A method for constructing natural resource remote sensing monitoring process based on multidimensional knowledge reasoning

Abstract

The present application relates to a method for constructing a natural resource remote sensing monitoring process based on multi-dimensional knowledge reasoning, including: classifying expert knowledge in the fields of remote sensing image processing and natural resource monitoring into business knowledge, data knowledge, algorithm knowledge, and geospatial knowledge. Categorize expert knowledge, sort out a unified knowledge system from multi-dimensional perspectives such as business needs, multi-source data, algorithm models, and geospatial knowledge, integrate remote sensing multi-source data, multi-dimensional business analysis, multi-factor extraction, and multi-algorithm coupling construction processes and technologies, and construct the processes required to process user business needs by designing unified normative rules, so as to provide technical support for natural resource remote sensing monitoring.

Description

Technical Field

[0001] The present application relates to the field of computer science and technology, and in particular to a method for constructing a natural resource remote sensing monitoring process based on multi-dimensional knowledge reasoning. Background Art

[0002] Remote sensing monitoring of natural resources is a complex system engineering that requires the comprehensive use of satellite, aviation, ground and other remote sensing technologies to dynamically grasp the status and changes of various resources such as cultivated land, forests, grasslands, wetlands, water resources, surface matrix, underground space, and marine resources, and provide data guarantee and technical support for resource protection, system repair, and comprehensive management. Dynamic remote sensing monitoring of natural resources involves the collaboration of core contents such as automatic screening of multi-source data, automatic processing of multi-dimensional information, and spontaneous collaboration of multi-level operators. It has the characteristics of diverse surface types, complex and changeable scenes, and cumbersome processing procedures. How to realize the computer's automatic, fast, efficient, and accurate dynamic monitoring of the range of changes and automatic extraction of characteristic objects and automatic analysis of monitoring services is crucial to responding to the "Overall Plan for the Construction of Natural Resources Survey and Monitoring System" and improving the intelligence of monitoring.

[0003] The existing natural resource remote sensing monitoring process construction mainly relies on the traditional expert interactive decision-making method. However, due to the complexity of the business, the many fields involved, and the lack of unified expert process construction standards between different fields and departments, the existing natural resource remote sensing monitoring process construction method is complex, time-consuming and laborious. Summary of the invention

[0004] In order to at least to some extent overcome the problem that the natural resource remote sensing monitoring process construction method in the related technology is complicated and laborious, the present application provides a natural resource remote sensing monitoring process construction method based on multi-dimensional knowledge reasoning.

[0005] The scheme of this application is as follows:

[0006] A method for constructing a natural resource remote sensing monitoring process based on multi-dimensional knowledge reasoning, comprising:

[0007] The expert knowledge in the fields of remote sensing image processing and natural resource monitoring is classified into business knowledge, data knowledge, algorithm knowledge and geospatial knowledge;

[0008] Receive user business requirements and parse them into formal representations corresponding to knowledge;

[0009] Select the business node with the highest matching degree with the user's business needs from the business knowledge as the required business node;

[0010] Selecting data with a matching degree higher than a preset value in the data knowledge as candidate data sets;

[0011] Selecting algorithms whose matching degree with the demand service node is higher than a preset value from the algorithm knowledge as a candidate algorithm set;

[0012] Selecting the geospatial knowledge with the highest matching degree with the user's business demand and the business node of the demand from the geospatial knowledge as the auxiliary decision-making geospatial knowledge;

[0013] Establishing disjunctive matching rules between the candidate data sets, candidate algorithm sets and auxiliary decision-making geospatial nodes, and outputting data, algorithms and geospatial knowledge corresponding to the maximum matching score;

[0014] Build the processes required to handle user business needs based on the data, algorithms, and geospatial knowledge corresponding to the maximum matching score.

[0015] Preferably, the method further comprises:

[0016] Structural sorting of various types of knowledge;

[0017] Formalize the structured knowledge to form a remote sensing monitoring knowledge base;

[0018] The matching rules are sorted out to form an inference rule knowledge base.

[0019] Preferably, various types of knowledge are structured and sorted out, including:

[0020] Grading of various types of knowledge;

[0021] The hierarchical knowledge is analyzed from top to bottom to extract the organizational structure and attribute information of nodes at each level, as well as the connection relationship between nodes at each level.

[0022] Preferably, the structured knowledge is formally represented, including:

[0023] The extraction results of various types of knowledge are formally defined to obtain the unique identifiers corresponding to each type of knowledge as the basis for node matching.

