Data storage method and electronic device for a vehicle

Abstract

The application discloses a data storage method and an electronic device of a vehicle. The method comprises the following steps: in response to a data processing instruction of the vehicle, determining a plurality of scene data matched with a driving scene in which the vehicle is located; determining original address information of the plurality of scene data respectively to obtain a plurality of original address information, wherein the original address information is used to represent an original address to which the scene data is stored; in response to the existence of the same original address information in the plurality of original address information, adjusting the plurality of scene data to obtain adjusted plurality of scene data, wherein the data coincidence degree of the adjusted plurality of scene data is less than the data coincidence degree of the plurality of scene data before adjustment; and respectively storing the adjusted plurality of scene data from the original address to a target address of the vehicle. The application solves the technical problem of low data scalability of the vehicle.

Description

Technical Field

[0001] This application relates to the field of vehicles, and more specifically, to a data storage method and electronic device for vehicles. Background Technology

[0002] Currently, existing methods for mining raw vehicle data often rely too heavily on the quality and completeness of data preprocessing.

[0003] However, for complex scenarios that are difficult to describe exhaustively with limited labels or fixed rules, existing mining methods are unable to adapt to undefined complex scenarios, and also struggle to mine and store scenario data that matches such complex scenarios, resulting in the technical problem of low data scalability for vehicles.

[0004] There is currently no effective solution to the technical problem of low data scalability of the aforementioned vehicles. Summary of the Invention

[0005] This application provides a vehicle data storage method and electronic device to at least solve the technical problem of low data scalability in vehicles.

[0006] According to one aspect of the embodiments of this application, a vehicle data storage method is provided. The method includes: in response to a data processing instruction of the vehicle, determining multiple scene data matching the driving scenario in which the vehicle is located, wherein the data processing instruction is used to instruct the vehicle to process the scene data, and the scene data is used to represent the state of at least one object in the driving scenario when the vehicle is in the driving scenario; determining the original address information of the multiple scene data respectively to obtain multiple original address information, wherein the original address information is used to represent the original address where the scene data is stored; in response to the existence of the same original address information in the multiple original address information, adjusting the multiple scene data to obtain adjusted multiple scene data, wherein the data overlap of the adjusted multiple scene data is less than the data overlap of the original scene data; and storing the adjusted multiple scene data from the original address to the target address of the vehicle respectively.

[0007] Furthermore, in response to the vehicle's data processing instructions, multiple scenario data matching the vehicle's driving scenario are determined, including: in response to the data processing instructions, determining the vehicle's description data, wherein the description data describes the driving scenario in a preset data format; and based on the description data, determining multiple scenario data matching the driving scenario.

[0008] Furthermore, based on the descriptive data, multiple scenario data matching the driving scenario are determined, including: determining the driving scenario corresponding to the descriptive data; and determining multiple scenario data matching the driving scenario from the scenario dataset according to the data determination strategy and the driving scenario. The data determination strategy is used to represent the rules for determining multiple scenario data matching the driving scenario, and the scenario dataset includes multiple scenario data.

[0009] Furthermore, the data determination strategy includes: a first data determination strategy, a second data determination strategy, and a third data determination strategy. The first data determination strategy represents the rules for determining first scene data matching the driving scenario using various trajectory information of the vehicle. The trajectory information represents the vehicle's trajectory points and their locations. The first scene data represents the state of fixed objects within the scene. The second data determination strategy represents the rules for determining second scene data matching the driving scenario using query statements associated with the driving scenario. The second scene data represents the regions where fixed objects are located in different dimensions. The third data determination strategy represents the rules for determining third scene data matching the driving scenario using first prompt data associated with the driving scenario. The third scene data represents the state of fixed objects and the states of other objects within the scene besides the fixed objects. The scene dataset includes: a first scene dataset, a second scene dataset, and a third scene dataset. The system comprises a first scene dataset and a third scene dataset. The first scene dataset includes multiple types of first scene data, the second scene dataset includes multiple types of second scene data, and the third scene dataset includes multiple types of third scene data. Based on a data determination strategy and the driving scenario, multiple scene data matching the driving scenario are determined from the scene datasets. This includes: determining first scene data matching the driving scenario from the first scene dataset according to the first data determination strategy and the driving scenario; determining second scene data matching the driving scenario from the second scene dataset according to the second data determination strategy and the driving scenario; and determining third scene data matching the driving scenario from the third scene dataset according to the third data determination strategy and the driving scenario. The first scene data, the second scene data, and the third scene data matching the driving scenario are then identified as multiple scene data matching the driving scenario.

[0010] Further, according to the first data determination strategy and the driving scenario, the first scenario data matching the driving scenario is determined from the first scenario dataset, including: generating an entity list corresponding to the driving scenario according to the first data determination strategy, wherein the entity list includes: multiple driving scenario entities; determining the position information of multiple driving scenario entities in the entity list to obtain multiple position information, wherein the position information is used to represent the position of the driving scenario entity; determining the distance between the position corresponding to each of the multiple trajectory information and the position corresponding to each of the multiple position information to obtain multiple distances; determining the distance less than a distance threshold from the multiple distances as the first distance, and determining the trajectory information corresponding to the first distance as the first trajectory information; and determining the first scenario data corresponding to the first trajectory information from the first scenario dataset as the first scenario data matching the driving scenario.

[0011] Furthermore, according to the second data determination strategy and driving scenario, second scenario data matching the driving scenario is determined from the second scenario dataset, including: according to the second data determination strategy and query statement, multiple first reference scenario data are determined from the second scenario dataset, wherein the clarity of the first reference scenario data is higher than the clarity of other second scenario data in the second scenario dataset besides the first reference scenario data; using the multiple first reference scenario data as seeds, a search task is performed on the search platform to obtain multiple search scenario data; the multiple first reference scenario data and the multiple search scenario data are determined as second scenario data matching the driving scenario.

[0012] Further, according to the third data determination strategy and driving scenario, third scenario data matching the driving scenario is determined from the third scenario dataset, including: determining second reference scenario data from the third scenario dataset according to the third data determination strategy, wherein the comprehensibility of the second reference scenario data relative to the driving scenario is higher than the comprehensibility of other third scenario data in the third scenario dataset besides the second reference scenario data; generating first prompt data for the second reference scenario data, wherein the first prompt data is used to prompt the second reference scenario data; combining the second reference scenario data and the first prompt data to obtain combined data; inputting the combined data and other third scenario data into the target prompt model for semantic prompting to obtain second prompt data, wherein the second prompt data is used to prompt other third scenario data; and determining third scenario data matching the driving scenario based on the first prompt data and the second prompt data.

[0013] Furthermore, based on the first prompt data and the second prompt data, a third scene data matching the driving scenario is determined, including: determining the data similarity between the first prompt data and the second prompt data; and in response to the data similarity being higher than the data similarity threshold, determining other third scene data corresponding to the second prompt data as third scene data matching the driving scenario.

[0014] Furthermore, the multiple scene data includes: first scene data, second scene data, and third scene data. The first scene data represents the state of a fixed object within the object set; the second scene data represents the region where the fixed object is located in different dimensions; and the third scene data represents the state of the fixed object and the states of other objects within the object set besides the fixed object. The original address information of each of the multiple scene data sets is determined, resulting in multiple original address information sets, including: the first original address information where the matching first scene data is located, the second original address information where the matching second scene data is located, and the third original address information where the matching third scene data is located. The first original address information indicates the location where the matching first scene data is stored. The original address, the second original address information is used to represent the original address where the matched second scene data is stored, and the third original address information is used to represent the original address where the matched third scene data is stored; in response to the existence of the same original address information among multiple original address information, the multiple scene data are adjusted to obtain the adjusted multiple scene data, including: in response to the existence of the same original address information among the first original address information, the second original address information and the third original address information, deduplication is performed on the matched first scene data, the matched second scene data and the matched third scene data to obtain multiple deduplicated scene data; the multiple deduplicated scene data are fused, and the fused multiple deduplicated scene data is determined as the adjusted multiple scene data.

[0015] According to another aspect of the embodiments of this application, a vehicle data storage device is also provided. The device includes: a first determining unit, configured to determine multiple scene data matching the driving scenario of the vehicle in response to a data processing instruction of the vehicle, wherein the data processing instruction is used to instruct the vehicle to process the scene data, and the scene data is used to represent the state of at least one object in the driving scenario when the vehicle is in the driving scenario; a second determining unit, configured to determine the original address information of the multiple scene data respectively to obtain multiple original address information, wherein the original address information is used to represent the original address where the scene data is stored; an adjusting unit, configured to adjust the multiple scene data in response to the existence of the same original address information in the multiple original address information to obtain adjusted multiple scene data, wherein the data overlap of the adjusted multiple scene data is less than the data overlap of the original scene data; and a storage unit, configured to store the adjusted multiple scene data from the original address to the target address of the vehicle respectively.

