Species monitoring and identification method and system based on multimodal data fusion

Through multimodal data fusion method, data mining and analysis is carried out using climate, soil, vegetation, landform and water quality information, the problems of blindness and low efficiency of species monitoring in the existing technology are solved, and the accuracy and stability of species monitoring are improved. sex.

CN118364252BActive Publication Date: 2024-11-22CHINA NAT INST OF STANDARDIZATION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410519195.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-28
Publication Date
2024-11-22
Estimated Expiration
2044-04-28

AI Technical Summary

Technical Problem

Due to the lack of monitoring reference data for existing species monitoring and identification methods, users are more blind in species detection, resulting in low monitoring and identification efficiency and poor monitoring effect and stability.

Method used

Multi-constraint data mining is carried out by obtaining climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information and water source structure information of the area to be monitored, combining genetic analysis and food chain Analysis, generate species type prediction results.

Benefits of technology

The guidance and directionality of actual species detection have been improved, the technical goal of effectively reducing the blindness of species monitoring has been achieved, and the technical effect of improving the stability, quality and efficiency of species monitoring and identification has been achieved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118364252B_ABST
    Figure CN118364252B_ABST
Patent Text Reader

Abstract

The present application provides a species monitoring and identification method and system based on multimodal data fusion, which relates to the field of species monitoring and identification technology. The method includes: obtaining basic information of the area to be monitored, including climate type, soil composition, vegetation structure, landform type, water quality composition, water source structure and detected species type; performing multi-constraint data mining to obtain the first modal predicted species type; performing genetic analysis to obtain the second modal predicted species type; performing food chain analysis to obtain the third modal predicted species type; performing multimodal data fusion to generate species type prediction results. This application can solve the technical problem that due to the lack of monitoring reference data, users are more blind in species detection, resulting in low monitoring and identification efficiency, and poor monitoring effect and stability. It can improve the guidance and direction of actual species detection, and achieve the technical effect of improving the stability, quality and efficiency of species monitoring and identification.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of species monitoring and identification, and in particular to a species monitoring and identification method and system based on multimodal data fusion. Background Art

[0002] Species monitoring and identification is a key link in biodiversity conservation and research, and mainly relies on a variety of technologies and methods to achieve a comprehensive understanding and assessment of biological communities in a specific area. At present, traditional species monitoring and identification methods lack expected data, resulting in users lacking reference data for species monitoring, causing species monitoring and identification to be relatively blind, resulting in low monitoring efficiency and unclear monitoring effects.

[0003] In summary, the existing species monitoring and identification methods, due to the lack of monitoring reference data, lead to a large degree of blindness in species detection, resulting in low monitoring and identification efficiency, and poor monitoring effect and stability. Summary of the invention

[0004] The purpose of this application is to provide a species monitoring and identification method and system based on multimodal data fusion, so as to solve the technical problems that the existing species monitoring and identification methods, due to the lack of monitoring reference data, lead to users' blindness in species detection, resulting in low monitoring and identification efficiency, and poor monitoring effect and stability.

[0005] In view of the above problems, the present application provides a species monitoring and identification method and system based on multimodal data fusion.

[0006] In a first aspect, the present application provides a species monitoring and identification method based on multimodal data fusion, and the method is implemented by a species monitoring and identification system based on multimodal data fusion, wherein the method comprises: S10: obtaining basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types; S20: based on the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to perform multi-constraint data mining to obtain a first modal predicted species type; S30: perform genetic analysis based on the first modal predicted species type and the detected species type to obtain a second modal predicted species type; S40: perform food chain analysis based on the first modal predicted species type and the detected species type to obtain a third modal predicted species type; S50: perform multimodal data fusion on the first modal predicted species type, the second modal predicted species type and the third modal predicted species type to generate a species type prediction result and send it to the species monitoring user end.

[0007] In the second aspect, the present application also provides a species monitoring and identification system based on multimodal data fusion, which is used to execute a species monitoring and identification method based on multimodal data fusion as described in the first aspect, wherein the system includes: a basic information acquisition module for the area to be monitored, the basic information acquisition module for the area to be monitored is used to obtain basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types; a first modal prediction species type acquisition module, the first modal prediction species type acquisition module is used to predict the species type based on the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information. Multi-constraint data mining to obtain the first modality predicted species type; a second modality predicted species type acquisition module, the second modality predicted species type acquisition module is used to perform genetic analysis based on the first modality predicted species type and the detected species type to obtain the second modality predicted species type; a third modality predicted species type acquisition module, the third modality predicted species type acquisition module is used to perform food chain analysis based on the first modality predicted species type and the detected species type to obtain the third modality predicted species type; a species type prediction result generation module, the species type prediction result generation module is used to perform multimodal data fusion on the first modality predicted species type, the second modality predicted species type and the third modality predicted species type, generate a species type prediction result and send it to the species monitoring user end.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] By obtaining basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types; performing multi-constraint data mining according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information, to obtain a first modal predicted species type; performing genetic analysis according to the first modal predicted species type and the detected species type, to obtain a second modal predicted species type; performing food chain analysis according to the first modal predicted species type and the detected species type, to obtain a third modal predicted species type; performing multimodal data fusion on the first modal predicted species type, the second modal predicted species type and the third modal predicted species type, to generate a species type prediction result and send it to a species monitoring user end. That is to say, by taking the climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information and water source structure information of the area to be monitored as multiple constraints, networked data mining is carried out to obtain the first modal predicted species type; further genetic analysis is carried out based on the first modal predicted species type and the detected species type to obtain the second modal predicted species type; on the other hand, food chain analysis is carried out based on the first modal predicted species type and the detected species type to obtain the third modal predicted species type; finally, multimodal data fusion is performed on the first modal predicted species type, the second modal predicted species type and the third modal predicted species type to generate a species type prediction result, which can improve the guidance and direction of actual species detection, achieve the technical goal of effectively reducing the blindness of species monitoring, and achieve the technical effect of improving the stability, quality and efficiency of species monitoring and identification.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0012] Figure 1 A schematic diagram of a process of species monitoring and identification method based on multimodal data fusion in this application;

