Data collection method, device, server, and medium

Abstract

The present application provides a data collection method and device, server, and medium; the method includes: summarizing the collected first scene data to obtain the current total collection volume of each first scene data in different scenes; in different scenes, determining the first scene in which the current total collection volume corresponding to the first scene data is less than the corresponding target volume; for the first scene, determining a target position for collecting scene data; sending the target position to a target vehicle so that the target vehicle collects the missing scene data of the first scene; continuing to obtain the next total collection volume for the first scene, executing the next collection of the next missing scene data for the first scene, until the final total collection volume of the scene data for the first scene is not less than the corresponding target volume, thereby obtaining second scene data; the second scene data is all the collected scene data for each scene; and using the second scene data for different scenes to train a model for intelligent driving.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and specifically to a data collection method and device, server, and medium. Background Art

[0002] With the continuous development of Advanced Driver Assistance Systems (ADAS) and autonomous driving technologies, the demand for high-quality, diverse road scene data is growing. Data collection is fundamental to building intelligent driving artificial intelligence (AI) models, and its quality and coverage directly impact the model's training effectiveness and generalization capabilities. Traditional vehicle data collection methods often use fixed-frequency or fixed-point collection strategies, which can lead to uneven data distribution and poor AI model performance and generalization capabilities. Summary of the Invention

[0003] The present application provides a data collection method and device, server, and medium, which are beneficial to solving the problem of uneven data distribution caused by traditional collection methods and improving the effect and generalization ability of the model for realizing vehicle intelligent driving obtained by training based on the collected scene data.

[0004] In a first aspect, an embodiment of the present application provides a data collection method, which is applied to a cloud server and includes: summarizing first scene data collected by at least one vehicle to obtain the current total collection volume of each first scene data in different scenes; the first scene data is used to characterize road information and / or environmental information; in different scenes, determining the first scene in which the current total collection volume corresponding to the first scene data is less than the corresponding target volume; for the first scene, determining a target position for collecting scene data; sending the target position to a target vehicle so that the target vehicle collects gap scene data corresponding to the first scene at the target position; wherein the target vehicle includes vehicles in the area where the target position is located; continuing to obtain the next total collection volume under the first scene, executing the next collection of gap scene data for the first scene, until the final total collection volume of the scene data corresponding to the first scene is not less than the corresponding target volume, thereby obtaining the second scene data under the first scene; the second scene data is all collected scene data for each scene; and using the second scene data under different scenes to train a model for realizing intelligent driving of a vehicle.

[0005] It can be understood that in the data collection method provided in the embodiment of the present application, the cloud server can collect statistics on the number of first scene data collected by multiple vehicles and determine the first scene with insufficient total collection volume; then, for the first scene, determine the target location where scene data can be collected, and send the location to the target vehicle, so that the target vehicle can collect the missing scene data of the first scene at the target location; and continue to obtain the next total collection volume under the first scene, and perform the next collection of the next missing scene data for the first scene, until the final total collection volume of the scene data corresponding to the first scene is not less than the corresponding target volume; so that the scenes corresponding to the collected second scene data respectively meet the target volume, and then the scene data under each scene collected meet the target ratio; and use the second scene data to train a model for realizing intelligent driving of the vehicle. In this way, the missing scene data under the first scene can be collected through a closed-loop feedback mechanism, so that the final total collection volume of the scenes corresponding to the collected second scene data respectively is basically consistent with the target volume, effectively solving the problem of data imbalance caused by traditional collection methods; and thus improving the effect and generalization ability of the model for realizing intelligent driving of the vehicle obtained by training based on the collected scene data.

[0006] In some embodiments, the method further includes: when the type of the first data tag corresponding to the scene data in the first scene is greater than a first value, prioritizing the first data tag to obtain a priority sorting result of the first data tag; the first value is greater than or equal to 1; sending the priority sorting result to the target vehicle, so that the target vehicle plans a navigation path based on the priority sorting result and the target location, so as to navigate to the target location based on the navigation path to collect the gap scene data corresponding to the first data tag.

[0007] It can be understood that in the data collection method provided in the embodiment of the present application, when the cloud server identifies multiple gap scene data, that is, when there is more than one type of first data tag, the cloud server will prioritize the first data tags corresponding to the gap scene data and send the priority sorting results to the target vehicle. In this way, the target vehicle can reasonably arrange the collection order and path planning according to user needs based on the priority sorting results and target location. This is beneficial to improving the intelligence level of data collection and enhancing user experience.

[0008] In some embodiments, the prioritizing the first data tag to obtain the priority ranking result of the first data tag includes: determining a first priority weight of the first data tag; wherein the first priority weight is used to characterize the completion deviation of the scene data corresponding to the first data tag; determining a second priority weight of the first data tag based on the collection period of the gap scene data corresponding to the first data tag; wherein the collection period is inversely proportional to the second priority weight; and prioritizing the first data tag based on the first priority weight and the second priority weight to obtain the priority ranking result of the first data tag.

[0009] It is understood that in the data collection method provided in the embodiments of the present application, the first data tags are prioritized based on the completion deviation of the scene data corresponding to the first data tag and the collection deadline of the missing scene data corresponding to the first data tag, thereby obtaining a priority ranking result for the first data tags. This is beneficial for enabling the vehicle to prioritize the collection of missing scene data with large gaps and urgent deadlines, thereby improving the intelligent level of data collection and enhancing the user experience.

[0010] In some embodiments, determining the first priority weight of the first data tag includes: determining a first ratio between the collection quantity of the first data tag and the target quantity of the first data tag; determining a second ratio between the target quantity of the first data tag and the total target quantity of multiple first scene data; determining the data quality of the scene data corresponding to the first data tag; and determining the first priority weight of the first data tag based on the first ratio, the second ratio and the data quality.

[0011] It will be appreciated that in the data collection method provided in the embodiments of the present application, the first priority weight is determined based on the completed proportion of scene data for the first scenario (i.e., the first ratio), the target proportion of scene data for the first scenario (i.e., the second ratio), and the data quality of the scene data for the first scenario. This helps identify the collection gaps, importance, and data quality of the scene data corresponding to the first data tag, thereby making the priority sorting results more reasonable, thereby improving the intelligent level of data collection and enhancing the user experience.

[0012] In some embodiments, determining the target location of the collectible scene data includes: determining the target location of the gap scene data corresponding to the first data tag in the map data based on the first data tag corresponding to the gap scene data.

[0013] It is understood that in the data collection method provided in the embodiments of the present application, the first data tag corresponding to the gap scene data is used to match the map data to obtain the target location of the gap scene data. In this way, the target location of the gap scene data is determined in a targeted manner, which is beneficial for quickly and accurately determining the target location.

[0014] In some embodiments, based on the first data tag corresponding to the gap scene data, determining the target position of the gap scene data corresponding to the first data tag in the map data includes: constructing a first feature set of the gap scene data corresponding to the first data tag; determining a second feature set of the map data; wherein the second feature set is obtained by parsing the geometric information of the roads in the map data, the topological relationship of the road sections in the map data, and the identification information in the map data; matching the first feature set and the second feature set to determine the target position of the gap scene data corresponding to the first data tag in the map data.

[0015] It can be understood that in the data collection method provided in the embodiments of this application, by constructing a first feature set for gap scene data corresponding to a first data tag and matching it with a second feature set in the map data, target locations from which gap scene data can be collected can be efficiently identified in the map data. This allows for precise positioning of target locations from which gap scene data can be collected, thereby improving vehicle collection efficiency.

[0016] In some embodiments, determining the target location of the collectible scene data includes: based on the second data tag of the first scene data, determining the geographical location of the first scene data corresponding to the second data tag in the map data; marking the corresponding second data tag at the geographical location to obtain a scene distribution map; based on the first data tag corresponding to the gap scene data, querying the scene distribution map to determine the target location of the gap scene data corresponding to the first data tag.

[0017] It can be understood that in the data collection method provided in the embodiments of the present application, the geographic location of the first scene data corresponding to the second data tag is first determined in the map data based on the second data tag; the second data tag is annotated at the geographic location of the first scene data corresponding to the second data tag to obtain a scene distribution map; thus, when the target location corresponding to the gap scene data is determined, the target location corresponding to the gap scene data can be queried in the scene distribution map based on the first data tag. This is beneficial to improving the efficiency of determining the target location corresponding to the gap scene data, and further beneficial to improving the efficiency of data collection.

[0018] In some embodiments, querying the scene distribution map to determine the target location of the gap scene data corresponding to the first data tag includes: querying the scene distribution map to determine the candidate target location of the gap scene data corresponding to the first data tag; and determining the target location from the candidate target location that meets the first condition.

