CN111767440A

CN111767440A - Vehicle portrayal method based on knowledge graph, computer equipment and storage medium

Info

Publication number: CN111767440A
Application number: CN202010912546.XA
Authority: CN
Inventors: 刘建林
Original assignee: Ping An International Smart City Technology Co Ltd
Current assignee: Ping An International Smart City Technology Co Ltd
Priority date: 2020-09-03
Filing date: 2020-09-03
Publication date: 2020-10-13
Anticipated expiration: 2040-09-03
Also published as: CN111767440B

Abstract

The application relates to a vehicle portrayal method based on knowledge graph, a computer device and a storage medium, wherein the method comprises the following steps: the method comprises the steps of obtaining the data structuralization degree of original vehicle data, determining the classification of the original vehicle data according to the data structuralization degree, and carrying out data processing on the corresponding original vehicle data according to the preprocessing modes corresponding to different classifications to generate corresponding triple vehicle data. And carrying out knowledge fusion on the triple vehicle data to generate complete description knowledge of the vehicle, and constructing a vehicle map knowledge base based on the complete description knowledge. And performing knowledge calculation to determine the missing labels of the vehicles based on each vehicle map knowledge base, and generating vehicle image results by combining the missing labels and the initial vehicle labels. The method is adopted to realize the expansion of the vehicle data and the associated information, the complete label of the vehicle data is obtained by combining the existing label and the missing label, the complete vehicle portrait is obtained, and the accuracy of vehicle classification according to the complete vehicle portrait result is improved.

Description

Vehicle portrayal method based on knowledge graph, computer equipment and storage medium

Technical Field

The present application relates to the field of big data processing technology, and in particular, to a vehicle portrayal method based on an intellectual graph, a computer device, and a storage medium.

Background

With the wide application of the internet and big data technology, in order to analyze and process mass vehicle data resources and improve the efficiency of vehicle management and analysis performed by traffic police, vehicle data needs to be classified quickly in view of the situation that vehicle data and picture data are increasing day by day.

Conventionally, a tagged vehicle model is obtained by abstracting different data by adopting a portrait technology, so that vehicle classification and search are realized by classifying and extracting vehicles based on tags. In the conventional label-based object imaging method, it can be understood that a plurality of labels are firstly required to be allocated to each object, each label is provided with a corresponding weight, and the weight represents the interest intensity of the object in this aspect, so as to obtain the image data for the object.

However, in the conventional portrait mode, there is a certain deviation in the source data for generating each portrait, and the data is not comprehensive enough, so that the obtained portrait is not complete enough, and the characteristics of the objects cannot be described clearly, resulting in an inaccurate classification result for classifying each object.

Disclosure of Invention

In view of the above, it is desirable to provide a vehicle portrayal method based on an intellectual graph, a computer device and a storage medium, which can improve the accuracy of vehicle classification.

A method of vehicle portrayal based on an intellectual graph, the method comprising:

acquiring the data structuralization degree of original vehicle data, and determining the classification of the original vehicle data according to the data structuralization degree;

according to the preprocessing modes corresponding to the original vehicle data of different classifications, data processing is carried out on the corresponding original vehicle data to generate corresponding triple vehicle data;

carrying out knowledge fusion processing on the triple vehicle data to generate complete description knowledge of the vehicle;

constructing a vehicle map knowledge base based on the complete description knowledge;

performing knowledge calculation based on each constructed vehicle map knowledge base to determine missing labels of corresponding vehicles;

and combining the determined missing label of the vehicle and the pre-acquired initial vehicle label to generate a vehicle image result.

In one embodiment, the classification of the raw vehicle data includes semi-structured vehicle data, and unstructured vehicle data; the data processing of the corresponding original vehicle data according to the preprocessing mode corresponding to the original vehicle data of different classifications to generate the corresponding triple vehicle data includes:

acquiring an extraction object corresponding to the unstructured vehicle data; the extraction object comprises entities, attribute information of the entities and association relations among the entities;

determining a first preset processing mode corresponding to the non-structural vehicle data according to the extraction object;

extracting the named entities, the attribute information of the entities and the incidence relation among the entities from the unstructured vehicle data according to the first preset processing mode;

generating triple vehicle data corresponding to the unstructured vehicle data according to the named entities, attribute information respectively corresponding to the entities and the incidence relation among the entities;

extracting the semi-structured vehicle data from a webpage according to a second preprocessing mode corresponding to the semi-structured vehicle data, calling a preset wrapper, and restoring the extracted semi-structured vehicle data into structured vehicle data;

and converting the structured vehicle data into triple vehicle data according to a third preprocessing mode corresponding to the structured vehicle data.

In one embodiment, the knowledge fusion process comprises a data integration process, a knowledge disambiguation process, and a data fusion process; the knowledge disambiguation process comprises an entity alignment process and an attribute alignment process; the knowledge fusion processing is carried out on the triple vehicle data to generate complete description knowledge of the vehicle, and the method comprises the following steps:

performing data integration processing on the triple vehicle data to generate a complete data description corresponding to the triple vehicle data;

carrying out entity alignment processing and attribute alignment processing on the complete data description of the triple vehicle data to generate a standard knowledge representation;

and carrying out data fusion processing on the standard knowledge representation to generate complete description knowledge of the vehicle.

In one embodiment, the knowledge calculation includes similarity calculation and correlation analysis; the step of performing knowledge calculation based on the constructed vehicle map knowledge bases to determine missing labels of corresponding vehicles comprises the following steps:

calculating the similarity based on the constructed vehicle map knowledge bases to determine candidate vehicles;

semantic association is carried out on each candidate vehicle, random walk is carried out on the basis of vehicle basic information, and a random walk result is generated;

performing correlation analysis based on the random walk result, and determining entities potentially associated with the vehicle basic information;

and extracting entity labels of entities which are potentially associated with the vehicle basic information, and determining the extracted entity labels as missing labels of corresponding vehicles.