[0013] Figure 2 A schematic diagram of a process for obtaining a first modal predicted species type in a species monitoring and identification method based on multimodal data fusion in this application;

[0014] Figure 3 This is a structural diagram of a species monitoring and identification system based on multimodal data fusion in this application.

[0015] Description of reference numerals:

[0016] A module 11 for obtaining basic information of the area to be monitored, a module 12 for obtaining species types predicted by the first modality, a module 13 for obtaining species types predicted by the second modality, a module 14 for obtaining species types predicted by the third modality, and a module 15 for generating species type prediction results. DETAILED DESCRIPTION

[0017] This application provides a species monitoring and identification method and system based on multimodal data fusion, which solves the technical problems of existing species monitoring and identification methods, such as the lack of monitoring reference data, which leads to users' blindness in species detection, low monitoring and identification efficiency, and poor monitoring effect and stability. It can improve the guidance and directionality of actual species detection, achieve the technical goal of effectively reducing the blindness of species monitoring, and achieve the technical effect of improving the stability, quality and efficiency of species monitoring and identification.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] Embodiment 1:

[0020] Please refer to the attached Figure 1 The present application provides a species monitoring and identification method based on multimodal data fusion, wherein the method is applied to a species monitoring and identification system based on multimodal data fusion, and the method specifically comprises the following steps:

[0021] S10: Obtaining basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types;

[0022] Specifically, basic information of the area to be monitored is obtained, where the area to be monitored refers to the target area where species monitoring and identification will be carried out, which can be set based on actual conditions; the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types. The climate type information is an important feature that describes the natural conditions of a region. It is affected by a variety of factors such as latitude, atmospheric circulation, sea and land location, topography and ocean currents, such as temperate monsoon climate, temperate continental climate, etc.; the soil composition information includes the types and proportions of solid, liquid and gaseous substances, such as solid substances including minerals, organic matter, etc.; the vegetation structure information refers to the spatial and temporal configuration of plant species that constitute vegetation and their biological and ecological traits, such as the distribution and combination of understory plants, ground cover plants and shrubs; the landform type information refers to the morphological signs of the landform, which is affected by geological , climate, hydrology and other factors, such as plains, hills, plateaus, etc.; the water quality composition information refers to the mixed state of various substances in the water body, such as dissolved substances, colloidal substances and suspended matter; the water source structure information refers to the various sources of water resources and their relative proportions, including surface water, groundwater, etc., of which surface water includes rivers, lakes, reservoirs, ditches and ponds, and streams, etc., and groundwater includes deep well water, spring water and shallow well water, etc.; the detected species types refer to the species types that have been identified in the area to be monitored, including animals, plants, microorganisms, etc.; the above information can be set according to the actual situation of the area to be monitored. By obtaining the basic information of the area to be monitored, basic data support is provided for the next step of relevant information retrieval and species type prediction.

[0023] S20: performing multi-constraint data mining according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to obtain a first modal predicted species type;

[0024] Specifically, biological big data is a series of technologies and methods that enable the value contained in massive biological data to be mined and displayed, including data planning, data collection, analysis and mining and other technical means. Based on biological big data technology, the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information are used as multi-constraint retrieval conditions, and data mining is performed online to obtain the first modal prediction of species types. By setting multi-constraint retrieval conditions and performing data mining online based on biological big data technology, the adaptability of the mined data to the basic information of the monitored area can be improved, thereby improving the accuracy and rationality of data mining, and then improving the accuracy and authenticity of the first modal prediction of species types.

[0025] S30: performing genetic analysis based on the first modality predicted species type and the detected species type to obtain a second modality predicted species type;

[0026] Specifically, a genetic analysis is performed based on the first modality predicted species type and the detected species type, wherein the genetic analysis refers to calculating the genetic similarity of the first modality predicted species type and the detected species type based on the genomic characteristics of the organism, and screening out the predicted species type that meets the predetermined genetic similarity threshold and setting it as the second modality predicted species type. By judging the first modality predicted species type and the detected species type based on genetic analysis, the species type prediction result can be determined from the perspective of genetic similarity, providing a basis for the comprehensive prediction of species types.

[0027] S40: performing food chain analysis according to the first modality predicted species type and the detected species type to obtain a third modality predicted species type;

[0028] Specifically, a food chain analysis is performed based on the first modal predicted species type and the detected species type, wherein the food chain analysis refers to tracing the upstream and downstream of the food chain based on the food chain law of species, that is, the connection between different species formed through food relationships and the law of material circulation and energy flow, and the reliability of the species type is determined based on the food chain tracing results, and the species type prediction results that meet the expected reliability index are extracted and set as the third modal predicted species type. By judging the first modal predicted species type and the detected species type based on the food chain analysis, the species type prediction results can be judged from the perspective of food chain analysis, providing a basis for the comprehensive prediction of species types.

