semantic data in the second format is generated based on the data of each dimension in the vector semantic data in the second format.

[0010] An embodiment of the present specification provides a device for generating vector semantic data in a set format, including:

[0011] An acquisition unit, which acquires vector semantic data in a first format generated based on raw data collected by the sensor;

[0012] A data processing unit processes the dimension data of each dimension in the vector semantic data in the first format to obtain the dimension data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimension information and attribute data for describing attribute dimension information;

[0013] The vector semantic data generating unit generates the vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format.

[0014] An embodiment of the present specification provides a device for generating vector semantic data in a set format, including:

[0015] at least one processor; and,

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to:

[0018] Acquire vector semantic data in a first format generated based on raw data collected by the sensor;

[0019] Processing the dimensional data of each dimension in the vector semantic data in the first format to obtain the dimensional data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information;

[0020] The vector semantic data in the second format is generated based on the data of each dimension in the vector semantic data in the second format.

[0021] An embodiment of the present specification provides a computer-readable medium on which computer-readable instructions are stored. The computer-readable instructions can be executed by a processor to implement a method for generating vector semantic data in a set format.

[0022] An embodiment of the present specification realizes the following beneficial effects: by acquiring vector semantic data in a first format generated based on raw data collected by a sensor, dimensional data of each dimension in the vector semantic data in the first format is processed to obtain dimensional data of each dimension in the vector semantic data in a second format, and then vector semantic data in a second format is generated based on the data of each dimension in the vector semantic data in the second format.

[0023] In this way, by analyzing and processing the vector semantic data in the first format generated based on the original data, the dimensional data of each dimension of the vector semantic data in the second format is obtained, thereby realizing the unified conversion of the multi-source sensor data format, and being able to effectively eliminate the expression differences between multi-source heterogeneous data, thus laying the foundation for the processing, fusion, and map updating of complex heterogeneous multi-source sensor semantic data. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of this specification 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 recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0025] Figure 1 It is an application scenario diagram of a method for generating vector semantic data in a set format provided by an embodiment of this specification;

[0026] Figure 2 is a flow chart of a method for generating vector semantic data in a set format provided by an embodiment of this specification;

[0027] Figure 3 It is a tree-like multi-granularity semantic model expression diagram of the lane marking type provided in the embodiment of this specification;

[0028] Figure 4 It is a tree-like multi-granularity semantic model expression diagram of the obstacle type provided in the embodiment of this specification;

[0029] Figure 5 It is a flowchart of a method for generating vector semantic data in a set format provided by an embodiment of this specification;

[0030] Figure 6It is a tree structure definition flow chart of a method for generating vector semantic data in a set format provided by an embodiment of this specification;

[0031] Figures 7 to 10 It is a schematic diagram of a specific tree structure at each level provided in the embodiments of this specification;

[0032] Fig.11 It is a structural schematic diagram of a device for generating vector semantic data in a set format provided by an embodiment of this specification;

[0033] Fig.12 It is a structural diagram of a device for generating vector semantic data in a set format provided in an embodiment of this specification. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of one or more embodiments of this specification clearer, the technical solutions of one or more embodiments of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this description, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of one or more embodiments of this specification.

[0035] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0036] In the existing technology, with the in-depth research and development of autonomous driving technology, higher requirements and challenges have been raised for the quality, reliability and freshness of high-precision maps. The production and update of high-precision maps based on traditional self-collected equipment requires a large amount of data collection and processing, with large data production input and high cost, low output efficiency and long cycle, which can no longer meet the demand of autonomous driving for nationwide wide-area high-precision maps.

[0037] At present, crowdsourcing is usually used to update high-precision maps. In the map field, crowdsourcing refers to the use of users' intelligent connected vehicles and roadside equipment to collect road information and generate high-precision map crowdsourcing information, thereby achieving rapid and effective updates to high-precision maps.

[0038] However, sensor data obtained through crowdsourcing, on the one hand, may have data components of undefined dimensions due to differences in capabilities, data quality, and understanding of objective reality among sensor data from different sources, which may lead to reduced sensor data accuracy. On the other hand, sensor data from different sources may have differences in the granularity of feature attribute expression, multiple confidence levels, and uneven freshness, which may result in the data collected by the sensor being coarse or fine granularity. In this case, there may also be data components of undefined dimensions, making it incompatible with coarse or fine granularity data.

[0039] However, there is currently no unified method for sensor data formats to accommodate differences in understanding of multi-source sensor data, so as to normalize and solve the difficulties in using data caused by these differences. However, due to the wide range of sources of high-precision map crowdsourcing information, it is necessary to uniformly analyze and define a large number of data components of various granularities and dimensions, and it is necessary to consider the richness of the attributes of the identified elements and the granularity of expression, semantic consistency, and the unification of data models. Undoubtedly, this process is complex and difficult.

[0040] Therefore, in order to solve the defects in the prior art, this solution provides the following embodiments:

[0041] Figure 1 A scenario application diagram corresponding to the method for generating vector semantic data in a set format provided in an embodiment of this specification.

[0042] like Figure 1 The figure shows a map data update flow chart, which mainly includes data interaction between the vehicle-side processing system and the back-end processing system. The vehicle-side processing system includes local sensors and control sensors for collecting sensor data, an environmental model for preliminary processing of sensor data, a localization unit and a data acquisition unit, as well as a local real-time map server, a high-definition map server, and a car network map server. The back-end processing system includes a data storage unit, a shared data storage unit, a system data acquisition unit, a dynamic data generation unit, a sensor data fusion unit, a data collection strategy unit, a map update unit, a real-time map update unit, a real-time map transmission unit, a high-definition map transmission unit, and a car network map transmission unit.