In one embodiment, the vehicle map knowledge base comprises a vehicle basic information base, a vehicle violation business base and a vehicle event base; the building of a vehicle map knowledge base based on the complete description knowledge comprises:

determining vehicle basic information and vehicle violation information corresponding to the triple vehicle data based on the complete description knowledge;

according to the vehicle basic information, a vehicle basic information base corresponding to the triple vehicle data is constructed;

determining a vehicle violation business corresponding to the triple vehicle data according to the vehicle violation information, and constructing a vehicle violation business library according to the vehicle violation business;

and extracting vehicle events corresponding to the triple vehicle data from the complete knowledge description, and constructing a corresponding vehicle event library based on the vehicle events.

In one embodiment, the vehicle map knowledge base further comprises a vehicle relationship base, a vehicle tag base and a geographic location base; the building of a vehicle map knowledge base based on the complete description knowledge comprises:

semantic mapping is carried out on the basis of the vehicle event library, and a vehicle relation library is constructed; the vehicle relation library comprises vehicle relations which are subject-object relations, business process relations and membership relations;

dynamically labeling based on the vehicle event library, and establishing a vehicle label library; the vehicle tags comprise vehicle base tags, user tags, enterprise tags and event tags corresponding to vehicles, users and enterprises;

acquiring geographical position information corresponding to the complete description knowledge based on a preset third-party data interface, and constructing a corresponding geographical position library according to the geographical position information; the geographic location information includes vehicle location, route location, and zone division.

In one embodiment, the generating a vehicle image result by combining the determined missing tag of the vehicle and the pre-acquired initial vehicle tag includes:

acquiring a predetermined initial vehicle tag;

generating a plurality of vehicle tags associated with the triplet vehicle data in conjunction with the missing tag of the vehicle and the initial vehicle tag;

acquiring a preset weight corresponding to each vehicle label;

and generating a complete vehicle portrait result based on each vehicle label and the corresponding preset weight.

In one embodiment, after the generating the vehicle imaging result by combining the determined missing tag of the vehicle and the pre-acquired initial vehicle tag, the method further includes:

classifying each vehicle according to the vehicle portrait result to generate a corresponding vehicle classification result;

and generating corresponding recommendation information according to the vehicle classification result, and displaying and broadcasting the recommendation information.

A knowledge-graph based vehicle portrayal apparatus, the apparatus comprising:

the vehicle data classification module is used for acquiring the data structuralization degree of the original vehicle data and determining the classification of the original vehicle data according to the data structuralization degree;

the triple vehicle data generation module is used for carrying out data processing on corresponding original vehicle data according to the preprocessing modes corresponding to the original vehicle data of different classifications to generate corresponding triple vehicle data;

the complete description knowledge generation module is used for carrying out knowledge fusion processing on the triple vehicle data to generate complete description knowledge of the vehicle;

the vehicle map knowledge base construction module is used for constructing a vehicle map knowledge base based on the complete description knowledge;

the missing label determining module is used for performing knowledge calculation based on the constructed vehicle map knowledge bases to determine the missing labels of the corresponding vehicles;

and the vehicle portrait result generation module is used for combining the determined missing label of the vehicle and the pre-acquired initial vehicle label to generate a vehicle portrait result.

A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:

A computer storage medium having a computer program stored thereon, the computer program when executed by a processor implementing the steps of:

The vehicle portrayal method based on the knowledge graph, the computer equipment and the storage medium acquire the data structuring degree of the original vehicle data, determine the classification of the original vehicle data according to the data structuring degree, and perform data processing on the corresponding original vehicle data according to the preprocessing mode corresponding to the original vehicle data of different classifications to generate the corresponding triple vehicle data. Knowledge fusion processing is carried out on the triple vehicle data to generate complete description knowledge of the vehicle, and then a vehicle map knowledge base can be constructed on the basis of the complete description knowledge. And performing knowledge calculation based on the constructed knowledge base of each vehicle map, determining the missing labels of the corresponding vehicles, and generating a complete vehicle portrait result by combining the determined missing labels of the vehicles and the pre-acquired initial vehicle labels. The behavior data of the vehicle are expanded by the aid of the knowledge graph, the range of vehicle information related to the vehicle data is widened, the existing labels and the missing labels determined based on the knowledge graph are combined to obtain the complete labels of the vehicle data, and further the complete vehicle portrait can be obtained, so that the accuracy of vehicle classification according to the complete vehicle portrait result is improved.

Drawings

FIG. 1 is a diagram illustrating an example of an application of a method for generating a vehicle representation based on an knowledgegraph;

FIG. 2 is a flow diagram of a method for knowledge-graph based vehicle portrayal in accordance with one embodiment;

FIG. 3 is a schematic flow diagram illustrating the generation of triple vehicle data in one embodiment;

FIG. 4 is a schematic flow diagram of building a vehicle map knowledge base in one embodiment;

FIG. 5 is a block diagram of the overall architecture of the vehicle map knowledge base building system in one embodiment;

FIG. 6 is a schematic diagram of a basic knowledge graph building process in one embodiment;

FIG. 7 is a block diagram of a knowledge-based vehicle representation apparatus in accordance with one embodiment;

FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The knowledge graph-based vehicle portrayal method can be applied to the application environment shown in the figure 1. Wherein the terminal 102 communicates with the server 104 via a network. The data structuring degree of the original vehicle data is obtained, the classification of the original vehicle data stored locally from the terminal 102 or stored in the cloud of the server 104 is determined according to the data structuring degree, and then the corresponding original vehicle data is subjected to data processing according to the preprocessing modes corresponding to the original vehicle data of different classifications, so that the corresponding triple vehicle data is generated. Knowledge fusion processing is carried out on the triple vehicle data to generate complete description knowledge of the vehicle, and a vehicle map knowledge base is constructed on the basis of the complete description knowledge. And then, based on the constructed knowledge base of each vehicle map, performing knowledge calculation, determining the missing label of the corresponding vehicle, and generating a vehicle image result by combining the determined missing label of the vehicle and the pre-acquired initial vehicle label. Wherein the vehicle image result can be fed back to the terminal 102. The terminal 102 may be, but is not limited to, various personal computers, notebook computers, smart phones, and tablet computers, and the server 104 may be implemented by an independent server or a server cluster formed by a plurality of servers.