[0024] Preferably, the business knowledge includes: an abstract and complete summary of basic business demand information for monitoring behavior;

[0025] The data knowledge includes: information and features of time, region, land types, and sensor types contained in the remote sensing data itself;

[0026] The algorithm knowledge includes: basic capability description of the algorithm and adaptive execution conditions;

[0027] The geographic spatial knowledge includes: the inherent characteristics of the monitored object in the real space, its location, the abstract summary of the spatial information such as the shape of the object, etc.

[0028] Preferably, the formalized representation of the user's business requirements after analysis includes:

[0029] Time, location, monitoring elements, business type, monitoring content, algorithm requirements and auxiliary feature requirements.

[0030] Preferably, the algorithm for calculating the matching degree includes:

[0031] Define matching relationships for different matching items and assign corresponding weights;

[0032] Calculate the fuzzy matching score or the complete matching score of the item to be matched;

[0033] According to the fuzzy matching score or the complete matching score of the items to be matched, the sum of the matching degrees of all the items to be matched that have been assigned weights is obtained.

[0034] Preferably, the method further comprises:

[0035] When multi-step processing is required, matching rules between algorithm nodes in algorithm knowledge are established;

[0036] Screen multiple algorithm nodes that match the front and back functions, interfaces, data knowledge, and geographic space knowledge in the algorithm knowledge;

[0037] By assisting decision making, geographic spatial knowledge is embedded into each link of the process as supplementary information for each link.

[0038] Preferably, the method further comprises:

[0039] Based on the new knowledge acquired and user feedback, the remote sensing monitoring knowledge base and inference rule knowledge base are updated and expanded.

[0040] Preferably, the remote sensing monitoring knowledge base and matching rules are updated and expanded based on the acquired new knowledge and user feedback, including:

[0041] Construct a machine-readable model reasoning analysis layer to update and expand the remote sensing monitoring knowledge base and the reasoning rule knowledge base; including: according to the scene requirements, call the remote sensing monitoring knowledge base and the reasoning rule knowledge base for reasoning, analyze the adaptation operator model and data according to the specific needs of the user, and make behavioral decisions for each process step in the monitoring process;

[0042] Collect and process new data sources in real time to acquire new knowledge. When new knowledge is acquired, if the corresponding knowledge does not exist in the remote sensing monitoring knowledge base, the new knowledge will be stored, and expired or erroneous knowledge in the remote sensing monitoring knowledge base will be processed regularly.

[0043] Adjust the remote sensing monitoring knowledge base based on user feedback and historical interaction data.

[0044] The technical solution provided by the present application may include the following beneficial effects: The method for constructing a natural resource remote sensing monitoring process based on multidimensional knowledge reasoning in the present application includes: classifying expert knowledge in the field of remote sensing image processing and natural resource monitoring into business knowledge, data knowledge, algorithm knowledge and geospatial knowledge. Receive user business needs, and parse the user business needs into a formalized representation corresponding to the knowledge; select the business node with the highest matching degree with the user business needs in the business knowledge as the demand business node; select the data with a matching degree with the demand business node higher than a preset value in the data knowledge as the candidate data set; select the algorithm with a matching degree with the demand business node higher than a preset value in the algorithm knowledge as the candidate algorithm set; select the geospatial knowledge with the highest matching degree with the user business needs and the demand business node in the geospatial knowledge as the auxiliary decision-making geospatial knowledge; establish the disjunctive matching rules between the candidate data set, the candidate algorithm set and the auxiliary decision-making geospatial node, and output the data, algorithm and geospatial knowledge corresponding to the maximum matching score; construct the process required to process the user business needs according to the data, algorithm and geospatial knowledge corresponding to the maximum matching score.

[0045] In this technical solution, expert knowledge is classified, and a unified knowledge system is sorted out from multiple dimensions such as business needs, multi-source data, algorithm models, and geospatial knowledge. Remote sensing multi-source data, multi-dimensional business analysis, multi-factor extraction, and multi-algorithm coupling construction processes and technologies are integrated. By designing unified standard rules, the processes required to process user business needs are constructed to provide technical support for remote sensing monitoring of natural resources.

