Basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags

Abstract

The present invention discloses a basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags, including the steps of spatiotemporal tagging of basic data; establishing association relationships between tags; and one-click retrieval of basic data based on spatiotemporal tags. The advantages of the present invention are that: through the present invention, people, places, objects, organizations, and events in smart city management are labeled and classified, and logical relationships between data are constructed, and spatial relationships and movement trajectory relationships between related elements are visualized using maps, effectively supporting smart cities to manage targets throughout their life cycle.

Description

Technical Field

[0001] The invention relates to a data retrieval and spatial information display method, in particular to a basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags, belonging to the field of data retrieval and spatial information display. Background Art

[0002] In the field of smart city management, it is necessary to monitor various components, people, buildings, events, and legal entities in city management in real time. In the existing technology, basic data such as people, components, buildings, events, and legal entities are supported for query and displayed on the map based on coordinate annotation. Data association query is slow and does not support large-scale data map display. Therefore, it is the key to technology and development to establish spatiotemporal labels for the basic data model, obtain the spatiotemporal relationship of the basic data, and retrieve based on the spatiotemporal labels and use a specific spatial map visualization display method to solve the current problems.

[0003] In the existing technology, there are some problems in basic data query and retrieval and large-scale data map display methods:

[0004] (1) Search and retrieval methods for people, places, objects, organizations, and events:

[0005] By establishing primary and foreign key associations between people and houses, people and events, people and organizations, and places and objects, data query is performed through table links, which supports searching people by house, searching people by events, and searching events by people. However, it is impossible to fully establish the logical relationship between people, houses, events, organizations, and objects. On the other hand, as the amount of data increases, the query speed of two or more tables is very slow.

[0006] (2) Large-scale data map display:

[0007] When a large-scale map is displayed and the loading exceeds a certain amount, the map dragging and zooming functions are almost ineffective, the lag is very serious, and the browser is under great pressure. Summary of the invention

[0008] The purpose of the present invention is to design a basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags, establish spatiotemporal tags for basic data, improve the efficiency of user retrieval of basic data and establish associations between basic data, and integrate geographic spatial maps for visual display.

[0009] The technical solution of the present invention is:

[0010] The basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags includes the following steps:

[0011] Step S100: input basic urban management data such as people, places, objects, organizations, events, etc.;

[0012] Step S200: extracting entity features and attribute features, spatial geographic location features, time features, and inter-entity relationship features based on the basic data subject data table definition and data entity field information;

[0013] Step S300: Organize the features of step S200 to assist in manual selection to form basic data classification features such as people, places, objects, organizations, and events, and then form a spatiotemporal classification label system;

[0014] Step S400: perform sampling cleaning on the basic data, use the category mapping rules, and abstract the cleaning process into business rules, execute rule cleaning, perform global cleaning and iteration, and form sample data;

[0015] Step S500: Based on the sample data in step S400 and the spatiotemporal classification labels in step S300, Chinese text classification based on the Text-CNN model is used to integrate the basic data logical relationship and label the basic data entities. An entity may include multiple labels to describe the association relationship between entities and the entity time trajectory relationship;

[0016] Step S600: providing a data retrieval method for tag logical association relationships and business scenario tag groups, querying basic data, and returning a basic data result set;

[0017] Step S700: Based on the business logic relationship between the spatiotemporal tags, the basic data spatial association relationship found out by the query is displayed on the geographic spatial information;

[0018] Step S800: Based on the temporal trajectory relationship between the spatiotemporal tag entities, that is, the movement trajectory tag information of the basic data entities within the time period, the movement trajectory of the basic data entities is displayed on the geographic space information;

[0019] Step S900: Provide a geographic space information range, use a data interface to obtain basic data entities for query within the space range, use HTML CANVAS canvas to render output results in real time, filter out points that cannot be displayed, and achieve rapid display on the map.

[0020] Furthermore, the step S500 includes the following sub-steps:

[0021] (1) Based on the input basic data, extract the text fields in each record and combine them to form a text segment of the record. Through text preprocessing, the main purpose is to eliminate meaningless symbol information or other redundant information;

[0022] (2) Further digitize the text, that is, use numbers to represent specific words, because computers cannot directly process words created by humans. In order for computers to understand words, it is necessary to map the word information to a numerical semantic space, which is called the word vector space; use word2vec, an unsupervised learning model tool, to achieve the mapping of word information to the semantic space, and finally obtain a word vector model;

[0023] (3) Using the text convolutional neural network model, we can extract primary features from the input fixed-length text sequence using local word order information and combine the primary features into high-level features to solve the problem of multi-label scenarios of basic data and non-text feature access. At the same time, we can use deep models to achieve better generalization effects.