[0029] S50: Perform multimodal data fusion on the first modality predicted species type, the second modality predicted species type and the third modality predicted species type to generate a species type prediction result and send it to a species monitoring user terminal.

[0030] Specifically, first, weights are set for the first modal predicted species type, the second modal predicted species type, and the third modal predicted species type, respectively, where the weight value can be set according to the influence of the indicator on the overall species type prediction result, such as the influence analysis can be performed based on historical data, and the weight configuration can be performed using the coefficient of variation method, where the greater the influence of the indicator, the greater the corresponding weight value. Then, according to the weight value, weighted calculations are performed on the first modal predicted species type, the second modal predicted species type, and the third modal predicted species type to obtain the weight sum of each predicted species, and the weight sum is set as the credibility. Based on the preset credibility threshold, the credibility of the predicted species type is judged respectively, and the predicted species type with a credibility greater than the preset credibility threshold is set as the species type prediction result, where the preset credibility threshold is the standard credibility, and technicians in this field can set it based on the actual situation to obtain the species type prediction result, and send the species type prediction result to the species monitoring user end.

[0031] By obtaining the first modality based on data mining to predict species types, the second modality based on genetic analysis to predict species types, and the third modality based on food chain analysis to predict species types, it is possible to predict species types from multiple aspects. Finally, the species type prediction results are obtained by fusing the first modality predicted species type, the second modality predicted species type, and the third modality predicted species type, which can improve the scientificity, accuracy, and reliability of species type prediction.

[0032] The species monitoring and identification method based on multimodal data fusion is applied to a species monitoring and identification system based on multimodal data fusion, which can solve the technical problems of existing species monitoring and identification methods, such as lack of monitoring reference data, which leads to users being relatively blind in performing species detection, resulting in low monitoring and identification efficiency, and poor monitoring effect and stability. First, basic information of the area to be monitored is obtained, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types; then, multi-constraint data mining is performed according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to obtain a first modal predicted species type; then, genetic analysis is performed according to the first modal predicted species type and the detected species type to obtain a second modal predicted species type; next, food chain analysis is performed according to the first modal predicted species type and the detected species type to obtain a third modal predicted species type; finally, multi-modal data fusion is performed on the first modal predicted species type, the second modal predicted species type and the third modal predicted species type to generate a species type prediction result and send it to the species monitoring user end. By taking the climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information and water source structure information of the area to be monitored as multiple constraints, networked data mining is carried out to obtain the first modal predicted species type; further genetic analysis is carried out based on the first modal predicted species type and the detected species type to obtain the second modal predicted species type; on the other hand, food chain analysis is carried out based on the first modal predicted species type and the detected species type to obtain the third modal predicted species type; finally, multimodal data fusion is performed on the first modal predicted species type, the second modal predicted species type and the third modal predicted species type to generate a species type prediction result, which can improve the guidance and direction of actual species detection, achieve the technical goal of effectively reducing the blindness of species monitoring, and achieve the technical effect of improving the stability, quality and efficiency of species monitoring and identification.

[0033] Further, as attached Figure 2 As shown, step S20 of the present application includes:

[0034] Constructing a 6-dimensional constraint condition according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information, wherein the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information are in a logical AND relationship in the 6-dimensional constraint condition;

[0035] According to the six-dimensional constraint conditions, a screening distance threshold is configured;

[0036] Specifically, according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information, a 6-dimensional constraint condition is constructed, in which the constraint condition is to satisfy the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information, and the water source structure information. Then, a screening distance threshold is configured according to the 6-dimensional constraint condition, and the screening distance threshold is used to evaluate and judge the network-mined data.

[0037] Furthermore, the present application also includes the following steps:

[0038] Construct a distance constraint evaluation function:

[0039] ,

[0040] in, D Characterize the distance constraints of network retrieval data and 6-dimensional constraints, The j-th attribute feature value of the i-th dimension of the 6-dimensional feature representing the online retrieval data, The j-th attribute eigenvalue of the i-th dimension representing the 6-dimensional constraint condition, N Represents the number of characteristic attributes of the i-th dimension, Characterize the preset weight of the i-th dimension;

[0041] The screening distance threshold is configured according to the distance constraint evaluation function.

[0042] Specifically, the method for configuring the screening distance threshold according to the 6-dimensional constraint condition is as follows: first, construct a distance constraint evaluation function, and the distance constraint evaluation function expression is:

[0043] ,

[0044] In the distance constraint evaluation function, D A distance constraint representing the online search data and the six-dimensional constraint condition, wherein the smaller the distance constraint is, the higher the correlation between the representing data is; The j-th attribute feature value of the i-th dimension of the 6-dimensional feature representing the online search data, wherein the i-th dimension is any one of the multiple dimensions in the online search data, and the j-th attribute is any one of the multiple attribute features in the i-th dimension; The jth attribute eigenvalue of the i-th dimension represents the 6-dimensional constraint; N represents the number of characteristic attributes of the i-th dimension, N The value of can be set based on actual conditions; represents the preset weight of the i-th dimension, where The value of can be set based on the importance of the i-th dimension to multiple dimensions, where the greater the importance, The larger the value of .