[0046] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0048] Figure 1 It is a flowchart of a method for constructing a natural resource remote sensing monitoring process based on multi-dimensional knowledge reasoning provided by an embodiment of the present application;

[0049] Figure 2 It is a data knowledge organization structure table diagram provided by an embodiment of the present application;

[0050] Figure 3 It is an algorithm knowledge organization structure table diagram provided by an embodiment of the present application;

[0051] Figure 4 It is a business knowledge organization structure chart provided by an embodiment of the present application;

[0052] Figure 5 It is a geographic spatial knowledge organization structure table diagram provided by an embodiment of the present application;

[0053] Figure 6 It is a structural diagram of a natural resource remote sensing monitoring process construction system based on multi-dimensional knowledge reasoning provided by an embodiment of the present application. DETAILED DESCRIPTION

[0054] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0055] Embodiment 1

[0056] A method for constructing a natural resource remote sensing monitoring process based on multi-dimensional knowledge reasoning, comprising:

[0057] S11: Expert knowledge in the fields of remote sensing image processing and natural resource monitoring is classified into business knowledge, data knowledge, algorithm knowledge, and geospatial knowledge;

[0058] S12: receiving user business requirements and parsing the user business requirements into a formal representation corresponding to the knowledge;

[0059] S13: Select the business node with the highest matching degree with the user's business demand from the business knowledge as the demand business node;

[0060] S14: Filtering data with a matching degree with the required business node higher than a preset value in the data knowledge as a candidate data set;

[0061] S15: selecting algorithms whose matching degree with the required business node is higher than a preset value from the algorithm knowledge as a candidate algorithm set;

[0062] S16: Select the geospatial knowledge with the highest matching degree with the user's business requirements and the required business nodes from the geospatial knowledge as the auxiliary decision-making geospatial knowledge;

[0063] S17: Establishing disjunctive matching rules between candidate data sets, candidate algorithm sets and auxiliary decision-making geospatial nodes, and outputting data, algorithms and geospatial knowledge corresponding to the maximum matching score;

[0064] S18: Build the processes required to handle user business needs based on the data, algorithms, and geospatial knowledge corresponding to the maximum matching score.

[0065] In this technical solution, expert knowledge is classified, and a unified knowledge system is sorted out from multiple dimensions such as business needs, multi-source data, algorithm models, and geospatial knowledge. Remote sensing multi-source data, multi-dimensional business analysis, multi-factor extraction, and multi-algorithm coupling construction processes and technologies are integrated. By designing unified standard rules, the processes required to process user business needs are constructed to provide technical support for remote sensing monitoring of natural resources.

[0066] Embodiment 2

[0067] It should be noted that the method also includes:

[0068] Structural sorting of various types of knowledge;

[0069] Formalize the structured knowledge to form a remote sensing monitoring knowledge base;

[0070] The matching rules are sorted out to form an inference rule knowledge base.

[0071] In this technical solution, the expert knowledge in the field of remote sensing image processing and the expert knowledge in the field of natural resource monitoring are structured, including the organizational structure of nodes and the extraction of their attributes, the extraction of the connection relationships between nodes at all levels, and the use of mathematical symbols to formally represent the structured expert knowledge, establish an expression model for remote sensing monitoring knowledge, and use neo4j to store structured and formalized expert knowledge to form a remote sensing monitoring knowledge base as prior knowledge for subsequent reasoning and decision-making recommendations.

[0072] Structural sorting of various types of knowledge, including:

[0073] Grading of various types of knowledge;

[0074] The hierarchical knowledge is analyzed from top to bottom to extract the organizational structure and attribute information of nodes at each level, as well as the connection relationship between nodes at each level.

[0075] Summarize the abstract concepts in the domain knowledge and refine them step by step from top to bottom, abstractly summarize the monitoring process, classify the knowledge into four categories: business knowledge, data knowledge, algorithm knowledge and geographic space knowledge, analyze the concepts, attributes and hierarchical semantic relationships from top to bottom, form a monitoring process knowledge model, and extract the entities and attributes of the domain standard knowledge based on the established knowledge model;

[0076] The monitoring process is defined as:

[0077] process=f{A,K}

[0078] A represents rules, and K represents knowledge including business knowledge, data knowledge, algorithm knowledge and geographic space knowledge, which means that the process is obtained by matching knowledge and rules.

[0079] Four types of knowledge are defined as quadruples:

[0080] {K d ,K a ,K r ,K b}

[0081] They represent data knowledge, algorithm knowledge, business knowledge and geospatial knowledge respectively.