In one embodiment, as shown in FIG. 2, a method for knowledge-graph based vehicle representation is provided, which is illustrated by the example of the method applied to the server in FIG. 1, and comprises the following steps:

step S202, acquiring the data structuring degree of the original vehicle data, and determining the classification of the original vehicle data according to the data structuring degree.

Specifically, the degree of data structuring includes structured, semi-structured, unstructured, and the corresponding data classification includes structured data, semi-structured data, and unstructured data. After the data structuring degree of the original vehicle data is obtained, the original vehicle data is classified according to the data structuring degree, and different classified original vehicle data including structured vehicle data, semi-structured vehicle data and unstructured vehicle data are generated.

Furthermore, for the structured data, the data sources thereof comprise a manual input mode and a database docking mode, and the vehicle business process and data related to the vehicle business process can be manually input, or corresponding vehicle and vehicle type data can be acquired by docking a GIS service interface, a service system and the like. For semi-structured data, fields of the data need to be unified, wherein sources of the semi-structured data can include an alert data interface, a business data interface and a civil data interface. Aiming at unstructured data, namely text data, the knowledge to be extracted comprises entities, relations and attributes, the sources of the knowledge comprise data sources and labeled content, wherein the data sources comprise laws and regulations, external announcements, news reports and public opinion data, and the labeled content comprises portrait tags, case tags, proper nouns and relation refinements.

And step S204, performing data processing on corresponding original vehicle data according to the preprocessing modes corresponding to the different classified original vehicle data to generate corresponding triple vehicle data.

Specifically, the structured data generally represents data of the relational data, the data structure of the structured data is clear, and the data in the relational database can be directly converted into RDF data, namely ternary data. The semi-structured data representation has a certain data structure, but needs to be further extracted and sorted, for example, data in a web page can be extracted from the web page by using a preset wrapper, and the data can be restored into structured data.

The unstructured data, i.e., text data, is divided into entity extraction, attribute extraction, and relationship extraction according to the difference of extraction objects. For the entity extraction, also referred to as named entity recognition, the entities in this embodiment include concepts, persons, organizations, place names, time, and the like. Regarding relation extraction, namely representing the relation between the entities to be extracted and the important knowledge in the text, certain technical means are required to extract the relation information. The attribute extraction, namely the extraction of the attribute information of the entity, is expressed, and is similar to the relationship. The relationship reflects the external contact of the entity, and the attribute reflects the internal characteristics of the entity.

And further, processing the vehicle data with different structuralization degrees according to a corresponding preset processing mode, and converting the corresponding vehicle data into a triple form to obtain triple vehicle data.

And step S206, carrying out knowledge fusion processing on the triple vehicle data to generate complete description knowledge of the vehicle.

Specifically, knowledge fusion processing is performed on the obtained triple vehicle data, and the knowledge fusion processing comprises data integration processing, knowledge disambiguation processing and data fusion processing. The data integration processing is carried out on the triple vehicle data to generate complete data description corresponding to the triple vehicle data, and then entity alignment processing and attribute alignment processing are carried out on the complete data description of the triple vehicle data to generate the standard knowledge representation, so that data fusion processing can be carried out on the standard knowledge representation to generate complete description knowledge of the vehicle.

The knowledge fusion represents a process of integrating knowledge in a plurality of knowledge bases to form one knowledge base. Because different knowledge bases have different emphasis points on knowledge collection, the different knowledge bases have different descriptions of the same entity, one knowledge base may focus on the description of a certain aspect of the knowledge base, and the other knowledge base may focus on describing the relationship between the entity and other entities. Through knowledge fusion, the knowledge in different knowledge bases can be complementarily fused to form comprehensive, accurate and complete entity description, and the fused knowledge is stored. The knowledge disambiguation processing in the knowledge fusion processing process comprises entity alignment processing and attribute alignment processing.

And S208, constructing a vehicle map knowledge base based on the complete description knowledge.

Based on complete description knowledge, the constructed vehicle map knowledge base comprises the following components: the system comprises a vehicle basic information base, a vehicle violation business base, a vehicle event base, a vehicle relation base, a vehicle label base and a geographical position base.

Specifically, the vehicle basic information and the vehicle violation information corresponding to the triple vehicle data can be determined based on the complete description knowledge, and then the vehicle basic information base and the vehicle violation service base corresponding to the triple vehicle data are respectively constructed according to the vehicle basic information and the vehicle violation information.

Similarly, by extracting vehicle events corresponding to the triple vehicle data from the complete knowledge description, a corresponding vehicle event library is constructed based on the vehicle events.

Furthermore, a vehicle relation library can be constructed by performing semantic mapping based on the vehicle event library, and a vehicle label library can be constructed by performing dynamic labeling based on the vehicle event library.

And step S210, performing knowledge calculation based on the constructed vehicle map knowledge bases, and determining the missing labels of the corresponding vehicles.

Specifically, similarity calculation is carried out based on each constructed vehicle map knowledge base, candidate vehicles are determined, semantic association is carried out on each candidate vehicle, random walk is carried out based on vehicle basic information, and a random walk result is generated. And performing association analysis based on the random walk result, determining an entity which is potentially associated with the vehicle basic information, extracting an entity tag of the entity which is potentially associated with the vehicle basic information, and determining the extracted entity tag as a missing tag of the corresponding vehicle.

The knowledge calculation mode comprises rule management, analysis mining, entity marking and dynamic marking. The label propagation can be realized based on the vehicle map knowledge base, the missing label of the vehicle is further determined by combining a label propagation algorithm, the label inference is used for improving the understanding level of the machine to the label, and the accurate recommendation based on the label and the knowledge map is realized.

And step S212, combining the determined missing label of the vehicle and the pre-acquired initial vehicle label to generate a vehicle image result.

Specifically, a predetermined initial vehicle tag is obtained, and a plurality of vehicle tags associated with the triple vehicle data are generated in combination with the missing tag and the initial vehicle tag of the vehicle. And generating a complete vehicle portrait result by acquiring the preset weight corresponding to each vehicle label and based on each vehicle label and the corresponding preset weight.

