CN111666313A - Correlation construction and multi-user data matching method based on multi-source heterogeneous remote sensing data - Google Patents
Correlation construction and multi-user data matching method based on multi-source heterogeneous remote sensing data Download PDFInfo
- Publication number
- CN111666313A CN111666313A CN202010447854.XA CN202010447854A CN111666313A CN 111666313 A CN111666313 A CN 111666313A CN 202010447854 A CN202010447854 A CN 202010447854A CN 111666313 A CN111666313 A CN 111666313A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- remote sensing
- similarity
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
- G06F16/316—Indexing structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Abstract
The invention discloses a multi-source heterogeneous remote sensing data association construction and multi-user data matching method, which comprises the following steps: the method comprises the following steps: inputting a remote sensing service data set and a remote sensing product data set; step two: carrying out data cleaning and pretreatment; step three: carrying out data classification modeling; step four: extracting knowledge, including attribute extraction of remote sensing service data and feature extraction of remote sensing product data; step five: performing multi-source heterogeneous knowledge fusion, constructing a link relation of a plurality of knowledge systems, and fusing the attribute similarity of different body data according to metadata information extracted from remote sensing data to construct the link relation of the plurality of knowledge systems; step six: constructing an incidence relation model; step seven: mining and analyzing a large amount of collected user identity and behavior data to construct a user portrait label; step eight: carrying out similarity calculation on the identity and the behavior of the user; step nine: and constructing a personalized user portrait model, and matching the personalized user portrait model with a result in the incidence relation construction model.
Description
Technical Field
The invention relates to the field of data processing, in particular to a multi-source heterogeneous remote sensing data association construction and multi-user data matching method.
Background
In recent years, the remote sensing earth observation technology in China is rapidly developed, some fields are ascending internationally, and remote sensing data is widely applied to various fields of national economy and social development. In order to meet the continuously deepened application requirements, the number of satellites is continuously increased, the resolution/width of remote sensing images is continuously improved, a large amount of visible light, hyperspectral data, SAR and other various remote sensing data are obtained, the data can reach more than 10TB a day, users in different fields can accurately and intelligently recognize the user requirements, accurate recommendation of remote sensing data products is timely carried out, the service requirements of different users are met, and the method is a great challenge for application and popularization of the remote sensing data and realization of commercial value.
In the related field of remote sensing data popularization service, the existing system browses, downloads and makes remote sensing data in a way that a user screens out data from a huge database in a way of interacting with a computer visual terminal; in the field of remote sensing data user portrait construction, the existing system has great force for dividing users, and generally performs coarse-grained division from industries or business modes, such as troops, scientific research institutes, government departments and higher colleges, and cannot specifically construct user portrait on the application demand level.
The remote sensing big data era has various sources, huge data volume and complex incidence relation. When a user faces a data growth rate of a daily TB level magnitude, the size of the amount of monoscopic remote sensing data is also GB level, most of the existing systems call the remote sensing data through human-computer interaction, on one hand, available data cannot be found in time, on the other hand, the useful information of the extracted data is judged by means of manual interpretation, the efficiency is low, and no unified measurement standard exists; in addition, in the aspect of accurate positioning of users, the variety of remote sensing data users is considered, the existing system lacks multi-dimensional understanding of the users, cannot carry out conjecture and modeling analysis of data use intention by combining user information, and is difficult to provide accurate recommendation service guided by user requirements.
Disclosure of Invention
The invention aims to solve the problem of conversion from abstract remote sensing data to concrete information knowledge, establish an incidence relation between multisource heterogeneous data through means such as data characteristic fusion and the like, and provide a data basis for intelligent data recommendation; and further, user information is deeply mined, mining modeling is carried out on the user, and accurate recommendation service which is guided by requirements is provided for the user.
The mining and knowledge graph technology is applied to the technical field of remote sensing, the data mining technology is applied to feature extraction, a feature fusion and knowledge reasoning framework oriented to multi-source heterogeneous remote sensing data is constructed, organization association of the multi-source data is realized, and support is provided for subsequent data application; meanwhile, by means of a mining analysis technology, user information resources are comprehensively and systematically collected, main characteristics such as user background information and behavior habits are mined and analyzed, a personalized user portrait model is built, and data matching is carried out in combination with data association modeling. The method maximally mines the available information of the data and timely shares the multi-source heterogeneous data, so that a user comprehensively and generally grasps all information of interested data, and the use value of the remote sensing data is maximally realized.
The invention provides a multi-source heterogeneous remote sensing data association construction and multi-user data matching method, which comprises the following steps:
the method comprises the following steps: inputting a remote sensing service data set and a remote sensing product data set, wherein the remote sensing service data set comprises target, track, text and picture data; the remote sensing product data comprises Sar images, panchromatic images, multispectral image data and visible light data remote sensing image data;
step two: performing data cleansing and preprocessing, the data cleansing comprising: carrying out grammar check, spelling check, missing value processing, duplicate removal processing, invalid character removal and noise processing on the remote sensing service data; the preprocessing comprises the steps of carrying out image quality inspection on remote sensing product data by means of manual interpretation, and firstly carrying out image preprocessing on images with deformation and cloud and fog interference; removing unnecessary thick cloud areas and removing cloud and mist; carrying out deformation correction, aiming at the situation that an airport and a road are bent due to the geometric deformation of SAR imaging, carrying out deformation correction by using an image non-uniform sampling method, and resampling the position or the distance of pixels by using a bilinear interpolation or cubic interpolation method;
step three: carrying out data classification modeling, inputting the data classification modeling into the cleaned data set in the second step, and mining all class association rules with specified confidence degrees from the training data body sample set by using an association rule mining algorithm based on the classification of the association rules by using a classifier by adopting an association rule mining method in data mining; then iterating optimal rules from the mined class association rules for classification, finally gathering the entities representing the same object together by using the optimal rules, and combining a plurality of heterogeneous entities into a global unified entity;
step four: extracting knowledge, including attribute extraction of remote sensing service data and feature extraction of remote sensing product data; the attribute extraction of the remote sensing service data comprises the extraction of name, time, space, uploader, country, data format, keyword and metadata information; the metadata information comprises data types, data names, resolutions, satellites, sensors, including targets, scene numbers, positions, ranges, product levels and thumb maps; firstly, extracting an interested region according to a phase spectrum saliency map of a remote sensing image by the feature extraction; carrying out homogeneous filtering in the region of interest, and confirming a target region to be detected by combining a phase spectrum significance map; in the target area, calculating the main shaft direction of each target, and extracting an S-HOG feature descriptor of the target; judging an S-HOG feature descriptor of the target according to the shape knowledge of the target, and judging, identifying and extracting the target;
step five: performing multi-source heterogeneous knowledge fusion, constructing a link relation of a plurality of knowledge systems by fusing the attribute similarity of different ontology data according to metadata information extracted from remote sensing data;
step six: constructing an incidence relation model, namely establishing an index in a database through extraction and fusion of ontology knowledge, and constructing a knowledge map topological relation by using a map database to obtain an incidence relation construction model;
step seven: mining and analyzing a large amount of collected user identity and behavior data to construct user portrait tags, integrating and screening original data, and dividing the data into two types of tag data according to the attribute and importance of the data, namely static tag data and dynamic tag data;
step eight: similarity calculation is carried out according to the identities and behaviors of different types of users, mining analysis is carried out through static tag data and dynamic tag data, similarity calculation is carried out, and a huge user group is classified;
step nine: and constructing a personalized user portrait model, outputting the similarity calculation in the step eight as a class of behavior habits of the user, and matching the behavior habits with the result in the association relationship construction model.