[0043] The local sensor and the control sensor collect sensor data and send them to the environment model, the localization unit and the data collection unit for preprocessing. The preprocessed data are sent to the data storage unit and the shared data storage unit. The shared data storage unit sends the received data and the system data collection unit sends the collected system data to the dynamic data generation unit, the sensor data fusion unit and the data collection strategy unit respectively.

[0044] After the dynamic data generation unit, the sensor data fusion unit, and the data collection strategy unit process the received data, the processed data is sent to the real-time map update unit. The real-time map update unit sends the updated data to the real-time map transmission unit to transmit the updated data to the local real-time map server. The map update unit sends the updated data to the high-definition map transmission unit and the Internet of Vehicles map transmission unit respectively to transmit the updated data to the high-definition map server and the Internet of Vehicles map server respectively.

[0045] In this process, the sensor data fusion unit may execute the method for generating vector semantic data in a set format provided in the embodiment of this specification, and the specific implementation process is as follows: Figure 2 shown.

[0046] Figure 2 A schematic diagram of the overall solution flow of a method for generating vector semantic data in a set format provided in an embodiment of this specification. From a program perspective, the execution subject of the process can be a server, or an application installed on the server, wherein the server can be implemented as a cloud server or a transit server, which is not specifically limited here.

[0047] like Figure 2 As shown, the method may specifically include the following steps:

[0048] Step 201: Acquire vector semantic data in a first format generated based on raw data collected by a sensor;

[0049] Step 203: Processing the dimension data of each dimension in the vector semantic data in the first format to obtain the dimension data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimension information and attribute data for describing attribute dimension information;

[0050] Step 205: Generate vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format.

[0051] In the embodiments of this specification, the sensor may be a vehicle-mounted sensor, roadside equipment or other monitoring equipment that can collect road information. The sensor data collected by the sensor may refer to the classification result data expressed with spatial vector information of the vehicle-side sensor or roadside equipment's perception and recognition of the real world. It is the data generated during autonomous driving applications such as perception and positioning, and has data characteristics such as small data volume, high real-time performance, and good local relative relationships.

[0052] Specifically, the raw data collected by the sensor may refer to the native data collected by the sensor, and its data format is a format preset by the sensor itself. In one case, the server can directly analyze the raw data collected by the sensor to obtain the dimensional data of each dimension in the vector semantic data of the set format; in another case, the vehicle-side or roadside equipment can generate vector semantic data in a first format based on the raw data collected by the sensor, and the server processes the dimensional data of each dimension in the vector semantic data of the first format to obtain the dimensional data of each dimension in the vector semantic data of the second format. Based on the dimensional data of each dimension in the vector semantic data of the second format, vector semantic data of the second format is generated to generate or update high-precision map data based on the vector semantic data of the second format.

[0053] Further, processing the dimension data of each dimension in the vector semantic data in the first format to obtain the dimension data of each dimension in the vector semantic data in the second format may specifically include:

[0054] Obtaining a first configuration file corresponding to the vector semantic data in the first format;

[0055] The second configuration file is determined based on the first configuration file by utilizing a mapping relationship between the first configuration file and a second configuration file corresponding to the vector semantic data in the second format.

[0056] For example, the first configuration file corresponding to the vector semantic data in the first format corresponding to data A is 010203, where 01 represents a square, 02 represents white, and 03 represents a no-parking sign; according to the mapping relationship between the first configuration file and the second configuration file (a1 represents a square, b2 represents white, and c3 represents a no-parking sign), the second configuration file corresponding to the vector semantic data in the second format corresponding to data A is a1b2c3.

[0057] In the embodiments of this specification, in order to reduce the difficulty of fusion processing of data expressed from multiple sources and at different data granularities, the original data collected by the sensors is redefined comprehensively from the aspects of analysis of multi-source heterogeneous sensor data and conversion and generation methods of multi-source heterogeneous sensor data, which can effectively eliminate the expression differences between multi-source heterogeneous data models, thereby laying the foundation for processing, fusion and map updating based on complex multi-source heterogeneous data.

[0058] Among them, multi-granularity semantics can refer to the unified expression of different understandings of entities or attributes in the real world from different sources.

[0059] Specifically, this application is based on a systematic analysis of the crowdsourcing information update needs of high-precision maps, designs a comprehensive, inclusive and unified tree expression model, and then performs semantic analysis on the acquired multi-source heterogeneous data, including data semantic attribute classification, semantic attribute value domain analysis, etc. By designing the conversion principles of multi-source heterogeneous data to a unified tree structure, and template-izing the conversion principles to form a conversion template, it is possible to achieve unified conversion and fusion of models of multi-source heterogeneous data based on the conversion template.

[0060] That is, the raw data collected by the sensor or the first format data generated by the raw data collected by the sensor is analyzed and processed, and converted into dimensional data of each dimension in the high-precision map data of a set format, so as to realize the unified conversion and fusion of multi-source heterogeneous data formats.

[0061] based on Figure 2 The present specification also provides some specific implementation plans of the method, which are described below.

[0062] As an application embodiment, the method for generating vector semantic data in a set format may further include:

[0063] Aligning the vector semantic data in the second format with each element in the high-precision map data to obtain aligned data, wherein the aligned data has a unique corresponding map element;

[0064] Aggregating the aligned data of the same map element to obtain aggregated data, wherein the aggregated data of the same map element includes dimension data of each dimension of the same map element, and for the same dimension, the aggregated data includes the finest-grained dimension data;

[0065] Based on the aggregated data, the changed map elements are determined and the high-precision map data corresponding to the changed map elements are updated.