[0082] Formal representation of structured knowledge, including:

[0083] The extraction results of various types of knowledge are formally defined to obtain the unique identifiers corresponding to each type of knowledge as the basis for node matching.

[0084] It should be noted that business knowledge includes: an abstract and complete summary of the basic business demand information for monitoring behavior;

[0085] Data knowledge includes: information and features of time, region, land types, and sensor types contained in the remote sensing data itself;

[0086] Algorithm knowledge includes: description of the basic capabilities of the algorithm and the conditions under which it is adapted for execution;

[0087] Geographical spatial knowledge includes: the abstract summary of spatial information such as the inherent characteristics of the monitored objects in real space, their location, and the shape of the objects.

[0088] Specifically, the abstract and structured descriptions of the four types of knowledge, namely data knowledge, algorithm knowledge, geographic spatial knowledge, and business knowledge, are as follows:

[0089] K d (Data knowledge): refers to the information and features of the remote sensing data itself, such as time, region, land types, sensor types, etc. The sensor type determines the different characteristics of the sensor. For example, when the sensor type is synthetic aperture radar, the polarization mode and imaging mode attributes can be described; Adaptive objects include the category information of objects contained in the image. For example, the image data of Poyang Lake is more suitable for the study and extraction of water bodies; Extended semantic information uses semantic supplementary information that has not been abstracted and generalized to describe specific information in the image, which can provide an interface for the recognition and application of large models, such as the distribution and status of buildings, roads, vegetation, etc. For example: In a remote sensing image, there are canals in the middle of the rice fields for irrigation, and there are some exposed soil areas around them. The specific organizational structure, formal expression and description of data knowledge are as follows: Figure 2 .

[0090] K a (Algorithm knowledge): refers to the basic capability description of the algorithm, such as input, output, parameter settings, algorithm functions, and information and attributes that adapt to execution conditions. The algorithm input specifies the input data format and algorithm parameter settings, and the algorithm output describes the content and form of the output results. In the adaptability conditions, the adaptability to the data source, area, time phase, and type of land features need to be matched with data knowledge and business knowledge. The adaptability to data volume refers to the adaptability of the algorithm to small sample data, etc. The comprehensive accuracy includes efficiency indicators and efficiency labels (such as duration, accuracy, etc.) used to evaluate the model. The predecessor algorithm and successor algorithm describe the interfaces, types, or functions of the previous and subsequent algorithms that the algorithm adapts to in the process. The specific organizational structure, formal expression, and description of algorithmic knowledge are as follows: Figure 3 .

[0091] K b (Business knowledge): refers to the abstract and complete summary of the basic business demand information for monitoring behavior, which involves the description of monitoring elements and monitoring events in practice. b_factor represents the type of natural resource monitoring elements, and b_event represents remote sensing monitoring events. Among them, remote sensing monitoring events reflect various event forms in natural space, such as "non-agricultural" and "non-grain" cultivated land, including the occupation of cultivated land for lake digging, landscaping, afforestation, illegal construction and other event types. However, no matter what kind of remote sensing monitoring event b_event, it can be expressed as a specific natural resource element type b_factor. In the specific attributes of the monitoring event, the data requirements reflect the accuracy or type requirements of the monitoring event for the monitoring data. The minimum upper map area is the minimum threshold of the map change area that can define the event stipulated by the event or policy. The monitoring content is the supplementary content that needs to be paid more attention to according to the business or event needs, such as the change area, change indicators, violations, etc. in non-agricultural monitoring, and the process limitation is the specific link requirements for monitoring events under the support of policy regulations or expert knowledge. The specific organizational structure, formal expression and description of business knowledge are as follows: Figure 4 .

[0092] K r(Geospatial knowledge): refers to the abstract summary of the inherent characteristics of the monitored object in the real space and its location and morphology, such as the phenology, slope, and spatial distribution morphology of the monitored object. The spectral characteristics can describe the difference characteristics of the spectral curves of different land types in different regions. The phenological characteristics express the change characteristics of the monitored object due to seasonal changes in environmental factors (mainly meteorological conditions). Radar characteristics refer to the reflection characteristics of the objects obtained by radar remote sensing systems such as synthetic aperture radar (SAR), including echo intensity, phase information, scattering characteristics, etc. These characteristics are used to describe the characteristics of different types of objects in different geographical locations on SAR images. Semantic features can describe the understanding of geographic spatial relationships and spatial dynamics and hierarchies, such as urban layout, distribution patterns of road networks, and dynamic changes over time. Deep features are features that have been learned based on deep learning models, such as a pre-trained deep learning model for farmland identification in Changsha.

