Geospatial intelligence acquisition method and device, electronic equipment and storage medium

Abstract

The application provides a geospatial intelligence acquisition method and device, electronic equipment and a storage medium, and relates to the technical field of geospatial intelligence. The method comprises the following steps: receiving description information related to to-be-acquired intelligence; extracting geospatial event information related to the description information from data sets of multiple data sources according to the description information, obtaining multiple geospatial event information corresponding to the multiple data sources one by one; and generating geospatial intelligence under each data source according to the multiple geospatial event information. The application can improve the efficiency of acquiring geospatial intelligence from multiple data sources, reduce manual errors, and make the obtained geospatial intelligence more accurate and reliable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geospatial intelligence, and in particular to a geospatial intelligence acquisition method and device, an electronic device and a storage medium. BACKGROUND

[0002] With the advancement of technology, the field of geospatial intelligence is facing the challenges of data diversity and massiveness. Data sources are diverse, including satellite remote sensing images, unmanned aerial vehicle aerial photography, social media content, sensor networks and traditional media information, etc. These data are not only huge in quantity, but also vary in format. Structured data represented by geographic information and target positioning, and unstructured data represented by video, image, audio and text, the differences in data format make data processing more difficult. At the same time, the demand for real-time is increasing, especially in the fields of military operations, disaster response, etc., which need fast and accurate intelligence support, and the traditional manual analysis method has been difficult to meet this requirement.

[0003] In the face of massive data, precision and accuracy are the key to geospatial intelligence, and false information may lead to serious consequences. For example, in target identification and terrain analysis, high-precision data are crucial for developing effective action plans. Therefore, the efficiency and accuracy of manually processing massive heterogeneous data to obtain geospatial intelligence have been unable to meet the current actual situation. SUMMARY

[0004] The present application provides a geospatial intelligence acquisition method and device, an electronic device and a storage medium, to solve the defects of low accuracy and efficiency in obtaining geospatial intelligence from multi-source data based on manual processing in the prior art, and to improve the efficiency of obtaining geospatial intelligence, reduce human errors, and make the obtained geospatial intelligence more accurate and reliable.

[0005] The present application provides a geospatial intelligence acquisition method, comprising the following steps.

[0006] Receiving description information related to the intelligence to be acquired; extracting geospatial event information related to the description information from the data set of a plurality of target data sources, obtaining a plurality of geospatial event information corresponding to each data source; generating geospatial intelligence under each data source according to the plurality of geospatial event information.

[0007] According to the geospatial intelligence acquisition method provided by the present application, after generating geospatial intelligence under each data source according to the plurality of geospatial event information, it further comprises: generating geospatial prediction intelligence according to the geospatial intelligence under each data source; outputting the geospatial prediction intelligence.

[0008] The method further comprises: analyzing the occurrence probability of the early warning event according to the geospatial prediction information and pre-set early warning event information; and issuing an alarm when the occurrence probability is greater than a set threshold.

[0009] The method further comprises: generating information recommendation information according to the geospatial information of each data source; and outputting the information recommendation information.

[0010] The method further comprises: selecting a plurality of target data sources from a plurality of preset data sources as the plurality of data sources according to received data source selection indication information; obtaining data sets of the plurality of data sources from the plurality of data sources according to a data acquisition manner corresponding to each data source; pre-processing the data set of each data source according to a pre-processing manner corresponding to the data source; extracting initial processing information from the pre-processing result of the data set of each data source; and performing entity linking on the initial processing information of each data source to obtain the plurality of geospatial event information.

[0011] The geospatial information comprises time, a geographical position, and an occurrence event.

[0012] The application further provides a geospatial information acquisition device comprising the following modules: a receiving module, an extracting module, a generating module, and an outputting module.

[0013] The receiving module is configured to receive description information related to information to be acquired.

[0014] The extracting module is configured to extract geospatial event information related to the description information from data sets of a plurality of data sources according to the description information, and obtain a plurality of geospatial event information corresponding to the plurality of data sources one by one.