Further, the target in the step one is an object that can be extracted from the remote sensing image, and includes a ship, a port airplane, an airport, and a landmark building, and the track is an object motion track, and includes: the image comprises a ship track and an unmanned aerial vehicle navigation track, wherein the image is a thumbnail related to the image and an open image related to the target.
Furthermore, in the fifth step, link relations of a plurality of knowledge systems are constructed by fusing attribute similarities of different ontology data, wherein the attribute of the data refers to the attribute of metadata information, for remote sensing image products, a ship target is extracted in a manner of feature extraction and semantic analysis, the ship target is used as metadata information and comprises a plurality of attribute values of position, model and size, and whether the ontology data are related or not is judged by comparing attribute value sets.
Further, the seventh step of dividing the data into two types of tag data, which are static tag data and dynamic tag data, specifically includes:
extracting static label data, wherein the content of the static label data is stable and has no change, and the storage format is a structured form;
and extracting dynamic label data, wherein the dynamic label data is behavior information which changes constantly in the interaction process of a user and a service system, and the dynamic label data mainly comprises browsing behavior information, downloading behavior information and evaluation behavior information.
Further, in the step eight, the calculating of the user preference similarity includes:
taking the behavior habits of the user as the basis of user preference consideration, selecting 3 tags representing the behavior depth of the user, wherein the tags are respectively browsed, downloaded and evaluated, and setting weights for the 3 tags, wherein the weights are respectively represented by v, d and e, and the weights are respectively 1/6, 1/2 and 1/3; if the current product has corresponding behavior, adding corresponding weight to the corresponding behavior value, otherwise adding 0, and using pref to represent the preference degree of the user to a product, namely the preference value, then:
pref＝v+d+e (1)
and repeating the steps for each user to obtain a user preference set.
Further, in the eighth step, the calculating of the similarity of the user identities includes:
selecting 5 labels which can represent social identities of users from the constructed user portrait labels, wherein the labels are respectively organizations, types, industries and nationalities, setting weights for the 5 labels, and respectively using s1、s2、s3、s4、s5Indicating that the weight values are set to 2/9, 2/9, 2/9, 2/9, 1/9, respectively; calculating the identity similarity of the users by comparing 5 identity labels between every two users, if the corresponding labels are the same, adding corresponding weight, otherwise, 0, and using simI (u, v) to represent the identity similarity of the user u and the user v, then:
simI(u,v)＝s1+s2+s3+s4+s5(2)
further, the constructing of the user portrait model in the ninth step includes:
the user-data model construction comprises the steps of constructing a user-data matrix which represents preference values of users to products;
calculating the user similarity by combining the user identity similarity and the user behavior similarity to obtain a preference value set of the user for the product;
and finally, carrying out user matching on the associated data, matching a user interested data list set according to the obtained user prediction preference value set, and obtaining all associated data topology sets related to the user interested data through an association relation construction layer.
Further, the user-data model construction comprises: constructing a user-data matrix, wherein R (n, m) ═ pref which represents the preference value of a user to data, namely the preference value of the nth user to the mth product;
measuring user preference similarity by cosine similarity on the basis of a user-data matrix, wherein the cosine similarity is used for measuring the user similarity by calculating cosine included angles among vectors; assuming that the similarity between the user u and the user v is simcoll (u, v), the preference value of the user u is represented by a vector u, and the preference value of the user v is represented by a vector v, the simcoll (u, v) is calculated as follows:
further, the user similarity is calculated, the user similarity combines the user identity similarity simI and the user behavior similarity simcoll (u, v), and λ is used for adjusting parameters, where λ is greater than or equal to 0 and less than or equal to 1, and the specific calculation is as shown in a formula:
sim(u,v)＝λ×simI(u，v)+(1-λ)×simcoll(u,v) (4)
after the user similarity is obtained, taking k users with the highest target user similarity value as nearest neighbor users of the target user, and predicting the preference of the target user according to the preference value of the nearest neighbor users to the product; the calculation formula is as follows:
Has the advantages that:
the method of the invention has the advantages over the prior art: the remote sensing mass archived data is mastered, the conversion from the remote sensing data to specific knowledge is solved, and the incidence relation between multi-source heterogeneous data is established through means of data attribute extraction, feature extraction, data fusion and the like; meanwhile, information such as user attributes, behaviors, habits and the like is mined and modeled, and accurate recommendation service guided by demands is provided for users in time. The method is feasible through test verification, has high processing speed on large data volume, and realizes accurate recommendation of remote sensing data resources. Compared with the prior art, the problem that different users acquire large multi-source heterogeneous remote sensing data and are difficult to accurately position is solved.
