A vehicle reputation evaluation method, system, processing device and storage medium

Abstract

The present application relates to a kind of vehicle reputation evaluation method, system, processing device and storage medium, the method comprises: obtaining the local map of a plurality of crowd-sourcing vehicles with map element confidence;Based on the local map of each crowd-sourcing vehicle with map element confidence, establish road section or link relationship;Based on the vehicle ID, vehicle reputation and total map element of each crowd-sourcing vehicle, according to the established road section or link relationship, map element is associated, and association result is obtained, and then determine specific member vehicle;According to association result and specific member vehicle, each crowd-sourcing vehicle votes for each map element, and updates the reputation value of crowd-sourcing vehicle, based on the updated reputation value obtained by the present application, the reliability of vehicle sensor data can be effectively evaluated, which can be widely applied in the field of automatic driving.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic driving, and in particular to a vehicle reputation evaluation method, system, processing device and storage medium. BACKGROUND

[0002] High-precision maps have become an important part of the landing of high-level automatic driving technology, but currently high-precision maps lack real-time dynamic updating capability, and crowdsourcing updating has become an important trend and method to solve real-time updating of maps. Although the centralized mapping method based on a professional mobile measurement system (MMS) has high precision, it has high cost, long cycle and large workload, and cannot meet the actual application requirements of high-precision map updating, thereby making it difficult to support large-scale landing applications of high-level automatic driving systems.

[0003] The crowdsourcing updating method relies more on sensors carried by mass-produced vehicles to obtain massive data for map updating, and with the rapid development and popularization of artificial intelligence methods, mass-produced intelligent vehicles are gradually equipped with ADAS driving assistance systems, which have basic ability to perceive map elements. However, due to the low-cost sensors such as cameras, GPS and computing units carried by mass-produced vehicles, the precision of the data obtained is limited. However, due to the different perception capabilities of mass-produced vehicles, there are various uncertainties such as mis-detection, missed detection, and insufficient positioning, and there may be vehicles that maliciously attack map services, resulting in massive crowdsourcing data conflicting with each other and making it difficult to distinguish true and false map changes.

[0004] However, there is currently no data quality reputation evaluation system for vehicle systems for low-cost massive data of mass-produced vehicles. In order to more fully use the massive vehicle-end data obtained by crowdsourcing mass-produced vehicles and ensure the accuracy and accuracy of crowdsourcing updating data, it is necessary to establish a data quality evaluation system for sensors carried by vehicles to determine the quality reputation of different vehicles, which provides better technical support for the fusion, mining and dynamic updating of later crowdsourcing data. SUMMARY

[0005] In view of the above problems, the purpose of the present application is to provide a vehicle reputation evaluation method, system, processing device and storage medium, which can effectively evaluate the reliability of vehicle sensor data.

[0006] To achieve the above purpose, the present application adopts the following technical solutions: in a first aspect, a vehicle reputation evaluation method is provided, comprising:

[0007] Obtaining a vehicle-end local map with map element confidence of a plurality of crowdsourcing vehicles;

[0008] Establishing a road section or link relationship based on the vehicle-end local map with map element confidence of each crowdsourcing vehicle;

[0009] According to the established road segment or link relationship, the map elements are associated to obtain an association result, and a specific member vehicle is determined;

[0010] According to the association result and the specific member vehicle, each crowd-sourcing vehicle votes for each map element, and the reputation value of the crowd-sourcing vehicle is updated.

[0011] Further, the vehicle-end local map of the plurality of crowd-sourcing vehicles is obtained, comprising:

[0012] The original image data of the crowd-sourcing vehicle driving is obtained through the vehicle-mounted camera;

[0013] The obtained original image data is subjected to map element detection and noise filtering;

[0014] The processed map elements are tracked to obtain object-level occupancy of the map elements;

[0015] According to the object-level occupancy of the map elements, a confidence model is established to calculate the confidence of the map elements;

[0016] According to the object-level occupancy of the map elements and the vehicle pose of the crowd-sourcing vehicle, the map elements are subjected to 3D map reconstruction to map the element positions in the 3D space;

[0017] According to the confidence of the map elements and the 3D map reconstruction result, the vehicle-end local map with the confidence of each map element is obtained.