[0016] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0017] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0018] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0019] According to another aspect of the embodiments of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods in various embodiments of this application.

[0020] According to another aspect of the embodiments of this application, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of this application.

[0021] According to another aspect of the embodiments of this application, a vehicle is also provided, which includes the electronic equipment described in this application.

[0022] In this embodiment, in response to a vehicle's data processing instruction, multiple scenario data matching the vehicle's driving scenario are determined; the original address information of each scenario data is determined, resulting in multiple original address information; in response to the presence of identical original address information among the multiple original address information, the multiple scenario data are adjusted to obtain adjusted scenario data; and the adjusted scenario data are stored from their original addresses to the vehicle's target address. Since this embodiment can determine the original address information of each scenario data based on the determination of multiple scenario data matching the vehicle's driving scenario, that is, it can determine the original address where each scenario data is stored. Then, in the case of identical original address information among the multiple original address information, the multiple scenario data are adjusted to obtain adjusted scenario data. Subsequently, the adjusted scenario data are stored from their original addresses to the vehicle's target address, thereby achieving the goal of adapting to complex and changing scenarios, solving the technical problem of low vehicle data scalability, and ultimately achieving the technical effect of improving vehicle data scalability. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0024] Figure 1(a) is a schematic diagram of an application scenario of a vehicle data storage method according to an embodiment of this application;

[0025] Figure 1(b) is a flowchart of a vehicle data storage method according to an embodiment of this application;

[0026] Figure 2(a) is a schematic diagram of a multi-skill intelligent scene data mining system according to an embodiment of this application;

[0027] Figure 2(b) is a flowchart of a multi-skill intelligent scene data mining method according to an embodiment of this application;

[0028] Figure 3 This is a flowchart of a map mining method according to an embodiment of this application;

[0029] Figure 4 This is a flowchart of a multimodal retrieval and mining method according to an embodiment of this application;

[0030] Figure 5 This is a flowchart of a large-model-based mining method according to an embodiment of this application;

[0031] Figure 6This is a structural block diagram of a vehicle data storage device according to an embodiment of this application;

[0032] Figure 7 This is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0033] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0034] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0035] According to an embodiment of this application, an embodiment of a vehicle data storage method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0036] As an optional implementation, the above-described vehicle data storage method can be applied, but is not limited to, the application scenario shown in Figure 1(a). Figure 1(a) is a schematic diagram of an application scenario of a vehicle data storage method according to an embodiment of this application. As shown in Figure 1(a), in the application scenario, the terminal device 10 can communicate with the server 13 via the network 11, but is not limited to. The server 13 can perform operations on the database, such as writing or reading data. The terminal device 10 can include, but is not limited to, a human-machine interface screen, a processor, and a memory. The human-machine interface screen can be used, but is not limited to, to display a virtual machine on the mobile terminal 10. The vehicle 12 can be used, but is not limited to, to respond to the above-described human-machine interface operation, execute corresponding operations, or generate corresponding instructions and send the generated instructions to the server 13.

[0037] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here. Specifically, the vehicle data storage method of this application may include: step S102, in response to a vehicle data processing instruction, determining multiple scene data matching the driving scenario in which the vehicle is located; step S104, determining the original address information of each of the multiple scene data to obtain multiple original address information; step S106, in response to the existence of identical original address information among the multiple original address information, adjusting the multiple scene data to obtain adjusted multiple scene data; and step S108, storing the adjusted multiple scene data from the original address to the vehicle's target address.

[0038] It should be noted that all information (including but not limited to original address information, etc.) and data (including but not limited to scene data, etc.) involved in this application are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0039] Figure 1(b) is a flowchart of a vehicle data storage method according to an embodiment of the present application. As shown in Figure 1(b), the method may include the following steps.

[0040] Step S112: In response to the vehicle's data processing instructions, determine multiple scenario data that match the driving scenario in which the vehicle is located.

[0041] In the technical solution provided in step S112 of this application, the data processing instruction can be used to instruct the vehicle to process scene data. For example, the data processing instruction can be a data mining instruction, which can be used to instruct the vehicle to mine scene data; or, the data processing instruction can be a data update instruction, which can be used to instruct the vehicle to update scene data. This is only an example and is not specifically limited.

[0042] In this embodiment, the aforementioned scene data can be used to represent the state of at least one object in a driving scene when the vehicle is in motion. The driving scene can be any of the following scenarios: school scene, hospital scene, park scene, and workplace scene, etc. For example, if the driving scene is a school scene, at least one object in the driving scene can be a student and a teacher; if the driving scene is a hospital scene, at least one object can be a patient, a nurse, and a doctor; if the driving scene is a park scene, at least one object can be a preschool child and a parent; if the driving scene is a workplace scene, at least one object can be a vehicle, a workplace employee, and a workplace visitor. These are merely illustrative examples and not specific limitations.

[0043] In this embodiment, in response to a data processing instruction from the vehicle, multiple scene data matching the driving scenario in which the vehicle is located are determined. Optionally, upon receiving a data processing instruction, this embodiment can determine multiple scene data matching the driving scenario in which the vehicle is located from a scene dataset, wherein the scene dataset may include multiple scene data.

[0044] Optionally, in the scene dataset, the degree of matching between various scene data and the driving scene is determined. Multiple matching degrees that exceed a matching degree threshold are then designated as target matching degrees. Subsequently, the scene data corresponding to each of the multiple target matching degrees are determined as the various scene data matching the aforementioned driving scene, thereby achieving the goal of identifying multiple scene data matching the aforementioned driving scene.

[0045] Step S114: Determine the original address information of various scenario data to obtain various original address information.

[0046] In the technical solution provided in step S114 of this application, the aforementioned original address information can be used to represent the original address where the scene data is stored. Optionally, the aforementioned original address information can be the original path or source storage space, etc.

[0047] In this embodiment, after determining multiple scene data matching the vehicle's driving scenario in response to the vehicle's data processing instructions, the original address information of each scene data is determined, resulting in multiple original address information. Optionally, based on determining the multiple scene data matching the aforementioned driving scenario, this embodiment locates each scene data separately to obtain multiple original address information, thereby achieving the purpose of determining the original address where the scene data is stored.

[0048] Optionally, source information of the scene dataset can be obtained. This source information can represent the source of the scene dataset; for example, the source can be local space or cloud space. If the source information corresponds to local space, multiple original address information can be obtained by locating various scene data within the local space. If the source information corresponds to cloud space, multiple original address information can be obtained by locating various scene data within the cloud space.

[0049] Step S116: In response to the existence of the same original address information in multiple original address information, adjust the multiple scenario data to obtain the adjusted multiple scenario data.

[0050] In the technical solution provided in step S116 of this application, the degree of data overlap of the adjusted multiple scene data is less than the degree of data overlap of the original multiple scene data. For example, if the scene data is a scene data fragment, the degree of data overlap of the adjusted scene data fragment is less than the degree of data overlap of the original scene data fragment.

[0051] In this embodiment, the aforementioned adjustments may include deduplication and fusion processing, and the adjusted multiple scenario data may be multiple result data after deduplication and fusion. For example, the multiple result data after deduplication and fusion may also be referred to as the final result data; this is merely an example and not a specific limitation.

[0052] In this embodiment, after determining the original address information of multiple scene data and obtaining multiple original address information, in response to the existence of identical original address information among the multiple original address information, the multiple scene data are adjusted to obtain adjusted multiple scene data. Optionally, this embodiment compares the multiple original address information one by one based on the obtained multiple original address information to obtain a comparison result, wherein the comparison result can be used to indicate whether there is identical original address information among the multiple original address information. If the comparison result indicates that there is identical original address information among the multiple original address information, then deduplication and fusion processing are performed on the multiple scene data to obtain the adjusted multiple scene data.

[0053] Optionally, if the above comparison results show that there are no identical original address information among the multiple original address information, that is, any two original address information are different, then the multiple scenario data are fused, and the fused multiple scenario data is determined as the adjusted multiple scenario data.

[0054] It should be noted that the methods described above for obtaining adjusted scenario data are merely illustrative examples and are not intended to impose specific limitations. Any method that can adjust multiple scenario data in response to the presence of identical original address information in multiple original address information to obtain adjusted scenario data is within the protection scope of the embodiments of this application, and will not be described in detail here.

[0055] Step S118: Store the adjusted scenario data from the original address to the vehicle's target address.