[0015] The generating module is configured to generate geospatial information of each data source according to the plurality of geospatial event information.

[0016] The application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the geospatial information acquisition method when executing the computer program.

[0017] The application further provides a non-transitory computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the geographic space information acquisition method according to any one of the above.

[0018] The application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the geographic space information acquisition method according to any one of the above.

[0019] The geographic space information acquisition method and device, the electronic equipment and the storage medium provided by the application can automatically extract geographic space event information corresponding to a plurality of data sources from the plurality of data sources, generate corresponding geographic space information according to the geographic space event information corresponding to each data source, and solve the problem of low accuracy and efficiency of acquiring geographic space information from multi-source data based on a manual processing manner in the prior art, thereby improving the efficiency of acquiring geographic space information from multi-source data, reducing manual errors, and making the acquired geographic space information more accurate and reliable. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0021] Figure 1 is one of the flowcharts of the geographic space information acquisition method provided by the application.

[0022] Figure 2 is the acquisition of a plurality of geographic space event information corresponding to a plurality of data sources in the geographic space information acquisition method provided by the application.

[0023] Figure 3 is a schematic diagram of outputting the geographic space information under each data source in the geographic space information acquisition method provided by the application.

[0024] Figure 4 is the second flowchart of the geographic space information acquisition method provided by the application.

[0025] Figure 5 is the third flowchart of the geographic space information acquisition method provided by the application.

[0026] Figure 6 This is the fourth flowchart of the geospatial intelligence acquisition method provided by the present invention.

[0027] Figure 7 This is a schematic diagram of the geospatial information acquisition device provided by the present invention.

[0028] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0030] Breakthroughs in deep learning, natural language processing (NLP), and computer vision have opened up new possibilities for the automatic extraction of multi-source heterogeneous data. These technologies can autonomously learn how to more effectively extract valuable knowledge points from different types of raw data, improving the accuracy and efficiency of data processing.

[0031] The following is combined Figures 1-6 The geospatial intelligence acquisition method of the present invention is described.

[0032] Figure 1 This is one of the flowcharts illustrating the geospatial intelligence acquisition method provided by the present invention, such as... Figure 1 As shown, the method includes the following steps S110~S130.

[0033] S110: Receive descriptive information related to the intelligence to be acquired.

[0034] Descriptions of the intelligence to be acquired include, for example, time information, geographic location descriptions, and geospatial event keywords.

[0035] Time information can be a time period, such as the start and end dates of a specific period to obtain geospatial intelligence. Alternatively, it can be a specific point in time, such as "Year, Month, Day, Hour".

[0036] Geographic location description information, such as latitude and longitude, or geographical region name (e.g., XX sea area).

[0037] The geographic space event keyword is a keyword capable of expressing a geographic space event, such as a keyword related to a military action such as convoy escort by a patrol ship, illegal entry into territorial waters, or a keyword of a natural disaster such as a tsunami, red tide, seawater intrusion, or oil spill disaster.

[0038] The type of the description information to be acquired can include any one of the following: text and image. Of course, it can be understood that the above-mentioned types of description information are only exemplary, and the embodiments of the present application include that the type of the description information to be acquired includes but is not limited to text and image, and can also be voice.

[0039] S120: According to the description information, geographic space event information related to the description information is extracted from the data set of the acquired multiple data sources, and multiple geographic space event information corresponding to the multiple data sources is obtained.

[0040] In some embodiments, the multiple target data sources can be selected from the multiple preset data sources according to the received data source selection indication information, and then the data set of the multiple data sources can be acquired according to the data acquisition manner corresponding to each data source in the multiple data sources, and then the geographic space event information can be extracted from the data set of each data source. The data source selection indication information carries the identification of the multiple target data sources, so that the multiple target data sources can be selected from the multiple preset data sources. The data source selection indication information can be generated according to the selection operation of the user on the multiple preset data sources.