[0019] It is understood that in the data collection method provided in the embodiment of the present application, rather than taking any candidate target location as the target location, a target location that meets the condition is selected from the candidate target locations based on the first condition. This is beneficial to further improve the intelligent level of data collection and enhance the user experience.

[0020] In some embodiments, the method further includes: sending a first strategy to the target vehicle; the first strategy is used to instruct the target vehicle to prioritize collecting gap scene data corresponding to the third data tag when there is a target position corresponding to the third data tag on the navigation path; wherein the priority of the third data tag is lower than the first data tag corresponding to the target position of the navigation path.

[0021] It is understood that in the data collection method provided in the embodiments of this application, if a target location on the navigation path contains missing scene data corresponding to a low-priority first data tag, the missing scene data corresponding to the low-priority first data tag will be collected first. This helps improve collection efficiency, reduce vehicle operating costs, and further enhance the intelligent level of data collection.

[0022] In a second aspect, an embodiment of the present application provides a data collection device, characterized in that the device includes: a summarization module, configured to summarize first scene data collected by at least one vehicle, and obtain the current total collection amount of each first scene data in different scenes; the first scene data is used to characterize road information and / or environmental information; a determination module, configured to determine, in different scenes, a first scene in which the current total collection amount corresponding to the first scene data is less than the corresponding target amount; a sending module, configured to determine a target position for collecting scene data for the first scene; send the target position to a target vehicle, so that the target vehicle collects gap scene data corresponding to the first scene at the target position; wherein the target vehicle includes a vehicle in the area where the target position is located; an iteration module, configured to continue to obtain the next total collection amount under the first scene, and perform the next collection of the next gap scene data for the first scene until the final total collection amount of the scene data corresponding to the first scene is not less than the corresponding target amount, thereby obtaining the second scene data under the first scene; the second scene data is all collected scene data for each scene; and a training module, configured to use each second scene data under different scenes to train a model for realizing intelligent driving of a vehicle.

[0023] In a third aspect, an embodiment of the present application provides a cloud server comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the program, the data collection method described in the first aspect is implemented.

[0024] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the data collection method described in the first aspect.

[0025] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program or instructions, which, when executed by a processor, implements the data collection method described in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings herein are incorporated into and constitute a part of this specification. These drawings illustrate embodiments consistent with the present application and, together with the specification, serve to illustrate the technical solutions of the present application. Obviously, the drawings described below are merely some embodiments of the present application. Those skilled in the art can, without inventive effort, derive other drawings from these drawings.

[0027] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0028] Figure 1 A schematic diagram of the implementation process of a data collection method provided in an embodiment of the present application Figure 1 ;

[0029] Figure 2 A schematic diagram of an implementation flow for prioritizing first data tags provided in an embodiment of the present application;

[0030] Figure 3 A schematic diagram of an implementation flow for determining a first priority weight provided in an embodiment of the present application;

[0031] Figure 4 A schematic diagram of an implementation process for determining a target location provided in an embodiment of the present application Figure 1 ;

[0032] Figure 5 A schematic diagram of an implementation process for determining a target location provided in an embodiment of the present application Figure 2 ;

[0033] Figure 6 A schematic diagram of the implementation process of a data collection method provided in an embodiment of the present application Figure 2 ;

[0034] Figure 7 A schematic diagram of the implementation process of a data collection method provided in an embodiment of the present application Figure 3 ;

[0035] Figure 8 A schematic diagram of the structure of a policy package provided in an embodiment of the present application;

[0036] Figure 9 A schematic diagram of a data collection device provided in an embodiment of the present application;

[0037] Figure 10 A schematic diagram of the structure of the cloud server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application.

[0039] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described below in conjunction with the accompanying drawings. The embodiments described below are only part of the embodiments of this application, not all of the embodiments. Therefore, the described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0040] In the following description, reference is made to “some embodiments\other embodiments”, which describe a subset of all possible embodiments, but it can be understood that “some embodiments\other embodiments” may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.

[0041] In the following description, the terms "first\second" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understandable that "first\second" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0043] In a related technology, a vehicle data collection method based on a data distribution service is provided. This method partitions data into different functional domains of the vehicle and performs themed collection and transmission based on a distributed service protocol. This method primarily focuses on the data collection and transmission mechanisms within the vehicle bus, with an emphasis on the efficiency and subscribing nature of the collected data within the system.

[0044] However, the inventors of this application have conducted research and analysis on the aforementioned related technologies and found that they do not address the dynamic identification and data collection control of different driving scenarios by the collection vehicle during operation. Furthermore, they are unable to enable the cloud system to formulate collection strategies based on data requirements and automatically dispatch collection vehicles. Therefore, in response to the Advanced Driver Assistance Systems (ADAS) demand for data collection in specific proportions and diverse scenarios, the existing technologies still have the following major deficiencies:

[0045] (1) There is a lack of a structured collection mechanism based on collection targets, making it impossible to dynamically adjust data for specific scenarios (such as straight roads, curves, and traffic light intersections) according to preset proportions;

[0046] (2) Collection decisions rely on manual configuration or static rules, with a low degree of automation and difficulty adapting to complex traffic environments;

[0047] (3) The vehicle-cloud collaboration mechanism is missing or weak, and the cloud lacks the ability to perceive the overall vehicle collection status and provide strategic feedback.

[0048] In view of this, in an embodiment of the present application, a data collection method is provided. Figure 1 A schematic diagram of the implementation process of a data collection method provided in an embodiment of the present application Figure 1 ,like Figure 1 As shown, the method includes steps 101 to 105:

[0049] Step 101: Summarize first scene data collected by at least one vehicle to obtain the current total collection amount of each first scene data in different scenes; the first scene data is used to represent road information and / or environmental information;

[0050] Step 102: Determine, among different scenarios, a first scenario in which the current total collection amount corresponding to the first scenario data is less than the corresponding target amount;

[0051] Step 103: determining a target location for collecting scene data for the first scene; sending the target location to a target vehicle so that the target vehicle collects gap scene data corresponding to the first scene at the target location; wherein the target vehicle includes vehicles in the area where the target location is located;

[0052] Step 104: Continue to obtain the next total collection amount for the first scene, and perform the next collection of the next gap scene data for the first scene until the final total collection amount of the scene data corresponding to the first scene is not less than the corresponding target amount, thereby obtaining the second scene data for the first scene; the second scene data is all the collected scene data for each scene;

[0053] Step 105: Use the second scenario data in different scenarios to train a model for realizing intelligent driving of the vehicle.

[0054] It can be understood that in the data collection method provided in the embodiment of the present application, the cloud server can collect statistics on the number of first scene data collected by multiple vehicles and determine the first scene with insufficient total collection volume; then, for the first scene, determine the target location where scene data can be collected, and send the location to the target vehicle, so that the target vehicle can collect the missing scene data of the first scene at the target location; and continue to obtain the next total collection volume under the first scene, and perform the next collection of the next missing scene data for the first scene, until the final total collection volume of the scene data corresponding to the first scene is not less than the corresponding target volume; so that the scenes corresponding to the collected second scene data respectively meet the target volume, and then the scene data under each scene collected meet the target ratio; and use the second scene data to train a model for realizing intelligent driving of the vehicle. In this way, the missing scene data under the first scene can be collected through a closed-loop feedback mechanism, so that the final total collection volume of the scenes corresponding to the collected second scene data respectively is basically consistent with the target volume, effectively solving the problem of data imbalance caused by traditional collection methods; and thus improving the effect and generalization ability of the model for realizing intelligent driving of the vehicle obtained by training based on the collected scene data.

[0055] The following describes further optional implementations and related terms of each of the above steps.

[0056] In step 101, first scene data collected by at least one vehicle are summarized to obtain the current total collection amount of each first scene data in different scenes; the first scene data is used to represent road information and / or environmental information.

[0057] It should be understood that in the embodiments of the present application, the first scenario data is not limited. The first scenario data refers to data collected by the vehicle during actual driving that can represent road information and / or environmental information. In some embodiments, the road information includes: road type and / or road segment type; wherein the road type includes but is not limited to at least one of the following: expressway, urban expressway, national highway, provincial highway, rural road, etc.; the road segment type includes but is not limited to at least one of the following: straight road segment, curved road segment, unsignalized intersection, signalized intersection, roundabout, etc.; the environmental information includes: natural environment and / or traffic environment; the natural environment includes but is not limited to at least one of the following: daytime, nighttime, dusk, early morning, weather type (sunny, rainy, snowy, foggy), etc.; the traffic environment includes but is not limited to at least one of the following: vehicle density, pedestrian density, road width, road surface quality, signs and markings, lighting conditions, etc.

[0058] In some embodiments, the first scene data collected from at least one vehicle are aggregated to obtain the current total collection volume of each first scene data in different scenarios, including: aggregating the first scene data uploaded by the at least one vehicle to obtain the current total collection volume of each first scene data in different scenarios.

