Data mining method and device, equipment and medium

Abstract

The invention discloses a data mining method which comprises the following steps: acquiring multiple pieces of map data, establishing map topology based on the multiple pieces of map data, and determining trajectory data of a target object in the multiple pieces of map data; extracting behaviors of the target object based on the map topology and the trajectory data of the target object; generating a behavior tag for the behavior of the target object; screening the behavior tags according to a preset screening rule to obtain candidate tags; combining the candidate tags according to a preset combination rule to obtain a combined tag; according to the method, through the steps of label screening, label combination and the like, high-value labels required by users are mined from real data, so that the characteristics have a clear coupling relationship, the diversified requirements of different users on scenes can be flexibly met, and the scene recognition precision is remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of data mining technology, and in particular to a data mining method, apparatus, device and medium. Background Art

[0002] Intelligent driving data mining refers to the use of data mining techniques to analyze and process the massive amounts of data generated during vehicle operation during the development and application of intelligent driving technology, in order to optimize the performance of autonomous driving systems. In this data mining process, related methods often use simple fixed rule matching or statistical analysis methods to filter out target data. The coupling relationships between these data are unclear, making it difficult to describe complex scenarios and unable to flexibly adapt to the diverse needs of different users. Summary of the Invention

[0003] In view of the above-mentioned shortcomings of the prior art, the present application provides a data mining method, apparatus, device and medium to solve at least one defect in the prior art.

[0004] To achieve the above and other purposes, the present application provides a data mining method, which includes: Acquire a plurality of map data, establish a map topology based on the plurality of map data, and determine trajectory data of a target object in the plurality of map data; extracting the behavior of the target object based on the map topology and the trajectory data of the target object; generating a behavior label for the behavior of the target object; Filter the behavior tags according to preset filtering rules to obtain candidate tags; The candidate labels are combined according to a preset combination rule to obtain a combined label.

[0005] In one embodiment of the present application, establishing a map topology based on a plurality of map data includes: Acquire a plurality of first map data, and convert the formats of the plurality of first map data to obtain a plurality of second map data; wherein the plurality of second map data have the same format; Extracting elements from the plurality of second map data to obtain map elements; the map elements include lane lines, traffic signs, and curbs; A topological connection relationship of lanes, roads and intersections is established based on the map elements to obtain a map topology.

[0006] In one embodiment of the present application, the method further includes: a map optimization step, wherein the map optimization step includes at least one of the following: converting the plurality of second map data from a coordinate system corresponding to the second map data to a target coordinate system; Converting the semantic labels in the plurality of second map data into target semantic labels; Downsampling is performed on the map elements corresponding to the plurality of second map data.

[0007] In one embodiment of the present application, the method further includes: a trajectory optimization step, wherein the trajectory optimization step includes: Extracting trajectory data of the target object within the time window, wherein the trajectory data includes a path and a speed of the target object; The trajectory data is smoothed to obtain trajectory optimized data of the target object.

[0008] In one embodiment of the present application, extracting the behavior of the target object based on the map topology and the trajectory data of the target object includes: Calculating the behavior characteristics of the target object according to the map topology and the trajectory data of the target object; The behavior of the target object is obtained based on the behavior characteristics.

[0009] In one embodiment of the present application, the preset screening criteria include at least one of the following: Lighting conditions, weather conditions, time period, visibility, road type, traffic flow status, and target object behavior.

[0010] In one embodiment of the present application, the preset combination rule includes at least one of the following: Logical combination, indicating a tag combination determined according to a logical relationship between behavior tags, wherein the logical relationship includes at least one of the following: AND, OR, NOT; Sequence combination, which means the label combination determined by the time sequence of the behavior labels; Weight distribution, which indicates the combination of tags that determines the importance of the behavior tags. Different behavior tags have different weights. Priority combination is used to represent a label combination determined by priority. Different behavior labels have different priorities.