[0056] In the technical solution provided by step S118 of this application, the target address is the address where the adjusted multi-scenario data is to be stored, and the target address can be used to indicate the target storage path or target storage space.

[0057] In this embodiment, in response to the presence of identical original address information among multiple original address information, the multiple scene data are adjusted to obtain adjusted multiple scene data. Then, the adjusted multiple scene data are stored from the original address to the vehicle's target address. Optionally, this embodiment may migrate the adjusted multiple scene data from the original storage path to the vehicle's target storage path, or migrate the adjusted multiple scene data from the original storage space to the vehicle's target storage space.

[0058] In steps S112 to S118 of this application, in response to the vehicle's data processing instructions, multiple scenario data matching the vehicle's driving scenario are determined; the original address information of each of the multiple scenario data is determined, resulting in multiple original address information; in response to the existence of identical original address information among the multiple original address information, the multiple scenario data are adjusted to obtain adjusted multiple scenario data; and the adjusted multiple scenario data are stored from their original addresses to the vehicle's target address. Since this embodiment of the application can determine the original address information of each of the multiple scenario data based on the determination of multiple scenario data matching the vehicle's driving scenario, that is, it can determine the original address where each of the multiple scenario data is stored. Then, in the case where identical original address information exists among the multiple original address information, the multiple scenario data are adjusted to obtain adjusted multiple scenario data. Subsequently, the adjusted multiple scenario data are stored from their original addresses to the vehicle's target address, thereby achieving the goal of adapting to complex and changing scenarios, thus solving the technical problem of low data scalability of vehicles, and ultimately achieving the technical effect of improving the data scalability of vehicles.

[0059] The following section further describes the steps of determining multiple scenario data matching the driving scenario of the vehicle in response to the data processing instructions of the vehicle in this embodiment.

[0060] As an optional embodiment, step S112, in response to the vehicle's data processing instruction, determines multiple scenario data matching the driving scenario in which the vehicle is located, including: in response to the data processing instruction, determining the vehicle's description data, wherein the description data describes the driving scenario in a preset data format; and based on the description data, determining multiple scenario data matching the driving scenario.

[0061] In this embodiment, the aforementioned descriptive data can describe the driving scenario using a preset data format. Specifically, the descriptive data can be text-based, and the preset data format can be text-based; that is, text-based descriptive data can describe the driving scenario using text format.

[0062] In this embodiment, in response to a data processing instruction, vehicle description data is determined. Optionally, upon receiving a data processing instruction, this embodiment parses the instruction to obtain vehicle description data. For example, upon receiving a data mining instruction regarding a target scene, parsing the instruction yields mined description data for the vehicle. This mined description data describes the target scene in text form, and the target scene can be any of the following: a school scene, a hospital scene, a park scene, or a workplace scene, etc.

[0063] In this embodiment, after determining the vehicle's description data in response to a data processing instruction, multiple scenario data matching the driving scenario are determined based on the description data. Optionally, this embodiment can determine the driving scenario corresponding to the description data based on the determined description data, and then determine multiple scenario data matching the driving scenario. This achieves the goal of determining multiple scenario data matching the driving scenario based on the description data, thereby realizing the technical effect of improving the matching degree of scenario data.

[0064] The following section further explains the steps of determining multiple scenario data matching the driving scenario based on the description data in this embodiment.

[0065] As an optional implementation method, based on the description data, multiple scenario data matching the driving scenario are determined, including: determining the driving scenario corresponding to the description data; and determining multiple scenario data matching the driving scenario from the scenario dataset according to the data determination strategy and the driving scenario, wherein the data determination strategy is used to represent the rules for determining multiple scenario data matching the driving scenario, and the scenario dataset includes multiple scenario data.

[0066] In this embodiment, after determining the vehicle's description data in response to a data processing instruction, the driving scenario corresponding to the description data is determined. Optionally, based on the determined description data, this embodiment performs scene recognition on the description data to obtain the driving scenario corresponding to the description data. For example, by inputting the mined description data into a scene recognition model for scene recognition, the target scene corresponding to the mined description data can be obtained, wherein the scene recognition model is constructed based on a semantic analysis model.

[0067] In this embodiment, the data determination strategy described above can be used to represent rules for determining multiple scenario data that match the driving scenario. For example, the data determination strategy can be a data mining strategy, which may include: map mining strategy, multimodal retrieval mining strategy, and large model mining strategy. Alternatively, the data determination strategy can be a data update strategy, which may include: map update strategy and periodic update strategy, etc., which are only illustrative examples and are not specifically limited.

[0068] In this embodiment, the aforementioned scene dataset may include multiple scene data. For example, the aforementioned scene dataset may include multiple scene data fragments.

[0069] In this embodiment, after determining the driving scenario corresponding to the descriptive data, various scenario data matching the driving scenario are determined from the scenario dataset according to the data determination strategy and the driving scenario. Optionally, based on determining the driving scenario corresponding to the descriptive data, this embodiment determines the degree of matching between various scenario data and the driving scenario in the scenario dataset according to the data determination strategy and the driving scenario. Multiple matching degrees that are greater than a matching degree threshold are taken as target matching degrees. Then, the scenario data corresponding to each of the multiple target matching degrees are determined as various scenario data matching the aforementioned driving scenario. This achieves the goal of determining various scenario data matching the driving scenario from the scenario dataset, thereby realizing the technical effect of improving the diversity of scenario data.

[0070] The following description further explains the steps of determining various scenario data matching the driving scenario from the scenario dataset according to the data determination strategy and driving scenario in this embodiment.

[0071] As an optional implementation method, according to a data determination strategy and a driving scenario, multiple scenario data matching the driving scenario are determined from a scenario dataset, including: according to a first data determination strategy and a driving scenario, determining first scenario data matching the driving scenario from a first scenario dataset; according to a second data determination strategy and a driving scenario, determining second scenario data matching the driving scenario from a second scenario dataset; according to a third data determination strategy and a driving scenario, determining third scenario data matching the driving scenario from a third scenario dataset; and determining the first scenario data matching the driving scenario, the second scenario data matching the driving scenario, and the third scenario data matching the driving scenario as multiple scenario data matching the driving scenario.

[0072] In this embodiment, the data determination strategy may include: a first data determination strategy, a second data determination strategy, and a third data determination strategy.

[0073] In this embodiment, the first data determination strategy described above can be used to represent rules for determining first scene data that matches the driving scenario using various trajectory information of the vehicle. For example, the first data determination strategy can be a map mining strategy or a map update strategy.

[0074] In this embodiment, the trajectory information can be used to represent the vehicle's trajectory points and their locations. For example, the trajectory information can be the Global Positioning System (GPS) log of the vehicle's trajectory.

[0075] In this embodiment, the aforementioned first scene data can be used to represent the state of a fixed object within an object, and the aforementioned first scene data can be a first scene data fragment. The fixed object can be an object with a fixed geographical location in a driving scene. For example, if the driving scene is a school scene, the fixed object can be the school in the school scene; if the driving scene is a hospital scene, the fixed object can be the hospital in the hospital scene; if the driving scene is a workplace scene, the fixed object can be the workplace in the workplace scene. This is merely an example and not a specific limitation.

[0076] In this embodiment, the aforementioned scene dataset may include: a first scene dataset, a second scene dataset, and a third scene dataset. For example, the first scene dataset may be a map database, the second scene dataset may be a preliminary image set, and the third scene dataset may be a scene data pool or a scene image set.

[0077] In this embodiment, the aforementioned first scene dataset may include multiple types of first scene data. For example, the aforementioned scene data pool may include multiple types of first scene data fragments.

[0078] In this embodiment, after determining the driving scenario corresponding to the descriptive data, first scenario data matching the driving scenario is determined from the first scenario dataset according to the first data determination strategy and the driving scenario. Optionally, based on determining the driving scenario corresponding to the descriptive data, this embodiment determines a first matching degree between multiple first scenario data and the driving scenario in the first scenario dataset according to the first data determination strategy and the driving scenario. Multiple first matching degrees that are greater than a matching degree threshold are respectively taken as first target matching degrees. Then, the first scenario data corresponding to each of the multiple first target matching degrees are determined as multiple first scenario data matching the aforementioned driving scenario. This achieves the goal of determining first scenario data matching the driving scenario from the first scenario dataset, thereby realizing the technical effect of improving the diversity of first scenario data.

[0079] In this embodiment, the aforementioned second data determination strategy can be used to represent rules for determining second scenario data matching the driving scenario using a query statement associated with the driving scenario. For example, the aforementioned second data determination strategy can be a multimodal retrieval mining strategy or a multimodal retrieval update strategy, and the aforementioned query statement can be query text.