Drawings
FIG. 1 is a flow chart of incidence construction of the present invention;
FIG. 2 is a topological relationship diagram of an association relationship according to the present invention;
FIG. 3 is a general flow chart of a multi-source heterogeneous remote sensing data association construction and multi-user data matching method.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments, and all other embodiments obtained by a person skilled in the art based on the embodiments of the present invention belong to the protection scope of the present invention without creative efforts.
According to one embodiment of the invention, a multi-source heterogeneous remote sensing data association construction and multi-user data matching method is provided, association relation construction is carried out, methods such as data mining, text analysis and semantic recognition are used for carrying out feature extraction, model abstraction and standardized description on various remote sensing body data, an association relation of the body data is constructed by using a knowledge graph technology, and a data modeling layer is a basic support for realizing intelligent recommendation. The main process comprises the following steps:
1. and inputting original data. Data input is the basis of modeling of the whole incidence relation, the project mainly researches multi-source data consisting of remote sensing product data such as SAR, visible light, multiple spectrums, three-dimensional models, key targets, texts, tracks and the like and remote sensing service data, and the data share characteristics such as position, target, attribute, time and the like.
2. And (5) data aggregation classification. In order to construct a global unified association relationship, after each business field constructs and completes a self map relationship model, the ontology representing the same object is gathered together through association mining of heterogeneous ontologies, and a plurality of heterogeneous ontologies are combined into a global unified ontology through cross-domain ontology integration.
3. And (5) extracting knowledge. The knowledge extraction layer extracts 3 types of knowledge elements of entities, attributes and relationships among the entities from the ontology data. The entity represents an object which exists objectively, the entity extraction is to extract the entity from original body data, the entity needs to be found firstly, and then attribute filling is continuously carried out on entity data, wherein the attribute filling comprises entity description, pictures, synonymous entity names and the like; the relation extraction is to establish relations among entities with different concepts, including 3 types of knowledge elements of the entities, attributes and relations among the entities, and solve the problem of semantic links among the entities.
4. And (4) knowledge fusion. Entities from different sources are identified through a cross-domain knowledge representation learning model and text semantics, an entity relation representation model is established, and cross-domain semantics fusion of a plurality of knowledge maps is achieved.
5. And (5) constructing an association relation. A relational model is abstracted on a data table (a target characteristic library, a test sample library and a target knowledge library) after the existing knowledge is fused, main entities, attributes and associated edges are created, and the organization and the association of multi-field, multi-source and multi-type information are realized.
Specifically, according to an embodiment of the present invention, the flowchart shown in fig. 1 is combined as follows:
the method comprises the following steps: and inputting a remote sensing service data set and a remote sensing product data set. The remote sensing service data set comprises data such as targets, tracks, texts, pictures and the like; the remote sensing product data comprises remote sensing image data such as Sar images, panchromatic images, multispectral image data, visible light data and the like.
Step two: the method comprises the steps of data cleaning and preprocessing, namely performing grammar check, spelling check, missing value processing, duplicate removal processing, invalid character removal and noise processing on remote sensing service data; carrying out rough inspection on remote sensing product data, carrying out image quality inspection by means of manual interpretation, and carrying out image preprocessing on images with deformation and cloud and fog interference;
removing the mist: unnecessary thick cloud areas are removed, cloud and fog are removed by using the existing cloud judging tool, and image quality is improved.
And (3) deformation correction: aiming at the problem that the geometrical deformation of SAR imaging causes the bending of airports and roads and influences the extraction precision of subsequent images, the method of image non-uniform sampling is used for deformation correction of Sar images, and the image resampling can adopt methods with higher precision such as bilinear interpolation and cubic interpolation.
Step three: data classification modeling
The classifier uses a method for mining association rules in data mining, and based on the classification of the association rules, firstly, the association rules of all specified confidence degrees are mined from a training data body sample set by using an association rule mining algorithm; and then iterating the optimal rule from the mined class association rules for classification. And finally, aggregating the ontologies representing the same object together by using an optimal rule, and combining a plurality of heterogeneous ontologies into a global unified ontology.
Step four: and (5) extracting knowledge.
The knowledge extraction comprises attribute extraction of remote sensing service data and feature extraction of remote sensing product data.
The knowledge extraction of the remote sensing service data mainly comprises the extraction of metadata information such as names, time, space, uploaders, belonged countries, data formats and keywords.
The knowledge extraction of the remote sensing product data comprises two parts of attribute extraction and feature extraction. The attribute extraction mainly comprises metadata information such as data types, data names, resolutions, affiliated satellites, sensors, inclusion targets, scene numbers, positions, ranges, product levels, thumb maps and the like. Extracting characteristics, namely extracting an interested area according to a phase spectrum saliency map of a remote sensing image; carrying out homogeneous filtering in the region of interest, and confirming a target region to be detected by combining a phase spectrum significance map; in the target area, calculating the main shaft direction of each target, and extracting an S-HOG feature descriptor of the target; and judging the S-HOG feature descriptor of the target according to the shape knowledge of the target, and judging, identifying and extracting the target.
Step five: and (4) knowledge fusion. Knowledge fusion is to construct a link relationship between two or more pieces of knowledge. According to metadata information extracted from the remote sensing data, link relations of a plurality of knowledge systems are constructed by fusing similarity of attributes (attributes of the metadata information) of different ontology data. For remote sensing image products, a ship target is extracted through modes of feature extraction, semantic analysis and the like, the ship target serves as metadata information and can have a plurality of attribute values such as positions, models, sizes and the like, and whether ontology data are related or not is judged through comparison of attribute value sets.
Step six: and (5) constructing an association relation. By extracting and fusing ontology knowledge, an index is established in a database, and a knowledge graph topological relation is established by using a graph database. The representation is shown in fig. 2. According to the extracted entity and entity relation, data used in the system are put in a warehouse, a graph database is used for storing a knowledge graph and can be used for directly and visually displaying the topology, and the display in the graph 2 is a topological relation graph example constructed after a system catalog selects a scene image.