[0059] In other embodiments, the first scene data collected by the at least one vehicle are summarized to obtain the current total collection volume of each of the first scene data in different scenarios, including: summarizing the second data tags corresponding to the first scene data uploaded by the at least one vehicle to obtain the current total collection volume of each of the first scene data in different scenarios.

[0060] It should be understood that in an embodiment of the present application, the vehicle can upload the collected first scene data in real time; or the vehicle can upload the collected first scene data and its corresponding second data tag in real time; or the vehicle can upload the second data tag corresponding to the collected first scene data in real time, and only upload the first scene data when the network is good.

[0061] In some embodiments, the first scene data is acquired in real time by onboard sensors installed on the vehicle, such as high-definition cameras, lidar, etc.; the first scene data is associated with a multi-dimensional data tag that describes the specific scene corresponding to the first scene data. It should be understood that in the embodiments of the present application, the second data tag is a combination of keywords for classifying the first scene data, such as curves, nighttime, and urban roads, which are used for subsequent statistical analysis of the amount of first scene data collected and for identifying missing scene data.

[0062] It should be understood that in the embodiment of the present application, the first scene data in any scene includes the corresponding second data tag.

[0063] In some embodiments, the same vehicle can collect scene data of the same scene, or can collect scene data of different scenes; the different vehicles can collect scene data of the same scene, or can collect scene data of different scenes; this is not limited in the embodiments of the present application.

[0064] In step 102, among different scenarios, a first scenario is determined in which the current total collection amount corresponding to the first scenario data is less than the corresponding target amount.

[0065] It should be understood that in the embodiments of the present application, the first scenario is not limited, and the first scenario refers to the situation where the total amount of first scenario data collected does not reach a preset target amount. In the embodiments of the present application, the specific implementation method for determining the first scenario is not limited. In some embodiments, the first scenario is determined based on the total amount of first scenario data collected under the first scenario and the target amount of first scenario data under the first scenario.

[0066] Further, in some embodiments, determining the first scene based on the total collected amount of first scene data under the first scene and the target amount of first scene data under the first scene includes: determining a third ratio between the total collected amount of first scene data under the first scene and the target amount of first scene data under the first scene; when the third ratio is less than 1, taking the scene corresponding to the first scene data as the first scene.

[0067] Furthermore, in other embodiments, determining the first scene based on the total collection amount of the first scene data under the first scene and the target amount of the first scene data under the first scene includes: when the total collection amount of the first scene data under the first scene is less than the target amount of the first scene data under the first scene, taking the scene corresponding to the first scene data as the first scene.

[0068] In step 103, for the first scene, a target location for collecting scene data is determined; the target location is sent to a target vehicle so that the target vehicle collects gap scene data corresponding to the first scene at the target location; wherein the target vehicle includes vehicles in the area where the target location is located.

[0069] It should be understood that in the embodiments of the present application, the area where the target location is located is not limited. In some embodiments, the area where the target location is located may be the street where the target location is located. In other embodiments, the area where the target location is located may be the administrative district or economic zone where the target location is located. In still other embodiments, the area where the target location is located may be the city, province, or country where the target location is located.

[0070] In some embodiments, the method further includes: when the type of the first data tag corresponding to the scene data in the first scene is greater than a first value, prioritizing the first data tag to obtain a priority sorting result of the first data tag; the first value is greater than or equal to 1; sending the priority sorting result to the target vehicle, so that the target vehicle plans a navigation path based on the priority sorting result and the target location, so as to navigate to the target location based on the navigation path to collect the gap scene data corresponding to the first data tag.

[0071] It can be understood that in the data collection method provided in the embodiment of the present application, when the cloud server identifies multiple gap scene data, that is, when there is more than one type of first data tag, the cloud server will prioritize the first data tags corresponding to the gap scene data and send the priority sorting results to the target vehicle. In this way, the target vehicle can reasonably arrange the collection order and path planning according to user needs based on the priority sorting results and target location. This is beneficial to improving the intelligence level of data collection and enhancing user experience.

[0072] It should be understood that in the embodiments of the present application, the first data tag is not limited. In some embodiments, the second data tag of the first scene data in the first scene is the first data tag of the scene data in the first scene.

[0073] In some embodiments, when the type of the first data tag corresponding to the gap scene data is greater than a first value, the first data tag is prioritized to obtain a priority ranking result of the first data tag, including: before determining the target position where the scene data can be collected, when the type of the first data tag corresponding to the gap scene data is greater than a first value, the first data tag is prioritized to obtain a priority ranking result of the first data tag.

[0074] In other embodiments, when the type of the first data tag corresponding to the gap scene data is greater than a first value, the first data tag is prioritized to obtain a priority ranking result of the first data tag, including: after determining the target position where the scene data can be collected, when the type of the first data tag corresponding to the gap scene data is greater than a first value, the first data tag is prioritized to obtain a priority ranking result of the first data tag.

[0075] In some embodiments, the method further includes: planning a navigation path based on the priority sorting result, the target position and the position of the target vehicle; and sending the navigation path and the target position to the target vehicle corresponding to the navigation path, so that the target vehicle navigates to the target position based on the navigation path to collect the gap scene data corresponding to the corresponding first data tag.

[0076] It should be understood that in the embodiments of the present application, the basis or specific implementation method for prioritizing the first data tags is not limited. In some embodiments, the first data tags can be prioritized based on the difference between the total amount of scene data collected in the first scenario and the target amount, thereby obtaining a priority ranking result for the first data tags; wherein, the larger the difference, the higher the corresponding priority.

[0077] In other embodiments, the first data tags may be prioritized based on the collection period of the scene data in the first scenario to obtain a priority ranking result of the first data tags; wherein, the shorter the collection period, the higher the corresponding priority.

[0078] In some further embodiments, the first data tags can be prioritized based on the completion deviation of the scene data under the first scenario to obtain a priority ranking result of the first data tags; wherein, the larger the completion deviation, the higher the corresponding priority; the completion deviation is determined based on the completion ratio of the gap scene data and the target ratio of the gap scene data.

[0079] In yet other embodiments, Figure 2 A schematic diagram of an implementation flow of prioritizing the first data tag provided in an embodiment of the present application is shown as follows: Figure 2 As shown, the first data tags may be prioritized through the following steps 201 to 203 to obtain a priority ranking result of the first data tags:

[0080] Step 201: Determine a first priority weight of the first data tag; wherein the first priority weight is used to characterize a completion deviation of the scene data corresponding to the first data tag;

[0081] Step 202: Determine a second priority weight of the first data tag based on the collection period of the gap scenario data corresponding to the first data tag; wherein the collection period is inversely proportional to the second priority weight;

[0082] Step 203: Prioritize the first data tags according to the first priority weight and the second priority weight to obtain a priority ranking result of the first data tags.

[0083] It is understood that in the data collection method provided in the embodiments of the present application, the first data tags are prioritized based on the completion deviation of the scene data corresponding to the first data tag and the collection deadline of the missing scene data corresponding to the first data tag, thereby obtaining a priority ranking result for the first data tags. This is beneficial for enabling the vehicle to prioritize the collection of missing scene data with large gaps and urgent deadlines, thereby improving the intelligent level of data collection and enhancing the user experience.

[0084] It should be understood that in the embodiments of the present application, there is no limitation on the first priority weight, and the first priority weight is a numerical indicator used to measure the completion of the current gap scene data relative to the preset target completion. The first priority weight reflects the degree of gap in the gap scene data. In some embodiments, the completion deviation is used to represent the gap between the current amount of collected data and the target amount of data. In other embodiments, the completion deviation is used to represent the gap between the current proportion of collected data and the target proportion. For example, in one possible implementation, if the target collection amount of the curve scene is 1 million, and only 500,000 have been completed so far, the completion deviation of the curve scene is larger, and at this time, the first priority weight corresponding to the first data tag of the curve scene is higher.

[0085] In this embodiment of the present application, the second priority weight is not limited and is a numerical indicator used to characterize the urgency of collecting the gap scenario data. The collection deadline refers to the time window in which the gap scenario data must be collected. For example, data on 1 million curved road sections must be collected within this week. The closer the collection time of a particular gap scenario data is to the deadline, the more urgent the need to collect the gap scenario data. Therefore, the second priority weight of the first data tag corresponding to the gap scenario data is higher.

[0086] In some embodiments, the first data tags are prioritized according to the first priority weight and the second priority weight to obtain a priority ranking result of the first data tags, including: determining a first product of the first priority weight and a first coefficient; determining a second product of the second priority weight and a second coefficient; adding the first product and the second product to obtain a first numerical value; and sorting the first numerical values ​​corresponding to each of the first data tags to obtain a priority ranking result of the first data tags.