[0018] Further, the obtained original image data is subjected to map element detection, noise filtering and three-dimensional high-precision map element detection processing, comprising:

[0019] The original image data is subjected to map element detection by using an artificial intelligence (AI) learning method to obtain pixels of the map elements existing on the original image data;

[0020] The pixels of the map elements are subjected to noise filtering and three-dimensional high-precision map element detection processing to obtain the processed map elements.

[0021] Further, the vehicle-end local map with the confidence of each map element of each crowd-sourcing vehicle is established to establish the road segment or link relationship, comprising:

[0022] The vehicle-end local map with the confidence of each map element of each crowd-sourcing vehicle is divided into road segments or links to associate the observation areas between the crowd-sourcing vehicles;

[0023] According to the divided road segments or links, the road segment or link relationship is established.

[0024] Further, each crowd-sourced vehicle votes on each map element according to the association result and the specific member vehicle, and updates the reputation value of the crowd-sourced vehicle, including:

[0025] Each crowd-sourced vehicle votes on each map element according to the association result and the specific member vehicle, and obtains a voting result;

[0026] According to the voting result, a consensus result is determined;

[0027] According to the consensus result, a reward mechanism is established, and the reputation value of the crowd-sourced vehicle is updated.

[0028] Further, the voting result of each crowd-sourced vehicle determines whether it supports or opposes the consensus result, and when the consensus result is supported, the reputation value of the vehicle is increased; when the consensus result is opposed, the reputation value of the vehicle is decreased, wherein only when the number of specific member vehicles is greater than a certain threshold, each crowd-sourced vehicle begins to vote, and the threshold depends on the vehicle density of the observation area.

[0029] Further, the reputation value R of the updated crowd-sourced vehicle is: i={m，c}

[0030] R i={m，c} = R i={m，c} + r {p，n}

[0031] Wherein, r {p，n} is a reward result, r is a reward result obtained from voting, and:

[0032]

[0033] Wherein, X(C m ) is a voting function with a confidence value as input; CS j＝{1，0} is a consensus result; r p is a positive reward; r n is a negative reward.

[0034] In a second aspect, a vehicle reputation evaluation system is provided, including:

[0035] A map acquisition module is configured to acquire a vehicle-end local map with map element confidence of a plurality of crowd-sourced vehicles;

[0036] A relationship establishment module is configured to establish a road segment or link relationship based on the vehicle-end local map with map element confidence of each crowd-sourced vehicle;

[0037] An association module is configured to associate map elements according to the established road segment or link relationship based on the vehicle ID, the vehicle reputation, and the total map elements of each crowd-sourced vehicle, to obtain an association result, and further determine a specific member vehicle; ​

[0038] a reputation updating module configured to vote each map element by the association result and the specific member vehicle and each crowdsourcing vehicle, and update the reputation value of the crowdsourcing vehicle.

[0039] In a third aspect, a processing device is provided, which comprises computer program instructions, wherein the computer program instructions are configured to implement the steps of the vehicle reputation evaluation method when executed by the processing device.

[0040] In a fourth aspect, a computer readable storage medium is provided, which stores computer program instructions, wherein the computer program instructions are configured to implement the steps of the vehicle reputation evaluation method when executed by a processor.

[0041] The present application has the following advantages due to the above technical solutions:

[0042] 1. The present application is mainly used for determining the reputation value of a crowdsourcing vehicle, so as to ensure that data of better quality is selected for crowdsourcing updating of a high-precision map, and to ensure the quality of high-precision map data and effectively evaluate the reliability of vehicle sensor data. Meanwhile, the reputation system can punish false information by severely reducing the reputation value, so as to prevent subsequent attacks.

[0043] 2. The present application establishes a reputation mechanism for map elements, proposes a voting mechanism based on map element existence to check and detect the quality, and updates the reputation value of the vehicle through the consensus decision of the crowdsourcing vehicle, so as to ensure the quality of data participating in crowdsourcing updating.

