Benthic organism sample identification analysis method and system based on data mining

Abstract

The application discloses a benthic organism sample identification analysis method and system based on data mining. The method first acquires standard image data of benthic organisms at different growth stages, extracts appearance features and constructs a feature database; an identification model is established based on the database, three-dimensional structure image data of target samples and communities are collected by using underwater robots, initial identification is carried out; if the result is not unique, interspecific relationships are mined, a community relationship network is constructed by combining a graph algorithm, the reliability of the identification result is calculated, secondary identification is carried out, and the final result is obtained. The application improves the identification accuracy and reliability, and is suitable for marine ecological monitoring and biodiversity research.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of biological identification, and in particular to a benthic organism sample identification analysis method and system based on data mining. BACKGROUND

[0002] Benthic organisms are an important component of marine, lake and river ecosystems, and have important significance for ecological environment monitoring, biodiversity protection and water health assessment. Traditional benthic organism identification methods mainly rely on manual classification and microscopic observation, which not only consumes time and effort, but is also easily affected by subjective factors, resulting in large errors in the identification results. In addition, the manual identification method is difficult to adapt to the needs of large-scale sample analysis, limiting the efficiency and accuracy of benthic organism research.

[0003] In recent years, the development of computer vision and data mining technology has provided a new solution for biological identification. Image recognition-based methods can automatically extract biological sample features, improving identification efficiency. However, existing image analysis-based benthic organism identification methods still have weak adaptability in complex environments, especially in cases where there are many types of biological samples and high similarity in morphological characteristics, which can easily lead to misjudgment. In addition, existing methods usually ignore the interspecific relationships of biological communities, which have important reference value in the biological identification process.

[0004] To solve the above problems, the present application proposes a benthic organism sample identification analysis method and system based on data mining, which builds a benthic organism feature database, uses an underwater robot to collect samples, analyzes interspecific relationships using a graph theory algorithm, and performs secondary identification through credibility calculation to improve the accuracy and reliability of identification. This method not only improves the automated identification capability of benthic organism samples, but also enhances the understanding of the overall structure of biological communities, providing a more scientific and effective tool for marine ecological monitoring and biodiversity research. SUMMARY

[0005] To solve at least one of the above technical problems, the present application proposes a benthic organism sample identification analysis method and system based on data mining.

[0006] The first aspect of the present application provides a benthic organism sample identification analysis method based on data mining, comprising:

[0007] Obtaining standard image data of benthic organisms at different growth stages, extracting appearance features of the benthic organisms according to the standard image data, and constructing a benthic organism feature database according to the appearance features;

[0008] A benthic organism identification model is constructed based on the benthic organism characteristic database. The target benthic organism sample is collected by an underwater robot, and three-dimensional structural image data of the benthic organism community within a preset range of the collection point is obtained. The appearance characteristics of the target benthic organism sample are identified based on the benthic organism identification model to obtain the initial identification result of the target benthic organism sample.

[0009] The initial identification results are tested for uniqueness. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community.

[0010] Each initial identification result is fused with the interspecific relationship, and a relationship network after the fusion of each initial identification result and the interspecific relationship is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks;

[0011] The credibility of each initial identification result in the community population relationship network is determined based on the community interspecific relationship network. The target benthic organism sample is then re-identified based on the credibility of the existence to obtain the final identification result.

[0012] In this scheme, the steps of acquiring standard image data of benthic organisms at different growth stages, extracting the appearance features of benthic organisms based on the standard image data, and constructing a benthic organism feature database based on the appearance features are as follows:

[0013] Standard image data of benthic organisms at different growth stages are obtained, and edge detection is performed on the standard image data based on the Canny operator to extract the contour features of benthic organisms at different growth stages.

[0014] Based on the outline features, the perimeter, area, length-to-width ratio, and roundness of benthic organisms at different growth stages are calculated to obtain the scale characteristics of benthic organisms.