[0093] Since the information obtained from image data is not the entire information of the geographic environment complex, but only a part of the information in the geographic environment complex that can be expressed on remote sensing images with limited viewing angles, knowledge about geospatial information can be used as a supplement to the knowledge in the monitoring process to help with accurate monitoring. Figure 5 .

[0094] Embodiment 3

[0095] It should be noted that the formal representation of the user's business requirements after analysis includes:

[0096] Time, location, monitoring elements, business type, monitoring content, algorithm requirements and auxiliary feature requirements.

[0097] In this embodiment, user business needs are analyzed, business need analysis rules are established, and remote sensing monitoring process construction reasoning rules based on first-order logic representation are established to analyze the adaptability of each process construction, optimize the monitoring business process, and form a reasoning rule knowledge base based on business needs.

[0098] First, establish the definition of user demand analysis rules, and analyze the user's business needs (User_Needs) into a formal expression that can correspond to knowledge. The user's business needs can be analyzed into time (T), location (R), monitoring elements (F), business type (C), monitoring content (D), algorithm requirements (A), and auxiliary feature requirements (S). Among them, monitoring content, algorithm requirements, and auxiliary feature requirements are all optional, mainly describing the user's special monitoring needs. For example, when the user needs to implement traditional machine learning algorithms to achieve the purpose of rapid monitoring or method comparison, it can be analyzed as algorithm requirements (A). When the user hopes to use land phenological characteristics as an auxiliary to achieve accurate identification in a certain monitoring business, it can be analyzed as auxiliary feature requirements (S). Therefore, the user's business needs are the following seven tuples:

[0099] User_Needs={T,R,F,C,D,A,S};

[0100] It should be noted that the algorithms for calculating the matching degree include:

[0101] Define matching relationships for different matching items and assign corresponding weights;

[0102] Calculate the fuzzy matching score or the complete matching score of the item to be matched;

[0103] According to the fuzzy matching score or the complete matching score of the items to be matched, the sum of the matching degrees of all the items to be matched that have been assigned weights is obtained.

[0104] In order to facilitate subsequent reasoning, we first define the matching relationship between nodes. There are two types of matching relationships. The first type of node matching relationship is a fuzzy matching relationship based on fuzzy logic. The matching relationship between two knowledge nodes x and y conforms to the fuzzy function. It is mainly used for matching between text description knowledge. Assume that the fuzzy matching relationship between knowledge nodes FuzzMatch(x,y) is a "complete match" of the fuzzy set A, then its membership function u A (x,y) is:

[0105]

[0106] therefore:

[0107] FuzzMatch(x,y)=u A (x, y)

[0108] The second type of node matching relationship is a complete matching relationship: the relationship between two knowledge nodes x and y is

[0109] Match(x, y) = {0, 1}

[0110] Each type of relationship is attached with a weight s, which determines the priority of the matching relationship. The matching score between two nodes is expressed as Match(x, y)*s, FuzzMatch(x, y)*s. The weight is learned by the random forest algorithm. Users can also define the priority by customizing the weight.

[0111] The process construction score is mainly the sum of the fuzzy matching score and the complete matching score between two knowledge instances. n and algorithm knowledge instance y m Take the time matching score calculation as an example:

[0112]

[0113] Among them, x n Represents data n, y m Denotes algorithm m, S i , S j Indicates the priority weight of the matching relationship, i indicates the fuzzy matching relationship order, and j indicates the complete matching relationship order.

[0114] Based on the above-established natural resource monitoring knowledge base, user demand analysis rules, and node matching relationships, the natural resource monitoring process construction rules are established. The user's business needs are taken as the starting point to establish matching rules between business knowledge, data knowledge, algorithm knowledge, and geographic space knowledge, so as to realize the automatic construction of the process:

[0115] (1) Select the business node with the highest matching degree with the user's business demand from the business knowledge as the required business node, and establish a matching rule between the user's business demand U and the business node x:

[0116] Match(S(U),b time (x))∨Match(R(U),b region (x)

[0117] ∨Match(F(U), b factor (x))∨Match(C(U), b evevttype (x)

[0118] ∨Match(D(U),b description (x)

[0119] ∨FuzzyMatch(A(U), a description (x)

[0120] →CalulateScore1∧Output(x)

[0121] The business demand U is matched with the time, location, monitoring content and monitoring category of the business knowledge, the matching score Score1 is calculated and the business node corresponding to the maximum matching score is output as the demand business node.