[0011] To achieve the above-mentioned and other related purposes, the present application provides a data mining device, comprising: a data acquisition module, configured to acquire a plurality of map data, establish a map topology based on the plurality of map data, and determine trajectory data of a target object in the plurality of map data; a behavior extraction module, configured to extract the behavior of the target object based on the map topology and the trajectory data of the target object; A label generation module, configured to generate a behavior label for the behavior of the target object; A label screening module is used to screen the behavior labels according to preset screening rules to obtain candidate labels; The label combination module is used to combine the candidate labels according to a preset combination rule to obtain a combined label.

[0012] To achieve the above objectives and other related objectives, the present application provides a data mining device, comprising: one or more processors; and The memory is used to store one or more programs. When the one or more programs are executed by the one or more processors, the memory implements the data mining method.

[0013] To achieve the above objectives and other related objectives, the present application provides one or more machine-readable media having instructions stored thereon, which, when executed by one or more processors, enable the processors to perform the data mining method.

[0014] Beneficial effects of this application: A data mining method of the present application includes: acquiring multiple map data, and establishing a map topology based on the multiple map data and determining trajectory data of a target object in the multiple map data; extracting the behavior of the target object based on the map topology and the trajectory data of the target object; generating a behavior label for the behavior of the target object; screening the behavior label according to preset screening rules to obtain candidate labels; combining the candidate labels according to preset combination rules to obtain a combined label; the present application mines high-value labels required by users from real data through steps such as label screening and label combination, so that there is a clear coupling relationship between each feature, which can flexibly adapt to the diverse needs of different users for scenes and significantly improve the scene recognition accuracy.

[0015] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings: Figure 1 is a flow chart of a data mining method according to an embodiment of the present application; Figure 2 A flowchart of a method for establishing a map topology according to an embodiment of the present application; Figure 3This is a principle block diagram of a data mining device according to an embodiment of the present application; Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the memory of an embodiment of the present application is shown. DETAILED DESCRIPTION

[0017] The following describes the embodiments of the present application through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present application from the content disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.

[0018] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application. Therefore, the illustrations only show components related to the present application and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0019] Although the terms "first," "second," "A," and "B," etc. may be used herein to describe various elements, these elements should not be limited by these terms and are merely used to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element without departing from the scope of the technology described below. The term "and / or" includes a combination of a plurality of related items or any of the plurality of related items.

[0020] As used herein, unless the context indicates otherwise, the singular form is intended to include the plural form, and it will be understood that the term "comprising" means the presence of stated features, quantities, steps, operations, elements, or combinations thereof, but does not preclude the presence or addition of one or more other features, quantities, steps, operations, elements, components, or combinations thereof.

[0021] Before describing the components in detail, it is intended to clarify that the components in this specification are divided only by the primary function of each component. That is, two or more components described below may be combined into one component, or may be divided into two or more components based on more detailed functions. In addition to the primary function of the component, each component described below may also perform some or all of the functions of other components, and some of the primary functions of each component may be exclusively performed by other components.

[0022] The embodiments of the present application respectively provide a data mining method, a data mining apparatus, a data mining device, and a computer-readable storage medium, and these embodiments will be described in detail below.

[0023] See also Figure 1 , Figure 1 This is a flow chart of a data mining method according to an embodiment of the present application. Figure 1 As shown, the data mining method includes steps S110 to S150: Step S110, acquiring a plurality of map data, establishing a map topology based on the plurality of map data, and determining trajectory data of a target object in the plurality of map data; Map data refers to road information from different suppliers or sensors, which can be achieved through lidar point clouds, camera images, and high-precision map vector data. Map topology is a structured model that describes the connection relationship of the road network. It can be achieved by extracting map elements such as lane lines, traffic signs, and curbs to establish a lane-road-intersection hierarchical relationship. The purpose of establishing map topology is to convert the original map data into a standardized format that can be processed by algorithms. Trajectory data refers to the movement path information of the target object on the road. It is extracted from the map data and can include trajectory point sequences, inertial navigation data, visual odometry data, etc.

[0024] Step S120 , extracting the behavior of the target object based on the map topology and the trajectory data of the target object; The target object's behavior is a driving behavior segment that is crucial for optimizing autonomous driving algorithms (e.g., dangerous lane changes, slowing down in traffic jams, etc.). Target object behavior can be analyzed in a scenario-specific manner using classification algorithms based on map topology and the target object's trajectory data. For example, combining map topology and trajectory data to distinguish between stopping and waiting and slowing down in traffic jams can be used.