[0045] Configure the screening distance threshold, which is the standard data distance and can be set based on the distance constraint evaluation function and actual needs. By constructing a distance constraint evaluation function to perform distance evaluation on the online search data, it is possible to quantitatively evaluate the relevance of the online search data with the six-dimensional constraint conditions, while improving the accuracy and efficiency of the quantitative evaluation.

[0046] According to the screening distance threshold and the six-dimensional constraint condition, searching online for the first species type detection record;

[0047] Strict frequent sequence mining is performed on the first species type detection records to obtain the first modal predicted species type.

[0048] Specifically, based on biological big data technology, data retrieval is performed online to obtain multiple species type detection data, and then the distance constraints are calculated for the multiple species type detection data according to the distance constraint evaluation function, and the species type detection data with distance constraints less than the screening distance threshold are extracted and set as the first species type detection data, and the first species type detection record is obtained according to the multiple first species type detection data. Finally, the first species type detection record is strictly frequently mined to obtain the first modal predicted species type.

[0049] Furthermore, the present application also includes the following steps:

[0050] According to the first species type detection record, extract the first detection species type set, the second detection species type set until the Mth detection species type set, where M represents the total number of records, M≥1;

[0051] Perform association frequency statistics according to the first detected species type set, the second detected species type set, and up to the Mth detected species type set to obtain a detected species type list and a species association frequency list;

[0052] Traversing the detected species type list, performing frequency proportion analysis based on the species association frequency list, and obtaining an association frequency proportion list, wherein the association frequency proportion list corresponds to the detected species type list in one-to-one correspondence;

[0053] According to the association frequency ratio list, the first modality predicted species type whose association frequency ratio is greater than or equal to a first association frequency ratio threshold is sorted from the detected species type list.

[0054] Specifically, the method for performing strict frequent sequence mining on the first species type detection record to obtain the first modal predicted species type is as follows: first, according to the first species type detection record, the first detection species type set, the second detection species type set, and the Mth detection species type set are obtained in sequence, where M represents the total number of records and M is an integer greater than 1. Then, according to the first detection species type set, the second detection species type set, and the Mth detection species type set, the same species type is statistically associated with the frequency, where the association frequency refers to the total number of times the species type appears in the first to Mth detection species type sets, a detection species type list is constructed based on different species types, and a species association frequency list is constructed based on the association frequency statistics, where the species type and the species association frequency correspond one to one.

[0055] The detected species type list is traversed, and the species types are subjected to frequency ratio analysis based on the species association frequency list, wherein the frequency ratio refers to the ratio of the association frequency of a single species type to the association frequency of all species types, and an association frequency ratio list is obtained, wherein the association frequency ratio in the association frequency ratio list corresponds to the species types in the detected species type list one by one. Then, based on the association frequency ratio list, species types whose association frequency ratio is greater than or equal to the first association frequency ratio threshold are screened from the detected species type list, and are set as the first modal predicted species types.

[0056] Furthermore, the present application also includes the following steps:

[0057] According to the association frequency ratio list, sorting the initial predicted species types whose association frequency ratio is greater than or equal to a first association frequency ratio threshold from the detected species type list;

[0058] Traversing the initial predicted species types to perform associated species matching to obtain associated species types;

[0059] Specifically, first, the first correlation frequency ratio threshold value can be set by a person skilled in the art based on the number of species types and the actual situation of the correlation frequency ratio; then, the correlation frequency ratio list is screened according to the first correlation frequency ratio threshold value, and the species types in the detected species type list whose correlation frequency ratio is greater than or equal to the first correlation frequency ratio threshold value are extracted and set as the initial predicted species types. Then, the initial predicted species type is matched with associated species, where the associated species matching refers to the species type that often appears with the initial predicted species type, that is, the species type with a high degree of association, to obtain the associated species type.

[0060] Furthermore, the present application also includes the following steps:

[0061] According to the initial predicted species type, extract the first initial predicted species type, and search the second species type detection record online, wherein the detected species type of any second species type detection record includes the first initial predicted species type;

[0062] According to the detection record of the second species type, the detection frequency ratios of several related species types are counted;

[0063] According to the detection frequency proportions of the plurality of associated species types, the associated species types that are greater than or equal to a second associated frequency proportion threshold are extracted.

[0064] Specifically, the method for obtaining the associated species type is as follows: first, a first initial predicted species type is randomly extracted from the initial predicted species types, and the first initial predicted species type is any one of the initial predicted species types; then, with the first initial predicted species type as a constraint, the second species type detection record is retrieved online based on biological big data technology, wherein the detected species type of any one of the second species type detection records includes the first initial predicted species type.

[0065] Then, based on the second species type detection record, the detection frequencies of several related species types are counted respectively, wherein the detection frequency refers to the frequency of occurrence of the related species type in the second species type detection record, and the detection frequency ratio of the related species type is calculated based on the detection frequency, and the detection frequency ratio of the related species type is the ratio of the detection frequency of a single related species type to the sum of the detection frequencies of all species types, and the detection frequency ratios of several related species types are obtained; a second associated frequency ratio threshold is configured, and the second associated frequency ratio threshold can be set based on actual conditions, and the detection frequency ratios of the several related species types are screened according to the second associated frequency ratio threshold, and the associated species types greater than or equal to the second associated frequency ratio threshold are extracted and set as the associated species types.

[0066] The initial predicted species type and the associated species type are combined to generate the first modal predicted species type.