[0066] In actual application scenarios, due to the huge number of vehicle-side sensors or roadside equipment, the amount of collected data is also large. The types and numbers of map elements corresponding to these sensor data are numerous. In order to improve data processing efficiency, the vector semantic data in the second format can be aligned with the elements in the high-precision map data. That is, the vector semantic data in the second format can be classified according to the elements in the high-precision map data. For example, the vector semantic data in the second format corresponding to the same road section or the same intersection can be aligned. In this way, the vector semantic data in the second format of the same element can be processed separately subsequently to improve data processing efficiency and accuracy.

[0067] Furthermore, the data for the same map element may come from multiple different sensors, each sensor has different data accuracy, and the resulting data granularity is also different.

[0068] For example, for the same road sign, the data from sensor A describes its geometric dimension as a quadrilateral and its color dimension as white;

[0069] The data of sensor B, whose geometric dimension is described as rectangle and whose functional dimension is no parking;

[0070] The data of sensor C describes the geometric dimension of a square and the location dimension of the sensor C being set on the sidewalk.

[0071] Then, after aggregating these data, the data of this element may include geometric dimension, color dimension, functional dimension and location dimension. As for the geometric dimension, only the most fine-grained square may be included.

[0072] By aggregating the data after alignment of the same element, on the one hand, the data formats of various sensors can be made compatible, and on the other hand, the amount of data in the configuration file can be reduced.

[0073] As an application embodiment, the dimensional data of the high-precision map data in the set format may also include: relationship data.

[0074] In the embodiments of this specification, relational data may specifically refer to logical relationships or attribution relationship data between physical entities, etc. Among them, physical entities may refer to the abstraction of entities that objectively exist in the real world, such as lane markings, traffic signs, etc. can all belong to physical entities.

[0075] Furthermore, further abstraction based on physical entities can form logical entities. Logical entities focus on the logical relationship between physical entities. For example, lane boundaries are logical entities, which may be composed of multiple marking physical entities, or they may be composed of markings and traffic guardrails.

[0076] For example, the relationship data may include the belonging relationship data with the lane, the belonging relationship data with the boundary reference line, etc., which are not specifically limited here.

[0077] As an application embodiment, the method may specifically include:

[0078] When the original data is data containing information of road reference lines,

[0079] The geometric data includes one or more of spatial position data and shape data;

[0080] The attribute data includes one or more of travel direction data and vehicle speed data.

[0081] In the embodiments of the present specification, the road reference line may specifically refer to an abstract expression of a road, that is, a reference line of a road represented on an electronic map.

[0082] Geometric data may specifically refer to attributes related to the spatial position and shape of physical objects in the real world. It may express only the spatial position or both the spatial position and shape data.

[0083] Attribute data may refer to the feature-related attributes that describe physical objects in the real world, and may specifically be composed of categorical attributes and numerical attributes. In the case where the original data is a road reference line, the attribute data may be the travel direction data, vehicle speed data, etc. related to the road, which are not specifically limited here.

[0084] Specifically, classification attributes are used to express the attributes of each type of real object, which are further divided into enumeration types and tree types. The tree type can flexibly accommodate attributes of multiple granularities of different perception systems.

[0085] Numeric attributes are used to express the numerical type attributes of real objects, such as length, diameter, etc.

[0086] Attribute data may also include geometric attributes, which are used to express the attributes of the shape of real objects, such as length, width, pitch angle, roll angle, heading angle, etc. The values ​​of such attributes may be expressed numerically.

[0087] As an application embodiment, the method may specifically include:

[0088] When the original data is data containing information about lane reference lines,

[0089] The geometric data includes one or more of spatial position data and shape data;

[0090] The attribute data includes one or more of travel direction data, lane type data and vehicle speed data.

[0091] In the embodiments of the present specification, a lane reference line may refer to an abstract expression of a lane on an electronic map, wherein a lane may refer to a portion of a lane for a single column of vehicles to travel.

[0092] When the original data is data containing information about lane reference lines, the attribute data may specifically refer to lane-related travel direction data, lane type data, and vehicle speed data, etc., which are not specifically limited here.

[0093] As an application embodiment, the method may specifically include:

[0094] When the original data is data containing information of a boundary reference line,

[0095] The attribute data includes one or more of boundary role data, boundary crossability data, and line marking quantity data.

[0096] In the embodiments of this specification, the boundary reference line may refer to a component of a lane, and each lane is composed of two lane boundaries on the left and right. The lane boundary is a logical entity, which may be composed of markings, guardrails, etc.

[0097] When the original data is data containing information about a boundary reference line, the attribute data may specifically refer to boundary role data, boundary crossability data, and marking line quantity data related to the boundary reference line, which are not specifically limited here.

[0098] As an application embodiment, the method may specifically include:

[0099] When the original data is data containing information about lane markings,

[0100] The attribute data includes one or more of marking type data, marking color data, marking width data, and marking material data.

[0101] In the embodiments of this specification, lane markings may specifically refer to information related to a paint line object printed on a road surface to separate driving lanes, including ID, geometry, attributes, and the relationship with a boundary reference line. The geometric expression of lane markings is to represent the shape of the lane markings through continuous 3D point coordinates.

[0102] When the original data is data containing lane marking information, the attribute data may specifically refer to lane marking type data, lane marking color data, lane marking width data, lane marking material data, etc. related to the lane marking, which is not specifically limited here.

[0103] Among them, the lane marking color indicates the color recorded according to the actual situation on site, such as white, yellow, red, black, unknown color, etc.

[0104] The lane marking width data represents the lane marking width in millimeters.

[0105] Lane marking material data represents the properties of lane marking materials, which are divided into unknown, paint, and other lane marking relationship expressions.

[0106] It should be noted that there is a many-to-one relationship between lane markings and boundary reference lines, and a boundary reference line may contain one or more lane markings.

[0107] In a specific application scenario, the data type of the marking line type data may specifically include:

[0108] One or more of unknown type, other type, solid line type, and dashed line type.

[0109] like Figure 3 As shown, a tree-like multi-granularity semantic model expression diagram of lane marking types is shown.