[0122] (2) Secondly, the corresponding candidate data sets and candidate algorithm sets are obtained based on the required business nodes.

[0123] (2-1) In the data knowledge, select the data with a matching degree higher than the preset value of the required business node as the candidate data set. Assuming that the required business node corresponding to the maximum match is x, the required business node x and the data knowledge d y Create matching rules:

[0124] Match(b time (x), d time (d y ))∨Match(b region (x), d region (d y ))

[0125] ∨Match(b factor , d object (d y ))

[0126] ∨FuzzyMatch(b description (x), a decription (d y ))

[0127] →CalulateScore2∧Output(d y )

[0128] Calculate the matching score Score2 and output the corresponding data d y , obtain the top n corresponding data with the highest scores as candidate data sets.

[0129] (2-2) Select the algorithms with matching degree higher than the preset value in the algorithm knowledge as the candidate algorithm set, and compare the matching degree between the required business node x and the algorithm knowledge a z Create matching rules:

[0130] Match(b time (x), a time (a z ))∨Match(b region (x), a region (a z ))

[0131] ∨Match(b factor , a class (az ))

[0132] ∨FuzzyMatch(b processlimit (x), a decription (a z ))

[0133] ∨FuzzyMatch(b factor (k i ), b event (e i ))

[0134] →CalulateScore3∧Output(a z )

[0135] Calculate the matching score Score3 and output the corresponding algorithm a z , obtain the top n corresponding algorithms with the highest scores as the candidate algorithm set.

[0136] (3) Select the geospatial knowledge with the highest matching degree with the user's business needs and the required business node from the geospatial knowledge as the auxiliary decision-making geospatial knowledge, and compare the business needs U, the required business node x and the geospatial knowledge r i To match:

[0137] FuzzyMatch(S(U), r spectral (r i ))∨FuzzyMatch(s(U),r semantic (r i ))

[0138] ∨FuzzyMatch(S(U), r phenology (r i ))

[0139] ∨FuzzyMatch(S(U), r radar (r i ))∨FuzzyMatch(S(U),r dl (r i ))

[0140] ∨FuzzyMatch(S(U), r grade (r i ))

[0141] ∨FuzzyMatch(S(U), r textures (r i ))

[0142] ∨FuzzyMatch(S(U), r distribu tion (r i ))

[0143] ∨FuzzyMatch(b_description(x),r distribution (r i ))

[0144] →CalulateScore4∧Output(r i )

[0145] Calculate the matching score Score4 and output the geographic spatial knowledge r corresponding to the maximum matching score i As a decision-making aid, geospatial knowledge is used to assist verification in the monitoring process.

[0146] (4) According to the candidate data set, candidate algorithm set and decision-making auxiliary geographic spatial knowledge, the disjunctive matching rules among the three are established. Suppose the candidate data set d y , candidate algorithm set a z , decision-making support geospatial knowledge i :

[0147]

[0148] The matching score Score5 is calculated and the matching data, algorithm and geographic space knowledge corresponding to the maximum matching score are output.

[0149] Finally, the process required to process user business needs can be built based on the data, algorithms and geospatial knowledge corresponding to the maximum matching score.

[0150] It should be noted that the method also includes:

[0151] When multi-step processing is required, matching rules between algorithm nodes in algorithm knowledge are established;

[0152] Screen multiple algorithm nodes that match the front and back functions, interfaces, data knowledge, and geographic space knowledge in the algorithm knowledge;

[0153] By assisting decision making, geographic spatial knowledge is embedded into each link of the process as supplementary information for each link.

[0154] When the process is complex and requires multiple steps, it is also necessary to establish matching rules between operators. s and operator y t The disjunctive matching rules are as follows:

[0155]

[0156] The matching score Score6 is calculated and the operator model corresponding to the maximum matching score is output. In the above rules, geographic spatial knowledge Matching analysis with various operators can make full use of geographic spatial knowledge and embed it into each link of the process, serving as supplementary information for each link to help improve monitoring results.

[0157] Finally, the above matching rules are organized to form an inference rule knowledge base.

[0158] Embodiment 4

[0159] It should be noted that the method also includes:

[0160] Based on the new knowledge acquired and user feedback, the remote sensing monitoring knowledge base and inference rule knowledge base are updated and expanded.