[0087] It should be understood that in the embodiments of the present application, the first coefficient and the second coefficient are not limited. In some embodiments, the first coefficient and the second coefficient are preset based on empirical values. In other embodiments, the first coefficient and the second coefficient are obtained by fitting.

[0088] It should be understood that in the embodiment of the present application, the collection period of the gap scenario data corresponding to the first data tag can be understood as the collection period of the first scenario data under the corresponding first scenario.

[0089] Furthermore, in some embodiments, Figure 3 A schematic diagram of an implementation flow for determining a first priority weight is provided in an embodiment of the present application, such as Figure 3As shown, the first priority weight of the first data tag can be determined through the following steps 301 to 304:

[0090] Step 301, determining a first ratio between the collected amount of the first data tag and the target amount of the first data tag;

[0091] Step 302, determining a second ratio between a target amount of the first data tag and a total target amount of the plurality of first scene data;

[0092] Step 303: Determine the data quality of the scene data corresponding to the first data tag;

[0093] Step 304: Determine a first priority weight of the first data tag based on the first ratio, the second ratio, and the data quality.

[0094] It will be appreciated that in the data collection method provided in the embodiments of the present application, a first priority weight is determined based on the completed ratio of scene data for a first scenario (i.e., the first ratio), the target ratio of scene data for the first scenario (i.e., the second ratio), and the data quality of the scene data for the first scenario. This helps identify collection gaps, importance, and data quality of the scene data corresponding to the first data tag, thereby making the priority sorting results more reasonable, thereby improving the intelligent level of data collection and enhancing the user experience.

[0095] It should be understood that in the embodiments of the present application, if the first ratio is low, it indicates that there is a large gap in the gap scenario data corresponding to the first ratio. Therefore, the cloud server can assign a higher collection priority to the first data tag corresponding to the gap scenario data. In this way, by introducing the first ratio, the cloud server can dynamically identify the extent of the gap in the gap scenario data. In this way, the cloud server can dispatch vehicles to the relevant areas to collect the gap scenario data in a targeted manner, which is beneficial for ensuring that the data distribution is basically consistent with expectations.

[0096] In the embodiment of the present application, the introduction of the second ratio helps to balance the collection ratio of various scene data on a global scale, avoids the neglect of certain high-value but low-frequency scenes, and helps to ensure the diversity and coverage of the collected scene data. For example, if the target number of first data labels is 1 million, and the total target number of all first scene data is 10 million, the second ratio is 0.1. This calculation result indicates that the first data label accounts for a small proportion of the entire data set and may belong to a scarce scene.

[0097] For example, in one possible implementation, the cloud server uses an engine such as Structured Query Language (SQL) or MapReduce to count the data quantity of each tag combination in real time and calculate the current actual proportion of data for each tag. Based on the current actual proportion, a scheduling weight function is constructed:

[0098] (1)

[0099] in, (i.e., an example of the first priority weight) represents the acquisition priority weight of the i-th category scene, (i.e., an example of the second ratio) represents the target ratio of the i-th type of scene, (i.e. an example of the first ratio) represents the current actual proportion, (i.e., an example of data quality) represents the quality assessment value of this type of data. and These are system configurable parameters.

[0100] Then, the system will calculate the difference between the current amount and the target value. and their urgency (such as the collection task deadline), to generate a scenario collection gap ranking (i.e., an example of a priority sorting result).

[0101] In some embodiments, determining the target location of the collectible scene data includes: determining the target location of the gap scene data corresponding to the first data tag in the map data based on the first data tag corresponding to the gap scene data.

[0102] It is understood that in the data collection method provided in the embodiments of the present application, the first data tag corresponding to the gap scene data is used to match the map data to obtain the target location of the gap scene data. In this way, the target location of the gap scene data is determined in a targeted manner, which is beneficial for quickly and accurately determining the target location.

[0103] It should be understood that in the embodiment of the present application, the second data tag corresponding to the gap scene data is used as the first data tag.

[0104] Furthermore, in some embodiments, Figure 4 A schematic diagram of an implementation process for determining a target location provided in an embodiment of the present application Figure 1 ,like Figure 4 The target position of the gap scene data corresponding to the first data tag can be determined in the map data based on the first data tag corresponding to the gap scene data through the following steps 401 to 403:

[0105] Step 401: construct a first feature set of gap scene data corresponding to the first data tag;

[0106] Step 402: determining a second feature set of the map data; the second feature set is obtained by parsing geometric information of roads in the map data, topological relationships of road segments in the map data, and identification information in the map data;

[0107] Step 403: Match the first feature set and the second feature set, and determine the target position of the gap scene data corresponding to the first data tag in the map data.

[0108] It can be understood that in the data collection method provided in the embodiments of this application, by constructing a first feature set for gap scene data corresponding to a first data tag and matching it with a second feature set in the map data, target locations from which gap scene data can be collected can be efficiently identified in the map data. This allows for precise positioning of target locations from which gap scene data can be collected, thereby improving vehicle collection efficiency.

[0109] It should be understood that in the embodiments of the present application, the first feature set is not limited. In some embodiments, the first feature set refers to a set of structured features extracted based on the scene type described by the first data tag (such as a curve, a signalized intersection, etc.), and the first feature set is used to represent the typical attributes and semantic information of the gap scene data corresponding to the first feature set. For example, for the label of the curve scene type, the first feature set may include physical attributes such as the road curvature range, the length of the road section, and whether there are guardrails; while for the label of the signalized intersection scene type, it may include topological and identification information such as the number of node connections, traffic light configuration, and intersection shape.

[0110] For example, in one possible implementation, the first feature set is obtained by the cloud server through a preset scene feature dictionary and stored in a structured form. The cloud server stores the first feature set for subsequent matching with the second feature set in the map data.

[0111] It should be understood that in the embodiments of this application, the second feature set is not limited. In some embodiments, the second feature set refers to a set of structured features extracted from map data that reflect the actual road environment, including road geometry (such as coordinate point sequence, road width, and slope), topological relationships between road sections (such as road connection methods and intersection types), and road identification information (such as traffic signs, signal light types, speed limit information, etc.). All structured feature information in the second feature set is typically derived from a high-precision map database or open map service interface and is preprocessed into a unified data structure to facilitate efficient calculation and comparison by the system.

[0112] For example, in one possible implementation, the cloud server traverses all road segments in the map data and extracts their geometric attributes (e.g., polylines constructed from GPS coordinates, road curvature), topological attributes (e.g., road junctions, intersection types), and identification attributes (e.g., traffic light locations, number of lanes). The cloud server then combines these attributes into a second feature set that matches the data structure of the first feature set. By constructing this second feature set, the cloud server can fully capture the true state and structural characteristics of the roads in the map data, providing the foundation for subsequent matching of the second feature set with the first feature set, thereby improving the accuracy and reliability of scene positioning.

[0113] In some embodiments, the first feature set and the second feature set have the same format, which is beneficial for matching the first feature set with the second feature set, thereby determining the target location of the gap scene data corresponding to the first data tag in the map data.

[0114] It should be understood that in the embodiments of the present application, the specific implementation method for matching the first feature set with the second feature set is not limited. In some embodiments, a feature similarity algorithm can be used to match the first feature set with the second feature set to determine the geographical area closest to the gap scene data corresponding to the first data tag as the target location of the gap scene data; wherein the feature similarity algorithm includes but is not limited to at least one of the following: Euclidean distance, cosine similarity, fuzzy matching, etc.

[0115] In some embodiments, determining the target position of the scene data corresponding to the first data tag in the map data includes: determining the candidate target position of the scene data corresponding to the first data tag in the map data; and determining the target position from the candidate target position that meets the first condition.

[0116] In some embodiments, Figure 5 A schematic diagram of an implementation process for determining a target location provided in an embodiment of the present application Figure 2 ,like Figure 5 The target location for collecting scene data can be determined by the following steps 501 to 503:

[0117] Step 501: Based on the second data tag of the first scene data, determine the geographical location of the first scene data corresponding to the second data tag in the map data;

[0118] Step 502: Mark the corresponding second data tag at the geographic location to obtain a scene distribution map;

[0119] Step 503: Based on the first data tag corresponding to the gap scene data, a query is performed on the scene distribution map to determine the target location of the gap scene data corresponding to the first data tag.

[0120] It can be understood that in the data collection method provided in the embodiments of the present application, the geographic location of the first scene data corresponding to the second data tag is first determined in the map data based on the second data tag; the second data tag is annotated at the geographic location of the first scene data corresponding to the second data tag to obtain a scene distribution map; thus, when the target location corresponding to the gap scene data is determined, the target location corresponding to the gap scene data can be queried in the scene distribution map based on the first data tag. This is beneficial to improving the efficiency of determining the target location corresponding to the gap scene data, and further beneficial to improving the efficiency of data collection.