[0041] In some embodiments, in order to obtain more accurate data, after acquiring the data set of the multiple data sources, the corresponding data set can be preprocessed according to the preprocessing manner corresponding to each data source; then the initial processing information is extracted from the preprocessing result of each data set; finally, the initial processing information of each data source is entity linked to obtain the geographic space event information corresponding to each data source.

[0042] The above-mentioned data set acquisition process, preprocessing process of the data set of the corresponding data source, information extraction process and entity linking process will be described below.

[0043] As shown in Figure 2 The multiple preset data sources can include news webpage data, social media data, remote sensing image data, unmanned aerial vehicle aerial photography data, and the like.

[0044] The multiple data sources can be selected by the user from the multiple preset data sources according to the actual situation, and of course, the multiple preset data sources can be the multiple target data sources by default.

[0045] Assume that the multiple target data sources selected from multiple preset data sources include news webpage data, social media data, remote sensing image data, and unmanned aerial vehicle aerial photography data. The following takes the four data sources as examples to illustrate the process of obtaining multiple geographic spatial event information.

[0046] Specifically, the news webpage data may include, for example, text, pictures, videos, location information, and other data in news webpages issued on various news platforms on the network. The social media data may include, for example, text, pictures, videos, location information, and other data on chat software and social media platforms. The remote sensing image data may include, for example, optical images, synthetic aperture radar images, multispectral / hyperspectral images, and other image data. The unmanned aerial vehicle aerial photography data may include, for example, infrared images and laser radar data.

[0047] The data sets of the multiple data sources may be obtained in different ways. For example, for the social media data and the news webpage data, refer to Figure 2 The data sets may be obtained by a multi-modal webpage data extraction technology based on rule template information extraction (a pre-defined rule framework used to standardize the obtained information content, information type (for example, picture or text), and the like) and text information and visual information extraction.

[0048] For the news webpage data and the social media data, a general first information extraction model may be constructed by using a deep neural network. The visual information and the text information are combined, for example, the text information of a webpage DOM tree (Document Object Model Tree) and the visual structure information generated by a WEB rendering engine are combined and mapped to a deep neural network for coding. The key information related to the description information (title, content, time, author, category, and the like) in the news webpage is uniformly identified and extracted to obtain an intelligent universal parser of heterogeneous data in different webpages, that is, the first information extraction model. The news webpage data is intelligently extracted based on the first information extraction module, and the data sets related to the description information obtained in S110 are extracted from the news webpage data and the social media data.

[0049] Refer to Figure 2For remote sensing image data and UAV aerial photography data, the corresponding data sets can be obtained through an Application Programming Interface (API) interface or a transmission private line. For open data sets such as Landsat, Sentinel, Moderate Resolution Imaging Spectroradiometer (MODIS), etc., the API interface is used to call the relevant equipment; for national or regional data centers, remote sensing image data and UAV aerial photography data sets can be obtained from the relevant equipment through a transmission private line.

[0050] After obtaining the data sets of various data sources, in order to filter out unnecessary, useless or mixed noise data, the data sets of each data source obtained can be preprocessed.

[0051] Different preprocessing methods can be used to preprocess the data sets of different data sources.

[0052] Referring to Figure 2 For the data sets obtained from news webpage data and social media data, the preprocessing methods that can be used include at least one of the following: missing value processing, removing out-of-bound values, consistency checking, deleting duplicate values, labeling processing, and formatting.

[0053] Specifically, the missing value processing method is used to process missing values in the data set, for example, using mean, median, interpolation or prediction model to fill in missing values. The out-of-bound value removal method detects data points outside the range in the data set by defining a reasonable threshold, and processes the data points outside the range by deleting, replacing or marking. The consistency checking method is used to ensure the consistency of all data in the data set in terms of format, unit and logic. For example, for dates in the data set, all dates are unified to the format YYYY-MM-DD, the value unit in the data set is checked to ensure the consistency of cross-table associated fields, etc. The duplicate value deletion method identifies and deletes duplicate records in the data set to ensure the uniqueness of the data and delete redundant data. The labeling processing method converts classification variables in data analysis and machine learning into numerical form. For text data, perform word segmentation, stem extraction and word embedding, etc. Formatting includes unifying data formats (such as date and value formats), standardizing or normalizing data, and saving data in a set file format.