After the association relationship is built, a user portrait building layer is further carried out, user information resources are comprehensively and systematically collected through a mining analysis technology, main characteristics such as user background information and behavior habits are mined and analyzed, and a personalized user portrait model is built. Referring to fig. 3, the main process includes the following steps:
step seven: a user profile label is constructed. And mining and analyzing a large amount of collected user identity and behavior data. After the original data are integrated and screened, the data are divided into two types of label data according to the attribute and the importance of the data, namely static label data and dynamic label data.
1.1 static tag data extraction. The content of the static label data is stable and has no change, and the storage format is a structured form.
1.2 dynamic tag data extraction. The dynamic label data is the behavior information which changes constantly in the interaction process of the user and the service system. The dynamic label data mainly comprises browsing behavior information, downloading behavior information and evaluation behavior information.
Step eight: and calculating the similarity. The similarity calculation aims at the identities and behaviors of different types of users, mining analysis is carried out through static tag data and dynamic tag data, similarity statistics is carried out, and the purpose is to classify huge user groups and improve the efficiency of intelligent recommendation. The invention provides two strategies of identity similarity calculation and user preference similarity calculation.
Step nine: and constructing a personalized user portrait model. The output of the similarity calculation is the behavior habit of a class of users, and is matched with the result in the incidence relation construction model. The specific flow chart is shown in fig. 3.
The seventh step specifically comprises: attribute extraction, namely after the original data are integrated and screened, dividing the original data into static data tags and dynamic data tags according to the attributes of the data;
extracting static data labels, wherein the static data labels mainly comprise user attribute information, and mainly comprise user names, user IDs (identities), passwords, mailboxes, contact ways, organizations, types, industries, nationalities and the like;
and extracting dynamic label data, which mainly comprises browsing behavior information, downloading behavior information and evaluation behavior information. The access behavior information mainly comprises: the unit where the user is located, user access time, user access times, IP accessed by the user and a user access module; the downloading behavior information mainly comprises a downloading unit, a downloading data type, the downloading data frequency, the downloading data size, a downloading data satellite load, a downloading data type and a downloading data resolution; the evaluation behavior information mainly comprises evaluation grading, evaluation content and evaluation users.
In the step eight, two strategies, namely user identity similarity calculation and user preference similarity calculation, are adopted.
The user identity similarity calculation comprises the following steps: selecting 5 labels which can represent social identities of users from the constructed user portrait labels, wherein the labels are respectively organizations, types, industries and nationalities, setting weights for the 5 labels, and respectively using s1、s2、s3、s4、s5Indicating that the weight values are set to 2/9, 2/9, 2/9, 2/9, 1/9, respectively. Calculating the identity similarity of the users by comparing 5 identity labels between every two users, adding corresponding weight if the corresponding labels are the same, otherwise, 0, and representing the identity phase of the user u and the user v by simI (u, v)Similarity, then:
simI(u,v)＝s1+s2+s3+s4+s5(1)
wherein the user preference similarity calculation comprises: taking the behavior habits of the user as the basis of user preference, selecting 3 tags which can represent the behavior depth of the user better, browsing, downloading and evaluating respectively, and setting weights for the 3 tags, wherein the weights are represented by v, d and e respectively, and the weights are 1/6, 1/2 and 1/3 respectively. If the current product has corresponding behavior, adding corresponding weight to the corresponding behavior value, otherwise adding 0, and using pref to represent the preference degree of the user to a product, namely the preference value, then:
pref＝v+d+e (2)
repeating the steps for each user to obtain a user preference set;
and step nine, constructing a user portrait model, specifically comprising constructing a user-data model, calculating user similarity and matching associated data with a user.
1. And constructing a user-data model. A user-data matrix is constructed as shown in table 1 below, where R (n, m) ═ pref denotes the preference value of the user for data, i.e., the preference value of the user n for the product m.
TABLE 1 user-data matrix
| p1 | p2 | …… | pm | |
| u1 | R(1,1) | R(1,2) | …… | R(1,m) |
| u2 | R(2,1) | R(2,2) | …… | R(2,m) |
| …… | … | …… | …… | …… |
| un | R(n,1) | R(n,2) | …… | R(n,m) |
pmRepresents the m-th product; u. ofnRepresenting the nth user.
And finally, measuring the preference similarity of the users by cosine similarity on the basis of the user-data matrix, wherein the cosine similarity is used for measuring the similarity of the users by calculating cosine included angles among vectors. Assuming that the similarity between the user u and the user v is simcoll (u, v), the preference value of the user u is represented by a vector u, and the preference value of the user v is represented by a vector v, the simcoll (u, v) is calculated as follows:
2. calculating user similarity, wherein the user similarity combines user identity similarity simI and user behavior similarity simcoll (u, v), and lambda is used for adjusting parameters, wherein lambda is more than or equal to 0 and less than or equal to 1, and the specific calculation is shown as a formula:
sim(u,v)＝λ×simI(u，v)+(1-λ)×simcoll(u,v)(4)
after the user similarity is obtained, taking k users with the highest target user similarity value as nearest neighbor users of the target user, and then predicting the preference of the target user according to the preference value of the nearest neighbor users to the product; the calculation formula is as follows:
3. User matching of associated data. According to the obtained user prediction preference value set, a user interest data list set can be matched, all relevant data topology sets related to the user interest data can be obtained through an incidence relation building layer, and the sets are automatically pushed to a user side in a topology or data set mode for the user to check.
The method maximally mines the available information of the data and timely shares the multi-source heterogeneous data, so that a user comprehensively and generally grasps all information of interested data, and the use value of the remote sensing data is maximally realized.
Although illustrative embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, but various changes may be apparent to those skilled in the art, and it is intended that all inventive concepts utilizing the inventive concepts set forth herein be protected without departing from the spirit and scope of the present invention as defined and limited by the appended claims.