[0080] In this embodiment, the aforementioned second scene data can be used to represent the area where a fixed object is located in different dimensions. For example, if the driving scene is a school scene, the aforementioned second scene data can be used to represent the area where the school is located at different angles and times in the school scene; if the driving scene is a hospital scene, the aforementioned second scene data can be used to represent the area where the hospital is located at different angles and times in the hospital scene; if the driving scene is a unit scene, the aforementioned second scene data can be used to represent the area where the unit is located at different angles and times in the unit scene. This is only an example and is not a specific limitation.

[0081] In this embodiment, the second scene dataset may include multiple types of second scene data. For example, the scene image set may include multiple fragments of second scene data.

[0082] In this embodiment, after determining the driving scenario corresponding to the descriptive data, second scenario data matching the driving scenario is determined from the second scenario dataset according to the second data determination strategy and the driving scenario. Optionally, based on determining the driving scenario corresponding to the descriptive data, this embodiment determines a second matching degree between multiple types of second scenario data and the driving scenario in the second scenario dataset according to the second data determination strategy and the driving scenario. Multiple second matching degrees that are greater than a matching degree threshold are taken as second target matching degrees. Then, the second scenario data corresponding to each of the multiple second target matching degrees are determined as multiple types of second scenario data matching the aforementioned driving scenario. This achieves the goal of determining second scenario data matching the driving scenario from the second scenario dataset, thereby realizing the technical effect of improving the diversity of second scenario data.

[0083] In this embodiment, the aforementioned third data determination strategy can be used to represent a rule for determining third scenario data that matches the driving scenario using first prompt data associated with the driving scenario. For example, the aforementioned third data determination strategy can be a large model mining strategy or a large model update strategy.

[0084] In this embodiment, the aforementioned third scene data can be used to represent the state of a fixed object and the state of other objects within that object besides the fixed object. For example, if the driving scenario is a school scenario, the aforementioned third scene data can be used to represent the state of the school within the school scenario and the state of other objects within that object besides the school. If the driving scenario is a hospital scenario, the aforementioned third scene data can be used to represent the state of the hospital within the hospital scenario and the state of other objects within that object besides the hospital. If the driving scenario is a workplace scenario, the aforementioned third scene data can be used to represent the state of the workplace within the workplace scenario and the state of other objects within that object besides the workplace. This is merely an example and is not intended to impose any specific limitations.

[0085] In this embodiment, the aforementioned third scene dataset may include multiple types of third scene data. For example, the aforementioned scene image set may include multiple types of third scene data fragments.

[0086] In this embodiment, after determining the driving scenario corresponding to the descriptive data, according to the third data determination strategy and the driving scenario, third scenario data matching the driving scenario is determined from the third scenario dataset; the first scenario data matching the driving scenario, the second scenario data matching the driving scenario, and the third scenario data matching the driving scenario are determined as multiple scenario data matching the driving scenario. Optionally, based on determining the driving scenario corresponding to the descriptive data, this embodiment, according to the third data determination strategy and the driving scenario, determines the third matching degree between multiple third scenario data and the driving scenario in the third scenario dataset. Multiple third matching degrees that are greater than a matching degree threshold are taken as third target matching degrees. Then, the third scenario data corresponding to each of the multiple third target matching degrees are determined as multiple third scenario data matching the aforementioned driving scenario, and the first scenario data matching the driving scenario, the second scenario data matching the driving scenario, and the third scenario data matching the driving scenario are taken as multiple scenario data matching the driving scenario. This achieves the goal of determining third scenario data matching the driving scenario from the third scenario dataset, thereby realizing the technical effect of improving the diversity of third scenario data.

[0087] The following description further explains the steps of determining the first scenario data matching the driving scenario from the first scenario dataset according to the first data determination strategy and driving scenario in this embodiment.

[0088] As an optional embodiment, according to a first data determination strategy and a driving scenario, determining first scenario data matching the driving scenario from a first scenario dataset includes: generating an entity list corresponding to the driving scenario according to the first data determination strategy, wherein the entity list includes: multiple driving scenario entities; determining the location information of multiple driving scenario entities in the entity list to obtain multiple location information, wherein the location information is used to represent the location of the driving scenario entities; determining the distance between the location corresponding to each of the multiple trajectory information and the location corresponding to each of the multiple location information to obtain multiple distances; determining the distance less than a distance threshold from the multiple distances as a first distance, and determining the trajectory information corresponding to the first distance as first trajectory information; and determining the first scenario data corresponding to the first trajectory information from the first scenario dataset as the first scenario data matching the driving scenario.

[0089] In this embodiment, the entity list may include multiple driving scenario entities. This entity list can also be referred to as an entity name list. For example, if the driving scenario is a school scenario, the entity name list is a school name list; if the driving scenario is a hospital scenario, the entity name list is a hospital name list. This is merely an example and not a specific limitation.

[0090] In this embodiment, the location information described above can be used to represent the location of entities in the driving scene. For example, the location of the entities in the driving scene can be represented by their latitude and longitude coordinates.

[0091] In this embodiment, after determining the driving scenario corresponding to the description data, a list of entities related to the driving scenario is generated using a large language model according to the first data determination strategy. Then, the map application interface is called to query the location information of multiple driving scenario entities in the entity list.

[0092] In this embodiment, after obtaining multiple location information, the distances between the locations corresponding to each of the multiple trajectory information and their respective corresponding locations are determined, resulting in multiple distances. Each distance is compared with a distance threshold, and the distances smaller than the threshold are defined as the first distance, and the trajectory information corresponding to the first distance is defined as the first trajectory information. Then, from the first scene dataset, the first scene data corresponding to the first trajectory information is determined as the first scene data matching the driving scene. This achieves the goal of determining the first scene data corresponding to the first trajectory information as the first scene data matching the driving scene, thereby improving the accuracy of the first scene data.

[0093] Optionally, a list of entity names related to the target scene is generated using a large language model. A map application programming interface (API) is then called to query the latitude and longitude coordinates corresponding to each entity in the list. The spatial distance between the point sequence of the vehicle's trajectory and the latitude and longitude coordinates is calculated. Data segments with spatial distances less than a preset threshold are then filtered out. For example, the preset threshold could be, but is not limited to, 200 meters or 210 meters; these values ​​are merely illustrative and not specifically limited.

[0094] The following description further explains the steps of determining second scenario data matching the driving scenario from the second scenario dataset according to the second data determination strategy and driving scenario in this embodiment.

[0095] As an optional embodiment, according to the second data determination strategy and the driving scenario, determining second scenario data matching the driving scenario from the second scenario dataset includes: determining multiple first reference scenario data from the second scenario dataset according to the second data determination strategy and the query statement, wherein the clarity of the first reference scenario data is higher than the clarity of other second scenario data in the second scenario dataset besides the first reference scenario data; using the multiple first reference scenario data as seeds, performing a search task on a search platform to obtain multiple search scenario data; and determining the multiple first reference scenario data and the multiple search scenario data as second scenario data matching the driving scenario.

[0096] In this embodiment, the aforementioned first reference scene data can be a search seed. The clarity of the first reference scene data is higher than the clarity of other second scene data in the second scene dataset besides the first reference scene data.

[0097] In this embodiment, after determining the driving scenario corresponding to the descriptive data, multiple first reference scenario data are determined from the second scenario dataset according to the second data determination strategy and query statement. Optionally, based on determining the driving scenario corresponding to the descriptive data, this embodiment uses the text description of the driving scenario as the query statement, according to the second data determination strategy, to query a preliminary image set related to the driving scenario from the second scenario dataset. Then, multiple first reference scenario data are determined from the preliminary image set, wherein the clarity of the first reference scenario data is higher than the clarity of other second scenario data in the preliminary image set besides the first reference scenario data.

[0098] In this embodiment, the search platform can be a data mining website (Portal) or a data update platform.

[0099] In this embodiment, the search task described above can be an image search task.

[0100] In this embodiment, after determining multiple first reference scenario data from the second scenario dataset according to the second data determination strategy and query statement, a search task is performed on a search platform using these multiple first reference scenario data as seeds to obtain multiple search scenario data. The multiple first reference scenario data and the multiple search scenario data are then determined as second scenario data matching the driving scenario. Optionally, this embodiment uses the determined multiple first reference scenario data as search seeds to perform a search task on a data mining website to obtain multiple search scenario data. Next, the multiple first reference scenario data and the searched multiple search scenario data are merged into second scenario data matching the driving scenario, thereby achieving the goal of expanding the second scenario data and thus realizing the technical effect of improving the accuracy of the second scenario data.