[0054] Referring to Figure 2 For the data sets obtained from remote sensing image data and UAV aerial photography data, the preprocessing methods that can be used include at least one of the following: image orthorectification, image registration, image fusion, and image mosaic process.

[0055] Specifically, the orthorectification of the image adopts a data automatic processing platform, selects control points from the data set, performs regional network joint adjustment on the full-color data in the work area, removes and adjusts control points with large residual errors, automatically extracts homonymic points in the overlapping area between scenes on the basis that the control points in each scene meet the error requirements of the regional network, and removes homonymic points with large residual errors until the homonymic point error meets the specified requirements. The image registration takes the orthorectified full-color band image as the control basis, uses the homonymic points of the full-color image to correct and register the multispectral data. Image fusion is to fuse the high-resolution full-color remote sensing image and the low-resolution multispectral remote sensing image to obtain a fused image with rich color information and high resolution, which highlights the information of ground feature elements. Image mosaicking performs mosaicking processing on the fused digital orthophoto map (DOM) image through mosaicking lines, performs color uniformization processing on the images between adjacent mosaicking areas, keeps the color transition natural at the joint between scenes, and reasonably joints the ground features without ghosting and virtual phenomena.

[0056] When extracting and describing the geospatial event information from the data set obtained from multiple data sources, the following methods can be used for extraction.

[0057] Referring to Figure 2 For the data set obtained from news webpage data and social media data, entity recognition, relationship extraction and feature extraction methods can be used to extract geospatial event information.

[0058] For example, a joint model combining a statistical machine learning method based on a stacked conditional random field and a rule-based probability recognition method automatically recognizes the named entity of the required category from the input text (i.e., the data set input into the joint model), and then determines the category to which the entity belongs, thereby completing the entity recognition process. A remote supervision relationship extraction method combining a gated recurrent unit (GRU) neural network and an attention mechanism is used to extract entity relationships, thereby completing the relationship extraction process. Then, term frequency-inverse document frequency (TF-IDF) is used for feature extraction, including lexical features, syntactic features and semantic features, thereby completing the feature extraction process, and obtaining the geospatial event information corresponding to the news webpage data and the social media data.

[0059] Referring to Figure 2 For the data set obtained from remote sensing image data and unmanned aerial vehicle aerial photography data, a feature pyramid construction and target recognition method can be used to extract geospatial event information.

[0060] For example, a dynamic multi-scale feature pyramid structure can be constructed, the input feature map is trained by using an anchor mechanism and a regression frame mechanism to obtain a candidate frame, and a multi-scale feature map obtained by adding high-level semantic information and low-level detail information can take into account the category information and the position information, and the category information and the position information are determined as the geographic spatial event information corresponding to the remote sensing image data.

[0061] For each kind of geographic spatial event information extracted above, in order to further improve the accuracy of the geographic spatial event information, the geographic spatial event information of each kind of data source can be taken as initial processing information of the kind of data source, and the initial processing information is subjected to entity linking to obtain final geographic spatial event information of the kind of data source.

[0062] Referring to Figure 3 The multi-modal entity linking can supplement the context information missing in the text mode and help eliminate the ambiguity of the text entity by using implicit reasoning information. The processing mode of entity linking includes at least one of the following: candidate entity generation, entity disambiguation, and fuzzy reference recognition. The candidate entity generation obtains a multi-modal knowledge graph of the geographic spatial event information by a multi-modal clustering algorithm, divides the data samples similar to the search data (description information related to the intelligence to be obtained) in the multi-modal knowledge graph into the same cluster, so that the similarity between the data points in the same cluster is maximum and the similarity between the data points in different clusters is minimum, and the entity data in the same cluster is the candidate entity set. The purpose is to calculate the similarity between the entity to be obtained (the entity of the description information related to the intelligence to be obtained) and the entity in the candidate entity set (the geographic spatial event information). The entity disambiguation converts the multi-modal entity into a vector representation, fuses the feature vectors of each mode into a comprehensive multi-modal feature vector, and finally maps the fused multi-modal feature vector into a unified vector space using a deep neural network. Through similarity calculation, the candidate entity with the highest similarity score is selected as the final disambiguation result, the entity linking work is completed, and the geographic spatial event information corresponding to each kind of data source is finally obtained.