[0067] Specifically, the initial predicted species type and the associated species type are finally merged. If the same species type exists in the two, only one is retained to obtain the first modal predicted species type. By setting multiple constraints for data retrieval, the species type with high frequency is set as the initial predicted species type; and based on the initial predicted species type, the associated species type with high correlation is retrieved again; finally, the initial predicted species type and the associated species type are merged to obtain the first modal predicted species type, which can improve the comprehensiveness and accuracy of the first modal predicted species type setting, and at the same time improve the comprehensiveness and rationality of subsequent species type predictions.

[0068] Further, genetic analysis is performed based on the first modality predicted species type and the detected species type to obtain the second modality predicted species type. Step S30 of the present application includes:

[0069] Configure species type gene database;

[0070] Traversing the first modality predicted species type and the detected species type, performing genome average genetic similarity calculation on the species type gene database, and obtaining several groups of genome average genetic similarity calculation results;

[0071] Traversing the plurality of groups of genome average genetic similarity calculation results, and sorting genome-similar species types whose genome average genetic similarity calculation results are greater than or equal to a genome average genetic similarity threshold;

[0072] The genome-similar species types are added to the second modality predicted species types.

[0073] Specifically, first, a species type gene database is configured, where the species type gene database is a database specifically used to store and manage genetic information of specific species. This database usually contains a large amount of multi-dimensional data such as the gene sequence, structure, variation and function of the species. The data in the species type gene database in this application are all genomic characteristics of existing known species that have been shared and can be obtained through data collection.

[0074] Then, based on the species type gene database, the first genome feature set of the first modal predicted species type and the detected genome feature set of the detected species type are obtained respectively; and the genome features in the first genome feature set and the detected genome feature set are respectively calculated for average genome genetic similarity, wherein average genome genetic similarity (Average nucleotide identity, ANI) refers to the similarity of homologous genes between the genomes of two species, and their genetic similarity is evaluated by comparing the gene sequences between the two species. The similarity can be calculated by the existing MUMmer algorithm or the BLASTn method to obtain several groups of genome average genetic similarity calculation results.

[0075] A genome average genetic similarity threshold is configured, and the genome average genetic similarity threshold can be set by a person skilled in the art based on actual conditions; the calculation results of the several groups of genome average genetic similarities are judged according to the genome average genetic similarity threshold, and genome similar species types whose genome average genetic similarity calculation results are greater than or equal to the genome average genetic similarity threshold are extracted, and the genome similar species types are added to the second modality predicted species type to obtain the second modality predicted species type.

[0076] By judging the species type predicted by the first modality and the detected species type based on genetic analysis, the species type prediction results can be judged from the perspective of genetic similarity, and the species type predicted by the second modality can be obtained, providing a basis for the subsequent comprehensive prediction of species types.

[0077] Further, a food chain analysis is performed according to the first modal predicted species type and the detected species type to obtain a third modal predicted species type. Step S40 of the present application further includes:

[0078] Traversing the first modal predicted species type and the detected species type to perform upstream and downstream food chain tracing to obtain the first food chain tracing species type;

[0079] Traverse the upstream tracing species types of the food chain of the first food chain tracing species type and perform downstream tracing to obtain the second food chain tracing species type;

[0080] Traverse the food chain downstream tracing species types of the first food chain tracing species type to perform upstream tracing to obtain the third food chain tracing species type;

[0081] Obtaining a first to-be-analyzed traceability species type of the first food chain traceability species type;

[0082] When the first traceable species type to be analyzed belongs to the upstream traceable species type in the food chain, extracting the second traceable species type to be analyzed from the second food chain traceable species type;

[0083] Calculate the proportion of the first modal predicted species type and the detected species type in the second traceable species type to be analyzed, and set it as the species type trigger probability;

[0084] When the first traceability species type to be analyzed belongs to the traceability species type in the downstream of the food chain, extracting the fourth traceability species type to be analyzed from the third food chain traceability species type;

[0085] Calculate the proportion of the first modal predicted species type and the detected species type in the fourth traceable species type to be analyzed, and set it as the species type trigger probability;

[0086] When the species type trigger probability is greater than or equal to the trigger probability threshold, the first traceable species type to be analyzed is added to the third modality predicted species type.

[0087] Specifically, first, the first modal predicted species type and the detected species type are traced in the upstream and downstream food chains, wherein the upstream and downstream food chain tracing refers to determining the relationship between the species type in the food chain and the adjacent upstream and downstream species. The food chain is a chain relationship in which various organisms in the ecosystem use other organisms as food to maintain their own life activities. The upstream food chain refers to the species that provide food sources for it, and the downstream food chain refers to other organisms that rely on the species as a food source. Taking rabbits as an example, the upstream food chain includes various plants, such as grass and vegetables, and the downstream food chain includes foxes, owls and other carnivores that prey on rabbits; the first food chain tracing species type is obtained, and the first food chain tracing species type includes the upstream tracing species type of the food chain and the downstream tracing species type of the food chain. Then, based on the food chain, the upstream tracing species type of the first food chain tracing species type is traced downstream of the food chain to obtain the second food chain tracing species type; the downstream tracing species type of the first food chain tracing species type is traced upstream to obtain the third food chain tracing species type.

[0088] The first traceability species type to be analyzed of the first food chain traceability species type is obtained. The first traceability species type to be analyzed refers to a species type that cannot be determined whether it is in the upstream and downstream food chains. Since the food chain of species may be diverse, when there are many species types in the upstream and downstream food chains corresponding to the species, individual species types cannot be accurately set as food chain-associated species.