[0101] Optionally, a preliminary image set can be retrieved from the scene data pool using the text description of the target scene as the query condition. For example, a new text-to-image search task can be created on a data mining platform, with the query text being "school area, school gate, dismissal time, students crossing the road". After executing the new text-to-image search task, a preliminary image set can be obtained. Then, representative images are selected from the preliminary image set as query seeds. Using the query seeds as input, a similarity retrieval based on visual features can be performed to mine more relevant data fragments. For example, several of the most representative images can be selected from the text-to-image search. Among these selected images, the clarity of the school area scene at different angles and times is higher than the preset clarity, while the unselected images show that the clarity of the school area scene at different angles and times is lower than or equal to the preset clarity. Using the selected images as seeds, multiple image search tasks can be created on the data mining platform. After executing the image search tasks, all the results obtained can be merged to obtain an expanded image set, and the corresponding data paths can be parsed out.

[0102] The following description further explains the steps of determining the third scenario data matching the driving scenario from the third scenario dataset according to the third data determination strategy and driving scenario in this embodiment.

[0103] As an optional implementation, according to a third data determination strategy and a driving scenario, determining third scenario data matching the driving scenario from a third scenario dataset includes: determining second reference scenario data from the third scenario dataset according to the third data determination strategy, wherein the comprehensibility of the second reference scenario data relative to the driving scenario is higher than the comprehensibility of other third scenario data in the third scenario dataset besides the second reference scenario data; generating first prompt data for the second reference scenario data, wherein the first prompt data is used to prompt the second reference scenario data; combining the second reference scenario data and the first prompt data to obtain combined data; inputting the combined data and other third scenario data into a target prompt model for semantic prompting to obtain second prompt data, wherein the second prompt data is used to prompt other third scenario data; and determining third scenario data matching the driving scenario based on the first prompt data and the second prompt data.

[0104] In this embodiment, the comprehensibility of the second reference scene data relative to the driving scene is higher than the comprehensibility of other third scene data in the third scene dataset besides the second reference scene data. For example, the second reference scene data can be a keyframe or a short video clip.

[0105] In this embodiment, the first prompt data can be used to prompt the second reference scene data. For example, the first prompt data can be a prompt word from a keyframe or a prompt word from a short video clip.

[0106] In this embodiment, after determining the driving scenario corresponding to the description data, according to the third data determination strategy, in the third scenario dataset, multiple third scenario data are sorted according to the comprehensibility of each third scenario data relative to the driving scenario. From the sorted multiple third scenario data, second reference scenario data is determined, and then first prompt data is generated for the determined second reference scenario data.

[0107] In this embodiment, the second prompt data can be used to prompt other third scene data, and the second prompt data can be the natural language output of the target prompt model. For example, the target prompt model can be constructed based on a large language model.

[0108] In this embodiment, after generating first prompt data for the second reference scene data, the second reference scene data and the first prompt data are combined to obtain multimodal combined data. Next, the combined data and other third scene data are input into the target prompt model for semantic prompting to obtain second prompt data. Then, the data similarity between the first and second prompt data is determined. Based on the determined data similarity, third scene data matching the driving scenario can be determined. This achieves the goal of combining the first and second prompt data to determine the third scene data matching the driving scenario, thereby improving the accuracy of the third scene data.

[0109] Optionally, a large data mining model is loaded onto a cloud server. Keyframes or short video clips are read in batches from a data pool. Clue words, as shown below, are constructed for each image frame or short video clip and combined into a multimodal input, which is then transmitted to the large data mining model. Model inference is performed, and the inference results are obtained. It is then determined whether the inference results match the constructed clue words. Based on the determination results, the corresponding keyframe or short video clip can be output, or the next data can be processed.

[0110] The following description further explains the steps of determining the third scenario data matching the driving scenario based on the first and second prompt data in this embodiment.

[0111] As an optional embodiment, determining third scenario data matching the driving scenario based on first and second prompt data includes: determining the data similarity between the first and second prompt data; and, in response to a data similarity higher than a data similarity threshold, determining other third scenario data corresponding to the second prompt data as third scenario data matching the driving scenario.

[0112] In this embodiment, after inputting the combined data and other third-scene data into the target prompt model for semantic prompting to obtain the second prompt data, a similarity calculation is performed on the first and second prompt data to obtain the data similarity between them. This data similarity is compared with a data similarity threshold. If the data similarity is higher than the threshold, it indicates that the second prompt data matches the first prompt data, and the other third-scene data corresponding to the second prompt data can be identified as the third-scene data matching the driving scenario. If the data similarity is lower than or equal to the threshold, it indicates that the second prompt data does not match the first prompt data, and the combined data and the next other third-scene data are input into the target prompt model for semantic prompting.

[0113] Optionally, it determines whether the inference result matches the constructed prompt. If the inference result matches the constructed prompt, it records the keyframe or short video clip corresponding to the matching inference result and outputs the corresponding keyframe or short video clip. If the inference result does not match the constructed prompt, it continues processing the next piece of data.

[0114] The following section further explains the steps of adjusting multiple scenario data to obtain adjusted scenario data in response to the existence of the same original address information in multiple original address information in the above embodiment.

[0115] As an optional embodiment, step S114 involves determining the original address information of multiple scene data to obtain multiple original address information, including: determining the first original address information where the matched first scene data is located, the second original address information where the matched second scene data is located, and the third original address information where the matched third scene data is located, wherein the first original address information is used to represent the original address where the matched first scene data is stored, the second original address information is used to represent the original address where the matched second scene data is stored, and the third original address information is used to represent the original address where the matched third scene data is stored; step S116 involves adjusting the multiple scene data in response to the existence of the same original address information among the multiple original address information to obtain adjusted multiple scene data, including: in response to the existence of the same original address information among the first, second, and third original address information, deduplicating the matched first scene data, the matched second scene data, and the matched third scene data to obtain multiple deduplicated scene data; fusing the multiple deduplicated scene data; and determining the fused multiple deduplicated scene data as the adjusted multiple scene data.

[0116] In this embodiment, the aforementioned multiple scene data may include: first scene data, second scene data, and third scene data.

[0117] In this embodiment, the aforementioned first scene data can be used to represent the state of a fixed object within the object.

[0118] In this embodiment, the aforementioned second scene data can be used to represent the regions where a fixed object is located in different dimensions.

[0119] In this embodiment, the aforementioned third scene data can be used to represent the state of a fixed object, as well as the states of other objects in the object besides the fixed object.

[0120] In this embodiment, the aforementioned first original address information can be used to represent the original address where the matched first scene data is stored.

[0121] In this embodiment, the aforementioned second original address information can be used to represent the original address where the matched second scene data is stored.

[0122] In this embodiment, the aforementioned third original address information can be used to represent the original address where the matched third scene data is stored.

[0123] In this embodiment, the first original address information is obtained by locating the matched first scene data, the second original address information is obtained by locating the matched second scene data, and the third original address information is obtained by locating the matched third scene data. Then, the first, second, and third original address information are compared one by one to obtain comparison results. If the comparison results show that there is a common original address information among the first, second, and third original address information, deduplication processing is performed on the matched first, second, and third scene data to obtain multiple deduplicated scene data sets.

[0124] In this embodiment, after deduplicating the matched first scene data, the matched second scene data, and the matched third scene data to obtain multiple deduplicated scene data, the multiple deduplicated scene data are fused together, and the fused multiple deduplicated scene data is determined as the adjusted multiple scene data. This achieves the technical objective of fusing deduplicated scene data, thereby realizing the technical effect of reducing the data overlap of the adjusted multiple scene data.

[0125] Optionally, the mining results of the map mining module, the multimodal retrieval mining module, and the large model-based mining module are deduplicated and merged according to the data paths in the three files, and the merged mining results are automatically migrated to the target storage space.

[0126] In this embodiment, in response to a vehicle's data processing instruction, multiple scenario data matching the vehicle's driving scenario are determined; the original address information of each scenario data is determined, resulting in multiple original address information; in response to the presence of identical original address information among the multiple original address information, the multiple scenario data are adjusted to obtain adjusted scenario data; and the adjusted scenario data are stored from their original addresses to the vehicle's target address. Since this embodiment can determine the original address information of each scenario data based on the determination of multiple scenario data matching the vehicle's driving scenario, that is, it can determine the original address where each scenario data is stored. Then, in the case of identical original address information among the multiple original address information, the multiple scenario data are adjusted to obtain adjusted scenario data. Subsequently, the adjusted scenario data are stored from their original addresses to the vehicle's target address, thereby achieving the goal of adapting to complex and changing scenarios, solving the technical problem of low vehicle data scalability, and ultimately achieving the technical effect of improving vehicle data scalability.