[0025] Step S130, generating a behavior label for the behavior of the target object; The behavior label is a semantic description of the target object's driving action. The extracted target object's behavior will be matched with the corresponding behavior label and stored in the database. The database will also store some other label types, such as weather, time, lighting, visibility, road shape, etc.

[0026] Step S140, filtering the behavior tags according to preset filtering rules to obtain candidate tags; Preset filtering rules: These are established criteria or conditions that are used to determine which tags meet the requirements. Filtering behavior tags based on preset filtering rules involves filtering and selecting certain behavior tags according to pre-set rules, ultimately retaining the tags that meet the criteria as candidate tags for further analysis, modeling, or other purposes.

[0027] Step S150: combining candidate tags according to a preset combination rule to obtain a combined tag.

[0028] Combination rules refer to the logical associations between multiple tags, and can include strategies such as logical operations, time series matching, and weight assignment. Combining candidate tags according to preset combination rules involves matching or combining a set of pre-defined "candidate tags" according to the combination rules to generate a new "combined tag." For example, certain tags can only be combined with tags of specific types, or the combined tag must meet certain grammatical or semantic requirements. By combining tags, complex information can be described more efficiently and comprehensively, improving retrieval and management efficiency.

[0029] The embodiments of the present application generate composite scene labels based on a dynamic label combination mechanism to enhance the ability to describe continuous driving actions and complex environmental conditions; by combining screening criteria with combination rules, high-value scene data sets can be quickly generated, thereby improving the efficiency of autonomous driving algorithm training.

[0030] See also Figure 2 , Figure 2 This is a flow chart of establishing a map topology based on multiple map data according to an embodiment of the present application. Figure 2 In , a map topology is established based on multiple map data, including; Step S210: Acquire a plurality of first map data and convert the format of the plurality of first map data to obtain a plurality of second map data; wherein the plurality of second map data have the same format; Generally speaking, raw map data—also known as primary map data—usually uses industry-standard formats (such as OpenDrive and NDS). However, data structures vary between vendors. Therefore, format conversion is necessary to unify map data from different data sources into a standardized data format (such as Shapefile) to generate multiple secondary map data sets for subsequent processing.

[0031] In the implementation of this application, by converting the format of original map data from multiple sources into a map dataset with a unified coordinate system and data structure, data standardization eliminates compatibility issues between map data formats of different manufacturers and reduces computing resource consumption.

[0032] Step S220: extracting elements from the plurality of second map data to obtain map elements; the map elements include lane lines, traffic signs, and curbs; The converted map data, i.e., the second map data, contains a variety of elements (such as lane lines, traffic signs, curbs, etc.), which need to be extracted one by one according to the preset element list to form a structured feature set containing geometric attributes and semantic labels.

[0033] Step S230 : establishing a topological connection relationship of lanes, roads, and intersections based on the map elements to obtain a map topology.

[0034] Among them, the map topology includes: ① Lane connectivity (such as lane merging and bifurcation): By establishing lane connectivity, we can accurately understand the vehicle's driving path under complex road conditions and improve the accuracy of scenario analysis.

[0035] ②Road hierarchy structure (such as main roads and secondary roads); by establishing a road hierarchy structure, different types of roads can be distinguished, thereby more accurately identifying traffic scenarios.

[0036] ③ Traffic rule association (such as binding traffic lights to lanes); by considering traffic control factors, the completeness and accuracy of scene understanding are improved.

[0037] The relationship between lanes and roads, and the relationship between roads and intersections are established through the extracted map elements to form a lane-road-intersection topological connection relationship, providing rich contextual information for subsequent high-value label extraction.

[0038] In the embodiments of the present application, data standardization preprocessing is achieved through format conversion, combined with structured extraction of key map elements such as lane lines and traffic signs, which can improve the accuracy of map topology.

[0039] In one embodiment, the method further includes: a map optimization step, wherein the map optimization step includes at least one of the following: converting the plurality of second map data from a coordinate system corresponding to the second map data to a target coordinate system; Converting the semantic labels in the plurality of second map data into target semantic labels; Downsampling is performed on the map elements corresponding to the plurality of second map data.