[0063] S130: generating geographic spatial intelligence under each kind of data source according to the plurality of geographic spatial event information.

[0064] The geographic spatial intelligence can include time, geographic location, and event. The geographic location can include latitude and longitude, or geographic coordinate information positioned in other ways.

[0065] The time, the geographic location, and the event occurring at the geographic location at the time are extracted from the plurality of geographic spatial event information as the event, and the time, the geographic location, and the event are associated to obtain the geographic spatial intelligence.

[0066] Through S110-S130, a plurality of geographic spatial event information corresponding to a plurality of data sources can be automatically extracted from a wide range of sources, complex structure and a large amount of implicit information. Finally, according to the plurality of geographic spatial event information, the geographic spatial intelligence under each data source is generated.

[0067] In some embodiments, after generating the geographic spatial intelligence under each data source, the geographic spatial intelligence under each data source can be directly outputted.

[0068] For example, the geographic spatial intelligence under each data source can be displayed on the terminal in a set display manner. For example, a table as shown in Figure 3 Each row of the table includes time, geographic location and event, corresponding to a group of intelligence information. A plurality of groups of intelligence information are generated in chronological order and arranged in chronological order.

[0069] Of course, it can be understood that, Figure 3 Only one display manner of the geographic spatial intelligence is exemplarily shown, and in the specific implementation, the display manner of the geographic spatial intelligence includes but is not limited to Figure 4 The manners listed in

[0070] In some embodiments, as shown in Figure 5 After S130 is executed, S410-S420 can also be executed.

[0071] S410: generating geographic spatial prediction intelligence according to the geographic spatial intelligence under each data source.

[0072] The initial intelligence prediction model can be trained based on the previously obtained geographic spatial intelligence under all data sources to obtain a target intelligence prediction model.

[0073] The currently obtained geographic spatial intelligence under each data source is inputted into the target intelligence prediction model to obtain the geographic spatial prediction intelligence outputted by the target intelligence prediction model.

[0074] S420: outputting the geographic spatial prediction intelligence.

[0075] The geographic spatial prediction intelligence is displayed in the display interface of the terminal.

[0076] In some embodiments, as shown in Figure 6 After S130 is executed, S510-S520 can also be executed.

[0077] S510: According to the geographical space prediction information and the pre-set early warning event information, the occurrence probability of the early warning event is analyzed.

[0078] The early warning event information may include, for example, early warning keywords, early warning pictures, etc., and the occurrence probability of the early warning event may be calculated, for example, by calculating the similarity between the geographical space prediction information and the early warning event information, and determining the similarity as the occurrence probability of the early warning event.

[0079] S520: In the case where the occurrence probability is greater than a set threshold, an alarm is issued.

[0080] The set threshold may be set by a person skilled in the art according to actual conditions, and the embodiments of the present application are not limited in this regard.

[0081] In some implementations, as shown in FIG. 13B, after S130 is performed, the following S610-S620 can also be performed. Figure 7

[0082] S610: According to the geographical space information under each data source, information recommendation information is generated.

[0083] The information recommendation information may include, for example, information about events that occur frequently at a geographical location of the geographical space information, or events that are highly concerned at the geographical location, etc.

[0084] S620: The information recommendation information is output.

[0085] The present application proposes to use deep learning, natural language processing (NLP), knowledge graph, computer vision and other technologies to construct a multi-source heterogeneous data knowledge automatic extraction method in the field of geographical space information based on remote sensing image data, unmanned aerial vehicle aerial photography data, social media data, sensor network data and other multi-source heterogeneous data, to integrate different types of data such as text, pictures and videos across modalities, to improve data processing efficiency and quality, and to enhance the analysis capability of geographical space information.