[0127] The technical solutions of the embodiments of this application will be illustrated below with reference to preferred embodiments.

[0128] Currently, existing methods for mining raw vehicle data often rely too heavily on the quality and completeness of data preprocessing.

[0129] However, for complex scenarios that are difficult to describe exhaustively with limited labels or fixed rules, existing mining methods are unable to adapt to undefined complex scenarios, and also struggle to mine and store scenario data that matches such complex scenarios, resulting in the technical problem of low data scalability for vehicles.

[0130] However, this application proposes a vehicle data storage method. Based on identifying multiple scenario data matching the vehicle's driving scenario, the original address information of each scenario data point can be determined; that is, the original address where each scenario data point is stored can be determined. Then, when the same original address information exists among the multiple original address information points, the multiple scenario data points are adjusted to obtain adjusted scenario data points. These adjusted scenario data points are then stored from their original addresses to the vehicle's target address. This achieves the goal of adapting to complex and changing scenarios, thereby solving the technical problem of low data scalability in vehicles and ultimately improving the technical effect of vehicle data scalability.

[0131] In this embodiment, the raw data of the vehicle can be mined through the multi-mode intelligent scene data mining system. For example, Figure 2(a) is a schematic diagram of a multi-mode intelligent scene data mining system according to an embodiment of this application. As shown in Figure 2(a), the multi-mode intelligent scene data mining system 200 may include: an instruction receiving module 201, a map mining module 2021, a multimodal retrieval mining module 2022, a large model-based mining module 2023, a result fusion module 203, and an output module 204.

[0132] In this embodiment, the instruction receiving module 202 can be used to receive a mining instruction for the target scene input by the user. The mining instruction may include a text description of the scene.

[0133] In this embodiment, the map mining module 2021 can be used to generate a list of entity names related to the target scene using a large language model, call the map application programming interface (API) to query the latitude and longitude coordinates corresponding to each entity in the entity name list, and perform spatial distance calculation between the GPS point sequence of the vehicle trajectory and the latitude and longitude coordinates, and filter out data segments with a distance less than a preset threshold as map mining results.

[0134] In this embodiment, the multimodal retrieval and mining module 2022 can be used to perform text-to-image search, seed selection, and image search steps to obtain multimodal retrieval and mining results. Specifically, the text-to-image search step can be implemented by retrieving a preliminary image set from the data pool using a text description of the target scene as the query condition. The seed selection step can be implemented by selecting representative images from the preliminary image set as query seeds. The image search step can be implemented by using the query seeds as input and performing a similarity retrieval based on visual features to mine more relevant data fragments.

[0135] In this embodiment, the large-scale model-based mining module 2023 can be used to input video or keyframe data segments from the data pool into the large-scale mining model, issue recognition instructions including a description of the target scene to the large-scale mining model, and parse the output results of the large-scale mining model to obtain mining results based on the large-scale model. Specifically, when the large-scale mining model identifies that a data segment includes the target scene, it adds the data segment to the results list.

[0136] In this embodiment, the result fusion module 203 can be used to deduplicate and merge the mining results of the map mining module 2021, the multimodal retrieval mining module 2022, and the large model-based mining module 2023.

[0137] In this embodiment, the output module 204 can be used to output the merged mining results.

[0138] For example, when receiving a mining instruction with the target scene being "school area," the aforementioned multi-intelligence scene data mining system can simultaneously initiate map mining, multimodal retrieval mining, and large-model mining processes. Subsequently, the system deduplicates and merges the mining results output from each of the three processes, and automatically downloads the merged results to a specified output directory.

[0139] In this embodiment, the multi-intelligence scene data mining method is executed by the multi-intelligence scene data mining system to mine the raw data of the vehicle. For example, Figure 2(b) is a flowchart of a multi-intelligence scene data mining method according to an embodiment of this application. As shown in Figure 2(b), the method may include the following steps.

[0140] Step S210: Receive the target scene mining instructions input by the user.

[0141] After receiving the user's input of the target scene mining instruction, step S211 is executed to perform the map mining process.

[0142] After receiving the target scene mining instructions input by the user, step S212 is executed to perform the multimodal retrieval mining process.

[0143] After receiving the target scene mining instruction input by the user, step S213 is executed to carry out the mining process based on the large model.

[0144] After executing the map mining process, the multimodal retrieval mining process, and the large model-based mining process, step S214 is executed to deduplicate and merge the mining results of the map mining module, the multimodal retrieval mining module, and the large model-based mining module.

[0145] In the technical solution provided in step S214 of this application, a fusion script is executed to deduplicate the data in the three files according to their data paths. Then, the deduplicated final list is used as input to execute a unified download script and set the target directory, completing the automated data migration.

[0146] After deduplication and merging of the mining results from the map mining module, the multimodal retrieval mining module, and the large model-based mining module, step S215 is executed to automatically migrate the merged mining results to the target storage space.

[0147] In this embodiment, map mining can be performed on the raw vehicle data by executing a map mining method. For example, Figure 3 This is a flowchart of a map mining method according to an embodiment of this application, such as... Figure 3 As shown, the method may include the following steps.

[0148] Step S301: Use the large language model to generate a list of entity names related to the target scene.

[0149] In the technical solution provided in step S301 of this application, by calling a large language model (e.g., DeepSeek) and inputting the prompt "Please list the names of common primary schools, middle schools and universities in major Chinese cities", a list of school names can be generated.

[0150] After generating a list of entity names related to the target scene using a large language model, step S302 is executed to call the map application interface and query the latitude and longitude coordinates corresponding to each entity in the entity name list.

[0151] In the technical solution provided in step S302 of this application, the script is executed to call the map API and query the latitude and longitude of these schools in batches.

[0152] After calling the map application interface to query the latitude and longitude coordinates corresponding to each entity in the entity name list, step S303 is executed to calculate the spatial distance between the point sequence of the vehicle trajectory and the latitude and longitude coordinates.

[0153] In the technical solution provided by step S303 of this application, the script is executed to input the GPS log of the vehicle trajectory, and a distance matching threshold is set (e.g., but not limited to, 200 meters), and the distance between each trajectory point and all schools is sampled and calculated.

[0154] After calculating the spatial distance between the point sequence of the vehicle trajectory and the latitude and longitude coordinates, step S304 is executed to determine whether the calculated spatial distance is less than a preset threshold.

[0155] In the technical solution provided in step S304 of this application, if the preset threshold is 200 meters, the data corresponding to the trajectory segments with a distance of less than 200 meters will be output. If the preset threshold is 250 meters, the data corresponding to the trajectory segments with a distance of less than 250 meters will be output. The values ​​here are only illustrative examples and are not specifically limited.

[0156] If the calculated spatial distance is determined to be less than a preset threshold, then step S305 is executed to mark the data segments whose spatial distance is less than the preset threshold as matching data.

[0157] If the calculated spatial distance is greater than or equal to a preset threshold, then step S306 is executed to continue processing the next trajectory point.

[0158] After processing the next trajectory point, proceed to step S303.

[0159] In this embodiment, multimodal retrieval and mining methods can be used to perform multimodal retrieval and mining on the raw vehicle data. For example, Figure 4 This is a flowchart of a multimodal retrieval and mining method according to an embodiment of this application, such as... Figure 4 As shown, the method may include the following steps.

[0160] Step S401: Using the text description of the target scene as the query condition, retrieve a preliminary image set from the data pool.

[0161] In the technical solution provided in step S401 of this application, a text search image task is created on the data mining platform (Portal), and the query text is "school area, school gate, dismissal time, students crossing the road". After executing the newly created text search image task, a preliminary image set can be obtained.

[0162] After retrieving a preliminary image set from the data pool using the text description of the target scene as the query condition, step S402 is executed to obtain the retrieved image set.

[0163] After obtaining the retrieved image set, step S403 is executed to filter seed images.

[0164] In the technical solution provided by step S403 of this application, several of the most representative images are selected from the preliminary image set. Among them, the selected images show that the clarity of the school area scene at different angles and times is higher than the preset clarity, while the unselected images show that the clarity of the school area scene at different angles and times is lower than or equal to the preset clarity.

[0165] After screening the seed images, step S404 is executed, using the seed images as input, to perform a similarity retrieval based on visual features and uncover more relevant data fragments.

[0166] In the technical solution provided in step S404 of this application, multiple image search tasks are created on the Portal using the selected images as seeds. After executing the multiple image search tasks, multiple search image sets can be obtained. By merging the multiple search image sets, an extended image set can be obtained, and the corresponding data paths can be parsed out.