[0044] In summary, the present application can be widely applied in the field of autonomous driving. BRIEF DESCRIPTION OF DRAWINGS

[0045] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included to provide a description of the preferred embodiments and are not intended to limit the scope of the present application. Throughout the drawings, the same reference designations are used to represent the same elements. In the drawings:

[0046] Figure 1 is a method flowchart provided by an embodiment of the present application;

[0047] Figure 2 is a flowchart of vehicle-side local map acquisition provided by an embodiment of the present application;

[0048] Figure 3 is a schematic diagram of describing an expected observation area with a crowdsourcing vehicle as a reference point provided by an embodiment of the present application;

[0049] Figure 4is a road segment or link association map of a crowd-sourced vehicle overlap area provided by an embodiment of the present application;

[0050] Figure 5 is a map element association map of a vehicle in a road segment area provided by an embodiment of the present application;

[0051] Figure 6 is a flowchart of updating a reputation value provided by an embodiment of the present application;

[0052] Figure 7 is a reward and punishment system model of positive and negative rewards and punishments provided by an embodiment of the present application. DETAILED DESCRIPTION

[0053] Example embodiments of the present application will be described herein below with reference to the accompanying drawings. While example embodiments of the present application are illustrated, it is to be understood that the present application can be carried out in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0054] It is to be understood that the terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The terms "comprises", "comprising", "includes", "including" and the like are to be construed to be inclusive (i.e., to include both instances of open and closed transitions) unless otherwise indicated. The methods described herein can be implemented in a number of ways. Embodiments of the application described herein can be directed to methods comprising the steps of the methods described herein, compositions comprising the compositions described herein, and computer program products comprising computer code to implement the methods described herein.

[0055] Although the terms first, second, third, etc. can be used herein to describe various elements, components, regions, layers and / or sections, these elements, components, regions, layers and / or sections should not be limited by these terms. These terms can be only used to distinguish one element, component, region, layer or section from another region, layer or section. Terms such as "first", "second", and other numerical terms when used herein do not imply a sequence or order unless clearly indicated by the context. Thus, a first element, component, region, layer or section discussed below could be termed a second element, component, region, layer or section without departing from the teachings of the example embodiments.

[0056] This invention provides a method, system, processing device, and storage medium for evaluating vehicle reputation of mass-produced intelligent vehicle data for crowdsourced updates of autonomous driving maps. The location of map elements detected by the vehicle is first obtained during the high-precision 3D map reconstruction process. Artificial intelligence deep learning algorithms are used to detect map elements, and positioning algorithms are used to obtain the vehicle's location at that point in time. This location is crucial for associating each map element with corresponding detections from other crowdsourced vehicles. When creating associations, these crowdsourced vehicles are considered members of the voting process. Since each crowdsourced vehicle has its own visual range, this membership will vary depending on whether one map element differs from another. The decision to vote for the existence of a map element is automatically generated based on the confidence level of the map element detection and the vehicle reputation itself. Vehicle reputation is crucial for determining which vehicle data is reliable for constructing crowdsourced high-precision maps. This value is generated by the consensus reached by networked vehicles passing through a certain area. This protocol, called reputation proof, uses vehicle reputation to vote on the state of high-definition map elements. The reputation value is updated each time a map element reaches a consensus with vehicles in that area regarding its existence.

[0057] Example 1

[0058] like Figure 1 As shown, this embodiment provides a vehicle credit assessment method, including the following steps:

[0059] 1) such as Figure 2 As shown, the process involves obtaining local vehicle-side maps with map element confidence scores for several crowdsourced vehicles, specifically:

[0060] 1.1) Obtain raw image data of the crowdsourced vehicle's movement through the vehicle-mounted camera, including raw image sequences or videos.

[0061] 1.2) Map element detection and noise filtering are performed on the acquired raw image data to obtain processed map elements, which include road markings, traffic lights, traffic poles, and traffic signs, etc.

[0062] 1.2.1) Using artificial intelligence (AI) learning methods, map element detection is performed on the acquired raw image data to obtain the pixels of the map elements present in the raw image data.