Claims (9)
1. A multi-source heterogeneous remote sensing data association construction and multi-user data matching method is characterized by comprising the following steps:
the method comprises the following steps: inputting a remote sensing service data set and a remote sensing product data set, wherein the remote sensing service data set comprises target, track, text and picture data; the remote sensing product data comprises Sar images, panchromatic images, multispectral image data and visible light data remote sensing image data;
step two: performing data cleansing and preprocessing, the data cleansing comprising: carrying out grammar check, spelling check, missing value processing, duplicate removal processing, invalid character removal and noise processing on the remote sensing service data; the preprocessing comprises the steps of carrying out image quality inspection on remote sensing product data by means of manual interpretation, and firstly carrying out image preprocessing on images with deformation and cloud and fog interference; removing unnecessary thick cloud areas and removing cloud and mist; carrying out deformation correction, aiming at the situation that an airport and a road are bent due to the geometric deformation of SAR imaging, carrying out deformation correction by using an image non-uniform sampling method, and resampling the position or the distance of pixels by using a bilinear interpolation or cubic interpolation method;
step three: carrying out data classification modeling, inputting the data classification modeling into the cleaned data set in the second step, and mining all class association rules with specified confidence degrees from the training data body sample set by using an association rule mining algorithm based on the classification of the association rules by using a classifier by adopting an association rule mining method in data mining; then iterating optimal rules from the mined class association rules for classification, finally gathering the entities representing the same object together by using the optimal rules, and combining a plurality of heterogeneous entities into a global unified entity;
step four: extracting knowledge, including attribute extraction of remote sensing service data and feature extraction of remote sensing product data; the attribute extraction of the remote sensing service data comprises the extraction of name, time, space, uploader, country, data format, keyword and metadata information; the metadata information comprises data types, data names, resolutions, satellites, sensors, including targets, scene numbers, positions, ranges, product levels and thumb maps; firstly, extracting an interested region according to a phase spectrum saliency map of a remote sensing image by the feature extraction; carrying out homogeneous filtering in the region of interest, and confirming a target region to be detected by combining a phase spectrum significance map; in the target area, calculating the main shaft direction of each target, and extracting an S-HOG feature descriptor of the target; judging an S-HOG feature descriptor of the target according to the shape knowledge of the target, and judging, identifying and extracting the target;
step five: performing multi-source heterogeneous knowledge fusion, constructing a link relation of a plurality of knowledge systems by fusing the attribute similarity of different ontology data according to metadata information extracted from remote sensing data;
step six: constructing an incidence relation model, namely establishing an index in a database through extraction and fusion of ontology knowledge, and constructing a knowledge map topological relation by using a map database to obtain an incidence relation construction model;
step seven: mining and analyzing a large amount of collected user identity and behavior data to construct user portrait tags, integrating and screening original data, and dividing the data into two types of tag data according to the attribute and importance of the data, namely static tag data and dynamic tag data;
step eight: similarity calculation is carried out according to the identities and behaviors of different types of users, mining analysis is carried out through static tag data and dynamic tag data, similarity calculation is carried out, and a huge user group is classified;
step nine: and constructing a personalized user portrait model, outputting the similarity calculation in the step eight as a class of behavior habits of the user, and matching the behavior habits with the result in the association relationship construction model.
2. The multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on the claim 1 is characterized in that:
the target in the step one is an object which can be extracted from the remote sensing image, and comprises a ship, a port airplane, an airport and a landmark building, wherein the track is an object motion track, and the method comprises the following steps: the image comprises a ship track and an unmanned aerial vehicle navigation track, wherein the image is a thumbnail related to the image and an open image related to the target.
3. The multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on the claim 1 is characterized in that:
and in the fifth step, the link relations of a plurality of knowledge systems are constructed by fusing the attribute similarity of different ontology data, the attribute of the data refers to the attribute of metadata information, for the remote sensing image product, a ship target is extracted in a manner of feature extraction and semantic analysis, the ship target is used as the metadata information and comprises a plurality of attribute values of position, model and size, and whether the ontology data are related or not is judged by comparing the attribute value sets.
4. The multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on the claim 1 is characterized in that:
seventhly, dividing the data into two types of tag data, namely static tag data and dynamic tag data, and specifically comprising:
extracting static label data, wherein the content of the static label data is stable and has no change, and the storage format is a structured form;
and extracting dynamic label data, wherein the dynamic label data is behavior information which changes constantly in the interaction process of a user and a service system, and the dynamic label data mainly comprises browsing behavior information, downloading behavior information and evaluation behavior information.
5. The multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on the claim 1 is characterized in that: in the eighth step, the calculating of the similarity of the user preferences includes:
taking the behavior habits of the user as the basis of user preference consideration, selecting 3 tags representing the behavior depth of the user, wherein the tags are respectively browsed, downloaded and evaluated, and setting weights for the 3 tags, wherein the weights are respectively represented by v, d and e, and the weights are respectively 1/6, 1/2 and 1/3; if the current product has corresponding behavior, adding corresponding weight to the corresponding behavior value, otherwise adding 0, and using pref to represent the preference degree of the user to a product, namely the preference value, then:
pref＝v+d+e (1)
and repeating the steps for each user to obtain a user preference set.
6. The multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on claim 1, wherein in the eighth step, the user identity similarity calculation comprises:
selecting 5 labels which can represent social identities of users from the constructed user portrait labels, wherein the labels are respectively organizations, types, industries and nationalities, setting weights for the 5 labels, and respectively using s1、s2、s3、s4、s5Indicating that the weight values are set to 2/9, 2/9, 2/9, 2/9, 1/9, respectively; calculating the identity similarity of the users by comparing 5 identity labels between every two users, if the corresponding labels are the same, adding corresponding weight, otherwise, 0, and using simI (u, v) to represent the identity similarity of the user u and the user v, then:
simI(u,v)＝s1+s2+s3+s4+s5(2)。
7. the multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on the claim 1 is characterized in that the user portrait model construction in the ninth step comprises the following steps:
the user-data model construction comprises the steps of constructing a user-data matrix which represents preference values of users to products;
calculating the user similarity by combining the user identity similarity and the user behavior similarity to obtain a preference value set of the user for the product;
and finally, carrying out user matching on the associated data, matching a user interested data list set according to the obtained user prediction preference value set, and obtaining all associated data topology sets related to the user interested data through an association relation construction layer.