[0161] Update and expand the remote sensing monitoring knowledge base and matching rules based on new knowledge and user feedback, including:

[0162] Construct a machine-readable model reasoning analysis layer to update and expand the remote sensing monitoring knowledge base and the reasoning rule knowledge base; including: according to the scene requirements, call the remote sensing monitoring knowledge base and the reasoning rule knowledge base for reasoning, analyze the adaptation algorithm model and data according to the specific needs of the user, and make behavioral decisions for each process step in the monitoring process;

[0163] Collect and process new data sources in real time to acquire new knowledge. When new knowledge is acquired, if the corresponding knowledge does not exist in the remote sensing monitoring knowledge base, the new knowledge will be stored, and expired or erroneous knowledge in the remote sensing monitoring knowledge base will be processed regularly.

[0164] Adjust the remote sensing monitoring knowledge base based on user feedback and historical interaction data.

[0165] In specific practice, this technical solution also has the ability of intelligent knowledge input and dynamic rule adjustment. By collecting new knowledge and analyzing user feedback, it automatically updates and expands the knowledge base and rules to ensure that the latest user needs are met.

[0166] Construct a machine-readable model reasoning and analysis layer and involve the dynamic update and maintenance of the knowledge base and rule base. This layer can call the remote sensing monitoring knowledge base and the reasoning rule knowledge base for reasoning according to the scene requirements, and derive the adaptation operator model and data according to the specific needs of the user, and make behavioral decisions for each process step in the monitoring process;

[0167] Specifically, according to the knowledge reasoning rule construction and management module, the existing data and operator instance knowledge in the database are matched with each other, and the matching relationship between data and operators, and between demand and operators is established. The highest matching score is obtained, and the actual optimal process is constructed, and the process chain is output.

[0168] The technical implementation of dynamic knowledge update relies on real-time data update, maintenance and expansion strategies. By collecting and processing new data sources in real time, such as new operators, new data, new geospatial knowledge, etc., new entities, relationships and attributes are extracted, such as data types, data formats, operator description information, monitoring tasks, etc. If these corresponding knowledge do not exist in the knowledge base, they are updated to the knowledge base, and new rules are constructed based on the above form and stored in the rule knowledge base. The knowledge base needs to be regularly maintained and optimized, and expired or erroneous information is processed to ensure the accuracy and timeliness of the knowledge base;

[0169] Furthermore, by continuously adjusting according to user feedback and historical interaction data, new monitoring process knowledge is obtained to improve the functionality of the knowledge base, trigger model fine-tuning, and pass the optimized decision-making plan to the knowledge reasoning and rule construction modules for dynamic rule expansion.

[0170] Embodiment 5

[0171] This embodiment provides a natural resource remote sensing monitoring process construction system based on multi-dimensional knowledge reasoning, referring to Figure 6 ,include:

[0172] Multi-dimensional knowledge structured expression and storage module, knowledge reasoning rule construction module, business process automatic construction and dynamic expansion module and user interaction and application evaluation module.

[0173] The multi-dimensional knowledge structured expression and storage module is mainly used to build a remote sensing monitoring knowledge base, that is, a natural resources monitoring knowledge base.

[0174] The knowledge inference rule building module is mainly used to build the inference rule knowledge base.

[0175] The automatic construction and dynamic expansion module of the business process is mainly used to realize the automatic update function of the knowledge base in the fourth embodiment.

[0176] The user interaction and application evaluation module is used to receive user business needs and display the final monitoring results, determine the relevant knowledge entities and levels, and receive user feedback. Based on the automatic construction and dynamic expansion module of the upper-level business process, this level uses the user interaction interface to obtain the user's business needs, convert the user's business needs into knowledge, and map the demand knowledge to the upper-level module. In the rule knowledge base, combined with the scenario requirements, it pushes the process construction decision for the user, evaluates the adaptability of different data, operators and businesses based on user feedback, and provides a better solution;

[0177] The technical implementation of the user interaction interface mainly focuses on user experience and interaction design. It is usually a front-end interface that allows users to enter queries and display results. Through front-end and back-end interaction technologies such as RESTful API and WebSocket, the user's query operation is simplified and intuitive feedback is provided. At the same time, user feedback and behavior data are collected and analyzed, and the interaction interface is continuously optimized to improve user experience and satisfaction.