[0110] Among them, the dashed line types include ordinary dashed lines and special dashed lines. Ordinary dashed lines include long dashed lines and short dashed lines. Special dashed lines include other special dashed lines, dashed line blocks and diamond-shaped speed reduction lines.

[0111] As an application embodiment, the method may specifically include:

[0112] When the original data is data containing information about roadside obstacles,

[0113] The geometric data includes one or more of representative line data and geometric shape data;

[0114] The attribute data includes obstacle type data.

[0115] In the embodiments of this specification, roadside obstacles may specifically refer to features at the boundaries of road pavement that serve as boundaries between carriageways, sidewalks, green spaces, isolation zones, and other parts of the road. The main information includes ID, geometry, core attributes, and relationships with other objects.

[0116] Due to the integrity of observation and the subjectivity of expression, the geometry of the same roadside obstacle is diverse. In order to better accommodate the geometric expression of roadside obstacles from multiple heterogeneous sources without losing or deforming information, the geometric data of roadside obstacles are divided into representative lines and geometric shapes.

[0117] The representative line may refer to a three-dimensional vector line that abstractly expresses the existence and geometric position of a roadside obstacle. The geometric shape may refer to a three-dimensional surface that describes the existence, geometric shape and position of a roadside obstacle, etc., which is not specifically limited here.

[0118] For obstacle type data, such as Figure 4 As shown, it is expressed through a tree-like multi-granularity semantic model.

[0119] As an application embodiment, the data type of the obstacle type data may specifically include:

[0120] One or more of Unknown Type, Other Type, Curb Type, Guardrail Type, and Wall Type.

[0121] Among them, the curb types may include other curbs, ordinary curbs and no-parking curbs, etc., while the no-parking curbs may include ordinary no-parking curbs and no-long-term-stopping curbs, etc.

[0122] Guardrail types may include other guardrails, New Jersey guardrails, safety guardrails, fences, movable guardrails and water horse guardrails, etc., and movable guardrails may include movable guardrail teeth, plug-in movable guardrail teeth, push-pull movable guardrail teeth, folding movable guardrails, etc., which are not specifically limited here.

[0123] In addition, it should be noted that there is a many-to-one relationship between roadside obstacles and boundary reference lines, and one boundary reference line may contain one or more roadside obstacles.

[0124] As an application embodiment, the method may specifically include:

[0125] When the original data is data containing information of traffic signs,

[0126] The geometric data includes one or more of representative point data and geometric shape data;

[0127] The attribute data includes one or more of identification type data and identification content data.

[0128] In the embodiments of this specification, traffic signs may refer to text arrow symbols on the road, which are printed on the road lane surface and are used to remind vehicles to drive, instruct or restrict vehicle driving. They may include printed objects of information such as maximum speed, large vehicles, small vehicles, and overtaking lanes used to indicate the direction of vehicle travel. The main information includes ID, geometry, attributes, and relationships with other objects.

[0129] The geometric data of the text arrow symbol is expressed through modeling, including the text arrow symbol representative point data and geometric shape data, etc. Among them, the representative point data is expressed by three-dimensional points, and the geometric shape data is expressed by the length and width of the circumscribed rectangle of the text arrow symbol.

[0130] The attribute data of text arrow symbols can be expressed through a tree-like multi-granularity semantic model.

[0131] In a specific application scenario, the identification type data may specifically include:

[0132] One or more of unknown type, other type, text type, arrow type, and symbol type.

[0133] The arrow types may include other arrows, straight arrows, right turn arrows, straight or right turn arrows, left turn arrows, straight or left turn arrows, left turn or right turn arrows, straight or left turn or right turn arrows, U-turn arrows, straight or U-turn arrows, left turn or U-turn arrows, etc. The right turn arrows may also include other right turn arrows, ordinary right turn arrows and small angle right turn arrows, etc. The left turn arrows may also include other left turn arrows, ordinary left turn arrows and small angle left turn arrows, etc.

[0134] Symbol types may include other symbols, lane type symbols, deceleration symbols, etc., while lane type symbols may also include other lane type symbols, bicycle lane symbols, bus lane symbols, and multi-occupant marking symbols, etc., and deceleration symbols may also include other deceleration marking symbols, diamond-shaped deceleration marking symbols, and triangular deceleration marking symbols, etc.

[0135] In addition, it should be noted that there is a many-to-one relationship between text arrow symbols and lanes and roads, and a lane or road may contain one or more text arrow symbols.

[0136] As an application embodiment, the method may specifically include:

[0137] When the original data is data containing information of traffic signs,

[0138] The geometric data includes one or more of position data and shape data;

[0139] The attribute data includes one or more of sign type data and color data.

[0140] In the embodiments of this specification, a traffic sign is a physical entity erected on the side of a road or above a road to provide information to road users.

[0141] The geometry of traffic signs is expressed through modeling, which can include the coordinates of the representative point of the sign, the size of the sign, the posture of the sign, the shape and heading of the sign. Among them, traffic signs can be represented by the position of the representative point, including the quality factor of the existence of the sign, the size of the sign is expressed by the length and width, the posture of the sign is expressed by the pitch angle / yaw angle / roll angle, and the shape of the sign is recorded according to the actual shape on site. Each attribute contains a quality factor of the attribute accuracy. In specific application scenarios, for different sensor feedback information, the expression may not be complete and is recorded based on the sensor feedback information. These geometric attributes are optional and it is not mandatory to reflect all geometric attributes.

[0142] The representative point position can indicate the existence of the traffic sign and can be expressed by a three-dimensional point.

[0143] The shape attribute of the sign records the geometric shape type of the lane traffic sign. All attributes include attribute precision quality factors. Among them, the shape type can specifically include unknown shape type, irregular shape type, rectangle, triangle, circle, etc., which are not specifically limited here.