[0024] Furthermore, the step S600 includes the following sub-steps:

[0025] (1) Exact match query, which retrieves results by accurately matching the specified fields with the input query tag conditions;

[0026] (2) Fuzzy query, where the search field contains query tag conditions that match the input;

[0027] (3) Based on the label logical relationship, it allows any number of queries to be combined in a single query, and specifies which parts of the query statement table name must match, should match, or cannot match, to perform combined queries and nested relationship queries.

[0028] Furthermore, the step S700 includes the following sub-steps:

[0029] (1) Rendering query results, showing the queried people, places, objects, organizations, and event information on a map;

[0030] (2) Select a person, place, object, organization, or event to view the logical relationships between tags.

[0031] Furthermore, the step S800 includes the following sub-steps:

[0032] (1) Trajectory data visualization: Draw a trajectory path diagram on the map by moving the basic data trajectory to highlight the spatial position of the trajectory;

[0033] (2) Using the movement of time attributes to dynamically display the spatial position of the trajectory;

[0034] (3) When the amount of trajectory data is large, the trajectory data is classified and aggregated to obtain aggregated data, and then the aggregated data is visualized.

[0035] The present invention provides a method for visualizing and integrating geographic information based on massive data, which displays the correlation relationship and movement trajectory between basic data.

[0036] Data association relationships are associated through database primary and foreign keys, and data retrieval is completed by adding indexes to the key fields of the association relationships and basic data and displayed on the map. This traditional technical implementation method makes it difficult to ensure that the data is accurately presented in a meaningful and operable way for end users when displaying large amounts of data in GIS; in addition, large amounts of data cause a huge load on the browser, and it takes a long time to see the results. Therefore, the multi-level multi-branch tree labeling system provided by the present invention implements labeling of basic data, realizes the association relationship between data, establishes a label index retrieval system, supports one-click data retrieval services, and uses the memory canvas method to dynamically update the points on the map using the map rendering input, update, and exit modes. After the map is zoomed and roamed, the data is retrieved again according to the visible range of the map, and the points that cannot be effectively rendered are filtered out, thereby reducing the number of points in SVG and realizing the rapid display of massive data on the map.

[0037] The present invention supports the integrated display of spatial relationships of massive smart city basic data on people, land, objects, organizations, and events. The basic data is used to provide users with rich information and support decision-making. On the one hand, richer information is displayed to users on the front end, and on the other hand, the category search needs of the search are supported. It mainly has rich polysemous expressions in map scene query and decision support. It is difficult to achieve the purpose through traditional text matching engines or simple synonym generalization; therefore, by mining labels to form a label category system of associations between people, land, objects, events, and organizations, the user's actual query scenario is realized to support the user's search needs; establish the objective category distribution of the real world, and the cognition of this distribution; support the subordinate and parallel relationships between different labels. Finally, each major category will construct a multi-layer multi-branch tree system.

[0038] The beneficial effects of the present invention are as follows: through the present invention, people, places, objects, organizations, and events in smart city management are labeled and classified, and the logical relationship between data is constructed, and the spatial relationship and movement trajectory relationship between related elements are displayed visually using maps, effectively supporting the smart city to manage the entire life cycle for the target.

[0039] The present invention will be further described below in conjunction with the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 : A flowchart of an example of basic data retrieval and integrated display based on spatiotemporal association of spatiotemporal tags of the present invention.

[0041] Figure 2: A system structure diagram of an example of basic data retrieval and integrated display based on spatiotemporal association of spatiotemporal tags in the present invention.

[0042] Figure 3 : The present invention is a result diagram of the spatiotemporal label system that is based on the basic data retrieval and integrated display of the spatiotemporal association of spatiotemporal labels. DETAILED DESCRIPTION

[0043] The preferred embodiments of the present invention are described below. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0044] Example 1

[0045] like Figure 1 As shown, a basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags includes the following steps:

[0046] Step S100: input basic urban management data such as people, places, objects, organizations, events, etc.;

[0047] Step S200: extracting entity features and attribute features, spatial geographic location features, time features, and inter-entity relationship features based on the basic data subject data table definition and data entity field information;

[0048] Step S300: Organize the features of step S200 to assist in manual selection to form basic data classification features such as people, places, objects, organizations, and events, and then form a spatiotemporal classification label system;

[0049] Step S400: First, perform sampling cleaning on the basic data, use the category mapping rules, and abstract the cleaning process into business rules, execute rule cleaning, perform global cleaning and iteration, and form sample data;

[0050] Step S500: Based on the sample data in step S400 and the spatiotemporal classification labels in step S300, Chinese text classification based on the Text-CNN model is used to integrate the basic data logical relationship and label the basic data entities. An entity may include multiple labels to describe the association relationship between entities and the entity time trajectory relationship;

[0051] The specific steps include the following:

[0052] (1) Based on the input basic data, extract the text fields in each record and combine them to form a text segment of the record. Through text preprocessing, the main purpose is to eliminate meaningless symbol information or other redundant information;

[0053] (2) Further digitize the text, that is, use numbers to represent specific words, because computers cannot directly process words created by humans. In order for computers to understand words, it is necessary to map the word information to a numerical semantic space, which is called the word vector space; use word2vec, an unsupervised learning model tool, to achieve the mapping of word information to the semantic space, and finally obtain a word vector model;

[0054] (3) Using the text convolutional neural network (CNN) model, we extract primary features from the input fixed-length text sequence using local word order information and combine the primary features into high-level features to solve the problem of multi-label scenarios of basic data and non-text feature access. At the same time, we use deep models to achieve better generalization effects.