[0015] Shape features of benthic organisms in the standard image data are extracted based on Hu invariant moments, and texture features of benthic organisms in the standard image data are extracted based on gray-level co-occurrence matrix.

[0016] Histogram analysis was performed on the standard image data to determine the color characteristics of benthic organisms at different growth stages, and the outline features, scale features, shape features, texture features, and color features were labeled as the appearance features of benthic organisms.

[0017] A benthic organism characteristic database is constructed, the appearance characteristics are imported into the benthic organism characteristic database for storage, and the appearance characteristics in the benthic organism characteristic database are associated with the growth stages of the benthic organisms.

[0018] In this scheme, the benthic organism identification model is constructed based on the benthic organism characteristic database. An underwater robot collects target benthic organism samples and acquires three-dimensional structural image data of the benthic organism community within a preset range of the collection point. The appearance characteristics of the target benthic organism samples are then identified based on the benthic organism identification model to obtain the initial identification results of the target benthic organism samples. Specifically:

[0019] The appearance characteristics of benthic organisms at different growth stages in the benthic organism characteristic database are converted into numerical expression formats, and a benthic organism identification model is constructed based on the appearance characteristics in the data expression formats.

[0020] The underwater robot collects target benthic organism samples, obtains the location information of the target benthic organism sample collection point, and obtains three-dimensional point cloud data within a preset range of the collection point based on the location information of the collection point.

[0021] The benthic community within a preset range of the collection point is modeled based on the three-dimensional point cloud data to obtain three-dimensional structural image data of the benthic community.

[0022] Acquire image data of the target benthic organism sample, extract the appearance features of the target benthic organism sample from the image data, and convert the appearance features of the target benthic organism sample into a numerical expression format;

[0023] The appearance features of the target benthic organism sample in numerical expression format are imported into the benthic organism identification model for data matching, and the Euclidean distance between the appearance features of the target benthic organism sample and the appearance features of benthic organisms at different growth stages in the benthic organism feature database is calculated.

[0024] The similarity between the target benthic organism sample and benthic organisms at each different growth stage in the benthic organism characteristic database is assessed based on the Euclidean distance.

[0025] The benthic organisms whose similarity is greater than the preset value are output along with their growth stage to obtain the initial identification results of the benthic organism identification model for the target benthic organism sample.

[0026] In this scheme, the initial identification results are checked for uniqueness. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community. Specifically:

[0027] Obtain the benthic organism identification type data contained in the initial identification results, obtain the similarity of each benthic organism identification type, and construct a similarity distribution curve;

[0028] According to the similarity distribution curve, if the maximum similarity value is greater than the preset similarity value and there are no other data points in the similarity distribution curve that are greater than the preset similarity value, then the benthic organism type and the growth stage of the benthic organism corresponding to the maximum similarity value will be identified as the unique identification result.

[0029] If the initial identification result is the only identification result, the initial identification result shall be taken as the final identification result of the target benthic organism sample;

[0030] If multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, or if the maximum similarity value is less than the preset similarity value, the initial identification result of the target benthic organism sample will be marked as a non-unique identification result.

[0031] If the initial identification result is not unique, the name of each species in the three-dimensional structure image data of the benthic community is labeled, and data mining is performed on each species based on crawler technology to obtain the food web and nutrient relationship information of each species.

[0032] The interspecific relationships within the benthic community of the target benthic organism sample are determined based on the food web and nutrient relationship information.

[0033] In this scheme, the initial identification result is fused with the interspecific relationship, and a relationship network is constructed based on graph theory algorithm after the initial identification result and the interspecific relationship are fused, resulting in multiple community interspecific relationship networks. Specifically:

[0034] When the initial identification result is not unique and multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, the benthic organism identification type corresponding to the data points with a similarity value greater than the preset similarity value is used as the fusion species element.

[0035] When the initial identification result is not unique and the maximum similarity value in the similarity distribution curve is less than the preset similarity value, the benthic organism identification type corresponding to the preset number of data points with the highest similarity in the similarity distribution curve is used as the fusion species element.