[0144] The attribute of the color data of the traffic sign records the color of the traffic sign. The color types may specifically include white, yellow, red, blue, green, etc., which are not specifically limited here.

[0145] As an application embodiment, the method may specifically include:

[0146] When the original data is data containing information of a rod-shaped object,

[0147] The geometric data includes one or more of position data and shape data;

[0148] The attribute data includes rod type data.

[0149] In the embodiments of the present specification, the rod-shaped object may be a rod extending in the vertical direction (excluding rods constituting guardrails) or a vertical rod (excluding horizontal rods), which is expressed by geometric data, attribute data, and relational data.

[0150] The geometric data of the rod-shaped object is expressed through modeling, which can include the coordinates of the rod-shaped object's representative points and the rod-shaped object's dimensions. The rod-shaped object's existence is expressed through its representative points. At the same time, its accuracy index can also be expressed through the quality factor.

[0151] The rod type data may specifically include tree rods, contour rods, other rods, etc., which are not specifically limited here.

[0152] In addition, for the relationship data of the rods, there is a many-to-one affiliation between the rods and lanes and roads, and a lane or road may contain one or more rods.

[0153] As an application embodiment, the method may specifically include:

[0154] When the original data is data containing information about traffic lights,

[0155] The geometric data includes one or more of position data and shape data;

[0156] The attribute data includes traffic light type data.

[0157] In the embodiments of this specification, a traffic light may refer to a signal device installed at an intersection, a crosswalk, or other locations to control the passage of vehicles.

[0158] The geometric data of traffic lights are expressed through modeling, which can include the coordinates of the traffic light representative point, the size, posture, shape and heading of the traffic light. The traffic light is represented by the position of the representative point, and the traffic light representative point is expressed by a three-dimensional point; the size of the traffic light is expressed by the length and width of the circumscribed rectangle, and the posture is expressed by the pitch angle / yaw angle / Euler angle. The shape of the traffic light is all rectangular by default.

[0159] Traffic light type data may specifically include intersection traffic flow control lights, toll booth lights, lane status lights, pedestrian lights, non-motor vehicle lights, etc., without specific limitation here.

[0160] In addition, for the relational data of traffic lights, traffic lights need to be expressed in association with the lanes they control.

[0161] As an application embodiment, the method may specifically include:

[0162] When the original data is data containing information about obstacles above,

[0163] The geometric data includes one or more of representative point data and shape data;

[0164] The attribute data includes upper obstacle type data.

[0165] In the embodiments of this specification, the upper obstacle may specifically refer to the side of an obstacle located above the road, such as an overpass, a tunnel entrance, a pedestrian bridge, a tram bridge, a gantry and other obstacle surfaces across the road, etc., which are not specifically limited here.

[0166] The geometric data of the upper obstacles are expressed through modeling, which may include the coordinates, size, position, shape and heading of the representative points.

[0167] The upper obstacle type data may specifically include other obstacle types, height limit poles, overpasses, tunnels, etc., which are not specifically limited here.

[0168] In addition, for the relationship data of the upper obstacles, the upper obstacles need to be expressed in association with the controlled roads.

[0169] As an application embodiment, the method may specifically include:

[0170] When the original data is data containing information of a stop line,

[0171] The geometric data includes one or more of spatial position data and shape data;

[0172] The attribute data includes one or more of stop line type data and stop line color data.

[0173] In the embodiment of the present specification, the stop line may specifically be a parking position line indicating a vehicle waiting to be released.

[0174] The geometry of the stop line is expressed through modeling, which may include the coordinates, size, and posture of the stop line representative point. The stop line can be represented by the position of the representative point, which may include the existence quality factor of the stop line. The size of the stop line can be expressed by length and width, and the posture can be expressed by pitch angle, yaw angle, or roll angle.

[0175] The stop line type data may specifically include a stop line (single solid line), a yield line (double solid line), a slow-down yield line (double dashed line), a virtual stop line (derived from a crosswalk), etc., and is not specifically limited here.

[0176] The stop line color data is used to record the color of the stop line, which may include white, yellow, red, etc., and is not limited here.

[0177] The relationship data of the stop line may include lane information associated with the stop line, and specifically may include a regular lane and a diversion lane.

[0178] As an application embodiment, the method may specifically include:

[0179] When the original data is data containing information about speed bumps,

[0180] The geometric data includes one or more of spatial position data and shape data.

[0181] In the embodiments of this specification, the speed bump may specifically be a traffic facility installed on a road to slow down passing vehicles.

[0182] The geometric data of the speed bump is expressed through modeling, which may include the coordinates, size, and posture of the speed bump representative point. The speed bump is represented by the position of the representative point, which may include the quality factor of the speed bump's existence. The size of the stop line may be expressed by length and width, and the posture may be expressed by pitch angle, yaw angle, or roll angle, etc., which are not specifically limited here.

[0183] The relationship data of the speed bump may include an association relationship with the road and the lane.

[0184] As an application embodiment, the method may specifically include:

[0185] When the original data is data containing information of a specific area of ​​a road surface,

[0186] The geometric data includes one or more of representative point data and shape data;

[0187] The attribute data includes road surface specific area type data.

[0188] In the embodiments of this specification, the specific road surface area may specifically refer to an area of ​​the road surface set for special purposes, which may include crosswalks, bicycle lanes at intersections, warning areas, diversion areas, etc., which are not limited here.

[0189] The geometric data of a specific area of ​​the road surface can be expressed through modeling, and may include coordinates, size, position, shape, and heading of representative points, etc., which are not specifically limited here.

[0190] The road surface specific area type data may specifically include pedestrian crossings, bicycle lanes at intersections, warning areas, diversion areas, etc., which are not specifically limited here.