The preset weights corresponding to different vehicle labels are different, and the strong degree of the interest of the object in the directions to which the different labels belong can be represented by setting the different preset weights.

In one embodiment, after generating the vehicle imaging result by combining the determined missing tag of the vehicle and the pre-acquired initial vehicle tag, the method further includes:

classifying each vehicle according to the vehicle portrait result to generate a corresponding vehicle classification result; and generating corresponding recommendation information according to the vehicle classification result, and displaying and broadcasting the recommendation information.

Specifically, according to the obtained vehicle image result, a corresponding vehicle classification result can be generated, and according to the vehicle classification result, corresponding recommendation information can be generated. And the generated vehicle portrait result completes the corresponding missing label, so that the accuracy of the generated recommendation information is correspondingly improved.

According to the knowledge graph-based vehicle portrayal method, the data structuring degree of original vehicle data is obtained, the classification of the original vehicle data is determined according to the data structuring degree, and corresponding original vehicle data are subjected to data processing according to the preprocessing modes corresponding to the original vehicle data of different classifications, so that corresponding triple vehicle data are generated. Knowledge fusion processing is carried out on the triple vehicle data to generate complete description knowledge of the vehicle, and then a vehicle map knowledge base can be constructed on the basis of the complete description knowledge. And performing knowledge calculation based on the constructed knowledge base of each vehicle map, determining the missing labels of the corresponding vehicles, and generating a complete vehicle portrait result by combining the determined missing labels of the vehicles and the pre-acquired initial vehicle labels. The behavior data of the vehicle are expanded by the aid of the knowledge graph, the range of vehicle information related to the vehicle data is widened, the existing labels and the missing labels determined based on the knowledge graph are combined to obtain the complete labels of the vehicle data, and further the complete vehicle portrait can be obtained, so that the accuracy of vehicle classification according to the complete vehicle portrait result is improved.

In one embodiment, as shown in fig. 3, the step of generating triple vehicle data, that is, the step of performing data processing on corresponding original vehicle data according to a preprocessing manner corresponding to the original vehicle data of different classifications to generate corresponding triple vehicle data specifically includes:

step S302, an extraction object corresponding to the unstructured vehicle data is obtained, wherein the extraction object comprises entities, attribute information of the entities and incidence relations among the entities.

Specifically, the classification of the raw vehicle data includes semi-structured vehicle data, and unstructured vehicle data. The extraction object corresponding to the obtained unstructured vehicle data comprises entities, attribute information of the entities and incidence relations among the entities.

Step S304, according to the extraction object, determining a first preset processing mode corresponding to the unstructured vehicle data.

Specifically, different extraction objects correspond to different preset processing modes, the extraction objects include entities, attribute information of the entities and association relations among the entities, and the first preset processing mode corresponding to the different extraction objects includes entity extraction processing, attribute extraction processing and relation extraction processing.

In the embodiment, the entities include concepts, persons, organizations, place names, time, and the like. The relationship extraction process represents the relationship between the entity to be extracted and the entity, and the attribute extraction process represents the need to extract the attribute information of the entity. The relationship reflects the external contact of the entity, and the attribute reflects the internal characteristics of the entity.

Step S306, named entities, attribute information of the entities and association relations among the entities are extracted from the unstructured vehicle data according to a first preset processing mode.

Specifically, named entities are extracted from the unstructured vehicle data according to an entity extraction process. The attribute information of each entity is extracted from the unstructured vehicle data according to the attribute extraction process. And extracting the association relation among the entities from the unstructured vehicle data according to the relation extraction processing.

For example, for relationship extraction, a deep learning method is adopted, two entities, their relationship and a sentence from origin are used as training data, a model is trained, then relationship extraction is performed on test data, the test data needs to provide the two entities and the sentence from origin, and the model is searched in a known relationship obtained by training to obtain the relationship between the two entities in the test data.

Step S308, generating triple vehicle data corresponding to the unstructured vehicle data according to the named entities, the attribute information respectively corresponding to the entities and the incidence relation among the entities.

Specifically, the processing for the unstructured vehicle data includes entity extraction processing, attribute extraction processing, and relationship extraction processing, and the processing results obtained accordingly are named entities, attribute information corresponding to the entities, and association relationships among the entities, respectively, so that by combining the named entities, the attribute information corresponding to the entities, and the association relationships among the entities, triple vehicle data corresponding to the unstructured vehicle data can be generated.

Step S310, extracting the semi-structured vehicle data from the webpage according to a second preprocessing mode corresponding to the semi-structured vehicle data, calling a preset wrapper, and restoring the extracted semi-structured vehicle data into the structured vehicle data.

In particular, the semi-structured data representation has a certain data structure, but still needs to be further extracted and sorted, such as data in a web page. The preset wrapper can be called by adopting a second preprocessing mode corresponding to the semi-structured vehicle data, the data are extracted from the webpage, and the semi-structured vehicle data obtained by extraction are restored into the structured vehicle data.

Step S312, converting the structured vehicle data into triple vehicle data according to a third preprocessing manner corresponding to the structured vehicle data.

Specifically, the structured data generally represents data of relational data, the data structure of which is clear, and vehicle data in the relational database can be directly converted into RDF data, namely triple vehicle data.

In this embodiment, an extraction object corresponding to the non-structural vehicle data is obtained, and a first preset processing mode corresponding to the non-structural vehicle data is determined according to the extraction object. And then extracting the named entities, the attribute information of the entities and the incidence relation among the entities from the unstructured vehicle data according to a first preset processing mode, and generating the triple vehicle data corresponding to the unstructured vehicle data according to the named entities, the attribute information respectively corresponding to the entities and the incidence relation among the entities. And extracting the semi-structured vehicle data from the webpage according to a second preprocessing mode corresponding to the semi-structured vehicle data, calling a preset wrapper, restoring the extracted semi-structured vehicle data into structured vehicle data, and converting the structured vehicle data into triple vehicle data according to a third preprocessing mode corresponding to the structured vehicle data. The method and the device realize accurate classification and pretreatment of the original vehicle data, the generated triple vehicle data can be directly applied, repeated processing operation in the subsequent vehicle classification process is reduced, and the work efficiency of vehicle classification is further improved.