[0121] In some embodiments, determining the target position of the first scene data corresponding to the second data tag in the map data includes: constructing a third feature set of the first scene data corresponding to the second data tag; determining a second feature set of the map data; wherein the second feature set is obtained through the geometric information of the roads in the map data, the topological relationship of the road sections in the map data and the identification information in the map data; matching the third feature set and the second feature set to determine the target position of the first scene data corresponding to the second data tag in the map data.

[0122] It should be understood that in the embodiments of the present application, the labeling is not limited. In some embodiments, the labeling refers to adding a second data label to a corresponding location in the map data in a visual or structured manner. For example, a curve label is added to the location of a curve on the map, or a signalized intersection label is added to the location of a signalized intersection.

[0123] In the embodiments of the present application, the scene distribution map is not limited. In some embodiments, the scene distribution map refers to a visual map that integrates map data and scene labels, and the scene distribution map shows the distribution of various scenes in geographic space. The scene distribution map can serve as an important reference for data gap identification and path planning. It can be understood that in the embodiments of the present application, by intuitively presenting the distribution of various scenes on the scene distribution map, it can assist in the subsequent target location query and / or navigation path planning of the gap scene data, thereby further improving the degree of automation and intelligence of data collection.

[0124] In some embodiments, querying the scene distribution map to determine the target location of the gap scene data corresponding to the first data tag includes: querying the scene distribution map to determine the candidate target location of the gap scene data corresponding to the first data tag; and determining the target location from the candidate target location that meets the first condition.

[0125] It is understood that in the data collection method provided in the embodiment of the present application, rather than taking any candidate target location as the target location, a target location that meets the condition is selected from the candidate target locations based on the first condition. This is beneficial to further improve the intelligent level of data collection and enhance the user experience.

[0126] It should be understood that in the embodiments of the present application, the candidate target locations are not limited. In some embodiments, the candidate target locations refer to geographical areas that are preliminarily screened out in the scene distribution map and may meet the needs of collecting gap scene data. The candidate target locations are usually obtained by querying the first data tag corresponding to the gap scene data in the scene distribution map. The candidate target locations may be distributed in multiple cities or regions.

[0127] In the embodiments of the present application, the first condition is not limited. In some embodiments, the first condition can be determined based on user input. In other words, the user can set the first condition based on the current data collection situation. In other embodiments, the first condition is preset based on user needs or task requirements.

[0128] Furthermore, in some embodiments, the first condition includes, but is not limited to, at least one of the following: low traffic density, high collection frequency, easy passage, etc. It should be understood that in the embodiments of the present application, if the task requires collecting scene data with high traffic density or pedestrian flow, the first condition may also be: high traffic density and / or difficult passage.

[0129] It should be understood that in the embodiments of the present application, the target location is a specific geographical location that has been screened and confirmed and will be included in the acquisition path planning. In some embodiments, the target location is represented in the form of coordinate points or geo-fences and serves as the basis for subsequent navigation path generation and vehicle scheduling.

[0130] For example, in one possible implementation, when the cloud server identifies insufficient data for a certain type of scene, it identifies it as the first scene and searches the scene distribution map for candidate areas that match the first data tag corresponding to the missing scene data for the first scene. For example, if the data gap for unsignalized intersections is large, all roads at unsignalized intersections in the scene distribution map will be retrieved and identified as candidate target locations. Subsequently, the candidate target locations are evaluated, taking into account factors such as traffic conditions and data collection feasibility, and ultimately the most suitable locations are selected as the official task execution areas.

[0131] In some embodiments, the method further includes: sending a first strategy to the target vehicle; the first strategy is used to instruct the target vehicle to prioritize collecting gap scene data corresponding to the third data tag when there is a target position corresponding to the third data tag on the navigation path; wherein the priority of the third data tag is lower than the first data tag corresponding to the target position of the navigation path.

[0132] It is understood that in the data collection method provided in the embodiments of this application, if a target location on the navigation path contains missing scene data corresponding to a low-priority first data tag, the missing scene data corresponding to the low-priority first data tag will be collected first. This helps improve collection efficiency, reduce vehicle operating costs, and further enhance the intelligent level of data collection.

[0133] It should be understood that in the embodiments of the present application, the first strategy is not limited. In some embodiments, the first strategy is an intelligent collection control instruction issued to a target vehicle, which is used to instruct the vehicle, when navigating based on a navigation path, to prioritize the collection of the missing scene data corresponding to the low-priority first data tag if there is a target location on the navigation path that has missing scene data corresponding to the low-priority first data tag.

[0134] It should be understood that in the embodiment of the present application, the second data tag is not limited. In some embodiments, the priority of the second data tag is lower than the first data tag of the gap scene data corresponding to the target position of the navigation path.

[0135] In step 104, continue to obtain the next total collection amount under the first scene, and perform the next collection of the next gap scene data for the first scene until the final total collection amount of the corresponding scene data under the first scene is not less than the corresponding target amount, and obtain the second scene data under the first scene; the second scene data is all collected scene data for each scene.

[0136] In step 105, a model for realizing intelligent driving of a vehicle is obtained by training using the second scenario data in different scenarios.

[0137] It should be understood that in the embodiment of the present application, for a scenario in which the current total collection amount corresponding to the first scenario data is not less than the corresponding target amount, the second scenario data is the first scenario data for that scenario. For a scenario in which the current total collection amount corresponding to the first scenario data is less than the corresponding target amount, the second scenario data is the first scenario data and the gap scenario data for the first scenario.

[0138] It should be understood that in the embodiments of the present application, the model for realizing intelligent driving of the vehicle is not limited. In some embodiments, the model for realizing intelligent driving of the vehicle refers to an intelligent driving decision model built based on deep learning or other machine learning algorithms, such as target detection, behavior prediction, path planning and other modules. The second scene data provides rich and diverse training samples. These training samples cover scene data of various road types, multiple time periods and various environmental factors. This diverse content helps to improve the generalization ability and robustness of the model for realizing intelligent driving of the vehicle.

[0139] For example, in one possible implementation, the cloud server utilizes the second scenario data and continuously optimizes the model parameters of the model used to realize the intelligent driving of the vehicle through supervised learning or reinforcement learning, thereby improving the adaptability of the model used to realize the intelligent driving of the vehicle in complex traffic environments, and providing strong support for the safety and reliability of the intelligent driving system.

[0140] The following describes an exemplary application of the embodiments of the present application in a practical application scenario.

[0141] An embodiment of the present application provides a data collection method (i.e., an example of a data collection method) for collecting multi-dimensional scene data based on a preset label ratio and combining it with the execution of an automatic scheduling strategy, which is suitable for data training, modeling, and optimization scenarios of advanced driver assistance and autonomous driving systems.

[0142] With the rapid development of advanced driver assistance systems and autonomous driving technologies, automakers are increasingly demanding massive, accurate, and diverse data. Vehicle data collection systems typically rely on traditional methods such as fixed-frequency acquisition, fixed-point acquisition, or manual collection. This unstructured, unstrategic data collection approach can easily lead to uneven data distribution, resulting in a lack of data for certain typical scenarios, which in turn affects the integrity and generalization capabilities of subsequent model training.

[0143] The present application provides an automatic scheduling data collection method based on scene tag distribution (i.e., an example of a data collection method). This method is beneficial for resolving issues such as low efficiency in manual task issuance, redundant collection content, and insufficient coverage of specific scenarios in related data collection technologies.

[0144] The present application provides an automated, closed-loop data acquisition and scheduling system. In some embodiments, acquisition control is achieved through the following steps:

[0145] Step 11: The vehicle side (i.e., an example of a vehicle) automatically labels the scene and uploads the data;

[0146] Step 12: The cloud (i.e., an example of a cloud server) counts the proportion of data from various scenarios, compares it with the preset collection targets, and automatically identifies collection gaps;

[0147] In step 13, the cloud uses map semantic analysis technology to locate the geographic location of the scene corresponding to the missing label (such as a winding road or an unsignaled intersection). Based on the results, the navigation path and strategy parameters are collected and packaged into a task and sent to the target vehicle.

[0148] Step 14: The vehicle goes to the designated area to collect gap data based on the strategy and navigation information, and automatically pauses collection in non-target scenarios;

[0149] Step 15: All data is uploaded in real time, and closed-loop optimization is continuously iterated.

[0150] It can be understood that in the embodiments of the present application, the method can achieve the following beneficial effects: (1) realize the automated process from label analysis to task issuance; (2) accurately schedule collection based on data label gaps and map semantic positioning; (3) improve data collection efficiency and reduce data redundancy; (4) have strong scalability and multi-city deployment capabilities; (5) reduce manual intervention, and the system is suitable for data needs in multiple scenarios such as advanced driver assistance, autonomous driving simulation, and simulation training.