[0086] ​The application provides a geographical space information acquisition method. The method comprises the following steps: receiving description information related to information to be acquired; extracting geographical space event information related to the description information from a data set of a plurality of data sources, to obtain a plurality of geographical space event information corresponding to the plurality of data sources; and generating geographical space information under each data source according to the plurality of geographical space event information. Therefore, the application can automatically extract geographical space event information corresponding to each data source from a plurality of data sources, generate geographical space information of each data source according to the geographical space event information corresponding to each data source, and output geographical space information under each data source, so as to solve the problem of low accuracy and efficiency of acquiring geographical space information from multi-source data based on manual processing in the prior art, improve the efficiency of acquiring geographical space information from multi-source data, reduce manual errors, and make the acquired geographical space information more accurate and reliable.

[0087] The geographical space information acquisition device provided by the application is described below. The geographical space information acquisition device described below can be referred to in correspondence with the geographical space information acquisition method described above.

[0088] Figure 7 FIG. 1 is a structural schematic diagram of the geographical space information acquisition device provided by the application. As shown in FIG. 1, the geographical space information acquisition device 700 comprises a receiving module 701, an extracting module 702, and a generating module 703. Figure 7

[0089] The receiving module 701 is configured to receive description information related to information to be acquired.

[0090] The extracting module 702 is configured to extract geographical space event information related to the description information from a data set of a plurality of data sources according to the description information, to obtain a plurality of geographical space event information corresponding to the plurality of data sources.

[0091] The generating module 703 is configured to generate geographical space information under each data source according to the plurality of geographical space event information.

[0092] In some embodiments, the geographical space information acquisition device 700 further comprises an output module (not shown in the figure) configured to output the geographical space information under each data source. Figure 7

[0093] In some embodiments, the geographical space information acquisition device 700 further comprises a prediction information generating module (not shown in the figure) and a prediction information output module (not shown in the figure). Figure 7 Figure 7

[0094] ​​​​The prediction information generation module is configured to generate, by the generation module 703, the geospatial prediction information according to the geospatial intelligence of each data source after the generation module 703 generates the geospatial intelligence of each data source according to the plurality of geospatial event information.

[0095] The prediction information output module is configured to output the geospatial prediction information.

[0096] In some embodiments, the geospatial intelligence acquisition apparatus 700 further comprises an analysis module (not shown in the figure) and an alarm module (not shown in the figure). Figure 7 Figure 7

[0097] The analysis module is configured to analyze the occurrence probability of the early warning event according to the geospatial prediction information and the pre-set early warning event information after the prediction information generation module generates the geospatial prediction information according to the geospatial intelligence of each data source.

[0098] The alarm module is configured to issue an alarm when the occurrence probability is greater than a set threshold.

[0099] In some embodiments, the geospatial intelligence acquisition apparatus 700 further comprises a recommendation information generation module (not shown in the figure) and a recommendation information output module (not shown in the figure). Figure 8 Figure 8

[0100] The recommendation information generation module is configured to generate the intelligence recommendation information according to the geospatial intelligence of each data source after the generation module 703 generates the geospatial intelligence of each data source according to the plurality of geospatial event information.

[0101] The recommendation information output module is configured to output the intelligence recommendation information.

[0102] In some embodiments, the extraction module 702 is specifically configured to: select, according to the received data source selection indication information, a plurality of target data sources from a plurality of preset data sources as the plurality of data sources; acquire, according to a data acquisition manner corresponding to each data source in the plurality of data sources, a data set of the plurality of data sources; pre-process the data set of each data source according to a pre-processing manner corresponding to the data source; extract initial processing information from a pre-processing result of the data set of each data source; and perform entity linking on the initial processing information of each data source to obtain the plurality of geospatial event information.

[0103] In some embodiments, the geospatial intelligence comprises time, geographical position, and an occurrence event.