In one embodiment, as shown in fig. 4, the step of constructing a vehicle map knowledge base, that is, the step of constructing the vehicle map knowledge base based on the complete description knowledge specifically includes:

and S402, determining vehicle basic information and vehicle violation information corresponding to the triple vehicle data based on the complete description knowledge.

Specifically, vehicle basic information and vehicle violation information corresponding to the triple vehicle data are extracted from the complete description knowledge, wherein the vehicle basic information comprises license plate numbers, vehicle types, vehicle systems, places of production, brands, vehicle body colors, registration dates and the like, and the vehicle violation information comprises information of fake plate vehicles, night-day exits, foothold points, first city entrances and the like.

And step S404, constructing a vehicle basic information base corresponding to the triple vehicle data according to the vehicle basic information.

Specifically, a vehicle basic information base corresponding to the triple vehicle data is constructed according to vehicle basic information including license plate numbers, vehicle types, vehicle series, production places, brands, vehicle body colors, registration dates and the like.

And S406, determining the vehicle violation business corresponding to the triple vehicle data according to the vehicle violation information, and constructing a vehicle violation business library according to the vehicle violation business.

Specifically, according to vehicle violation information including information of fake plate vehicles, night-day and night-night exits, foothold points, first city entering and the like, vehicle violation services corresponding to triple vehicle data are determined, including accident handling services, violation handling services, vehicle driving management services and the like, and then according to the determined vehicle violation services, a vehicle violation service library corresponding to the triple vehicle data is constructed.

And step S408, extracting the vehicle events corresponding to the triple vehicle data from the complete knowledge description, and constructing a corresponding vehicle event library based on the vehicle events.

Specifically, vehicle events corresponding to triple vehicle data are extracted from complete knowledge description, and the vehicle events comprise an alarm condition event library, a business event library and a civil event library, wherein the alarm condition event library comprises alarm condition types, corresponding handling mechanisms and basic information of handling objects, the business event library comprises business types, handling mechanisms for completing businesses and object information for applying for businesses, and similarly, the civil event library comprises civil types, handling mechanisms for processing based on civil meanings and object information for applying for objects.

Further, a vehicle event library corresponding to the triple vehicle data is constructed according to the extracted vehicle events.

And step S410, semantic mapping is carried out based on the vehicle event library, and a vehicle relation library is constructed, wherein the vehicle relation included in the vehicle relation library is a subject-object relation, a business process relation and a membership relation.

Specifically, semantic mapping is carried out based on the constructed vehicle event library, and different objects with association relations among different vehicle events and the association relations which specifically exist among the different objects are determined. The related incidence relations, namely the vehicle relations, comprise a subject-object relation, a business process relation and a membership relation. And then, a vehicle relation library is constructed according to the subject-object relation, the business process relation, the membership relation and the like.

Step S412, dynamic labeling is carried out based on the vehicle event library, and a vehicle label library is established, wherein the vehicle labels comprise vehicle basic labels, user labels, enterprise labels and event labels corresponding to vehicles, users and enterprises.

Specifically, dynamic labeling is performed based on a vehicle event library, and labels related to different objects in different vehicle events are labeled, wherein the labels include vehicle base labels, user labels and enterprise labels. The vehicle base tag can be determined according to the base information of the vehicle, the user tag can be understood as a tag of a user to which the vehicle belongs, and the enterprise tag represents a development enterprise to which the vehicle belongs. The event labels corresponding to different objects are also included for different objects, including vehicles, users, and businesses.

And step S414, acquiring the geographical position information corresponding to the complete description knowledge based on a preset third-party data interface, and constructing a corresponding geographical position library according to the geographical position information. The geographic location information includes vehicle location, route location, and zone division.

Specifically, based on a third-party data interface, semi-self-building of a base can be achieved, wherein the semi-self-building of the base comprises the step of obtaining geographic position information, portrait information and mechanism information corresponding to complete description knowledge, and the corresponding knowledge base comprises a geographic position base, a portrait information base and a mechanism base. The geographic position library comprises vehicle positions, route positions and region division, the portrait information library comprises personal information, vehicle information and enterprise information, and the organization library comprises related traffic management organizations and corresponding staff.

In this embodiment, the vehicle basic information and the vehicle violation information corresponding to the triple vehicle data are determined based on the complete description knowledge, and the vehicle basic information base corresponding to the triple vehicle data is constructed according to the vehicle basic information. And determining the vehicle violation business corresponding to the triple vehicle data according to the vehicle violation information, and constructing a vehicle violation business library according to the vehicle violation business. And extracting vehicle events corresponding to the triple vehicle data from the complete knowledge description, and constructing a corresponding vehicle event library based on the vehicle events. Semantic mapping is carried out based on the vehicle event library, a vehicle relation library is constructed, dynamic marking is carried out based on the vehicle event library, and a vehicle label library is established. Meanwhile, based on a preset third-party data interface, geographic position information corresponding to the complete description knowledge is obtained, and a corresponding geographic position library is constructed according to the geographic position information. The vehicle data knowledge map is established comprehensively, missed associated vehicle data are reduced, and vehicle portrait results generated according to the triple vehicle data are more comprehensive and accurate.

In one embodiment, as shown in fig. 5, an overall architecture of a vehicle map knowledge base building system is provided, and referring to fig. 5, the overall architecture of the vehicle map knowledge base building system includes a data background, an application center, and an application foreground.

The data background part comprises a knowledge source for determining original vehicle data, and aiming at the structured data, the data source comprises a manual input mode and a database butt joint mode, and can manually input a vehicle business process and data related to the vehicle business process, or acquire corresponding vehicle and vehicle type data by butt joint of a GIS service interface, a service system and the like. For semi-structured data, fields of the data need to be unified, wherein sources of the semi-structured data can include an alert data interface, a business data interface and a civil data interface. Aiming at unstructured data, namely text data, the knowledge to be extracted comprises entities, relations and attributes, the sources of the knowledge comprise data sources and labeled content, wherein the data sources comprise laws and regulations, external announcements, news reports and public opinion data, and the labeled content comprises portrait tags, case tags, proper nouns and relation refinements.