[0167] After performing visual feature-based similarity retrieval with the seed image as input and mining more relevant data fragments, step S405 is executed to merge the mined data fragments with the seed image to obtain the retrieval results.

[0168] After merging the mined data fragments with the seed image to obtain the retrieval results, step S406 is executed to output the retrieval results.

[0169] In this embodiment, a large-model-based mining method can be implemented to mine the raw vehicle data. For example, Figure 5 This is a flowchart of a large-model-based mining method according to an embodiment of this application, such as... Figure 5 As shown, the method may include the following steps.

[0170] Step S501: Load the large mining model on the cloud server.

[0171] After loading the large mining model on the cloud server, step S502 is executed to read keyframes or short video clips from the data pool in batches.

[0172] After reading keyframes or short video clips in batches from the data pool, step S503 is executed to construct cue words as shown below for each image frame or short video clip, and combine them into a multimodal input for transmission to the large mining model. For example, the cue words are as follows: You are an autonomous driving data screening assistant. Please determine whether the given image / video contains a scene related to "school area". The judgment criteria include, but are not limited to: the presence of school signs, school gates, students in school uniforms, specific traffic signs near schools (e.g., speed bumps, children's warning signs), etc. If at least one of the above signs exists, please answer "Yes, this is a school area scene" and briefly explain the reason; if at least one of the above signs does not exist, please answer "No, this is not a school area scene".

[0173] After the multimodal input is combined and transmitted to the large mining model, step S504 is executed to perform model inference and obtain the inference result.

[0174] After performing model inference and obtaining the inference result, step S505 is executed to determine whether the inference result matches the constructed prompt words.

[0175] If the inference result is determined to match the constructed prompt word, then steps S506 and S507 are executed to record the key frame or short video clip corresponding to the matched inference result, and to output the corresponding key frame or short video clip.

[0176] If it is determined that the reasoning result does not match the constructed prompt, then step S508 is executed to continue processing the next data.

[0177] After processing the next data, proceed to step S504.

[0178] In this embodiment, based on identifying multiple scene data matching the vehicle's driving scenario, the original address information of each scene data can be determined; that is, the original address where each scene data is stored can be determined. Then, if the same original address information exists among the multiple original address information, the multiple scene data are adjusted to obtain adjusted scene data. Subsequently, the adjusted scene data is stored from its original address to the vehicle's target address. This achieves the goal of adapting to complex and changing scenarios, thereby solving the technical problem of low vehicle data scalability and ultimately improving the vehicle's data scalability.

[0179] According to another aspect of the embodiments of this application, corresponding to the embodiments of the above-described vehicle data storage method, the embodiments of this application also provide a vehicle data storage device. Figure 6 This is a structural block diagram of a vehicle data storage device according to an embodiment of this application, such as... Figure 6 As shown, the data storage device 600 of the vehicle may include: a first determining unit 602, a second determining unit 604, an adjusting unit 606, and a storage unit 608.

[0180] The first determining unit 602 is used to determine multiple scene data matching the driving scene in response to the data processing instruction of the vehicle. The data processing instruction is used to instruct the vehicle to process the scene data, and the scene data is used to represent the state of at least one object in the driving scene when the vehicle is in the driving scene.

[0181] The second determining unit 604 is used to determine the original address information of multiple scene data respectively, and obtain multiple original address information, wherein the original address information is used to represent the original address where the scene data is stored.

[0182] The adjustment unit 606 is used to adjust multiple scenario data in response to the existence of the same original address information in multiple original address information, so as to obtain the adjusted multiple scenario data, wherein the degree of data overlap of the adjusted multiple scenario data is less than the degree of data overlap of the original scenario data.

[0183] Storage unit 608 is used to store the adjusted multiple scene data from the original address to the vehicle's target address.

[0184] Optionally, the first determining unit 602 may include: a first determining module, configured to determine vehicle description data in response to a data processing instruction, wherein the description data describes the driving scenario in a preset data format; and a second determining module, configured to determine multiple scenario data matching the driving scenario based on the description data.

[0185] Optionally, the second determining module may include: a first determining submodule, used to determine the driving scenario corresponding to the description data; and a second determining submodule, used to determine multiple scenario data matching the driving scenario from the scenario dataset according to the data determining strategy and the driving scenario, wherein the data determining strategy is used to represent the rules for determining multiple scenario data matching the driving scenario, and the scenario dataset includes multiple scenario data.

[0186] Optionally, the second determining submodule can determine multiple scenario data matching the driving scenario from the scenario dataset by performing the following steps: determining first scenario data matching the driving scenario from the first scenario dataset according to the first data determining strategy and the driving scenario; determining second scenario data matching the driving scenario from the second scenario dataset according to the second data determining strategy and the driving scenario; determining third scenario data matching the driving scenario from the third scenario dataset according to the third data determining strategy and the driving scenario; and determining the first scenario data matching the driving scenario, the second scenario data matching the driving scenario, and the third scenario data matching the driving scenario as multiple scenario data matching the driving scenario.

[0187] Optionally, the second determining submodule can perform the following steps to determine first scenario data matching the driving scenario from the first scenario dataset according to the first data determining strategy and the driving scenario: generating an entity list corresponding to the driving scenario according to the first data determining strategy, wherein the entity list includes multiple driving scenario entities; determining the location information of multiple driving scenario entities in the entity list to obtain multiple types of location information, wherein the location information is used to represent the location of the driving scenario entities; determining the distance between the location corresponding to each of the multiple trajectory information and the location corresponding to each of the multiple location information to obtain multiple distances; determining the distance less than a distance threshold from the multiple distances as the first distance, and determining the trajectory information corresponding to the first distance as the first trajectory information; and determining the first scenario data corresponding to the first trajectory information from the first scenario dataset as the first scenario data matching the driving scenario.

[0188] Optionally, the second determining submodule can perform the following steps to determine second scenario data matching the driving scenario from the second scenario dataset according to the second data determining strategy and the driving scenario: determining multiple first reference scenario data from the second scenario dataset according to the second data determining strategy and the query statement, wherein the clarity of the first reference scenario data is higher than the clarity of other second scenario data in the second scenario dataset besides the first reference scenario data; using the multiple first reference scenario data as seeds, performing a search task on the search platform to obtain multiple search scenario data; and determining the multiple first reference scenario data and the multiple search scenario data as second scenario data matching the driving scenario.

[0189] Optionally, the second determining submodule can perform the following steps to determine third scenario data matching the driving scenario from the third scenario dataset according to the third data determining strategy and the driving scenario: determining second reference scenario data from the third scenario dataset according to the third data determining strategy, wherein the comprehensibility of the second reference scenario data relative to the driving scenario is higher than the comprehensibility of other third scenario data in the third scenario dataset besides the second reference scenario data; generating first prompt data for the second reference scenario data, wherein the first prompt data is used to prompt the second reference scenario data; combining the second reference scenario data and the first prompt data to obtain combined data; inputting the combined data and other third scenario data into the target prompt model for semantic prompting to obtain second prompt data, wherein the second prompt data is used to prompt other third scenario data; and determining third scenario data matching the driving scenario based on the first prompt data and the second prompt data.

[0190] Optionally, the second determining submodule can determine third scenario data matching the driving scenario based on the first prompt data and the second prompt data by performing the following steps: determining the data similarity between the first prompt data and the second prompt data; and in response to the data similarity being higher than the data similarity threshold, determining other third scenario data corresponding to the second prompt data as third scenario data matching the driving scenario.

[0191] Optionally, the second determining unit 604 may include: a third determining module, used to determine the first original address information where the matched first scene data is located, the second original address information where the matched second scene data is located, and the third original address information where the matched third scene data is located, wherein the first original address information is used to represent the original address where the matched first scene data is stored, the second original address information is used to represent the original address where the matched second scene data is stored, and the third original address information is used to represent the original address where the matched third scene data is stored; the adjusting unit 606 may include: a deduplication module, used to deduplicate the matched first scene data, the matched second scene data, and the matched third scene data in response to the existence of the same original address information in the first original address information, the second original address information, and the third original address information, to obtain multiple deduplicated scene data; a fourth determining module, used to fuse the multiple deduplicated scene data, and to determine the fused multiple deduplicated scene data as the adjusted multiple scene data.