[0178] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0179] It should be noted that, in the description of this application, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0180] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0181] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0182] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0183] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0184] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0185] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0186] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for constructing a natural resource remote sensing monitoring process based on multidimensional knowledge reasoning, characterized in that: include: The expert knowledge in the fields of remote sensing image processing and natural resource monitoring is classified into business knowledge, data knowledge, algorithm knowledge and geospatial knowledge; Receive user business requirements and parse them into formal representations corresponding to knowledge; Select the business node with the highest matching degree with the user's business needs from the business knowledge as the required business node; Selecting data with a matching degree higher than a preset value in the data knowledge as candidate data sets; Selecting algorithms whose matching degree with the demand service node is higher than a preset value from the algorithm knowledge as a candidate algorithm set; Selecting the geospatial knowledge with the highest matching degree with the user's business requirements and the business nodes of the requirements from the geospatial knowledge as auxiliary decision-making geospatial knowledge; Establishing disjunctive matching rules between the candidate data sets, candidate algorithm sets and auxiliary decision-making geospatial nodes, and outputting data, algorithms and geospatial knowledge corresponding to the maximum matching score; Build the processes required to handle user business needs based on the data, algorithms, and geospatial knowledge corresponding to the maximum matching score.

2. The method according to claim 1, characterized in that The method further comprises: Structural sorting of various types of knowledge; Formalize the structured knowledge to form a remote sensing monitoring knowledge base; The matching rules are sorted out to form an inference rule knowledge base.

3. The method according to claim 2, characterized in that Structural sorting of various types of knowledge, including: Grading of various types of knowledge; The hierarchical knowledge is analyzed from top to bottom to extract the organizational structure and attribute information of nodes at each level, as well as the connection relationship between nodes at each level.

4. The method according to claim 3, characterized in that Formal representation of structured knowledge, including: The extraction results of various types of knowledge are formally defined to obtain the unique identifiers corresponding to each type of knowledge as the basis for node matching.

5. The method according to claim 1, characterized in that The business knowledge includes: an abstract and complete summary of basic business demand information for monitoring behavior; The data knowledge includes: information and features of time, region, land types, and sensor types contained in the remote sensing data itself; The algorithm knowledge includes: basic capability description of the algorithm and adaptive execution conditions; The geographic spatial knowledge includes: the inherent characteristics of the monitored object in the real space, its location, the abstract summary of the spatial information such as the shape of the object, etc.

6. The method according to claim 1, characterized in that The formal representation of user business requirements after analysis includes: Time, location, monitoring elements, business type, monitoring content, algorithm requirements and auxiliary feature requirements.

7. The method according to claim 1, characterized in that The algorithms used to calculate the matching degree include: Define matching relationships for different matching items and assign corresponding weights; Calculate the fuzzy matching score or the complete matching score of the item to be matched; According to the fuzzy matching score or the complete matching score of the items to be matched, the sum of the matching degrees of all the items to be matched that have been assigned weights is obtained.

8. The method according to claim 1, characterized in that: The method further comprises: When multi-step processing is required, matching rules between algorithm nodes in algorithm knowledge are established; Screen multiple algorithm nodes that match the front and back functions, interfaces, data knowledge, and geographic space knowledge in the algorithm knowledge; By assisting decision making, geographic spatial knowledge is embedded into each link of the process as supplementary information for each link.

9. The method according to claim 1, characterized in that: The method further comprises: Based on the new knowledge acquired and user feedback, the remote sensing monitoring knowledge base and inference rule knowledge base are updated and expanded.

10. The method according to claim 9, characterized in that Update and expand the remote sensing monitoring knowledge base and matching rules based on new knowledge and user feedback, including: Construct a machine-readable model reasoning analysis layer to update and expand the remote sensing monitoring knowledge base and the reasoning rule knowledge base; including: according to the scene requirements, call the remote sensing monitoring knowledge base and the reasoning rule knowledge base for reasoning, analyze the adaptation operator model and data according to the specific needs of the user, and make behavioral decisions for each process step in the monitoring process; Collect and process new data sources in real time to acquire new knowledge. When new knowledge is acquired, if the corresponding knowledge does not exist in the remote sensing monitoring knowledge base, the new knowledge will be stored, and expired or erroneous knowledge in the remote sensing monitoring knowledge base will be processed regularly. Adjust the remote sensing monitoring knowledge base based on user feedback and historical interaction data.

Images