[0191] The relationship data of a specific area of ​​the road surface may include an association with the controlled road.

[0192] An embodiment of the present specification provides a method for generating vector semantic data in a set format, which obtains raw data collected by a sensor, analyzes the raw data, and obtains dimensional data of each dimension in high-precision map data in a set format, so as to generate high-precision map data in a set format based on the dimensional data of each dimension.

[0193] Alternatively, obtain first format data generated based on raw data collected by the sensor, process dimensional data of each dimension in the first format data, obtain dimensional data of each dimension in high-precision map data of a set format, and generate high-precision map data of a set format based on the dimensional data of each dimension.

[0194] In this way, by analyzing and processing the original data or the first format data generated from the original data, dimensional data of each dimension in the set format is obtained, thereby realizing a unified conversion of the multi-source sensor data format, and being able to effectively eliminate the expression differences between multi-source heterogeneous data, thus laying the foundation for the processing, fusion, and map updating of complex heterogeneous multi-source sensor semantic data.

[0195] Figure 5 A flowchart of a method for generating vector semantic data in a set format provided in an embodiment of this specification.

[0196] like Figure 5 As shown, the process may include the following steps:

[0197] Step 502: Define a tree structure for the multi-granularity semantic attributes of the acquired multi-source heterogeneous data.

[0198] Step 504: Perform semantic analysis on multi-source heterogeneous data.

[0199] Step 506: Generate multi-granularity semantic data.

[0200] Step 508: Generate vector semantic data in a set format based on the multi-granularity semantic data.

[0201] It should be understood that the order of some steps in the methods described in one or more embodiments of this specification can be interchanged according to actual needs, or some steps can be omitted or deleted.

[0202] Figure 5The method in the invention defines the multi-granularity semantic attributes of the acquired multi-source heterogeneous data in a tree structure, performs semantic analysis on the multi-source heterogeneous data, generates multi-granularity semantic data in a unified format, and thus generates vector semantic data in a set format.

[0203] based on Figure 5 The present specification also provides some specific implementation plans of the method, which are described below.

[0204] With respect to the above step 502, for various types of existing ground features, there are attributes of semantic expressions of different granularities, and the tree structure is defined, which may include the following steps:

[0205] Analyze the value range of multi-granularity semantic attributes;

[0206] A tree structure model of multi-granularity semantic attributes is defined.

[0207] For objects required for high-precision map updates, you can refer to the national standard definitions related to roads and analyze their semantic attributes.

[0208] Analyze the value range of multi-granularity semantic attributes, which may include:

[0209] The first step is to list the selected semantic attributes and relatively complete value range contents;

[0210] The second step is to determine whether there are semantic expressions of different granularities;

[0211] The judgment principle is that if the description granularity of the semantic attribute value can be abstracted into a coarser granularity, or refined into a finer granularity, and the semantic expression can be expressed correctly, but the degree of detail of the transmitted semantic information is different, that is, there are semantic expressions of different granularities.

[0212] Taking the shape attributes of road traffic signs as an example, we can refer to the national standard for road traffic signs. The shape attributes of road traffic signs generally include: regular triangle, circle, inverted triangle, octagon, square, rhombus and irregular shapes, etc.

[0213] Among them, "equilateral triangle" and "inverted triangle" can be abstracted into "triangle", and "equilateral triangle" and "inverted triangle" are both "triangles". There is no error in the semantic expression, but the level of detail of the semantic information is inconsistent. "Square" can be split into "rectangle" and "square" according to reality, and can also be abstracted into "quadrilateral". Therefore, the shape attributes of road traffic signs have semantic expressions of different granularities.

[0214] Taking the color attributes of road traffic signs as an example, we can refer to the national standard for road traffic signs. The color attributes of road traffic signs generally include: red, yellow, blue, green, brown, black, white, orange, etc.

[0215] Since no color can be combined into another color and still correctly express the semantics, there is no semantic expression of different granularities.

[0216] Defining the tree structure model of multi-granularity semantic attributes is to construct a tree model. It can be understood that a tree is a finite set consisting of n elements, where each element is called a node. In the tree model, each node is an expression of a value domain content.

[0217] In each node, there is a specific root node or root. Each root node is the first-level abstraction of its leaf nodes, and has a containment and being contained relationship with the leaf nodes. For example, the root node of "equilateral triangle" is "triangle".

[0218] Except for the root node, the remaining nodes are divided into m (m ≥ 0) mutually disjoint finite sets, each of which is a tree (called a subtree of the original tree). For example, under the root node "quadrilateral" there are two leaf nodes, "diamond" and "rectangle", and these two leaf nodes do not intersect with each other.

[0219] Specifically, Figure 6 As shown, the definition and construction of the tree structure includes the following steps:

[0220] Step 601: Select semantic attribute data;

[0221] Step 603: determining the original attribute value of the selected semantic attribute data;

[0222] Step 605: Determine whether each original attribute value can be split;

[0223] Step 607: If splitting is possible, splitting the original attribute value;

[0224] Step 609: Obtain the split attribute set;

[0225] Step 611: If the original attribute values ​​cannot be split, directly proceed to step 609;

[0226] Step 613: Repeat the above steps 305 to 309 for the split attribute set.

[0227] After selecting the semantic attribute data, the attribute values ​​contained therein are analyzed one by one according to the complete value domain of the selected semantic attributes with different granularities to determine whether they can be split into multiple non-overlapping finite sets, and the value domain that can be split twice is split to generate multiple two-layer tree structures.

[0228] The analysis and splitting steps are performed again for the split set until each value range cannot be split twice, and finally multiple multi-layer tree structures are generated.

[0229] Furthermore, the root nodes of multiple multi-layer trees are analyzed respectively to determine whether they can be abstracted into the same semantic expression again. If they can be abstracted into the same semantic expression, a new root node is generated, thereby generating a larger multi-layer tree structure.