[0089] When the first traceability species type to be analyzed belongs to the upstream traceability species type of the food chain, the second traceability species type to be analyzed is extracted from the second food chain traceability species type; then the proportion of the first modal predicted species type and the detected species type in the second traceability species type to be analyzed is calculated, for example: for the species that has been monitored is an upstream species type, that is, the downstream species of the species is a detected species type, then it is necessary to perform downstream species analysis on the species to determine the total number of downstream species of the species; and calculate the ratio of the total number of types to the number of downstream species suitable for the species in the detected species type, and set the ratio as the species type trigger probability.

[0090] When the first traceability species type to be analyzed belongs to the downstream traceability species type of the food chain, the fourth traceability species type to be analyzed is extracted from the third food chain traceability species type; then the proportion of the first modal predicted species type and the detected species type in the fourth traceability species type to be analyzed is calculated, for example: for the species that has been monitored is a downstream species type, that is, the upstream species of the species is a detected species type, then it is necessary to perform upstream species analysis on the species to determine the total number of upstream species of the species; and calculate the ratio of the total number of types to the number of upstream species suitable for the species in the detected species type, and set the ratio as the species type trigger probability.

[0091] Configure the trigger probability threshold, which can be set based on the actual situation. For example, the trigger probability threshold can be set to 0.8; the species type trigger probability is judged based on the trigger probability threshold. When the species type trigger probability is greater than or equal to the trigger probability threshold, the first traceable species type to be analyzed is added to the third modal predicted species type. In addition, if the traceable species type in the first food chain traceable species type does not belong to the traceable species type to be analyzed, that is, the species types in the upstream and downstream food chains of the species type all belong to the detected species type, then the species type is directly added to the third modal predicted species type.

[0092] By judging the first modal predicted species type and the detected species type based on food chain analysis, the species type prediction results can be judged from the perspective of food chain analysis, providing a basis for the subsequent comprehensive prediction of species types.

[0093] In summary, the species monitoring and identification method based on multimodal data fusion provided in this application has the following technical effects:

[0094] 1. By taking the climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information and water source structure information of the area to be monitored as multiple constraints, networked data mining is carried out to obtain the first modal predicted species type; further genetic analysis is carried out based on the first modal predicted species type and the detected species type to obtain the second modal predicted species type; on the other hand, food chain analysis is carried out based on the first modal predicted species type and the detected species type to obtain the third modal predicted species type; finally, multimodal data fusion is performed on the first modal predicted species type, the second modal predicted species type and the third modal predicted species type to generate species type prediction results, which can improve the guidance and direction of actual species detection, achieve the technical goal of effectively reducing the blindness of species monitoring, and achieve the technical effect of improving the stability, quality and efficiency of species monitoring and identification.

[0095] 2. By obtaining the first modality based on data mining to predict species types, the second modality based on genetic analysis to predict species types, and the third modality based on food chain analysis to predict species types, it is possible to predict species types from multiple aspects. Finally, the species type prediction results are obtained by integrating the first modality predicted species type, the second modality predicted species type, and the third modality predicted species type, which can improve the scientificity, accuracy, and reliability of species type prediction.

[0096] Embodiment 2:

[0097] Based on the same inventive concept as the species monitoring and identification method based on multimodal data fusion in the aforementioned embodiment, the present application also provides a species monitoring and identification system based on multimodal data fusion, please refer to the attached Figure 3 , the system comprising:

[0098] A module 11 for obtaining basic information of the area to be monitored, wherein the module 11 is used to obtain basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types;

[0099] A first modal prediction species type acquisition module 12, the first modal prediction species type acquisition module 12 is used to perform multi-constraint data mining according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to obtain a first modal prediction species type;

[0100] A second modality predicted species type obtaining module 13, wherein the second modality predicted species type obtaining module 13 is used to perform genetic analysis according to the first modality predicted species type and the detected species type to obtain a second modality predicted species type;

[0101] A third modality predicted species type obtaining module 14, wherein the third modality predicted species type obtaining module 14 is used to perform food chain analysis according to the first modality predicted species type and the detected species type to obtain a third modality predicted species type;

[0102] The species type prediction result generation module 15 is used to perform multimodal data fusion on the first modality predicted species type, the second modality predicted species type and the third modality predicted species type, generate a species type prediction result and send it to the species monitoring user end.

[0103] Furthermore, the first modality prediction species type obtaining module 12 in the system is also used for:

[0104] Constructing a 6-dimensional constraint condition according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information, wherein the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information are in a logical AND relationship in the 6-dimensional constraint condition;

[0105] According to the six-dimensional constraint conditions, a screening distance threshold is configured;

[0106] According to the screening distance threshold and the six-dimensional constraint condition, searching online for the first species type detection record;

[0107] Strict frequent sequence mining is performed on the first species type detection records to obtain the first modal predicted species type.

[0108] Furthermore, the first modality prediction species type obtaining module 12 in the system is also used for:

[0109] Construct a distance constraint evaluation function:

[0110] ,

[0111] in, D Characterize the distance constraints of network retrieval data and 6-dimensional constraints, The j-th attribute feature value of the i-th dimension of the 6-dimensional feature representing the online retrieval data, The j-th attribute eigenvalue of the i-th dimension representing the 6-dimensional constraint condition, NRepresents the number of characteristic attributes of the i-th dimension, Characterize the preset weight of the i-th dimension;

[0112] The screening distance threshold is configured according to the distance constraint evaluation function.