[0040] Converting multiple secondary map data from their corresponding coordinate systems to a target coordinate system involves converting the original map data from different sources into a unified relative coordinate system to eliminate differences in spatial datums between the different data sources. For example, converting map data from an original coordinate system (e.g., WGS84) to a local coordinate system (e.g., UTM or vehicle coordinate system) can be accomplished using a coordinate transformation matrix.

[0041] Among them, converting the semantic labels in multiple second map data into target semantic labels means uniformly mapping the semantic labels in different maps (such as lane type, traffic light status) to internally defined enumeration values, that is, completing the process of mapping the annotation information in heterogeneous data to a unified semantic system to solve the problem of inconsistent semantic descriptions of different data sources and establish semantic consistency across data sources. For example, the lane types in OpenDrive (such as "driving" and "parking") are mapped to internally defined enumeration values ​​(such as 1 for driving lanes and 2 for parking lanes). That is, the driving lanes in all map data are represented by "1" and the parking lanes are represented by "2".

[0042] Downsampling map elements corresponding to multiple pieces of second map data involves downsampling redundant points in the second map data to reduce duplicate or unnecessary coordinate points in the road geometry data. For example, more feature points are retained in areas of greater road curvature, while redundant points are reduced in straight sections. This reduces data volume and computational burden while maintaining road shape accuracy, while preserving key geometric features.

[0043] In the embodiments of this application, coordinate system conversion is used to reduce resource consumption for data alignment, semantic label conversion is used to reduce the computational complexity of subsequent feature matching, and downsampling is used to reduce data processing while ensuring that key elements are not lost. Furthermore, by eliminating coordinate system deviations and semantic ambiguity, the accuracy of matching trajectory data with map topology is significantly improved, providing reliable basic data support for subsequent behavioral analysis.

[0044] In one embodiment, the method further includes: a trajectory optimization step, which includes: extracting trajectory data of the target object within a time window, the trajectory data including the path and speed of the target object; and smoothing the trajectory data to obtain trajectory optimization data of the target object.

[0045] Raw trajectory data often contains noise (e.g., positioning drift and sensor detection errors) and requires smoothing and optimization to improve the accuracy of subsequent scene analysis. Trajectory data for all target objects (e.g., the vehicle itself, surrounding vehicles, pedestrians, etc.) within a time window of 10-60 seconds is typically extracted. The length of the time window can be dynamically adjusted based on the scenario (e.g., short windows for dense traffic and long windows for sparse traffic). The target object's path can be predicted based on its starting point using graph search algorithms (e.g., A* and Dijkstra).

[0046] After acquiring the target object's path and velocity, they are smoothed. Path smoothing uses the FemPosSmooth (finite element position smoothing) algorithm to optimize the trajectory points in the original trajectory data, eliminating position jumps (such as sudden changes caused by sensor jitter) while maintaining the physical plausibility of the trajectory (such as acceleration continuity). Velocity smoothing uses the Kalman filter to filter the velocity data, reducing noise interference and improving the stability of velocity estimation. The optimized trajectory data is then used for high-value scene mining to ensure the accuracy of feature calculation.

[0047] In one embodiment, extracting the behavior of the target object based on the map topology and the trajectory data of the target object includes: calculating the behavior characteristics of the target object based on the map topology and the trajectory data of the target object; and obtaining the behavior of the target object based on the behavior characteristics.

[0048] The behavioral characteristics of the target object are calculated based on the map topology and the trajectory data of the target object. The map topology and trajectory data can be associated and modeled through the attention mechanism to generate behavioral characteristics. Scenario-based analysis is then performed based on the behavioral characteristics. For example, the topological characteristics of the intersection and the duration of the trajectory can be combined to distinguish between parking and waiting and slow driving in congestion.