[0230] When all root nodes cannot be abstracted any further, the attribute category is used as the largest and final abstract root node, and other nodes are its leaf nodes.

[0231] Finally, each root node and leaf node of the entire tree structure are numbered separately.

[0232] Next, taking the shape attributes of traffic signs as an example, a tree structure is constructed.

[0233] The first step is to split the shape attribute values ​​of traffic signs.

[0234] Equilateral triangle → cannot be split

[0235] Circular → Indivisible

[0236] Inverted triangle → cannot be split

[0237] Octagon → can be divided into: regular octagon, irregular octagon

[0238] Square → can be divided into: rectangle, square

[0239] Rhombus → Indivisible

[0240] Irregular shape → cannot be split

[0241] The resulting tree structure is Figure 7 shown.

[0242] The second step is to split the secondary attribute values.

[0243] Regular octagon → cannot be split

[0244] Non-regular octagon → cannot be split

[0245] Rectangle → Cannot be split

[0246] Square → cannot be split

[0247] The third step is the abstract merging of root nodes.

[0248] Regular octagon + inverted triangle → can be combined: triangle

[0249] Square + Rhombus → Can be combined: Quadrilateral

[0250] Other non-mergeable

[0251] The resulting tree structure is Figure 8 shown.

[0252] Step 4: Tree merging.

[0253] This attribute is the shape of the traffic sign, so the shape can be used as the root node to finally complete the definition of the complete tree structure. The resulting tree structure is as follows Fig. 9 shown.

[0254] Step 5: Number each root node and leaf node of the entire tree structure.

[0255] Specific number such as Fig.10 shown.

[0256] Based on the same idea, the embodiments of this specification also provide a device corresponding to the above method. Fig.11 The embodiments of this specification provide corresponding to Figure 2 A schematic diagram of the structure of a device for generating vector semantic data in a set format. Fig.11 As shown, the device may include:

[0257] An acquiring unit 1101 acquires vector semantic data in a first format generated based on raw data collected by a sensor;

[0258] The data processing unit 1102 processes the dimension data of each dimension in the vector semantic data in the first format to obtain the dimension data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimension information and attribute data for describing attribute dimension information;

[0259] The vector semantic data generating unit 1103 generates the vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format.

[0260] based on Fig.11 The present specification also provides some specific implementation schemes of the device, which are described below.

[0261] Optionally, the dimensional data of the high-precision map data in the set format also includes: relationship data.

[0262] Optionally, the method specifically includes:

[0263] When the original data is data containing information of road reference lines,

[0264] The geometric data includes one or more of spatial position data and shape data;

[0265] The attribute data includes one or more of travel direction data and vehicle speed data.

[0266] Optionally, the method specifically includes:

[0267] When the original data is data containing information about lane reference lines,

[0268] The geometric data includes one or more of spatial position data and shape data;

[0269] The attribute data includes one or more of travel direction data, lane type data and vehicle speed data.

[0270] Optionally, the method specifically includes:

[0271] When the original data is data containing information of a boundary reference line,

[0272] The attribute data includes one or more of boundary role data, boundary crossability data, and line marking quantity data.

[0273] Optionally, the method specifically includes:

[0274] When the original data is data containing information about lane markings,

[0275] The attribute data includes one or more of marking type data, marking color data, marking width data, and marking material data.

[0276] Optionally, the data type of the marking line type data specifically includes:

[0277] One or more of unknown type, other type, solid line type, and dashed line type.

[0278] Optionally, the method specifically includes:

[0279] When the original data is data containing information about roadside obstacles,

[0280] The geometric data includes one or more of representative line data and geometric shape data;

[0281] The attribute data includes obstacle type data.

[0282] Optionally, the data type of the obstacle type data specifically includes:

[0283] One or more of Unknown Type, Other Type, Curb Type, Guardrail Type, and Wall Type.

[0284] Optionally, the method specifically includes:

[0285] When the original data is data containing information of traffic signs,

[0286] The geometric data includes one or more of representative point data and geometric shape data;

[0287] The attribute data includes one or more of identification type data and identification content data.

[0288] Optionally, the identification type data specifically includes:

[0289] One or more of unknown type, other type, text type, arrow type, and symbol type.

[0290] Optionally, the method specifically includes:

[0291] When the original data is data containing information of traffic signs,

[0292] The geometric data includes one or more of position data and shape data;

[0293] The attribute data includes one or more of sign type data and color data.

[0294] Based on the same idea, the embodiments of this specification also provide a device corresponding to the above method.

[0295] Fig.12 The embodiments of this specification provide corresponding to Figure 2 A schematic diagram of the structure of a device for generating vector semantic data in a set format. Fig.12 As shown, the device 1200 may include:

[0296] at least one processor 1210; and,

[0297] A memory 1230 is communicatively connected to the at least one processor; wherein,

[0298] The memory 1230 stores instructions 1220 executable by the at least one processor 1210. The instructions are executed by the at least one processor 1210 to enable the at least one processor 1210 to:

[0299] Acquire vector semantic data in a first format generated based on raw data collected by the sensor;

[0300] Processing the dimensional data of each dimension in the vector semantic data in the first format to obtain the dimensional data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information;

[0301] The vector semantic data in the second format is generated based on the data of each dimension in the vector semantic data in the second format.

[0302] Based on the same idea, the embodiment of this specification also provides a computer-readable medium corresponding to the above method. The computer-readable medium stores computer-readable instructions, which can be executed by a processor to implement the following method:

[0303] Acquire vector semantic data in a first format generated based on raw data collected by the sensor;

[0304] Processing the dimensional data of each dimension in the vector semantic data in the first format to obtain the dimensional data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information;

[0305] The vector semantic data in the second format is generated based on the data of each dimension in the vector semantic data in the second format.