[0055] Step S600: providing a data retrieval method for tag logical association relationships and business scenario tag groups, querying basic data, and returning a basic data result set;

[0056] The specific steps include the following:

[0057] (1) Exact match query, which retrieves results by accurately matching the specified fields with the input query tag conditions;

[0058] (2) Fuzzy query, where the search field contains query tag conditions that match the input;

[0059] (3) Based on the label logical relationship, it allows any number of queries to be combined in a single query, and specifies which parts of the query statement table name must match, should match, or cannot match, to perform combined queries and nested relationship queries.

[0060] Step S700: Based on the business logic relationship between the spatiotemporal tags, the basic data spatial association relationship found out by the query is displayed on the geographic spatial information;

[0061] The specific steps include the following:

[0062] (1) Rendering query results, showing the queried people, places, objects, organizations, and event information on a map;

[0063] (2) Select a person, place, object, organization, or event to view the logical relationships between tags.

[0064] Step S800: Based on the temporal trajectory relationship between the spatiotemporal tag entities, that is, the movement trajectory tag information of the basic data entities within the time period, the movement trajectory of the basic data entities is displayed on the geographic space information;

[0065] The specific steps include the following:

[0066] (1) Trajectory data visualization: Draw a trajectory path diagram on the map by moving the basic data trajectory to highlight the spatial position of the trajectory;

[0067] (2) Using the movement of time attributes to dynamically display the spatial position of the trajectory;

[0068] (3) When the amount of trajectory data is large, the trajectory data is classified and aggregated to obtain aggregated data, and then the aggregated data is visualized;

[0069] Step S900: Provide the geographic space information range, use the data interface to obtain the basic data entity of the query within the space range, use the canvas (HTML CANVAS) to render the output results in real time, filter the points that cannot be displayed, and realize fast display on the map, such as Figure 3 As shown, by labeling people, places, objects, and organizations, a spatiotemporal labeling system is formed, with time as the axis and events as the association, forming a spatiotemporal correlation relationship between people, places, objects, and organizations.

[0070] The people mentioned above are mainly labeled, their characteristics are extracted, and the group classification of people is obtained; based on the basic attributes of the population, people are labeled into registered population and floating population; based on the management and service attributes of the population, the population is labeled into key management and key service groups. For the service population, the service population is further refined into disabled people, low-income households, elderly people over 60 years old, women of childbearing age, teenagers under 18 years old, and idle youth in society through basic data information such as age, service content, and assistance in the population. Through the management attributes of the population, the population is divided into community correction personnel, prisoners released after serving their sentences, and patients with severe mental disorders.

[0071] The places mentioned above are mainly labeled and classified; based on the basic information and usage of the places, historical events that have occurred, etc., the places are divided into streets, lanes, communities, villages, buildings and places, houses, service places, parks and tourist attractions, vegetable markets, community public spaces, underground spaces, and key management places.

[0072] The objects are mainly labeled as components, and the labels of all objects are extracted; the equipment and facilities are divided into urban management component labels.

[0073] The organizations described are mainly for classifying and describing various social organizations and legal persons; classification is based on the basic information and relevant attributes of the organizations, and the organizations are labeled, and classified into: educational and training institutions, medical and health institutions, social work institutions, grassroots administrative agencies, mass autonomous organizations, social organizations, and non-public economic organizations.

[0074] The events mentioned above are mainly defined by time (a certain time), space (a certain place) and association (event) formed by associating people, places, objects and organizations based on the time axis. Classification labels are added based on time, event attributes and event content, and events are classified into difficult events, repeated events, overtime events, handled events, mass events, urban management events and environmental protection events.

[0075] like Figure 2 As shown, the present invention also provides a basic data display system based on spatiotemporal label spatial relationship, which includes four parts: basic data input, basic data spatiotemporal label system, one-click retrieval, and basic data spatial relationship display module.

[0076] The basic data input module is used for maintaining basic data of people, places, objects and organizations, and supports data import, entry, modification and query.

[0077] The basic data spatiotemporal tag system module is used for system spatiotemporal tag management, maintaining spatiotemporal tag basic metadata and system application algorithm to tag data to form a tag library.