[0063] Specifically, this step employs artificial intelligence (AI) learning methods for map element detection to provide accurate results. However, this step is not limited to a single algorithm. For example, a convolutional neural network (CNN) can be used to extract road marking segmentation results. After obtaining the segmentation results, the next step is to eliminate segmentation noise and determine the type and structure of the primitives.

[0064] 1.2.2) Noise filtering and 3D high-precision map element detection processing are performed on the pixels of the map elements to obtain processed map elements.

[0065] Specifically, the method of noise filtering and 3D high-precision map element detection processing in this step is not limited to one algorithm. For example, noise filtering can filter noise according to the absolute pixel size of the map element and the structure of the maximum boundary, and the structure used is designated as a line for a pole type and a lane line and a boundary box for other map elements.

[0066] 1.3) Tracking is performed on the processed map elements to obtain the tracking results of each map element, i.e., the object-level occupancy of the map elements.

[0067] Specifically, the method of map element tracking in this step is not limited to one algorithm. For example, optical flow method is used for tracking. Tracking is important for two reasons, including calculating the confidence of the existence of the map element and the purpose of map element reconstruction.

[0068] 1.4) A confidence model is established based on the object-level occupancy of the map elements using the Bayesian method to calculate the confidence of the map elements.

[0069] 1.5) 3D map reconstruction is performed on the map elements based on the object-level occupancy of the map elements and the vehicle pose of the crowd-sourced vehicles to map the element positions in the 3D space.

[0070] Specifically, an iterative optimization method is used to minimize the positioning error of the map elements relative to the previous positions.

[0071] Specifically, as shown in Figure 3 During the 3D map reconstruction process, this step adds an expected observation map, which is the area within a first distance threshold on the left and right sides of the vehicle trajectory. Given the trajectory passed by the vehicle, the expected observation map is important for determining the existence of the map elements. The closer the map elements are to the vehicle trajectory line, the higher the likelihood that the vehicle will observe the map elements. The expected observation map will serve as the member basis for the road segment association and map element association stage.

[0072] Specifically, the first distance threshold is affected by the field of view of the vehicle-mounted camera and its position relative to the center of gravity of the vehicle. For example, the first distance threshold is 6m.

[0073] 1.6) Based on the map element confidence and the 3D map reconstruction results, a local map with map element confidence is obtained for each crowd-sourced vehicle.

[0074] 2) The local map with map element confidence for each crowd-sourced vehicle is divided into road segments or links to associate the observation areas between the crowd-sourced vehicles.

[0075] Specifically, as shown in Figure 4 When the vehicle reputation is unknown, an initial value 30 is automatically assigned, which will be upgraded according to the consensus result of the next stage. The reputation value can be divided into two types, mapping vehicle reputation and crowd-sourcing vehicle reputation with ordinary sensors, among which the professional measurement system with high-precision laser radar and other sensors is called mapping vehicle. The mapping vehicle reputation is set to have a higher initial reputation value, and its maximum value is also set to the maximum number 100; for the crowd-sourcing vehicle reputation, the maximum reputation is set to 80, which is a lower maximum value. The reputation range is:

[0076] 70≥R m ≥100 (1)

[0077] 0≥R c ≥80 (2)

[0078] Wherein, R m is the mapping vehicle reputation, and R c is the crowd-sourcing vehicle reputation. The reputation values between map elements will be different, which means that although the vehicles are the same, the reputation for determining the existence of road markings may be different from the reputation for determining the traffic pole.

[0079] 3) According to the divided road sections or links, the road section or link relationship is established.

[0080] 4) Obtain the vehicle ID, vehicle reputation and total map elements of each crowd-sourcing vehicle, and based on the obtained vehicle ID, vehicle reputation and total map elements of each crowd-sourcing vehicle, according to the established road section or link relationship, the map elements are associated to obtain the association result, and then the specific member vehicle is determined.

[0081] Specifically, this process is important for collecting voting members for each map element. After the voting members, the map elements will also be associated, and the association is used to collect specific member vehicles through the definition of the map element position of the observation area.

[0082] Specifically, the total map elements include type, position and confidence.