[0151] In some embodiments, the technical solution described in the embodiments of the present application includes the following steps 21 to 28:

[0152] Step 21, label definition and target configuration: used to set the multi-dimensional label combination to be collected and its quantity target;

[0153] Step 22, data collection vehicles and labeling: Each vehicle labels the data during the collection process and uploads it to the cloud in real time;

[0154] Step 23: Real-time statistics and gap analysis: The cloud performs real-time statistics on the collected data by tag combination and compares it with the target ratio to identify collection gaps.

[0155] Step 24, data gap information generation: dynamically generate a new collection strategy package based on the gap situation, including a collection / stop tag combination;

[0156] Step 25: Map fusion and path planning: Combine the electronic map with the scene distribution model to determine the possible areas where the missing scenes may appear;

[0157] Step 26, dispatch control: dispatch appropriate vehicles to the target area for collection based on vehicle status, location, and collection requirements;

[0158] Step 27, vehicle-side execution: collect vehicle analysis strategy package and collect data only when it meets the target scenario;

[0159] Step 28, feedback and self-iteration module: upload the collected data for the cloud to update the statistical results, and realize closed-loop control until the collected data meets the collection target quantity.

[0160] In this embodiment of the present application, several data collection terminals (i.e., an example of a vehicle) are deployed nationwide. Each data collection vehicle (i.e., an example of a vehicle) is equipped with multi-source sensing equipment, including automotive-grade high-definition cameras, lidar, Global Navigation Satellite System (GNSS) / Inertial Measurement Unit (IMU) navigation modules, Controller Area Network (CAN) bus reader modules, millimeter-wave radar, ultrasonic sensors, etc. The data collection software system is deployed in the vehicle's edge computing unit and has localized image processing, scene recognition, and policy-driven capabilities. It can complete data preprocessing, label recognition, and policy execution at the edge, significantly reducing the computing burden in the cloud.

[0161] In some embodiments, an automated data collection and scheduling method based on multi-dimensional scene label ratio control can be implemented through the following steps 31 to 37.

[0162] Step 31: System initialization and label definition

[0163] In the early stages of system deployment, the cloud platform (an example of a cloud server) pre-establishes a multi-dimensional scenario tag library. This tag library includes but is not limited to the following dimensions:

[0164] Road type: expressways, urban expressways, national roads, provincial roads, rural roads, etc.;

[0165] Scenario types: straight road sections, curved road sections, unsignalized intersections, signalized intersections, roundabouts, etc.

[0166] Time period: daytime, nighttime, dusk, early morning;

[0167] Environmental factors (optional): weather type (sunny, rainy, snowy, foggy), traffic density, etc.

[0168] System operators can set the target collection volume through the configuration interface, as shown in Table 1.

[0169] The tag target information is stored in the cloud database and serves as the basis for scheduling data collection tasks.

[0170] Table 1: Label items and target collection volume corresponding to some scenarios

[0171]

[0172] Step 32: Data collection vehicle deployment and upload mechanism

[0173] The vehicle-side acquisition device (i.e., a vehicle) is an advanced acquisition terminal equipped with a positioning module, camera system, IMU, CAN interface, and AI computing unit. All acquisition vehicles are deployed with a unified software system and have the following capabilities:

[0174] (1) Real-time scene recognition of the current driving environment, collecting scene data when the scene is the required target label, and labeling the collected data according to the aforementioned labeling system;

[0175] (2) Support local cache and real-time upload mode. When the network is good, the data is uploaded to the cloud server in real time.

[0176] (3) It has the ability to accept and analyze strategies, and can automatically switch to the collection mode after receiving the new strategy package sent by the cloud server.

[0177] Step 33: Data Statistics and Gap Identification Process

[0178] The cloud platform (an example of a cloud server) deploys a real-time data statistical analysis and gap identification module, which continuously parses labels and summarizes the data uploaded by all vehicles to generate a statistical table, as shown in Table 2.

[0179] Table 2: Cloud data distribution

[0180]

[0181] The cloud server uses the engine to count the data quantity of each tag combination in real time and calculate the current actual proportion of the tag data. The scheduling weight function is constructed based on the current actual proportion:

[0182] (2)

[0183] in, (i.e., an example of the first priority weight) represents the acquisition priority weight of the i-th category scene, (i.e., an example of the second ratio) represents the target ratio of the i-th type of scene, (i.e. an example of the first ratio) represents the current actual proportion, (i.e., an example of data quality) represents the quality assessment value of this type of data. and These are system configurable parameters.

[0184] Then, the system will calculate the difference between the current amount and the target value. and their urgency (e.g., collection task deadline), generate a scenario collection gap ranking (i.e., an example of a priority sorting result), including:

[0185] (1) Stop collecting labels: such as "straight road";

[0186] (2) Prioritize label collection: such as "curve" and "signalized intersection";

[0187] (3) Time requirements: such as “add 200,000 bends within this week”;

[0188] (4) Recommended route type: such as “prioritize mountainous cities and old urban areas”;

[0189] (5) Road section filtering conditions: such as “avoid covered areas” or “avoid traffic congestion periods”.

[0190] The scene collection gap ranking information is presented in the form of a structured file. The storage format supports JavaScript Object Notation (JSON), Protocol Buffers (Protobuf), etc., and together with subsequent target scene location and other information, it constitutes the collection strategy.

[0191] Step 34: Map fusion and path planning

[0192] After the data gap information is generated, the automatic scheduling module will map the required label scenarios to the real geographical area and generate collection strategy information based on the vehicle formation information. The collection strategy information is sent to the vehicle end in the form of a strategy package. This process is mainly based on the following technologies:

[0193] Step 34.1: Scene semantic recognition in map information

[0194] Use high-precision maps or open maps, such as the AutoNavi Application Programming Interface (API), Baidu Maps API, and Open Street Map (OSM); these maps contain structured data such as road attributes, geometry, traffic facilities, slopes, and road markings.

[0195] Build a Scene Pattern Dictionary: For example, a "curve road section" can be defined as a road section whose geometric curvature is higher than a certain threshold and whose length is not less than X meters; a "roundabout" can be defined as multiple road connection points forming a closed circular topology; and an "unsignaled intersection" is defined as one with ≥3 node connections and no signal light attributes.

[0196] Analyze the geometric information and topological relationships in the map: use the GPS coordinates between road nodes to calculate curvature; extract the connection relationship between road sections to build the road graph structure; use traffic rules or traffic sign fields to confirm scene semantics.

[0197] Step 34.2, geolocation retrieval for tag-scarce scenarios

[0198] (1) Obtain statistical gaps in current label data (e.g., insufficient samples for “nighttime + roundabout + rural roads”).

[0199] (2) Construct multi-dimensional query conditions: Based on the scene feature dictionary, match road segments that meet specific semantic tags; add time conditions (such as nighttime light intensity or use historical data) and area type (rural, urban); support fuzzy matching and multi-condition cross-screening.

[0200] (3) Clustering or scoring sorting: Find a group of locations with the highest matching degree (such as urban road sections with multiple roundabouts within a radius of 5 km); the sorting criteria may include: low traffic density, low collection frequency, easy access, etc.

[0201] Step 34.3: Generate navigation route and send it to target vehicle

[0202] (1) Multi-point path planning (based on map API or self-developed navigation engine): Input the target area, scene point coordinates, and current vehicle location; plan the shortest path that "prioritizes passing through the target scene."

[0203] (2) Scene-based navigation: Multiple target points are issued as Waypoints, and the navigation path is automatically collected in sequence by the vehicle-side system; where Waypoint represents the coordinates of a pre-set address location, combined with the existing scene recognition module on the vehicle side, it can achieve "walking to the scene and automatically starting collection".

[0204] (3) Send the navigation information and strategy package to the target vehicle: the target vehicle is the collection vehicle deployed in the city where the navigation location is located.

[0205] Step 35: Vehicle-side strategy execution mechanism

[0206] After receiving the policy package, the vehicle drives to the target area under navigation guidance. The acquisition system will match the output of the tag recognition module with the "target tag" field in the policy.

[0207] If the current scene matches the target tag, the acquisition system is started, and the data is labeled and uploaded;

[0208] If the current scene does not match, the collection will stop and only navigation will be carried out.

[0209] This mechanism ensures that data is collected only in high-value scenarios, saving storage bandwidth and improving the effective data ratio.

[0210] In addition, the vehicle supports logging of collection status. If a certain section of collection fails or is interrupted (such as GPS loss), the system will automatically recollect or report to the cloud for supplementary scheduling.