[0078] The feature value extraction module mainly extracts basic attributes of people, places, objects, organizations and events.

[0079] The feature classification module mainly uses classification algorithms to extract various attributes and texts such as people, places, objects, organizations, events, etc.

[0080] The spatiotemporal tag system completes the addition, deletion, modification and maintenance of system spatiotemporal tags.

[0081] The attribute tags are used to tag people, places, objects, organizations, and events.

[0082] The classification labels are classified by using the algorithm results to label people, places, objects, organizations, and events.

[0083] The time tags are used to tag people, places, objects, organizations, and events with time attributes.

[0084] The spatial tags are used to label the spatial locations of people, places, objects, organizations, and events.

[0085] The one-click search module is used to provide a tag-based search for the formed basic data with tags and return the search results.

[0086] The basic data spatial relationship display module is used for basic data spatial relationship display, provides integrated display of multiple relationships of basic data based on GIS, and supports two situations: display of spatiotemporal relationship migration and display of current spatiotemporal correlation relationship.

[0087] The display of the migration of time and space relationships mainly displays the historical changes of people, places, objects and organizations.

[0088] The above-mentioned display of time-space correlation is mainly based on the comprehensive display of people, places, objects and organizations associated with events.

Claims

1. A basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags, characterized by: The following steps are involved: Step S100: input basic urban management data of people, places, objects, organizations and events; Step S200: extracting entity features and attribute features, spatial geographic location features, time features, and inter-entity relationship features based on the basic data subject data table definition and data entity field information; Step S300: Organize the features of step S200 to assist in manual selection to form basic data classification features of people, places, objects, organizations, and events, and then form a spatiotemporal classification label system; Step S400: First, perform sampling cleaning on the basic data, use the category mapping rules, and abstract the cleaning process into business rules, execute rule cleaning, perform global cleaning and iteration, and form sample data; Step S500: Based on the sample data in step S400 and the spatiotemporal classification labels in step S300, the Chinese text classification based on the Text-CNN model is used to integrate the basic data logical relationship and label the basic data entities. An entity includes multiple labels to describe the association relationship between entities and the entity time trajectory relationship; Step S600: providing a data retrieval method for tag logical association relationships and business scenario tag groups, querying basic data, and returning a basic data result set; Step S700: Based on the business logic relationship between the spatiotemporal tags, the basic data spatial association relationship found out by the query is displayed on the geographic spatial information; Step S800: Based on the temporal trajectory relationship between the spatiotemporal tag entities, that is, the movement trajectory tag information of the basic data entities within the time period, the movement trajectory of the basic data entities is displayed on the geographic space information; Step S900: Provide a geographic space information range, use a data interface to obtain basic data entities for query within the space range, use HTMLCANVAS canvas to render output results in real time, filter out points that cannot be displayed, and achieve rapid display on the map.

2. The basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags according to claim 1 is characterized in that: The step S500 includes the following sub-steps: (1) extracting text fields from each record based on the input basic data, combining them to form a text of the record, and removing meaningless symbol information or other redundant information through text preprocessing; (2) further digitizing the text, that is, using numbers to represent specific words, because computers cannot directly process words created by humans; in order for computers to understand words, it is necessary to map the word information to a digitized semantic space, which is called a word vector space; using word2vec, an unsupervised learning model tool, to achieve the mapping of word information to the semantic space, and finally obtain a word vector model; (3) using a text convolutional neural network model, from the input fixed-length text sequence, using local word order information, extracting primary features, and combining the primary features into advanced features, solving the problem of multi-label scenarios and non-text feature access of basic data, and using a deep model to achieve better generalization effect.

3. The basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags according to claim 1 is characterized in that: The step S600 includes the following sub-steps: (1) exact match query, retrieving results by exactly matching the input query tag condition to the specified field; (2) Fuzzy query, which includes query tag conditions that match the input in the search field; (3) Based on the logical relationship of tags, it allows any number of queries to be combined in a single query, and specifies which parts of the query statement table name must match, should match, or cannot match, to perform combined queries and nested relationship queries.

4. The basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags according to claim 1 is characterized in that: The step S700 includes the following sub-steps: (1) rendering the query results, displaying the queried person, place, object, organization, and event information on the map; (2) selecting a person, place, object, organization, or event, and viewing the logical association relationship between tags.

5. The basic data retrieval and integrated display method based on spatiotemporal association of spatiotemporal tags according to claim 1 is characterized in that: The step S800 includes the following sub-steps: (1) visualizing the trajectory data, drawing a trajectory path diagram of the basic data movement trajectory on the map to highlight the spatial position of the trajectory; (2) dynamically displaying the spatial position of the trajectory by using the movement of the time attribute; (3) when the amount of trajectory data is large, the trajectory data is obtained by classification and aggregation calculation to obtain aggregated data, and then the aggregated data is visualized.

Images