8. The multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on claim 7, wherein the user-data model construction comprises: constructing a user-data matrix, wherein R (n, m) ═ pref which represents the preference value of a user to data, namely the preference value of the nth user to the mth product;
measuring user preference similarity by cosine similarity on the basis of a user-data matrix, wherein the cosine similarity is used for measuring the user similarity by calculating cosine included angles among vectors; assuming that the similarity between the user u and the user v is simcoll (u, v), the preference value of the user u is represented by a vector u, and the preference value of the user v is represented by a vector v, the simcoll (u, v) is calculated as follows:
9. the multi-source heterogeneous remote sensing data association construction and multi-user data matching method based on claim 7, wherein the user similarity calculation combines user identity similarity simI and user behavior similarity simcoll (u, v), λ is used for adjusting parameters, λ is greater than or equal to 0 and less than or equal to 1, and the specific calculation is as shown in a formula:
sim(u,v)＝λ×simI(u，v)+(1-λ)×simcoll(u,v) (4)
after the user similarity is obtained, taking k users with the highest target user similarity value as nearest neighbor users of the target user, and predicting the preference of the target user according to the preference value of the nearest neighbor users to the product; the calculation formula is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010447854.XA CN111666313B (en) | 2020-05-25 | 2020-05-25 | Correlation construction and multi-user data matching method based on multi-source heterogeneous remote sensing data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010447854.XA CN111666313B (en) | 2020-05-25 | 2020-05-25 | Correlation construction and multi-user data matching method based on multi-source heterogeneous remote sensing data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111666313A true CN111666313A (en) | 2020-09-15 |
| CN111666313B CN111666313B (en) | 2023-02-07 |
Family
ID=72384489
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010447854.XA Active CN111666313B (en) | 2020-05-25 | 2020-05-25 | Correlation construction and multi-user data matching method based on multi-source heterogeneous remote sensing data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111666313B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112131275A (en) * | 2020-09-23 | 2020-12-25 | 中国科学技术大学智慧城市研究院(芜湖) | Enterprise portrait construction method of holographic city big data model and knowledge graph |
| CN112182131A (en) * | 2020-09-28 | 2021-01-05 | 中国电子科技集团公司第五十四研究所 | Remote sensing image recommendation method based on multi-attribute fusion |
| CN112256938A (en) * | 2020-12-23 | 2021-01-22 | 畅捷通信息技术股份有限公司 | Message metadata processing method, device and medium |
| CN112507185A (en) * | 2020-10-22 | 2021-03-16 | 复旦大学 | User portrait determination method and device |
| CN112966027A (en) * | 2021-03-22 | 2021-06-15 | 青岛科技大学 | Entity association mining method based on dynamic probe |
| CN112989063A (en) * | 2021-03-10 | 2021-06-18 | 青岛弯弓信息技术有限公司 | 3D modeling method and system based on knowledge graph |
| CN113076476A (en) * | 2021-04-01 | 2021-07-06 | 重庆邮电大学 | User portrait construction method of microblog heterogeneous information |
| CN113157678A (en) * | 2021-04-19 | 2021-07-23 | 中国人民解放军91977部队 | Multi-source heterogeneous data association method |
| CN113297395A (en) * | 2021-07-08 | 2021-08-24 | 中国人民解放军国防科技大学 | Spatio-temporal multi-modal mixed data processing method, correlation method and indexing method |
| CN113392320A (en) * | 2021-05-14 | 2021-09-14 | 国家卫星气象中心(国家空间天气监测预警中心) | Knowledge service-based satellite remote sensing service providing method |
| CN113987131A (en) * | 2021-11-11 | 2022-01-28 | 江苏天汇空间信息研究院有限公司 | Heterogeneous multi-source data correlation analysis system and method |
| CN114003956A (en) * | 2021-11-03 | 2022-02-01 | 江苏天汇空间信息研究院有限公司 | Spatial data analysis scheduling system and method applying big data analysis |
| CN114491108A (en) * | 2021-12-30 | 2022-05-13 | 中科星通(廊坊)信息技术有限公司 | Online classification system and method based on multi-source remote sensing application data |
| CN114565046A (en) * | 2022-03-01 | 2022-05-31 | 中科海慧（北京）科技有限公司 | Mineral resource detection system and detection method based on remote sensing hyperspectrum |
| CN113220894B (en) * | 2021-02-07 | 2023-08-18 | 国家卫星气象中心(国家空间天气监测预警中心) | Intelligent satellite remote sensing data acquisition method based on perception calculation |
| CN116756229A (en) * | 2023-06-01 | 2023-09-15 | 广州盛成妈妈网络科技股份有限公司 | Data mining and visualization method |
| CN117351257A (en) * | 2023-08-24 | 2024-01-05 | 长江水上交通监测与应急处置中心 | Multi-mode information-based shipping data extraction method and system |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060136462A1 (en) * | 2004-12-16 | 2006-06-22 | Campos Marcos M | Data-centric automatic data mining |
| US20130262483A1 (en) * | 2012-03-30 | 2013-10-03 | Nokia Corporation | Method and apparatus for providing intelligent processing of contextual information |
| CN104063421A (en) * | 2014-05-14 | 2014-09-24 | 中交宇科（北京）空间信息技术有限公司 | Retrieval method and device for mass transportation remote sensing data |
| CN104063555A (en) * | 2014-07-07 | 2014-09-24 | 成都理工大学 | User model establishing method for intelligent remote sensing information distribution |
| CN105095436A (en) * | 2015-07-23 | 2015-11-25 | 苏州国云数据科技有限公司 | Automatic modeling method for data of data sources |