[0049] Among them, behavioral features include: ① Lane affiliation: Based on the distance between the trajectory and the lane centerline, the lane and status of the target object (such as normal driving or driving on the line) are determined; ② Collision characteristics: Using the GJK (Gilbert-Johnson-Keerthi) collision detection algorithm, the collision information between the target object and other objects is calculated, and the collision time (TTC, Time-To-Collision, generally calculated by dividing the distance by the relative speed) is calculated; ③ Trajectory distance: Calculate the Euclidean distance (the distance between trajectory points) and contour distance (the shortest distance between contours) between the trajectories of two target objects; ④ Traffic speed: Statistically calculate the average speed and traffic volume of target objects belonging to the lane to identify congestion or abnormal speed distribution.

[0050] The target object's behavior includes the following: ① Lane change: Detecting lane change actions through lane attribution, calculating the number of lane changes and lateral acceleration, and combining collision characteristics to identify dangerous lane changes (such as cutting in too close); ② Slowing down in congestion: Determining the congestion level (such as low speed and high density) based on traffic speed distribution, and analyzing the cause of congestion (such as accidents, traffic lights) based on lane topology; ③ Cut-In: Quantifying the risk of cutting in using trajectory distance and Time to Count (TTC), such as aggressive cutting in (such as TTC < 2 seconds); ④ Obstacle detour: Identifying detour actions based on the target type (such as stationary vehicle, construction cone) and trajectory offset.

[0051] Of course, in real-world scenarios, the definition of target behavior, expert experience formulas, and threshold calibration must be adjusted based on actual circumstances (for example, different manufacturers and standards have different definitions of dangerous lane change distance and TTC). This requires a parameter self-learning mechanism based on historical accident data to continuously improve the calibration values ​​of the empirical formulas. Furthermore, it requires tailored parameter calibration and feature value selection based on user needs to better meet their specific requirements.

[0052] In one embodiment, the preset screening criteria include at least one of the following: lighting conditions, weather conditions, time period, visibility, road type, traffic flow status, and behavior of the target object.

[0053] During the label filtering phase, data behavior labels are filtered based on a series of predefined criteria, including lighting conditions, weather conditions, time of day, visibility, road type, traffic flow status, and target behavior. Each filtering criterion is further divided into multiple subcategories.

[0054] Lighting conditions refer to the quantitative assessment of the light intensity in the target environment. This can be achieved using on-board lighting sensor data or video image brightness analysis. This condition can be used to distinguish behaviors in different lighting environments. Lighting conditions can be categorized based on the intensity level of the light. Weather conditions refer to the identification of meteorological conditions such as precipitation and haze. This can be achieved using meteorological sensor data or image recognition technology. Weather conditions include sunny, cloudy, and rainy days. Time periods refer to the division of time into different intervals. This can be achieved using a 24-hour clock cycle. Time periods include morning, noon, evening, and midnight. Visibility refers to the quantitative parameter of the visible distance in the environment. This can be achieved using lidar echo intensity or camera image contrast analysis. Visibility is categorized based on visibility distance. Road type refers to the classification of road structure. This can be achieved using topological features such as the number of lane lines and barriers. Road types include intersections, T-junctions, straight roads, curves, and the number of lanes. Traffic flow status refers to the dynamic assessment of road conditions. Parameter calculation models such as vehicle density and average speed can be used to identify traffic congestion or smooth traffic.

[0055] In one embodiment, the preset combination rules include at least one of the following: logical combination, which represents a label combination determined by the logical relationship between behavior labels, wherein the logical relationship includes at least one of the following: and, or, not; sequence combination, which represents a label combination determined by the time sequence of the behavior labels; weight distribution, which represents a label combination determined by the importance of the behavior labels, and different behavior labels have different weights; priority combination, which is used to represent a label combination determined by priority, and different behavior labels have different priorities.

[0056] During the label combination phase, the selected behavior labels are combined according to a series of combination rules. These include logical combination (using logical AND, OR, and NOT conditions to determine the relationship between labels), sequence combination (sorting labels chronologically, such as the order of lane change followed by Cut-In and then lane change), weight allocation (assigning different weights to different behavior labels to reflect their influence in scene recognition), and priority combination (assigning different priorities based on their importance).

[0057] When combining scenarios, multiple filtering conditions can be activated at the same time, and filtering strategies that meet different needs can be formed through various combination rules.