[0083] Specifically, as shown in Figure 5 The clustering method is used to associate the clusters of map elements. The second distance threshold of the cluster is set to 4m for traffic signs, traffic poles and other vertical elements, and 2m for road markings. From these clusters, information such as vehicle ID, confidence, vehicle reputation and total map elements is obtained for the next step of voting.

[0084] 5) According to the association result and the specific member vehicle, each crowd-sourcing vehicle votes for each map element, and according to the voting result, the consensus result is determined.

[0085] Specifically, the voting result is the existence or non-existence of a map element, and the consensus result is obtained from each map element.

[0086] Specifically, the voting result of each crowdsourcing vehicle will determine whether it supports or opposes the consensus result, and the reputation value of the vehicle will increase when the consensus result is supported, and the reputation value of the vehicle will decrease when the consensus result is opposed. This process will be iterated for each map element.

[0087] Specifically, each crowdsourcing vehicle only starts voting when the number of specific member vehicles is greater than a certain threshold, which depends on the vehicle density of the observation area, for example: the threshold is 30% of the total vehicle density per hour.

[0088] Specifically, as shown in Figure 6 , the voting process is to determine the existence of a map element, and the existence voting will be collected from each specific member vehicle. When the minimum number of relevant vehicle constraints is met, the voting process begins. This process requires two pieces of information, vehicle reputation and confidence in the existence of a map element:

[0089]

[0090]

[0091] Where V j＝{1，0} is the vehicle vote; R i is the vehicle reputation; m is the mapping vehicle; c is the crowdsourcing vehicle; X(C m ) is the voting function with confidence value as input; C m is the confidence in the existence of a map element; C T is the confidence threshold; the number of confidence thresholds C T in the voting function can be modified according to the map element. It is important to ensure that each vehicle has one voting opportunity, in order to ensure fairness and encourage high reputation. After all the votes are cast, the consensus result is determined:

[0092]

[0093] Where CS j={1，0} is the consensus result; R is the voting parameter, which can be set according to actual conditions, for example R = 0.6; V j＝{1} represents the voting for the existence of a map element, and V j＝{0} represents the voting for the non-existence of a map element.

[0094] 6) According to the consensus result, establish a reward mechanism to update the reputation value of the crowdsourcing vehicle.

[0095] In particular, the reputation value of the crowd-sourced vehicle is updated when the map element and the vehicle of the observation area reach consensus on its existence. Within the road segment range, the reputation value of each crowd-sourced vehicle helps to vote.

[0096] In particular, this process will be carried out for each specific member vehicle that contributes to the voting phase, where the rewards include positive rewards and negative rewards:

[0097]

[0098]

[0099] where r p is the positive reward; r n is the negative reward; CS is the consensus result, i.e. match (1) or mismatch (-1); is the weight parameter, set to 1; v is the number of votes; n is the total number of members in the road segment / link.

[0100] In particular, the reputation value R i={m,c} of the updated crowd-sourced vehicle is:

[0101] R i={m，c} = R i={m，c} + r {p，n} (8)

[0102]

[0103] where r {p，n} is the reward result, r is the result of reward from voting.

[0104] In particular, as shown in Figure 7 , the reward model features are shown, where the higher the reputation of the vehicle, the lower the positive reward. However, the negative reward / punishment will be amplified by the high reputation value, and vice versa. This is to encourage vehicles to maintain the quality of their high-precision map mapping module.

[0105] Embodiment 2

[0106] The embodiment provides a vehicle reputation evaluation system, comprising:

[0107] a map acquisition module configured to acquire a vehicle-end local map with map element confidence of a plurality of crowd-sourced vehicles.

[0108] a relationship establishment module configured to establish a road segment or link relationship based on the vehicle-end local map with map element confidence of each crowd-sourced vehicle.

[0109] The association module is configured to associate the map elements according to the established road segment or link relationship based on the vehicle ID, the vehicle reputation and the total map elements of each crowdsourcing vehicle, to obtain an association result, and to determine a specific member vehicle.