[0104] ​​​​The geospatial intelligence acquisition device provided by this invention receives descriptive information related to the intelligence to be acquired; extracts geospatial event information related to the descriptive information from datasets of multiple data sources based on the descriptive information, obtaining multiple geospatial event information corresponding to each data source; and generates geospatial intelligence for each data source based on the multiple geospatial event information. Therefore, this invention can automatically extract geospatial event information corresponding to each data source from multiple data sources, and generate geospatial intelligence for that data source based on the geospatial event information corresponding to each data source. This solves the problem of low accuracy and efficiency in obtaining geospatial intelligence from multi-source data using manual processing methods in existing technologies, improving the efficiency of obtaining geospatial intelligence from multi-source data, reducing human error, and making the obtained geospatial intelligence more accurate and reliable.

[0105] ​ An example is a schematic diagram of the physical structure of an electronic device, such as... ​ As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840. The processor 810, communications interface 820, and memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a geospatial intelligence acquisition method. This method includes: receiving descriptive information related to the intelligence to be acquired; extracting geospatial event information related to the descriptive information from datasets of multiple data sources based on the descriptive information, obtaining multiple geospatial event information corresponding to each data source; and generating geospatial intelligence for each data source based on the multiple geospatial event information.

[0106] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0107] In another aspect, the present application also provides a computer program product comprising a computer program, the computer program being stored in a non-transitory computer readable storage medium, and the computer program being executable by a processor to cause a computer to perform the above-mentioned method for obtaining geospatial intelligence, the method comprising: receiving description information related to intelligence to be obtained; extracting geospatial event information related to the description information from data sets of a plurality of data sources according to the description information, to obtain a plurality of geospatial event information corresponding to the plurality of data sources one by one; and generating geospatial intelligence under each data source according to the plurality of geospatial event information.

[0108] In yet another aspect, the present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, the computer program being executable by a processor to implement a method for obtaining geospatial intelligence, the method comprising: receiving description information related to intelligence to be obtained; extracting geospatial event information related to the description information from data sets of a plurality of data sources according to the description information, to obtain a plurality of geospatial event information corresponding to the plurality of data sources one by one; and generating geospatial intelligence under each data source according to the plurality of geospatial event information.

[0109] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0110] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary universal hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0111] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for acquiring geospatial intelligence, characterized in that, include: Receive descriptive information related to the intelligence to be acquired; Based on the described information, geospatial event information related to the described information is extracted from datasets from multiple data sources to obtain multiple geospatial event information corresponding to each of the multiple data sources. The geospatial event information is obtained in the following way: initial processing information is extracted based on each dataset, and the initial processing information represents the initially obtained geospatial event information; entity linking is performed on the initial processing information of each data source to obtain the geospatial event information corresponding to each data source; wherein, the method of extracting the initial processing information is to use entity recognition, relation extraction and feature extraction, or to use feature pyramid construction and target recognition; the entity linking processing method includes at least one of the following: candidate entity generation, entity disambiguation, and fuzzy reference recognition; Based on the various geospatial event information, generate geospatial intelligence for each data source. The step of extracting geospatial event information related to the description information from datasets of multiple data sources, based on the description information, to obtain multiple geospatial event information corresponding to each of the multiple data sources, includes: Based on the received data source selection instruction information, select multiple target data sources from a variety of preset data sources as the multiple data sources, including news web page data, social media data, remote sensing image data, and drone aerial photography data; According to the data acquisition method corresponding to each of the multiple data sources, the datasets of the multiple data sources are obtained from the multiple data sources, including: for the news web page data and the social media data, the datasets are obtained through rule template information extraction, text and visual information extraction; for the remote sensing image data and the UAV aerial photography data, the datasets are obtained through application programming interfaces or dedicated transmission lines. According to the preprocessing method corresponding to each of the data sources, the datasets of the corresponding data sources are preprocessed, including: for the news web page data and the social media data, preprocessing is performed through missing value processing, removal of out-of-bounds values, consistency checks, deletion of duplicate values, labeling, and formatting; for the remote sensing image data and the UAV aerial photography data, preprocessing is performed through image orthorectification, image registration, image fusion, and image mosaicking. Initial processing information is extracted from the preprocessing results of the datasets of each of the aforementioned data sources. For the news webpage data and the social media data, initial processing information is obtained by means of entity recognition, relation extraction, and feature extraction. For the remote sensing image data and the UAV aerial photography data, initial processing information is obtained by means of feature pyramid construction and target recognition. Entity links are created for the initial processing information of each of the aforementioned data sources to obtain the various geospatial event information.