[0211] Step 36: Data Feedback and Closed-Loop Optimization

[0212] All data is uploaded to the cloud via edge servers or the public network. The upload process includes the following metadata: data item and corresponding tag; geographic coordinates and timestamp; vehicle identification number (VIN); execution task ID; and execution status (successful or aborted).

[0213] The cloud updates the label statistics table accordingly and enters the next round of gap identification and strategy update process, completing the closed loop of the entire process.

[0214] Step 37: Exception handling and manual intervention support

[0215] To ensure system stability, the cloud platform is designed with the following exception handling mechanisms:

[0216] If a persistent gap in the collection of a certain type of label is found, the system will automatically mark it as a "high-risk label" and push it for manual review;

[0217] Support operators to manually modify strategies, intervene in scheduling, and assign targeted collection tasks through the platform;

[0218] All operation logs are retained to support task traceability and responsibility analysis.

[0219] This application provides a novel, highly automated data collection system that achieves closed-loop self-optimization of the data collection process through multi-dimensional tag ratio settings and a collaborative mechanism between the cloud and vehicle. Intelligent fleet scheduling, implemented through a dispatch control module, not only addresses issues such as uneven data collection distribution and slow policy response in traditional systems but also reduces the need for manual intervention, offering broad commercial application prospects and engineering value.

[0220] Figure 6 A schematic diagram of the implementation process of a data collection method provided in an embodiment of the present application Figure 2 ,like Figure 6 As shown, the method includes the following steps 601 to 619:

[0221] Step 601: The cloud data pool receives data uploaded by the vehicle.

[0222] Step 602: Analyze the data in the cloud.

[0223] Step 603: Preset tags in the cloud;

[0224] Step 604: The cloud obtains the target configuration;

[0225] Step 605: The cloud determines data gap information;

[0226] Step 606: Obtain map information from the cloud;

[0227] Step 607: The cloud obtains vehicle formation information;

[0228] Step 608: The cloud sends the data gap information, map information, and vehicle formation information to the automatic dispatch module;

[0229] Step 609: The cloud generates collection strategy information;

[0230] Step 610: The cloud obtains the policy package and sends it to the vehicle.

[0231] Step 611: The vehicle obtains navigation information;

[0232] Step 612: The vehicle obtains the route;

[0233] Step 613: the vehicle moves to the target scene area;

[0234] Step 614: The vehicle obtains the collection strategy;

[0235] Step 615: The vehicle side analyzes the collection strategy;

[0236] Step 616: The vehicle obtains the collection priority;

[0237] Step 617: The vehicle side decides whether to collect the data;

[0238] Step 618: The vehicle collects data;

[0239] In step 619, the vehicle labels the data scene and uploads the data to the cloud.

[0240] Figure 7 A schematic diagram of the implementation process of a data collection method provided in an embodiment of the present application Figure 3 ,like Figure 7 As shown, the process includes the following steps 701 to 713:

[0241] Step 701: The cloud receives data uploaded by the vehicle.

[0242] Step 702: Cloud statistics tag data pool;

[0243] Step 703: Compare the collected targets in the cloud;

[0244] Step 704: The cloud identifies the missing tag class;

[0245] Step 705: The cloud generates a new collection strategy;

[0246] Step 706: The cloud calls the map semantic engine to obtain the location of the gap scene;

[0247] Step 707: The map semantic engine searches the map based on the gap tag feature.

[0248] Step 708: The map semantic engine returns the target scene location;

[0249] Step 709: The cloud sends the collection task;

[0250] Step 710: The vehicle obtains navigation information and strategy package;

[0251] Step 711: The vehicle-side collects data while the vehicle is driving and starts the collection task;

[0252] Step 712: The vehicle recognizes and labels the scene in real time;

[0253] Step 713: The vehicle uploads the tagged data.

[0254] Figure 8 A schematic diagram of the structure of a policy package provided in an embodiment of the present application is shown in FIG. Figure 8As shown, the strategy package 80 includes: a header 801, a collection priority 802, a map location 803, a collection behavior 804 and a reserved field 805; wherein, the header 801 includes the strategy package ID, the time of issuance, the target vehicle number, and the validity period; the collection priority 802 includes: the collection priority and the collection quantity of each scene tag; the map location 803 includes: a list of recommended collection areas and an estimated route corresponding to each area; the collection behavior 804 includes: when to collect, when to pause collection, and special behavior prompts.

[0255] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, ordinary technicians in this field should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions of the present application, which should be included in the scope of protection of the present application.

[0256] It should be noted that although the steps of the method in the present application are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.; or, the steps in different embodiments may be combined into a new technical solution. Based on the aforementioned embodiments, the embodiments of the present application provide a device, which includes the modules included and the units included in each module, which can be implemented by a processor; of course, it can also be implemented by a specific logic circuit; in the implementation process, the processor can be an AI acceleration engine (such as NPU, etc.), a graphics processing unit (GPU), a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP) or a field programmable gate array (FPGA), etc.

[0257] Figure 9 A schematic diagram of a data collection device provided in an embodiment of the present application is shown in FIG. Figure 9 As shown, the data collection device 90 includes: a summary module 901, a determination module 902, a sending module 903, an iteration module 904 and a training module 905; wherein,

[0258] Aggregation module 901 is configured to aggregate first scene data collected by at least one vehicle to obtain a current total amount of first scene data collected in different scenes; the first scene data is used to represent road information and / or environmental information;

[0259] A determination module 902 is configured to determine, in different scenarios, a first scenario in which the current total collection amount corresponding to the first scenario data is less than the corresponding target amount;

[0260] The sending module 903 is configured to determine a target location for collecting scene data for the first scene; send the target location to a target vehicle, so that the target vehicle collects the gap scene data corresponding to the first scene at the target location; wherein the target vehicle includes vehicles in the area where the target location is located;

[0261] Iteration module 904 is configured to continue acquiring the next total collection amount for the first scenario and perform the next collection of the next gap scenario data for the first scenario until the final total collection amount of the scene data corresponding to the first scenario is not less than the corresponding target amount, thereby obtaining the second scenario data for the first scenario; the second scenario data is all the collected scene data for each scenario;

[0262] The training module 905 is configured to use the second scenario data under different scenarios to train a model for realizing intelligent driving of the vehicle.

[0263] In some embodiments, the device further includes: a sorting module; wherein the sorting module is configured to prioritize the first data tag when the type of the first data tag corresponding to the scene data under the first scene is greater than a first value, and obtain a priority sorting result of the first data tag; the first value is greater than or equal to 1; the sending module 903 is also configured to send the priority sorting result to the target vehicle, so that the target vehicle plans a navigation path based on the priority sorting result and the target position, so as to navigate to the target position based on the navigation path to collect the gap scene data corresponding to the corresponding first data tag.

[0264] In some embodiments, the sorting module is configured to determine a first priority weight of the first data tag; wherein the first priority weight is used to characterize the completion deviation of the scene data corresponding to the first data tag; determine the second priority weight of the first data tag based on the collection period of the gap scene data corresponding to the first data tag; wherein the collection period is inversely proportional to the second priority weight; and prioritize the first data tag based on the first priority weight and the second priority weight to obtain a priority sorting result of the first data tag.

[0265] In some embodiments, the sorting module is configured to determine a first ratio between the collection quantity of the first data tag and the target quantity of the first data tag; determine a second ratio between the target quantity of the first data tag and the total target quantity of multiple first scene data; determine the data quality of the scene data corresponding to the first data tag; and determine a first priority weight of the first data tag based on the first ratio, the second ratio and the data quality.

[0266] In some embodiments, the sending module 903 is configured to determine, in the map data, a target location of the gap scene data corresponding to the first data tag based on the first data tag corresponding to the gap scene data.

[0267] In some embodiments, the sending module 903 is configured to construct a first feature set of the gap scene data corresponding to the first data tag; determine a second feature set of the map data; wherein the second feature set is obtained by parsing the geometric information of the roads in the map data, the topological relationship of the road sections in the map data and the identification information in the map data; match the first feature set and the second feature set to determine the target position of the gap scene data corresponding to the first data tag in the map data.

[0268] In some embodiments, the sending module 903 is configured to determine the geographical location of the first scene data corresponding to the second data tag in the map data based on the second data tag of the first scene data; mark the corresponding second data tag at the geographical location to obtain a scene distribution map; and query the scene distribution map based on the first data tag corresponding to the gap scene data to determine the target location of the gap scene data corresponding to the first data tag.

[0269] In some embodiments, the sending module 903 is configured to query the scene distribution map to determine the candidate target location of the gap scene data corresponding to the first data tag; and determine the target location as the candidate target location that meets the first condition.