[0110] The reputation updating module is configured to vote for each map element by each crowdsourcing vehicle according to the association result and the specific member vehicle, and to update the reputation value of the crowdsourcing vehicle.

[0111] The system provided in the embodiment is used for executing the above-mentioned method embodiments, and the specific process and detailed content are referred to the above-mentioned embodiments, which will not be described here.

[0112] Embodiment 3

[0113] The embodiment provides a processing device corresponding to the vehicle reputation evaluation method provided in the embodiment 1, and the processing device can be applied to a processing device of a client, such as a mobile phone, a notebook computer, a tablet computer, a desktop computer and the like, to execute the method of the embodiment 1.

[0114] The processing device includes a processor, a memory, a communication interface and a bus, the processor, the memory and the communication interface are connected through the bus to complete the communication among each other. The memory stores a computer program capable of running on the processing device, and the processing device runs the computer program to execute the vehicle reputation evaluation method provided in the embodiment 1.

[0115] In some implementations, the memory can be a high-speed random access memory (RAM), and can also include a non-volatile memory, such as at least one disk memory.

[0116] In other implementations, the processor can be a central processing unit (CPU), a digital signal processor (DSP) and various types of general-purpose processors, which are not limited here.

[0117] In addition, the logic instructions in the memory described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0118] Those skilled in the art can understand that the structure of the computing device described above is only part of the structure related to the scheme of the present application, and does not constitute a limitation on the computing device to which the scheme of the present application is applied. The specific computing device can include more or fewer components, or combine certain components, or have a different component arrangement.

[0119] Embodiment 4

[0120] The embodiment provides a computer program product corresponding to the vehicle reputation evaluation method provided in the embodiment 1. The computer program product can include a computer readable storage medium, which is loaded with computer readable program instructions for executing the vehicle reputation evaluation method described in the embodiment 1.

[0121] The computer readable storage medium can be a tangible device that maintains and stores instructions for use by an instruction execution device. The computer readable storage medium can be, for example but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination of the above.

[0122] The computer readable storage medium provided in the above embodiment has similar implementation principles and technical effects to the above method embodiments, and will not be described here.

[0123] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks. Figure 1 one or more flow or blocks.

[0124] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flow or blocks. Figure 1 one or more flow or blocks.

[0125] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks. Figure 1 one or more flow or blocks.

[0126] The above embodiments are only used for illustrating the present application, wherein the structure, connection mode and manufacturing process of each part can be changed, and any equivalent transformation and improvement based on the technical scheme of the present application should not be excluded from the protection scope of the present application.

Claims

1. A vehicle reputation assessment method, characterized by, The method comprises the following steps: obtaining a plurality of vehicle-end local maps with map element confidence of crowd-sourced vehicles; establishing a road segment or link relationship based on the vehicle-end local map with map element confidence of each crowd-sourced vehicle; based on the vehicle ID, vehicle reputation and total map element of each crowd-sourced vehicle, associating the map elements according to the established road segment or link relationship to obtain an association result, and then determining a specific member vehicle; based on the association result and the specific member vehicle, each crowd-sourced vehicle votes for each map element and updates the reputation value of the crowd-sourced vehicle; the method of obtaining a plurality of vehicle-end local maps of crowd-sourced vehicles comprises the following steps: obtaining original image data of the crowd-sourced vehicles through a vehicle-mounted camera; performing map element detection and noise filtering on the obtained original image data; tracking the processed map elements to obtain object-level occupancy of the map elements; establishing a confidence model according to the object-level occupancy of the map elements to calculate the map element confidence; performing 3D map reconstruction on the map elements according to the object-level occupancy of the map elements and the vehicle pose of the crowd-sourced vehicle to map the element position in the 3D space; obtaining a vehicle-end local map with map element confidence of each map element based on the map element confidence and the 3D map reconstruction result; the method of voting for each map element by each crowd-sourced vehicle based on the association result and the specific member vehicle, and updating the reputation value of the crowd-sourced vehicle comprises the following steps: voting for each map element by each crowd-sourced vehicle based on the association result and the specific member vehicle to obtain a voting result; determining a consensus result according to the voting result; establishing a reward mechanism according to the consensus result to update the reputation value of the crowd-sourced vehicle; Updating reputation values for crowd-sourced vehicles For: wherein, is the reputation value of the crowd-sourced vehicle after the update; is the reputation value of the crowd-sourced vehicle before the update; is the reward result, is the reward result from the vote, and: wherein, is a voting function taking as input the confidence values; is a consensus result; is a positive reward; is a negative reward.