[0058] For example, to screen high-value driving risk scenarios under certain weak perceptions, the combination rules may include: ① Use the logical combination "AND" to combine the "Light Condition (Low Light)", "Weather Condition (Rainy Day)", and "Time Period (Night)" tags to filter the data; ② Use sequence combinations to match the "target behavior (Cut-In)", "target behavior (lane change)", and "target behavior (Cut-In)" tags in chronological order to filter the data; ③ Use weight distribution to set weights from high to low for "road type (intersection area)", "visibility (low visibility)", and "traffic flow status (heavy traffic volume)", add up the weights, and filter the data according to the threshold.

[0059] ④ Prioritize the data using the priority combination in the order of "largest number (target behavior (Cut-In))" and "largest number (target behavior (Lane Change))".

[0060] Finally, the required combined label of "high-value driving risk scenario under certain weak perception" is obtained.

[0061] In the embodiments of the present application, by introducing logical combination, sequence combination, weight allocation and priority combination, the complex characteristics of long-tail scenarios can be coped with, and significant differences can be formed in data screening dimensions and combination flexibility.

[0062] This application significantly improves data processing efficiency and scene recognition accuracy through standardized map processing, multi-algorithm collaborative trajectory optimization, and intelligent behavior matching based on multi-dimensional features. It also supports multi-dimensional label screening and combination, addressing the shortcomings of traditional data mining methods in noise suppression, long-tail scene coverage, and implementation flexibility.

[0063] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0064] Figure 3 FIG. 1 is a block diagram of a data mining device according to an embodiment of the present application. Figure 3 As shown, a data mining device includes: A data acquisition module 310 is configured to acquire a plurality of map data, establish a map topology based on the plurality of map data, and determine trajectory data of a target object in the plurality of map data; A behavior extraction module 320 is used to extract the behavior of the target object based on the map topology and the trajectory data of the target object; A label generation module 330 is used to generate a behavior label for the behavior of the target object; The tag screening module 340 is used to screen the behavior tags according to preset screening rules to obtain candidate tags; The label combination module 350 is used to combine candidate labels according to preset combination rules to obtain a combined label.

[0065] It should be noted that the data mining apparatus provided in the above embodiments and the data mining method provided in the above embodiments are based on the same concept. The specific manner in which the various modules and units perform their operations has been described in detail in the method embodiments and will not be repeated here. In actual applications, the data mining apparatus provided in the above embodiments can, as needed, allocate the above functions to different functional modules, i.e., divide the internal structure of the apparatus into different functional modules to perform all or part of the functions described above. This is not a limitation herein.

[0066] An embodiment of the present application also provides a device, including: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the memory implements the data mining method in the above embodiment.

[0067] The embodiments of the present application further provide one or more machine-readable media having instructions stored thereon, which, when executed by one or more processors, enable the processors to execute the data mining method in the above embodiments.

[0068] Figure 4 FIG1 shows a schematic diagram of a computer system structure suitable for implementing a memory according to an embodiment of the present invention. It should be noted that Figure 4 The computer system of the memory shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0069] like Figure 4 As shown, computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on programs stored in read-only memory (ROM) 402 or programs loaded from storage into random access memory (RAM) 403. RAM also stores various programs and data required for system operation. CPU 401, ROM 402, and RAM 403 are interconnected via bus 404. An input / output (I / O) interface 405 is also connected to bus 404.

[0070] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, mouse, and the like; an output section 407 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 408 including devices such as a hard disk; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. Removable media 411, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 410 as needed, so that computer programs read from the media can be installed in the storage section 408 as needed.

[0071] In particular, according to an embodiment of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present invention includes a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the aforementioned data mining method. In such an embodiment, the computer program can be downloaded and installed from a network via a communication component and / or installed from removable media 411. When executed by the central processing unit (CPU) 401, the computer program performs the various functions defined in the system of the present invention.

[0072] It should be noted that the computer-readable medium described in the embodiments of the present invention may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may, for example, be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM) 403, read-only memory (ROM) 402, erasable programmable read-only memory (EPROM), flash memory, optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable computer program. This propagated data signal may take various forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0073] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the boxes can also occur in an order different from that marked in the accompanying drawings. For example, two boxes shown in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or can be implemented using a combination of dedicated hardware and computer instructions.