2) The application center part comprises three parts of knowledge acquisition, knowledge base construction and knowledge calculation.

a. For knowledge acquisition, the unstructured vehicle data includes entity extraction processing, attribute extraction processing, and relationship extraction processing. And directly converting the vehicle data in the relational database into the ternary group data aiming at the structured vehicle data. For semi-structured vehicle data, data is extracted from a webpage page by using a preset wrapper, and the data is restored into structured vehicle data. And aiming at knowledge acquisition, ontology construction is further included, and the ontology construction comprises data mapping processing, entity matching processing and entity fusion processing.

Further, for knowledge acquisition, knowledge fusion processing is carried out on the obtained triple vehicle data. The knowledge fusion processing comprises data integration processing, knowledge disambiguation processing and data fusion processing. The data integration processing is carried out on the triple vehicle data to generate complete data description corresponding to the triple vehicle data, and then entity alignment processing and attribute alignment processing are carried out on the complete data description of the triple vehicle data to generate the standard knowledge representation, so that data fusion processing can be carried out on the standard knowledge representation to generate complete description knowledge of the vehicle.

b. Aiming at knowledge base construction, the method comprises the following steps:

the method comprises the following steps of establishing an entity library by self, wherein the entity library comprises a business process library, a business document library and a special name word library, the business process library comprises accident handling business, illegal handling business, vehicle driving management business and the like, the business document library comprises certificates, business documents and certification materials, and the special name word library comprises contents of emerging technologies, field nouns, file specifications and the like.

And realizing a semi-self-built database based on a third-party data interface, wherein the semi-self-built database comprises a geographical position database, a portrait information database and an organization database, the geographical position database comprises a point position, a route position and a administrative region division condition, the portrait information database comprises personal information, vehicle information and enterprise information, and the organization database comprises related traffic management organizations and corresponding staff.

And establishing an event library, which comprises an alarm condition event library, a business event library and a civil event library, wherein the alarm condition event library comprises alarm condition types, corresponding disposal mechanisms and basic information of disposal objects, the business event library comprises business types, disposal mechanisms for completing businesses, object information for applying businesses, and similarly, the civil event library comprises civil types, disposal mechanisms for processing based on the civil and object information for applying for the objects.

And fourthly, performing semantic mapping based on the event library to establish a relational library, wherein the relational library comprises a subject-object relationship, a business process relationship and a membership relationship.

And fifthly, dynamically labeling based on the event library, and establishing a label library, wherein the label library comprises user labels, vehicle labels, enterprise labels and event labels corresponding to all the objects.

c. And aiming at the knowledge calculation and the knowledge calculation modes, which comprise rule management, analysis and mining, similarity calculation, association analysis, entity marking and dynamic marking, the knowledge calculation is carried out on the basis of the constructed vehicle map knowledge base, and the missing label of the vehicle is determined.

3) The application foreground part represents knowledge application, wherein the knowledge application comprises direct application such as service robot including online text processing and voice conversation, and offline service window robot guide such as citizen contact operation assistance including seat knowledge base search, police knowledge base search and the like. The method can also comprise key vehicle, crowd and enterprise portrayal, and police work research and judgment, including business object association analysis and business rule model mining.

The system comprises a client, a server, a client, a server and a server, wherein the server is connected with the client through a service interface, and the service interface comprises a metadata interface, a search interface, a data import interface, a rule mining interface, a visual display interface, a data authority management interface, a log audit interface, an application plug-in extension interface and the like.

In one embodiment, as shown in fig. 6, a basic knowledge graph building process is provided, and referring to fig. 6, the basic knowledge graph building process includes: and carrying out data integration and knowledge extraction on the original data consisting of the structured data, the semi-structured data and the unstructured time, and further realizing knowledge fusion. The knowledge fusion aims at entity alignment, and the entity alignment comprises entity disambiguation and coreference resolution to obtain standard knowledge representation. The standard knowledge representation can also be used for obtaining a data model by carrying out model construction according to the existing knowledge, so that data specification is realized, and then the standard knowledge representation is obtained. Wherein, the data model constructed according to the existing knowledge can be revised according to the generated standard knowledge representation.

The method comprises the steps of carrying out knowledge reasoning and knowledge discovery on standard knowledge, further evaluating obtained result data, and adding the result data meeting preset requirements into a knowledge graph. The knowledge inference is realized according to a description logic system, and means that new knowledge or conclusion is acquired according to the existing data model and data and an inference rule, and the new knowledge or conclusion is required to satisfy semantics. Quality assessment represents the assessment of the final result data, placing the qualified data in a knowledge graph. The quality evaluation method has differences on the difference of data requirements according to the difference of the constructed knowledge graph. The overall objective is to obtain satisfactory knowledge-graph data, the required criteria being determined on a case-by-case basis.

Furthermore, the data of the knowledge graph are divided into a data model and specific data, the data model represents a data organization frame of the knowledge graph, and different data models are adopted for different knowledge graphs. The construction of the knowledge graph can be completed by determining the data model of the knowledge graph firstly and supplementing data according to the frame agreed by the data model. The construction of the data model can integrate the requirements for data in the standard by referring to the relevant data standard of the industry to form a basic data model, and then the data model is perfected according to the data condition of actual collection. Or extracting from the public knowledge map data model, extracting the data model related to the industry from the public knowledge map data model, and then combining the industry knowledge for perfection.

In one embodiment, the knowledge-graph is constructed in a top-down and bottom-up manner. The top-down construction mode represents that a data model of the knowledge graph is determined firstly, and then specific data are filled according to the model to finally form the knowledge graph. The design of the data model is the top-level design of the knowledge graph, and the data model is determined according to the characteristics of the knowledge graph, which is equivalent to determining the range of data collected by the knowledge graph and the organization mode of the data. And the bottom-up construction mode means that specific data is collected in a triple mode, and then the data model is refined according to the data content. The knowledge graph is constructed by adopting the method, all data are collected to form a huge data set, and then the data are sorted, analyzed, summarized and summarized according to the data content and the characteristics of the data to form a framework so as to obtain a data model.