2. The vehicle reputation assessment method of claim 1, wherein, the method of performing map element detection on the obtained original image data, and performing noise filtering and three-dimensional high-precision map element detection processing comprises the following steps: using an artificial intelligence (AI) learning method to perform map element detection on the obtained original image data to obtain the pixels of the map elements existing on the original image data; performing noise filtering and three-dimensional high-precision map element detection processing on the pixels of the map elements to obtain the processed map elements.

3. The method of claim 1, wherein, the method of establishing a road segment or link relationship based on the vehicle-end local map with map element confidence of each crowd-sourced vehicle comprises the following steps: dividing the vehicle-end local map with map element confidence of each crowd-sourced vehicle into road segments or links to associate the observation areas between the crowd-sourced vehicles; establishing a road segment or link relationship according to the divided road segments or links.

4. A method of vehicle reputation assessment as claimed in claim 3, wherein, The voting result of each crowd-sourced vehicle will determine whether it supports or opposes the consensus result. When the consensus result is supported, the reputation value of the vehicle is increased; when the consensus result is opposed, the reputation value of the vehicle is decreased. Only when the number of specific member vehicles is greater than a certain threshold value, the crowd-sourced vehicles start to vote. The threshold value depends on the vehicle density of the observation area.

5. A vehicle reputation assessment system, characterized by, The method comprises the following steps: a map acquisition module is configured to obtain a plurality of vehicle-end local maps with map element confidence of crowd-sourced vehicles; a relationship establishment module is configured to establish a road segment or link relationship based on the vehicle-end local map with map element confidence of each crowd-sourced vehicle; The association module is configured to associate the map elements according to the established road segment or link relationship based on the vehicle ID, the vehicle reputation and the total map elements of each crowdsourcing vehicle, to obtain an association result, and to determine a specific member vehicle; The reputation updating module is configured to vote for each map element by each crowdsourcing vehicle according to the association result and the specific member vehicle, and to update the reputation value of the crowdsourcing vehicle; The method for obtaining the local map of the crowdsourcing vehicle at the vehicle end comprises: Obtaining original image data of the crowdsourcing vehicle through a vehicle-mounted camera; Performing map element detection and noise filtering on the obtained original image data; Tracking the processed map elements to obtain object-level occupancy of the map elements; Establishing a confidence model according to the object-level occupancy of the map elements to calculate the confidence of the map elements; Reconstructing a 3D map of the map elements according to the object-level occupancy of the map elements and the vehicle pose of the crowdsourcing vehicle to map the element positions in the 3D space; Obtaining the local map of the crowdsourcing vehicle at the vehicle end with the confidence of each map element according to the confidence of the map elements and the 3D map reconstruction result; The method for voting for each map element by each crowdsourcing vehicle according to the association result and the specific member vehicle, and updating the reputation value of the crowdsourcing vehicle comprises: Voting for each map element by each crowdsourcing vehicle according to the association result and the specific member vehicle to obtain a voting result; Determining a consensus result according to the voting result; Establishing a reward mechanism according to the consensus result to update the reputation value of the crowdsourcing vehicle; Updating reputation values of crowd-sourced vehicles For: wherein, is the reputation value of the crowd-sourced vehicle after the update; is the reputation value of the crowd-sourced vehicle before the update; is the reward result, is the reward result from the vote, and: wherein, is a voting function taking as input the confidence values; is a consensus result; is a positive reward; is a negative reward.

6. A processing device, characterized by The computer program instructions are executed by a processing device to implement the steps of the vehicle reputation evaluation method in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer program instructions are stored on the computer readable storage medium and are executed by a processor to implement the steps of the vehicle reputation evaluation method in any one of claims 1-4.

Images