[0074] The units involved in the embodiments of the present invention may be implemented in software or hardware, and the units described may also be provided in a processor. In some cases, the names of these units do not limit the units themselves.

[0075] Another aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When executed by a computer processor, the computer program causes the computer to perform the aforementioned data mining method. The computer-readable storage medium may be included in the memory described in the above embodiments, or may exist independently and not be incorporated into the memory.

[0076] Another aspect of the present invention provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the data mining method provided in each of the above embodiments.

[0077] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, any equivalent modifications or alterations accomplished by a person of ordinary skill in the art without departing from the spirit and technical concepts disclosed in this application shall be covered by the claims of this application.

Claims

1. A data mining method, characterized in that: The data mining method comprises: Acquire a plurality of map data, establish a map topology based on the plurality of map data, and determine trajectory data of a target object in the plurality of map data; extracting the behavior of the target object based on the map topology and the trajectory data of the target object; generating a behavior label for the behavior of the target object; Filter the behavior tags according to preset filtering rules to obtain candidate tags; The candidate labels are combined according to a preset combination rule to obtain a combined label.

2. The data mining method according to claim 1, characterized in that: The establishing of a map topology based on a plurality of map data includes: Acquire a plurality of first map data, and convert the formats of the plurality of first map data to obtain a plurality of second map data; wherein the plurality of second map data have the same format; Extracting elements from the plurality of second map data to obtain map elements; The map elements include: lane lines, traffic signs, and curbs; A topological connection relationship of lanes, roads and intersections is established based on the map elements to obtain a map topology.

3. The data mining method according to claim 2, characterized in that: The method further includes: a map optimization step, wherein the map optimization step includes at least one of the following: converting the plurality of second map data from a coordinate system corresponding to the second map data to a target coordinate system; Converting the semantic labels in the plurality of second map data into target semantic labels; Downsampling is performed on the map elements corresponding to the plurality of second map data.

4. The data mining method according to claim 1, wherein: The method further comprises: a trajectory optimization step, wherein the trajectory optimization step comprises: Extracting trajectory data of the target object within the time window, wherein the trajectory data includes a path and a speed of the target object; The trajectory data is smoothed to obtain trajectory optimized data of the target object.

5. The data mining method according to claim 1, wherein: The extracting the behavior of the target object based on the map topology and the trajectory data of the target object includes: Calculating the behavior characteristics of the target object according to the map topology and the trajectory data of the target object; The behavior of the target object is obtained based on the behavior characteristics.

6. The data mining method according to claim 1, characterized in that: The preset screening criteria include at least one of the following: Lighting conditions, weather conditions, time period, visibility, road type, traffic flow status, and target object behavior.

7. The data mining method according to claim 1, characterized in that: The preset combination rule includes at least one of the following: Logical combination, indicating a tag combination determined according to a logical relationship between behavior tags, wherein the logical relationship includes at least one of the following: AND, OR, NOT; Sequence combination, which means the label combination determined by the time sequence of the behavior labels; Weight distribution, which indicates the combination of tags that determines the importance of the behavior tags. Different behavior tags have different weights. Priority combination is used to represent a label combination determined by priority. Different behavior labels have different priorities.

8. A data mining device, characterized in that: The data mining device comprises: a data acquisition module, configured to acquire a plurality of map data, establish a map topology based on the plurality of map data, and determine trajectory data of a target object in the plurality of map data; a behavior extraction module, configured to extract the behavior of the target object based on the map topology and the trajectory data of the target object; A label generation module, configured to generate a behavior label for the behavior of the target object; A label screening module is used to screen the behavior labels according to preset screening rules to obtain candidate labels; The label combination module is used to combine the candidate labels according to a preset combination rule to obtain a combined label.

9. A data mining device, characterized in that: include: one or more processors; and A memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the memory implements the method according to any one of claims 1 to 7.

10. A machine-readable medium, characterized in that Instructions are stored thereon, which, when executed by one or more processors, cause the processors to perform the method according to any one of claims 1 to 7.

Images