| CN105654137A (en) * | 2015-12-31 | 2016-06-08 | 中国科学院电子学研究所 | Mass remote sensing data classification and organization method |
| CN106096663A (en) * | 2016-06-24 | 2016-11-09 | 长春工程学院 | A kind of grader construction method of the target in hyperspectral remotely sensed image being grouped based on sparse isomery |
| US20170124497A1 (en) * | 2015-10-28 | 2017-05-04 | Fractal Industries, Inc. | System for automated capture and analysis of business information for reliable business venture outcome prediction |
| CN106651424A (en) * | 2016-09-28 | 2017-05-10 | 国网山东省电力公司电力科学研究院 | Electric power user figure establishment and analysis method based on big data technology |
| CN106951433A (en) * | 2017-02-04 | 2017-07-14 | 百度在线网络技术（北京）有限公司 | A kind of search method and device |
| CN107633075A (en) * | 2017-09-22 | 2018-01-26 | 吉林大学 | A kind of multi-source heterogeneous data fusion platform and fusion method |
| WO2018098738A1 (en) * | 2016-11-30 | 2018-06-07 | 深圳益强信息科技有限公司 | Monitoring method and device base on aerospace multi-source remote sensing data |
| CN109767300A (en) * | 2019-01-14 | 2019-05-17 | 博拉网络股份有限公司 | Big data portrait and model building method based on user's habit |
| CN109784351A (en) * | 2017-11-10 | 2019-05-21 | 财付通支付科技有限公司 | Data classification method, disaggregated model training method and device |
| CN109800671A (en) * | 2018-12-28 | 2019-05-24 | 北京市遥感信息研究所 | The multi-source remote sensing information knowledge map construction method and system of object-oriented interpretation |
| CN109993220A (en) * | 2019-03-23 | 2019-07-09 | 西安电子科技大学 | Multi-source Remote Sensing Images Classification method based on two-way attention fused neural network |
| US20190264936A1 (en) * | 2016-10-21 | 2019-08-29 | Centrica Hive Limited | Hvac performance monitoring system |
| CN110245874A (en) * | 2019-03-27 | 2019-09-17 | 中国海洋大学 | A kind of Decision fusion method based on machine learning and knowledge reasoning |
-
2020
- 2020-05-25 CN CN202010447854.XA patent/CN111666313B/en active Active
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060136462A1 (en) * | 2004-12-16 | 2006-06-22 | Campos Marcos M | Data-centric automatic data mining |
| US20130262483A1 (en) * | 2012-03-30 | 2013-10-03 | Nokia Corporation | Method and apparatus for providing intelligent processing of contextual information |
| CN104063421A (en) * | 2014-05-14 | 2014-09-24 | 中交宇科（北京）空间信息技术有限公司 | Retrieval method and device for mass transportation remote sensing data |
| CN104063555A (en) * | 2014-07-07 | 2014-09-24 | 成都理工大学 | User model establishing method for intelligent remote sensing information distribution |
| CN105095436A (en) * | 2015-07-23 | 2015-11-25 | 苏州国云数据科技有限公司 | Automatic modeling method for data of data sources |
| US20170124497A1 (en) * | 2015-10-28 | 2017-05-04 | Fractal Industries, Inc. | System for automated capture and analysis of business information for reliable business venture outcome prediction |
| CN105654137A (en) * | 2015-12-31 | 2016-06-08 | 中国科学院电子学研究所 | Mass remote sensing data classification and organization method |
| CN106096663A (en) * | 2016-06-24 | 2016-11-09 | 长春工程学院 | A kind of grader construction method of the target in hyperspectral remotely sensed image being grouped based on sparse isomery |
| CN106651424A (en) * | 2016-09-28 | 2017-05-10 | 国网山东省电力公司电力科学研究院 | Electric power user figure establishment and analysis method based on big data technology |
| US20190264936A1 (en) * | 2016-10-21 | 2019-08-29 | Centrica Hive Limited | Hvac performance monitoring system |
| WO2018098738A1 (en) * | 2016-11-30 | 2018-06-07 | 深圳益强信息科技有限公司 | Monitoring method and device base on aerospace multi-source remote sensing data |
| CN106951433A (en) * | 2017-02-04 | 2017-07-14 | 百度在线网络技术（北京）有限公司 | A kind of search method and device |
| CN107633075A (en) * | 2017-09-22 | 2018-01-26 | 吉林大学 | A kind of multi-source heterogeneous data fusion platform and fusion method |
| CN109784351A (en) * | 2017-11-10 | 2019-05-21 | 财付通支付科技有限公司 | Data classification method, disaggregated model training method and device |
| CN109800671A (en) * | 2018-12-28 | 2019-05-24 | 北京市遥感信息研究所 | The multi-source remote sensing information knowledge map construction method and system of object-oriented interpretation |
| CN109767300A (en) * | 2019-01-14 | 2019-05-17 | 博拉网络股份有限公司 | Big data portrait and model building method based on user's habit |
| CN109993220A (en) * | 2019-03-23 | 2019-07-09 | 西安电子科技大学 | Multi-source Remote Sensing Images Classification method based on two-way attention fused neural network |
| CN110245874A (en) * | 2019-03-27 | 2019-09-17 | 中国海洋大学 | A kind of Decision fusion method based on machine learning and knowledge reasoning |
Non-Patent Citations (2)
| Title |
|---|
| 仇林遥 等: "地理时空数据关联与聚合服务方法综述", 《中国电子科学研究院学报》 * |
| 王婷等: "多传感器影像建模的海洋遥感数据管理应用", 《测绘科学》 * |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112131275A (en) * | 2020-09-23 | 2020-12-25 | 中国科学技术大学智慧城市研究院(芜湖) | Enterprise portrait construction method of holographic city big data model and knowledge graph |
| CN112182131A (en) * | 2020-09-28 | 2021-01-05 | 中国电子科技集团公司第五十四研究所 | Remote sensing image recommendation method based on multi-attribute fusion |
| CN112507185A (en) * | 2020-10-22 | 2021-03-16 | 复旦大学 | User portrait determination method and device |
| CN112256938A (en) * | 2020-12-23 | 2021-01-22 | 畅捷通信息技术股份有限公司 | Message metadata processing method, device and medium |
| CN113220894B (en) * | 2021-02-07 | 2023-08-18 | 国家卫星气象中心(国家空间天气监测预警中心) | Intelligent satellite remote sensing data acquisition method based on perception calculation |