[0113] Furthermore, the first modality prediction species type obtaining module 12 in the system is also used for:

[0114] According to the first species type detection record, extract the first detection species type set, the second detection species type set until the Mth detection species type set, where M represents the total number of records, M≥1;

[0115] Perform association frequency statistics according to the first detected species type set, the second detected species type set, and up to the Mth detected species type set to obtain a detected species type list and a species association frequency list;

[0116] Traversing the detected species type list, performing frequency proportion analysis based on the species association frequency list, and obtaining an association frequency proportion list, wherein the association frequency proportion list corresponds to the detected species type list in one-to-one correspondence;

[0117] According to the association frequency ratio list, the first modality predicted species type whose association frequency ratio is greater than or equal to a first association frequency ratio threshold is sorted from the detected species type list.

[0118] Furthermore, the first modality prediction species type obtaining module 12 in the system is also used for:

[0119] According to the association frequency ratio list, sorting the initial predicted species types whose association frequency ratio is greater than or equal to a first association frequency ratio threshold from the detected species type list;

[0120] Traversing the initial predicted species types to perform associated species matching to obtain associated species types;

[0121] The initial predicted species type and the associated species type are combined to generate the first modal predicted species type.

[0122] Furthermore, the first modality prediction species type obtaining module 12 in the system is also used for:

[0123] According to the initial predicted species type, extract the first initial predicted species type, and search the second species type detection record online, wherein the detected species type of any second species type detection record includes the first initial predicted species type;

[0124] According to the detection record of the second species type, the detection frequency ratios of several related species types are counted;

[0125] According to the detection frequency proportions of the plurality of associated species types, the associated species types that are greater than or equal to a second associated frequency proportion threshold are extracted.

[0126] Furthermore, the second modality prediction species type obtaining module 13 in the system is also used for:

[0127] Configure species type gene database;

[0128] Traversing the first modality predicted species type and the detected species type, performing genome average genetic similarity calculation on the species type gene database, and obtaining several groups of genome average genetic similarity calculation results;

[0129] Traversing the plurality of groups of genome average genetic similarity calculation results, and sorting genome-similar species types whose genome average genetic similarity calculation results are greater than or equal to a genome average genetic similarity threshold;

[0130] The genome-similar species types are added to the second modality predicted species types.

[0131] Furthermore, the third modality prediction species type acquisition module 14 in the system is also used for:

[0132] Traversing the first modal predicted species type and the detected species type to perform upstream and downstream food chain tracing to obtain the first food chain tracing species type;

[0133] Traverse the upstream tracing species types of the food chain of the first food chain tracing species type and perform downstream tracing to obtain the second food chain tracing species type;

[0134] Traverse the food chain downstream tracing species types of the first food chain tracing species type to perform upstream tracing to obtain the third food chain tracing species type;

[0135] Obtaining a first to-be-analyzed traceability species type of the first food chain traceability species type;

[0136] When the first traceable species type to be analyzed belongs to the upstream traceable species type in the food chain, extracting the second traceable species type to be analyzed from the second food chain traceable species type;

[0137] Calculate the proportion of the first modal predicted species type and the detected species type in the second traceable species type to be analyzed, and set it as the species type trigger probability;

[0138] When the first traceability species type to be analyzed belongs to the traceability species type in the downstream of the food chain, extracting the fourth traceability species type to be analyzed from the third food chain traceability species type;

[0139] Calculate the proportion of the first modal predicted species type and the detected species type in the fourth traceable species type to be analyzed, and set it as the species type trigger probability;

[0140] When the species type trigger probability is greater than or equal to the trigger probability threshold, the first traceable species type to be analyzed is added to the third modality predicted species type.

[0141] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The species monitoring and identification method based on multimodal data fusion and the specific examples in the aforementioned embodiment one are also applicable to the species monitoring and identification system based on multimodal data fusion in this embodiment. Through the aforementioned detailed description of the species monitoring and identification method based on multimodal data fusion, those skilled in the art can clearly understand the species monitoring and identification system based on multimodal data fusion in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method section description.

[0142] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0143] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A species monitoring and identification method based on multimodal data fusion, characterized in that: include: S10: Obtaining basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types; S20: performing multi-constraint data mining according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to obtain a first modal predicted species type; S30: performing genetic analysis based on the first modality predicted species type and the detected species type to obtain a second modality predicted species type; S40: performing food chain analysis according to the first modality predicted species type and the detected species type to obtain a third modality predicted species type; S50: performing multimodal data fusion on the species type predicted by the first modality, the species type predicted by the second modality, and the species type predicted by the third modality to generate a species type prediction result and send it to a species monitoring user terminal; Wherein, performing multi-constraint data mining according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to obtain the first modal prediction species type includes: Constructing a 6-dimensional constraint condition according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information, wherein the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information are in a logical AND relationship in the 6-dimensional constraint condition; According to the six-dimensional constraint conditions, a screening distance threshold is configured; According to the screening distance threshold and the six-dimensional constraint condition, searching online for the first species type detection record; Strict frequent sequence mining is performed on the first species type detection records to obtain the first modal predicted species type.