[0192] In this embodiment, the vehicle's data storage device includes the following units: a first determining unit, configured to determine multiple scene data matching the vehicle's driving scenario in response to the vehicle's data processing instructions, wherein the data processing instructions instruct the vehicle to process the scene data, and the scene data represents the state of at least one object in the driving scenario when the vehicle is in a driving scenario; a second determining unit, configured to determine the original address information of the multiple scene data respectively, obtaining multiple original address information, wherein the original address information represents the original address where the scene data is stored; an adjusting unit, configured to adjust the multiple scene data in response to the existence of the same original address information among the multiple original address information, obtaining adjusted multiple scene data, wherein the data overlap of the adjusted multiple scene data is less than the data overlap of the original scene data; and a storage unit, configured to store the adjusted multiple scene data from the original address to the vehicle's target address, thereby achieving the purpose of adapting to complex and changing scenarios, thus solving the technical problem of low data scalability of the vehicle, and thereby achieving the technical effect of improving the data scalability of the vehicle.

[0193] Embodiments of this application also provide an electronic device, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0194] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0195] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0196] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium for storing a computer program that, when executed by a processor, implements the methods in various embodiments of this application.

[0197] Embodiments of this application also provide a computer program that, when executed by a processor, implements the methods described in the various embodiments of this application.

[0198] Embodiments of this application also provide a vehicle that includes the electronic devices described in this application.

[0199] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0200] According to an embodiment of this application, an electronic device is also provided. Figure 7 This is a schematic diagram of an electronic device according to an embodiment of this application, such as... Figure 7 As shown, the electronic device 700 may include a memory 710 and a processor 720, wherein the memory 710 is used to store an executable program; and the processor 720 is used to run the program stored in the memory 710, and the program executes the method of this application when it runs.

[0201] In this application, "multiple" refers to two or more.

[0202] In this application, unless otherwise expressly defined, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0203] The terms “first,” “second,” “third,” “fourth,” etc., in this application (if present) are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0204] In this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, in this application, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0205] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided. The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the device control method for the vehicle in the embodiment.

[0206] Computer-readable storage media, also known as computer storage media, may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. These propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable storage media can transmit, propagate, or transfer programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0207] The program code contained in a computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, radio frequency, or any suitable combination thereof.

[0208] In the embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0209] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0210] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0211] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0212] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for storing vehicle data, characterized in that, include: In response to a data processing instruction from the vehicle, multiple scene data matching the driving scenario in which the vehicle is located are determined, wherein the data processing instruction is used to instruct the vehicle to process the scene data, and the scene data is used to represent the state of at least one object in the driving scenario when the vehicle is in the driving scenario; The original address information of various types of scene data is determined to obtain various types of original address information, wherein the original address information is used to represent the original address where the scene data is stored; In response to the presence of identical original address information among multiple types of original address information, the multiple types of scene data are adjusted to obtain adjusted multiple types of scene data, wherein the degree of data overlap of the adjusted multiple types of scene data is less than the degree of data overlap of the unadjusted multiple types of scene data; The adjusted scenario data are stored from the original address to the target address of the vehicle.

2. The method according to claim 1, characterized in that, In response to the data processing instructions of the vehicle, various scene data matching the driving scenario of the vehicle are determined, including: In response to the data processing instruction, the vehicle's description data is determined, wherein the description data describes the driving scenario in a preset data format; Based on the described data, various scenario data matching the driving scenario are determined.

3. The method according to claim 2, characterized in that, Based on the description data, various scenario data matching the driving scenario are determined, including: Determine the driving scenario corresponding to the described data; According to the data determination strategy and the driving scenario, the various scenario data that match the driving scenario are determined from the scenario dataset. The data determination strategy is used to represent the rules for determining the various scenario data that match the driving scenario. The scenario dataset includes the various scenario data.

4. The method according to claim 2, characterized in that, The data determination strategy includes: a first data determination strategy, a second data determination strategy, and a third data determination strategy. The first data determination strategy represents the rules for determining first scene data matching the driving scenario using various trajectory information of the vehicle. The trajectory information represents the trajectory points of the vehicle and the positions of those points. The first scene data represents the state of a fixed object within the vehicle. The second data determination strategy represents the rules for determining second scene data matching the driving scenario using query statements associated with the driving scenario. The second scene data represents the regions where the fixed object is located in different dimensions. The third data determination strategy represents... The method demonstrates the rules for determining third scene data matching the driving scenario using first prompt data associated with the driving scenario. The third scene data represents the state of the fixed object and the states of other objects within the fixed object. The scene dataset includes: a first scene dataset, a second scene dataset, and a third scene dataset. The first scene dataset includes multiple types of first scene data, the second scene dataset includes multiple types of second scene data, and the third scene dataset includes multiple types of third scene data. The method for determining the scene data matching the driving scenario from the scene datasets according to the data determination strategy and the driving scenario includes: Based on the strategy determined by the first data and the driving scenario, the first scenario data matching the driving scenario is determined from the first scenario dataset; Based on the second data determination strategy and the driving scenario, determine the second scenario data that matches the driving scenario from the second scenario dataset; According to the third data determination strategy and the driving scenario, the third scenario data that matches the driving scenario is determined from the third scenario dataset; The first scene data that matches the driving scenario, the second scene data that matches the driving scenario, and the third scene data that matches the driving scenario are determined as multiple scene data that match the driving scenario.

5. The method according to claim 4, characterized in that, Based on the first data determination strategy and the driving scenario, determine the first scenario data matching the driving scenario from the first scenario dataset, including: Based on the strategy determined by the first data, an entity list corresponding to the driving scenario is generated, wherein the entity list includes: multiple driving scenario entities; The location information of multiple driving scene entities in the entity list is determined to obtain multiple types of location information, wherein the location information is used to represent the location of the driving scene entity; The distances between the locations corresponding to each of the various trajectory information and the locations corresponding to each of the various location information are determined to obtain multiple distances; From the plurality of distances, the distances that are less than a distance threshold are determined as the first distance, and the trajectory information corresponding to the first distance is determined as the first trajectory information; From the first scene dataset, the first scene data corresponding to the first trajectory information is determined as the first scene data that matches the driving scene.

6. The method according to claim 4, characterized in that, Based on the second data determination strategy and the driving scenario, determine the second scenario data matching the driving scenario from the second scenario dataset, including: According to the second data determination strategy and the query statement, multiple first reference scene data are determined from the second scene dataset, wherein the clarity of the first reference scene data is higher than the clarity of other second scene data in the second scene dataset besides the first reference scene data. Using multiple sets of data from the first reference scenario as seeds, a search task is performed on the search platform to obtain multiple sets of search scenario data; Multiple first reference scenario data and multiple search scenario data are determined as the second scenario data that matches the driving scenario.

7. The method according to claim 4, characterized in that, According to the third data determination strategy and the driving scenario, the third scenario data matching the driving scenario is determined from the third scenario dataset, including: According to the third data determination strategy, a second reference scenario data is determined from the third scenario dataset, wherein the comprehensibility of the second reference scenario data relative to the driving scenario is higher than the comprehensibility of other third scenario data in the third scenario dataset excluding the second reference scenario data; For the second reference scene data, the first prompt data is generated, wherein the first prompt data is used to prompt the second reference scene data; The second reference scene data and the first prompt data are combined to obtain combined data; The combined data and the other third-scene data are input into the target prompting model for semantic prompting to obtain second prompting data, wherein the second prompting data is used to prompt the other third-scene data; Based on the first prompt data and the second prompt data, the third scene data that matches the driving scenario is determined.

8. The method according to claim 7, characterized in that, Based on the first prompt data and the second prompt data, the third scenario data matching the driving scenario is determined, including: Determine the data similarity between the first prompt data and the second prompt data; In response to the data similarity being higher than the data similarity threshold, the other third scene data corresponding to the second prompt data is determined as the third scene data matching the driving scene.

9. The method according to any one of claims 1 to 8, characterized in that, The various scene data include: first scene data, second scene data, and third scene data. The first scene data represents the state of a fixed object within the object; the second scene data represents the region where the fixed object is located in different dimensions; and the third scene data represents the state of the fixed object, as well as the states of other objects within the object besides the fixed object. The original address information of each of the various scene data is determined to obtain various original address information, including: The first original address information of the matched first scene data, the second original address information of the matched second scene data, and the third original address information of the matched third scene data are determined. The first original address information is used to represent the original address where the matched first scene data is stored, the second original address information is used to represent the original address where the matched second scene data is stored, and the third original address information is used to represent the original address where the matched third scene data is stored. In response to the presence of identical original address information among multiple types of original address information, the multiple types of scenario data are adjusted to obtain the adjusted multiple types of scenario data, including: in response to the presence of identical original address information among the first original address information, the second original address information, and the third original address information, deduplication is performed on the matched first scenario data, the matched second scenario data, and the matched third scenario data to obtain multiple deduplicated scenario data; The multiple deduplication scenario data are fused together, and the fused multiple deduplication scenario data are determined as the adjusted multiple scenario data.

10. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 9.

Images