| CN112989063A (en) * | 2021-03-10 | 2021-06-18 | 青岛弯弓信息技术有限公司 | 3D modeling method and system based on knowledge graph |
| CN112989063B (en) * | 2021-03-10 | 2023-02-10 | 青岛弯弓信息技术有限公司 | 3D modeling method and system based on knowledge graph |
| CN112966027B (en) * | 2021-03-22 | 2022-10-21 | 青岛科技大学 | Entity association mining method based on dynamic probe |
| CN112966027A (en) * | 2021-03-22 | 2021-06-15 | 青岛科技大学 | Entity association mining method based on dynamic probe |
| CN113076476A (en) * | 2021-04-01 | 2021-07-06 | 重庆邮电大学 | User portrait construction method of microblog heterogeneous information |
| CN113157678B (en) * | 2021-04-19 | 2022-03-15 | 中国人民解放军91977部队 | Multi-source heterogeneous data association method |
| CN113157678A (en) * | 2021-04-19 | 2021-07-23 | 中国人民解放军91977部队 | Multi-source heterogeneous data association method |
| CN113392320A (en) * | 2021-05-14 | 2021-09-14 | 国家卫星气象中心(国家空间天气监测预警中心) | Knowledge service-based satellite remote sensing service providing method |
| CN113297395A (en) * | 2021-07-08 | 2021-08-24 | 中国人民解放军国防科技大学 | Spatio-temporal multi-modal mixed data processing method, correlation method and indexing method |
| CN114003956B (en) * | 2021-11-03 | 2022-10-25 | 江苏天汇空间信息研究院有限公司 | Spatial data analysis scheduling system and method applying big data analysis |
| CN114003956A (en) * | 2021-11-03 | 2022-02-01 | 江苏天汇空间信息研究院有限公司 | Spatial data analysis scheduling system and method applying big data analysis |
| CN113987131B (en) * | 2021-11-11 | 2022-08-23 | 江苏天汇空间信息研究院有限公司 | Heterogeneous multi-source data correlation analysis system and method |
| CN113987131A (en) * | 2021-11-11 | 2022-01-28 | 江苏天汇空间信息研究院有限公司 | Heterogeneous multi-source data correlation analysis system and method |
| CN114491108A (en) * | 2021-12-30 | 2022-05-13 | 中科星通(廊坊)信息技术有限公司 | Online classification system and method based on multi-source remote sensing application data |
| CN114565046A (en) * | 2022-03-01 | 2022-05-31 | 中科海慧（北京）科技有限公司 | Mineral resource detection system and detection method based on remote sensing hyperspectrum |
| CN116756229A (en) * | 2023-06-01 | 2023-09-15 | 广州盛成妈妈网络科技股份有限公司 | Data mining and visualization method |
| CN116756229B (en) * | 2023-06-01 | 2023-11-28 | 广州盛成妈妈网络科技股份有限公司 | Data mining and visualization method |
| CN117351257A (en) * | 2023-08-24 | 2024-01-05 | 长江水上交通监测与应急处置中心 | Multi-mode information-based shipping data extraction method and system |
| CN117351257B (en) * | 2023-08-24 | 2024-04-02 | 长江水上交通监测与应急处置中心 | Multi-mode information-based shipping data extraction method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111666313B (en) | 2023-02-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111666313B (en) | Correlation construction and multi-user data matching method based on multi-source heterogeneous remote sensing data | |
| Miah et al. | A big data analytics method for tourist behaviour analysis | |
| Liu et al. | Classifying urban land use by integrating remote sensing and social media data | |
| KR101856120B1 (en) | Discovery of merchants from images | |
| KR101768521B1 (en) | Method and system providing informational data of object included in image | |
| Andrade et al. | POI mining for land use classification: A case study | |
| Hu et al. | GeoAI at ACM SIGSPATIAL: progress, challenges, and future directions | |
| CN103544216A (en) | Information recommendation method and system combining image content and keywords | |
| McArdle et al. | Classifying pedestrian movement behaviour from GPS trajectories using visualization and clustering | |
| CN114996488B (en) | Skynet big data decision-level fusion method | |
| CN115934990B (en) | Remote sensing image recommendation method based on content understanding | |
| Jasiewicz et al. | GeoPAT: A toolbox for pattern-based information retrieval from large geospatial databases | |
| Fang et al. | Synthesizing location semantics from street view images to improve urban land-use classification | |
| Singh et al. | Semantically guided geo-location and modeling in urban environments | |
| Lin et al. | Extracting urban landmarks from geographical datasets using a random forests classifier | |
| Chiang | Unlocking textual content from historical maps-potentials and applications, trends, and outlooks | |
| WO2022238873A1 (en) | Knowledge space analytics | |
| Djenouri et al. | Deep learning based decomposition for visual navigation in industrial platforms | |
| Samal et al. | Searching satellite imagery with integrated measures | |
| Ouchra et al. | An overview of GeoSpatial Artificial Intelligence technologies for city planning and development | |
| Guo et al. | Object discovery in high-resolution remote sensing images: a semantic perspective | |
| Xu et al. | Tourist Attraction Recommendation Method and Data Management Based on Big Data Analysis | |
| CN112015937B (en) | Picture geographic positioning method and system | |
| EP3343394A1 (en) | Method for and system for creating global thematic maps based on extraction of image descriptors | |
| Zhang et al. | Interactive mobile visual search for social activities completion using query image contextual model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230913 Address after: 101399 No. 2 East Airport Road, Shunyi Airport Economic Core Area, Beijing (1st, 5th and 7th floors of Industrial Park 1A-4) Patentee after: Zhongke Star Map Co.,Ltd. Patentee after: Zhongke Xingtu Intelligent Technology Co.,Ltd. Address before: 101399 No. 2 Airport East Road, Shunyi Airport Economic Core Area, Beijing Patentee before: Zhongke Star Map Co.,Ltd. |
|
| TR01 | Transfer of patent right |