[0270] In some embodiments, the sending module 903 is further configured to send a first strategy to the target vehicle; the first strategy is used to instruct the target vehicle to give priority to collecting the gap scene data corresponding to the third data tag when there is a target position corresponding to the third data tag on the navigation path; wherein, the priority of the third data tag is lower than the first data tag corresponding to the target position of the navigation path.

[0271] The description of the above device embodiment is similar to the description of the above method embodiment and has similar beneficial effects as the method embodiment. For technical details not disclosed in the device embodiment of this application, please refer to the description of the method embodiment of this application for understanding.

[0272] It should be noted that the division of modules in the embodiments of the present application is schematic and is merely a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units. They may also be implemented in the form of a combination of software and hardware.

[0273] It should be noted that in the embodiments of the present application, if the above-mentioned method is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present application, or the part that contributes to the relevant technology, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling the vehicle to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk or an optical disk. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0274] The embodiment of the present application provides a cloud server, Figure 10 A schematic diagram of the structure of the cloud server provided in the embodiment of the present application is shown in FIG. Figure 10 As shown, the cloud server 100 includes a memory 1001 and a processor 1002. The memory 1001 stores a computer program that can be run on the processor 1002. When the processor 1002 executes the program, the steps in the method provided in the above embodiment are implemented.

[0275] It should be noted that the memory 1001 is configured to store instructions and applications executable by the processor 1002, and can also cache data to be processed or processed by various modules in the processor 1002 and the cloud server 100 (for example, image data, audio data, voice communication data and video communication data), which can be implemented through flash memory (FLASH) or random access memory (RAM).

[0276] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method provided in the above embodiment are implemented.

[0277] An embodiment of the present application provides a computer program product containing instructions, which, when executed on a computer, enables the computer to execute the steps of the method provided in the above method embodiment.

[0278] It should be noted that the descriptions of the above storage medium and vehicle embodiments are similar to those of the above method embodiments and have similar beneficial effects as the method embodiments. For technical details not disclosed in the storage medium and vehicle embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0279] It should be understood that "one embodiment" or "an embodiment" or "some embodiments" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The above-mentioned serial numbers of the embodiments of the present application are for description only and do not represent the advantages and disadvantages of the embodiments. The above description of the various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other. For the sake of brevity, they will not be repeated here.

[0280] The term "and / or" in this article is only a description of the association relationship between associated objects, indicating that there can be three relationships. For example, object A and / or object B can mean: object A exists alone, object A and object B exist at the same time, and object B exists alone.

[0281] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a..." does not preclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0282] In the several embodiments provided in this application, it should be understood that the disclosed vehicles and methods can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of the modules is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the vehicle or module can be electrical, mechanical or other forms.

[0283] The modules described above as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules; they may be located in one place or distributed across multiple network units; some or all of the modules may be selected according to actual needs to achieve the purpose of this embodiment.

[0284] In addition, all functional modules in the embodiments of the present application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the above-mentioned integrated modules can be implemented in the form of hardware or in the form of hardware plus software functional units.

[0285] Those skilled in the art will understand that all or part of the steps of the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, and other media that can store program codes.

[0286] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application, or the part that contributes to the relevant technology, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling the vehicle to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks or optical disks.

[0287] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0288] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0289] The features disclosed in the several method or vehicle embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments or vehicle embodiments.

[0290] The above is only an implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A data collection method, characterized in that: The method is applied to a cloud server and includes: Summarizing first scene data collected by at least one vehicle to obtain current total collection amounts of each of the first scene data in different scenes; the first scene data is used to represent road information and / or environmental information; Determine, in different scenarios, a first scenario in which the current total collection amount corresponding to the first scenario data is less than the corresponding target amount; For the first scene, determining a target location for collecting scene data; sending the target location to a target vehicle so that the target vehicle collects gap scene data corresponding to the first scene at the target location; wherein the target vehicle includes vehicles in the area where the target location is located; Continue to obtain the next total collection amount for the first scene, and perform the next collection of the next gap scene data for the first scene until the final total collection amount of the scene data corresponding to the first scene is not less than the corresponding target amount, thereby obtaining the second scene data for the first scene; the second scene data is all the collected scene data for each scene; Using the second scenario data in different scenarios to train a model for realizing intelligent driving of a vehicle; The method further comprises: When the type of the first data tag corresponding to the scene data under the first scene is greater than a first value, determining a first ratio between the collection amount of the first data tag and the target amount of the first data tag; the first value is greater than or equal to 1; determining a second ratio between a target amount of the first data tag and a total target amount of the plurality of first scene data; determining data quality of scene data corresponding to the first data tag; Determining a first priority weight of the first data tag based on the first ratio, the second ratio, and the data quality; wherein the first priority weight is used to characterize a completion deviation of the scene data corresponding to the first data tag; Determining a second priority weight of the first data tag based on a collection period of the gap scenario data corresponding to the first data tag; wherein the collection period is inversely proportional to the second priority weight; Prioritizing the first data tags according to the first priority weight and the second priority weight to obtain a priority ranking result of the first data tags; The priority sorting result is sent to the target vehicle, so that the target vehicle plans a navigation path based on the priority sorting result and the target position, and navigates to the target position based on the navigation path to collect the gap scene data corresponding to the corresponding first data tag.

2. The data collection method according to claim 1, characterized in that: Determining a target location for collecting scene data includes: Based on the first data tag corresponding to the gap scene data, a target position of the gap scene data corresponding to the first data tag is determined in the map data.

3. The data collection method according to claim 2, characterized in that: The determining, based on the first data tag corresponding to the gap scene data, a target position of the gap scene data corresponding to the first data tag in the map data includes: Constructing a first feature set of gap scenario data corresponding to the first data label; Determining a second feature set of the map data; wherein the second feature set is obtained by parsing geometric information of roads in the map data, topological relationships of road sections in the map data, and identification information in the map data; The first feature set and the second feature set are matched to determine a target position of the gap scene data corresponding to the first data tag in the map data.

4. The data collection method according to claim 1, characterized in that: Determining a target location for collecting scene data includes: Based on the second data tag of the first scene data, determining, in the map data, a geographical location of the first scene data corresponding to the second data tag; Marking the corresponding second data tag at the geographic location to obtain a scene distribution map; Based on the first data tag corresponding to the gap scene data, a query is performed on the scene distribution map to determine the target location of the gap scene data corresponding to the first data tag.

5. The data collection method according to claim 4, characterized in that: The querying the scene distribution map to determine the target location of the missing scene data corresponding to the first data tag includes: Perform a query on the scene distribution map to determine a candidate target location of the missing scene data corresponding to the first data tag; The candidate target positions that meet the first condition are determined as the target position.

6. The data collection method according to claim 1, characterized in that: The method further comprises: A first strategy is sent to the target vehicle; the first strategy is used to instruct the target vehicle to prioritize collecting gap scene data corresponding to the third data tag when there is a target position corresponding to the third data tag on the navigation path; wherein the priority of the third data tag is lower than the first data tag corresponding to the target position of the navigation path.

7. A data collection device, characterized in that: The device comprises: a summarizing module configured to summarize first scene data collected by at least one vehicle to obtain a current total collected amount of each first scene data in different scenes; the first scene data is used to represent road information and / or environmental information; A determination module configured to determine, in different scenarios, a first scenario in which the current total collection amount corresponding to the first scenario data is less than the corresponding target amount; a sending module configured to determine a target location for collecting scene data for the first scene; and send the target location to a target vehicle so that the target vehicle collects gap scene data corresponding to the first scene at the target location; wherein the target vehicle includes vehicles in an area where the target location is located; an iteration module configured to continue acquiring a next total collection amount for the first scenario, and to perform a next collection of the next gap scene data for the first scenario, until a final total collection amount of the scene data corresponding to the first scenario is not less than the corresponding target amount, thereby obtaining second scene data for the first scenario; the second scene data being all collected scene data for each scenario; A training module configured to train a model for realizing intelligent driving of a vehicle using the second scenario data under different scenarios; A sorting module is configured to determine a first ratio between the collection quantity of the first data tag and the target quantity of the first data tag when the type of the first data tag corresponding to the scene data in the first scene is greater than a first value; the first value is greater than or equal to 1; determine a second ratio between the target quantity of the first data tag and the total target quantity of multiple first scene data; determine the data quality of the scene data corresponding to the first data tag; determine a first priority weight of the first data tag based on the first ratio, the second ratio and the data quality; wherein the first priority weight is used to characterize the completion deviation of the scene data corresponding to the first data tag; determine the second priority weight of the first data tag according to the collection period of the gap scene data corresponding to the first data tag; wherein the collection period is inversely proportional to the second priority weight; and prioritize the first data tags according to the first priority weight and the second priority weight to obtain the priority sorting result of the first data tags.

8. A cloud server comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images