[0306] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. Figure 3 As for the electronic voucher sending device shown, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0307] In the 1990s, it was very clear whether the improvement of a technology was hardware improvement (for example, improvement of the circuit structure of diodes, transistors, switches, etc.) or software improvement (improvement of the method flow). However, with the development of technology, many improvements of the method flow today can be regarded as direct improvements of the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that the improvement of a method flow cannot be implemented with a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming themselves, without having to ask chip manufacturers to design and make dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one kind of HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0308] The controller may be implemented in any suitable manner, for example, the controller may take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (e.g., software or firmware) executable by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, and the memory controller may also be implemented as part of the control logic of the memory. It is also known to those skilled in the art that, in addition to implementing the controller in a purely computer-readable program code manner, the controller may be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, such a controller may be considered as a hardware component, and the devices for implementing various functions included therein may also be considered as structures within the hardware component. Or even, the devices for implementing various functions may be considered as both software modules for implementing the method and structures within the hardware component.

[0309] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0310] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0311] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0312] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0313] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0314] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0315] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0316] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0317] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0318] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0319] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0320] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0321] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for generating vector semantic data in a set format, characterized in that: The method comprises: Acquire vector semantic data in a first format generated based on raw data collected by the sensor; Processing the dimensional data of each dimension in the vector semantic data in the first format to obtain the dimensional data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information; Generate the vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format; The method further comprises: Aligning the vector semantic data in the second format with each element in the high-precision map data to obtain aligned data, wherein the aligned data has a unique corresponding map element; Aggregating the aligned data of the same map element to obtain aggregated data, wherein the aggregated data of the same map element includes dimension data of each dimension of the same map element, and for the same dimension, the aggregated data includes the finest-grained dimension data; Based on the aggregated data, the changed map elements are determined and the high-precision map data corresponding to the changed map elements are updated.

2. The method according to claim 1, characterized in that The method specifically comprises: When the original data is data containing information of road reference lines, The geometric data includes one or more of spatial position data and shape data; The attribute data includes one or more of travel direction data and vehicle speed data.

3. The method according to claim 1, characterized in that The method specifically comprises: When the original data is data containing information about lane reference lines, The geometric data includes one or more of spatial position data and shape data; The attribute data includes one or more of travel direction data, lane type data and vehicle speed data.

4. The method according to claim 1, characterized in that: The method specifically comprises: When the original data is data containing information of a boundary reference line, The attribute data includes one or more of boundary role data, boundary crossability data, and line marking quantity data.

5. The method according to claim 1, characterized in that The method specifically comprises: When the original data is data containing information about lane markings, The attribute data includes one or more of marking type data, marking color data, marking width data, and marking material data. The data types of the marking type data specifically include: One or more of unknown type, other type, solid line type, and dashed line type.

6. The method according to claim 1, characterized in that The method specifically comprises: When the original data is data containing information about roadside obstacles, The geometric data includes one or more of representative line data and geometric shape data; The attribute data includes obstacle type data, and the data type of the obstacle type data specifically includes: One or more of Unknown Type, Other Type, Curb Type, Guardrail Type, and Wall Type.

7. The method according to claim 1, characterized in that The method specifically comprises: When the original data is data containing information of traffic signs, The geometric data includes one or more of representative point data and geometric shape data; The attribute data includes one or more of identification type data and identification content data, and the identification type data specifically includes: One or more of unknown type, other type, text type, arrow type, and symbol type.

8. The method according to claim 1, characterized in that The method specifically comprises: When the original data is data containing information of traffic signs, The geometric data includes one or more of position data and shape data; The attribute data includes one or more of sign type data and color data.

9. The method according to claim 1, characterized in that: The dimensional data of the high-precision map data in the set format also includes: relationship data.

10. A device for generating vector semantic data in a set format, characterized in that: The device comprises: An acquisition unit, which acquires vector semantic data in a first format generated based on raw data collected by the sensor; A data processing unit processes the dimension data of each dimension in the vector semantic data in the first format to obtain the dimension data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimension information and attribute data for describing attribute dimension information; a vector semantic data generating unit, which generates the vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format; The device is also used to: Aligning the vector semantic data in the second format with each element in the high-precision map data to obtain aligned data, wherein the aligned data has a unique corresponding map element; Aggregating the aligned data of the same map element to obtain aggregated data, wherein the aggregated data of the same map element includes dimension data of each dimension of the same map element, and for the same dimension, the aggregated data includes the finest-grained dimension data; Based on the aggregated data, the changed map elements are determined and the high-precision map data corresponding to the changed map elements are updated.

11. A device for generating vector semantic data in a set format, comprising: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to: Acquire vector semantic data in a first format generated based on raw data collected by the sensor; Processing the dimensional data of each dimension in the vector semantic data in the first format to obtain the dimensional data of each dimension in the vector semantic data in the second format, wherein the vector semantic data in the set format at least includes geometric data for describing geometric dimensional information and attribute data for describing attribute dimensional information; generating the vector semantic data in the second format based on the data of each dimension in the vector semantic data in the second format; and Aligning the vector semantic data in the second format with each element in the high-precision map data to obtain aligned data, wherein the aligned data has a unique corresponding map element; Aggregating the aligned data of the same map element to obtain aggregated data, wherein the aggregated data of the same map element includes dimension data of each dimension of the same map element, and for the same dimension, the aggregated data includes the finest-grained dimension data; Based on the aggregated data, the changed map elements are determined and the high-precision map data corresponding to the changed map elements are updated.

12. A computer-readable medium having computer-readable instructions stored thereon, wherein the computer-readable instructions can be executed by a processor to implement the method for generating high-precision map data as described in any one of claims 1 to 9.

Images