2. The geospatial intelligence acquisition method according to claim 1, characterized in that, After generating geospatial intelligence for each data source based on the various geospatial event information, the process further includes: Based on the geospatial intelligence under each of the aforementioned data sources, generate geospatial prediction intelligence; Output the geospatial prediction intelligence.

3. The geospatial intelligence acquisition method according to claim 2, characterized in that, After generating geospatial prediction intelligence based on the geospatial intelligence from each of the aforementioned data sources, the process further includes: Based on the geospatial prediction intelligence and pre-set early warning event information, analyze the probability of the occurrence of the early warning event; An alarm is issued if the probability of occurrence exceeds a set threshold.

4. The geospatial intelligence acquisition method according to claim 1, characterized in that, After generating geospatial intelligence for each data source based on the various geospatial event information, the process further includes: Based on the geospatial intelligence under each of the aforementioned data sources, intelligence recommendation information is generated; Output the intelligence recommendation information.

5. The geospatial intelligence acquisition method according to any one of claims 1-4, characterized in that, The geospatial intelligence includes: time, geographical location, and the event that occurred.

6. A geospatial intelligence acquisition device, characterized in that, include: The receiving module is used to receive descriptive information related to the intelligence to be acquired; An extraction module is used to extract geospatial event information related to the description information from datasets of multiple data sources, based on the description information, to obtain multiple geospatial event information corresponding to each of the multiple data sources. The geospatial event information is obtained in the following way: initial processing information is extracted based on each dataset, where the initial processing information represents the initially obtained geospatial event information; entity linking is performed on the initial processing information of each data source to obtain the geospatial event information corresponding to each data source; wherein, the method of extracting the initial processing information is to use entity recognition, relation extraction and feature extraction, or to use feature pyramid construction and target recognition; the entity linking processing method includes at least one of the following: candidate entity generation, entity disambiguation, and fuzzy reference recognition. The generation module is used to generate geospatial intelligence for each data source based on the various geospatial event information. Specifically, the extraction module is used to: select multiple target data sources from a variety of preset data sources as the multiple data sources according to the received data source selection instruction information, wherein the multiple data sources include news webpage data, social media data, remote sensing image data, and drone aerial photography data; obtain datasets from the multiple data sources according to the data acquisition method corresponding to each of the multiple data sources, including: obtaining datasets from the news webpage data and the social media data through rule template information extraction, text and visual information extraction; obtaining datasets from the remote sensing image data and the drone aerial photography data through application programming interfaces or dedicated transmission lines; and preprocessing the datasets from the corresponding data sources according to the preprocessing method corresponding to each of the data sources, including: The news webpage data and social media data are preprocessed through missing value handling, out-of-bounds value removal, consistency checks, duplicate value removal, tagging, and formatting. The remote sensing image data and UAV aerial photography data are preprocessed through image orthorectification, image registration, image fusion, and image mosaicking. Initial processing information is extracted from the preprocessing results of each data source dataset. For the news webpage data and social media data, initial processing information is obtained using entity recognition, relation extraction, and feature extraction. For the remote sensing image data and UAV aerial photography data, initial processing information is obtained using feature pyramid construction and target recognition. Entity links are established for the initial processing information of each data source to obtain the various geospatial event information.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the geospatial intelligence acquisition method as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the geospatial intelligence acquisition method as described in any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the geospatial intelligence acquisition method as described in any one of claims 1 to 5.

Images