It should be understood that although the various steps in the flow charts of fig. 2-4 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-4 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternating with other steps or at least some of the sub-steps or stages of other steps.

In one embodiment, as shown in FIG. 7, a knowledge-graph based vehicle representation apparatus is provided, comprising: a vehicle data classification module 702, a triple vehicle data generation module 704, a complete description knowledge generation module 706, a vehicle atlas knowledge base construction module 708, a missing tag determination module 710, and a vehicle representation results generation module 712, wherein:

the vehicle data classification module 702 is configured to obtain a data structuring degree of the original vehicle data, and determine a classification of the original vehicle data according to the data structuring degree.

The triple vehicle data generating module 704 is configured to perform data processing on corresponding original vehicle data according to a preprocessing manner corresponding to the original vehicle data of different classifications, so as to generate corresponding triple vehicle data.

And the complete description knowledge generation module 706 is configured to perform knowledge fusion processing on the triple vehicle data to generate complete description knowledge of the vehicle.

A vehicle map knowledge base construction module 708 configured to construct a vehicle map knowledge base based on the complete description knowledge.

And the missing tag determining module 710 is configured to perform knowledge calculation based on the constructed vehicle map knowledge bases to determine missing tags of corresponding vehicles.

A vehicle portrayal result generation module 712 for generating vehicle portrayal results in combination with the determined missing vehicle tag and the pre-acquired initial vehicle tag.

The vehicle portrayal device based on the knowledge map expands the behavior data of the vehicle by using the knowledge map, widens the range of vehicle information related to the vehicle data, obtains the complete labels of the vehicle data by combining the existing labels and the missing labels determined based on the knowledge map, and further can obtain the complete vehicle portrayal, so that the accuracy of vehicle classification according to the complete vehicle portrayal result is improved.

In one embodiment, the triple vehicle data generation module is further to:

acquiring an extraction object corresponding to the unstructured vehicle data; the extraction object comprises entities, attribute information of the entities and incidence relation among the entities; determining a first preset processing mode corresponding to the unstructured vehicle data according to the extraction object; extracting named entities, attribute information of the entities and incidence relations among the entities from unstructured vehicle data according to a first preset processing mode; generating triple vehicle data corresponding to the unstructured vehicle data according to the named entities, the attribute information respectively corresponding to the entities and the incidence relation among the entities; extracting the semi-structured vehicle data from the webpage according to a second preprocessing mode corresponding to the semi-structured vehicle data, calling a preset wrapper, and restoring the extracted semi-structured vehicle data into structured vehicle data; and converting the structured vehicle data into the triple vehicle data according to a third preprocessing mode corresponding to the structured vehicle data.

In the embodiment, the accurate classification and pretreatment of the original vehicle data are realized, the generated triple vehicle data can be directly applied, the repeated processing operation in the subsequent vehicle classification process is reduced, and the work efficiency of vehicle classification is further improved.

In one embodiment, the vehicle map knowledge base building module is further configured to:

determining vehicle basic information and vehicle violation information corresponding to the triple vehicle data based on the complete description knowledge; according to the vehicle basic information, a vehicle basic information base corresponding to the triple vehicle data is constructed; determining a vehicle violation business corresponding to the triple vehicle data according to the vehicle violation information, and constructing a vehicle violation business library according to the vehicle violation business; and extracting vehicle events corresponding to the triple vehicle data from the complete knowledge description, and constructing a corresponding vehicle event library based on the vehicle events.

In this embodiment, the vehicle events corresponding to the triple vehicle data are extracted from the complete knowledge description, and a corresponding vehicle event library is constructed based on the vehicle events. The vehicle data knowledge map is established comprehensively, missed associated vehicle data are reduced, and vehicle portrait results generated according to the triple vehicle data are more comprehensive and accurate.

semantic mapping is carried out on the basis of the vehicle event library, and a vehicle relation library is constructed; the vehicle relation library comprises vehicle relations including a subject-object relation, a business process relation and a membership relation; dynamically labeling based on the vehicle event library, and establishing a vehicle label library; the vehicle tags comprise vehicle base tags, user tags, enterprise tags and event tags corresponding to vehicles, users and enterprises; acquiring geographical position information corresponding to the complete description knowledge based on a preset third-party data interface, and constructing a corresponding geographical position library according to the geographical position information; the geographic location information includes vehicle location, route location, and zone division.

In the embodiment, the comprehensive vehicle data knowledge map is established, the missed associated vehicle data is reduced, and the vehicle portrait result generated according to the triple vehicle data is more comprehensive and accurate.

In one embodiment, the vehicle representation results generation module is further to:

acquiring a predetermined initial vehicle tag; generating a plurality of vehicle tags associated with the triple vehicle data in combination with the missing tag and the initial vehicle tag of the vehicle; acquiring preset weight corresponding to each vehicle label; and generating a complete vehicle portrait result based on each vehicle label and the corresponding preset weight.

In one embodiment, the full description knowledge generation module is further to:

carrying out data integration processing on the triple vehicle data to generate a complete data description corresponding to the triple vehicle data; carrying out entity alignment processing and attribute alignment processing on the complete data description of the triple vehicle data to generate a standard knowledge representation; and carrying out data fusion processing on the standard knowledge representation to generate complete description knowledge of the vehicle.

In one embodiment, the missing tag determination module is further configured to:

based on the constructed vehicle map knowledge base, similarity calculation is carried out to determine candidate vehicles; semantic association is carried out on each candidate vehicle, random walk is carried out on the basis of vehicle basic information, and a random walk result is generated; performing association analysis based on the random walk result, and determining entities potentially associated with the vehicle basic information; and extracting entity labels of entities which are potentially associated with the vehicle basic information, and determining the extracted entity labels as missing labels of the corresponding vehicles.

For specific limitations of the map-based vehicle-representation apparatus, reference may be made to the above limitations of the map-based vehicle-representation method, which are not described in detail herein. The modules of the aforementioned knowledgegraph-based vehicle representation apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 8. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store the triplet vehicle data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of knowledge-graph based vehicle portrayal.

Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, there is provided a computer device comprising a memory storing a computer program and a processor implementing the following steps when the processor executes the computer program:

acquiring the data structuralization degree of the original vehicle data, and determining the classification of the original vehicle data according to the data structuralization degree;

according to the preprocessing modes corresponding to the original vehicle data of different classifications, data processing is carried out on the corresponding original vehicle data to generate corresponding triple vehicle data; carrying out knowledge fusion processing on the triple vehicle data to generate complete description knowledge of the vehicle;

constructing a vehicle map knowledge base based on the complete description knowledge; performing knowledge calculation based on the constructed vehicle map knowledge base to determine missing labels of corresponding vehicles;

In one embodiment, the processor, when executing the computer program, further performs the steps of:

acquiring an extraction object corresponding to the unstructured vehicle data; the extraction object comprises entities, attribute information of the entities and incidence relation among the entities; determining a first preset processing mode corresponding to the unstructured vehicle data according to the extraction object;

extracting named entities, attribute information of the entities and incidence relations among the entities from unstructured vehicle data according to a first preset processing mode;

generating triple vehicle data corresponding to the unstructured vehicle data according to the named entities, the attribute information respectively corresponding to the entities and the incidence relation among the entities;

extracting the semi-structured vehicle data from the webpage according to a second preprocessing mode corresponding to the semi-structured vehicle data, calling a preset wrapper, and restoring the extracted semi-structured vehicle data into structured vehicle data;

and converting the structured vehicle data into the triple vehicle data according to a third preprocessing mode corresponding to the structured vehicle data.

carrying out data integration processing on the triple vehicle data to generate a complete data description corresponding to the triple vehicle data;

carrying out entity alignment processing and attribute alignment processing on the complete data description of the triple vehicle data to generate a standard knowledge representation; and carrying out data fusion processing on the standard knowledge representation to generate complete description knowledge of the vehicle.

based on the constructed vehicle map knowledge base, similarity calculation is carried out to determine candidate vehicles;

semantic association is carried out on each candidate vehicle, random walk is carried out on the basis of vehicle basic information, and a random walk result is generated; performing association analysis based on the random walk result, and determining entities potentially associated with the vehicle basic information;

and extracting entity labels of entities which are potentially associated with the vehicle basic information, and determining the extracted entity labels as missing labels of the corresponding vehicles.

determining vehicle basic information and vehicle violation information corresponding to the triple vehicle data based on the complete description knowledge; according to the vehicle basic information, a vehicle basic information base corresponding to the triple vehicle data is constructed;

semantic mapping is carried out on the basis of the vehicle event library, and a vehicle relation library is constructed; the vehicle relation library comprises vehicle relations including a subject-object relation, a business process relation and a membership relation;

In one embodiment, a computer storage medium is provided, having a computer program stored thereon, the computer program, when executed by a processor, implementing the steps of:

carrying out knowledge fusion processing on the triple vehicle data to generate complete description knowledge of the vehicle; constructing a vehicle map knowledge base based on the complete description knowledge;

performing knowledge calculation based on the constructed vehicle map knowledge base to determine missing labels of corresponding vehicles; and combining the determined missing label of the vehicle and the pre-acquired initial vehicle label to generate a vehicle image result.

In one embodiment, the computer program when executed by the processor further performs the steps of:

acquiring an extraction object corresponding to the unstructured vehicle data; the extraction object comprises entities, attribute information of the entities and incidence relation among the entities;

determining a first preset processing mode corresponding to the unstructured vehicle data according to the extraction object;

determining a vehicle violation business corresponding to the triple vehicle data according to the vehicle violation information, and constructing a vehicle violation business library according to the vehicle violation business; and extracting vehicle events corresponding to the triple vehicle data from the complete knowledge description, and constructing a corresponding vehicle event library based on the vehicle events.

acquiring a predetermined initial vehicle tag; generating a plurality of vehicle tags associated with the triple vehicle data in combination with the missing tag and the initial vehicle tag of the vehicle;

acquiring preset weight corresponding to each vehicle label; and generating a complete vehicle portrait result based on each vehicle label and the corresponding preset weight.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method of vehicle portrayal based on an intellectual graph, the method comprising:

2. The method of claim 1, wherein the classification of the raw vehicle data comprises semi-structured vehicle data, and unstructured vehicle data; the data processing of the corresponding original vehicle data according to the preprocessing mode corresponding to the original vehicle data of different classifications to generate the corresponding triple vehicle data includes:

3. The method of claim 1, wherein the knowledge fusion process comprises a data integration process, a knowledge disambiguation process, and a data fusion process; the knowledge disambiguation process comprises an entity alignment process and an attribute alignment process; the knowledge fusion processing is carried out on the triple vehicle data to generate complete description knowledge of the vehicle, and the method comprises the following steps:

4. The method of any one of claims 1 to 3, wherein the knowledge calculations include similarity calculations and correlation analyses; the step of performing knowledge calculation based on the constructed vehicle map knowledge bases to determine missing labels of corresponding vehicles comprises the following steps:

5. The method of any one of claims 1 to 3 wherein the vehicle map knowledge base comprises a vehicle basic information base, a vehicle violation services base and a vehicle event base; the building of a vehicle map knowledge base based on the complete description knowledge comprises:

6. The method of claim 5, wherein the vehicle map knowledge base further comprises a vehicle relationship base, a vehicle tag base, and a geographic location base; the building of a vehicle map knowledge base based on the complete description knowledge comprises:

7. The method of any one of claims 1 to 3, wherein the combining the determined missing tag of the vehicle and the pre-acquired initial vehicle tag to generate a vehicle imagery result comprises:

acquiring a predetermined initial vehicle tag;

acquiring a preset weight corresponding to each vehicle label;

8. The method of any one of claims 1 to 3, further comprising, after generating a vehicle imaging result in combination with the determined missing tag of the vehicle and the pre-acquired initial vehicle tag:

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 8.

10. A computer storage medium on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 8.