2. The method according to claim 1, characterized in that According to the six-dimensional constraints, the screening distance threshold is configured, including: Construct a distance constraint evaluation function: , in, D Characterize the distance constraints of network retrieval data and 6-dimensional constraints, The j-th attribute feature value of the i-th dimension of the 6-dimensional feature representing the online retrieval data, The j-th attribute eigenvalue of the i-th dimension represents the 6-dimensional constraint condition, N represents the number of characteristic attributes of the i-th dimension, Characterize the preset weight of the i-th dimension; The screening distance threshold is configured according to the distance constraint evaluation function.

3. The method according to claim 1, characterized in that Perform strict frequent sequence mining on the first species type detection records to obtain the first modal predicted species type, including: According to the first species type detection record, extract the first detection species type set, the second detection species type set until the Mth detection species type set, where M represents the total number of records, M≥1; Perform association frequency statistics according to the first detected species type set, the second detected species type set, and up to the Mth detected species type set to obtain a detected species type list and a species association frequency list; Traversing the detected species type list, performing frequency proportion analysis based on the species association frequency list, and obtaining an association frequency proportion list, wherein the association frequency proportion list corresponds to the detected species type list in one-to-one correspondence; According to the association frequency ratio list, the first modality predicted species type whose association frequency ratio is greater than or equal to a first association frequency ratio threshold is sorted from the detected species type list.

4. The method according to claim 3, characterized in that According to the association frequency ratio list, selecting the first modality predicted species type whose association frequency ratio is greater than or equal to a first association frequency ratio threshold from the detected species type list, further comprising: According to the association frequency ratio list, sorting the initial predicted species types whose association frequency ratio is greater than or equal to a first association frequency ratio threshold from the detected species type list; Traversing the initial predicted species types to perform associated species matching to obtain associated species types; The initial predicted species type and the associated species type are combined to generate the first modal predicted species type.

5. The method according to claim 4, characterized in that Traversing the initial predicted species types to perform associated species matching, obtaining associated species types, including: According to the initial predicted species type, extract the first initial predicted species type, and search the second species type detection record online, wherein the detected species type of any second species type detection record includes the first initial predicted species type; According to the detection record of the second species type, the detection frequency ratios of several related species types are counted; According to the detection frequency proportions of the plurality of associated species types, the associated species types that are greater than or equal to a second associated frequency proportion threshold are extracted.

6. The method according to claim 1, characterized in that Performing genetic analysis based on the first modality predicted species type and the detected species type to obtain a second modality predicted species type includes: Configure species type gene database; Traversing the first modality predicted species type and the detected species type, performing genome average genetic similarity calculation on the species type gene database, and obtaining several groups of genome average genetic similarity calculation results; Traversing the plurality of groups of genome average genetic similarity calculation results, and sorting genome-similar species types whose genome average genetic similarity calculation results are greater than or equal to a genome average genetic similarity threshold; The genome-similar species types are added to the second modality predicted species types.

7. The method according to claim 1, characterized in that Performing a food chain analysis based on the first modality predicted species type and the detected species type to obtain a third modality predicted species type includes: Traversing the first modal predicted species type and the detected species type to perform upstream and downstream food chain tracing to obtain the first food chain tracing species type; Traverse the upstream tracing species types of the food chain of the first food chain tracing species type and perform downstream tracing to obtain the second food chain tracing species type; Traverse the food chain downstream tracing species types of the first food chain tracing species type to perform upstream tracing to obtain the third food chain tracing species type; Obtaining a first to-be-analyzed traceability species type of the first food chain traceability species type; When the first traceable species type to be analyzed belongs to the upstream traceable species type in the food chain, extracting the second traceable species type to be analyzed from the second food chain traceable species type; Calculate the proportion of the first modal predicted species type and the detected species type in the second traceable species type to be analyzed, and set it as the species type trigger probability; When the first traceability species type to be analyzed belongs to the traceability species type in the downstream of the food chain, extracting the fourth traceability species type to be analyzed from the third food chain traceability species type; Calculate the proportion of the first modal predicted species type and the detected species type in the fourth traceable species type to be analyzed, and set it as the species type trigger probability; When the species type trigger probability is greater than or equal to the trigger probability threshold, the first traceable species type to be analyzed is added to the third modality predicted species type.

8. The species monitoring and identification system based on multimodal data fusion is characterized by: The steps for implementing the method according to any one of claims 1 to 7 include: A module for obtaining basic information of the area to be monitored, wherein the module is used to obtain basic information of the area to be monitored, wherein the basic information of the area to be monitored includes climate type information, soil composition information, vegetation structure information, landform type information, water quality composition information, water source structure information and detected species types; A first modal prediction species type acquisition module, the first modal prediction species type acquisition module is used to perform multi-constraint data mining according to the climate type information, the soil composition information, the vegetation structure information, the landform type information, the water quality composition information and the water source structure information to obtain a first modal prediction species type; A second modality predicted species type acquisition module, wherein the second modality predicted species type acquisition module is used to perform genetic analysis according to the first modality predicted species type and the detected species type to obtain a second modality predicted species type; A third modality predicted species type acquisition module, the third modality predicted species type acquisition module is used to perform food chain analysis according to the first modality predicted species type and the detected species type to obtain a third modality predicted species type; The species type prediction result generation module is used to perform multimodal data fusion on the first modality predicted species type, the second modality predicted species type and the third modality predicted species type, generate a species type prediction result and send it to the species monitoring user end.