[0036] Obtain food web and trophic relationship information for each fused species element, calculate the similarity between each fused species element and the food web and trophic relationship information of each species in the benthic community, and determine the niche overlap index between each fused species element and each species in the benthic community based on the similarity.

[0037] If the niche overlap index between the fused species element and a certain species in the benthic community is greater than a preset index, the fused species element and that species are identified as competing, and the interspecific relationships between the fused species element and other species in the benthic community are determined based on the interspecific relationships of the benthic community, thus obtaining the first fusion scheme.

[0038] If the niche overlap index between the fused species element and the species in the benthic community is not greater than the preset index, then the interspecific relationship between the fused species element and each species in the benthic community is determined based on the food web and trophic relationship information of the fused species element, and the second fusion scheme is obtained.

[0039] According to the first fusion scheme and the second fusion scheme, the fused species elements are fused with the interspecific relationships of the benthic community, and a relationship network after the fusion of fused species elements and interspecific relationships is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks.

[0040] In this scheme, the step of determining the credibility of each initial identification result in the community population relationship network based on the community interspecific relationship network, and then performing a secondary identification of the target benthic organism sample based on the credibility of the existence to obtain the final identification result, specifically involves:

[0041] The number of connections between each initial identification result and other network nodes in the community population relationship network is determined based on the community interspecific relationship network, and the degree of association of each initial identification result in the community interspecific relationship network is determined based on the number of connections.

[0042] The niche rationality of each initial identification result in the interspecific relationship network of the community is determined based on the niche index and interspecific relationship of each initial identification result in the interspecific relationship network of the community;

[0043] Based on the correlation and niche rationality, a comprehensive score is calculated to determine the credibility of each initial identification result in the interspecific relationship network of the community.

[0044] The target benthic organism sample was re-identified based on the credibility of the existence, and the initial identification result with the highest credibility was taken as the final identification result.

[0045] A second aspect of the present invention also provides a data mining-based benthic organism sample identification and analysis system, the system comprising: a memory and a processor, wherein the memory includes a data mining-based benthic organism sample identification and analysis method program, and when the data mining-based benthic organism sample identification and analysis method program is executed by the processor, the following steps are performed:

[0046] Obtain standard image data of benthic organisms at different growth stages, extract the appearance features of benthic organisms based on the standard image data, and construct a benthic organism feature database based on the appearance features;

[0047] A benthic organism identification model is constructed based on the benthic organism characteristic database. The target benthic organism sample is collected by an underwater robot, and three-dimensional structural image data of the benthic organism community within a preset range of the collection point is obtained. The appearance characteristics of the target benthic organism sample are identified based on the benthic organism identification model to obtain the initial identification result of the target benthic organism sample.

[0048] The initial identification results are tested for uniqueness. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community.

[0049] Each initial identification result is fused with the interspecific relationship, and a relationship network after the fusion of each initial identification result and the interspecific relationship is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks;

[0050] The credibility of each initial identification result in the community population relationship network is determined based on the community interspecific relationship network. The target benthic organism sample is then re-identified based on the credibility of the existence to obtain the final identification result.

[0051] This invention discloses a method and system for identifying and analyzing benthic organism samples based on data mining. The method first acquires standard image data of benthic organisms at different growth stages, extracts appearance features, and constructs a feature database. Based on this database, an identification model is established, and an underwater robot is used to collect target samples and three-dimensional structural image data of the community for initial identification. If the results are not unique, interspecific relationships are mined, and a community relationship network is constructed using graph theory algorithms. The reliability of the identification results is calculated, and a second identification is performed to obtain the final result. This invention improves the accuracy and reliability of identification and is applicable to marine ecological monitoring and biodiversity research. Attached Figure Description

[0052] Figure 1 A flowchart of a data mining-based method for identifying and analyzing benthic organism samples according to the present invention is shown;

[0053] Figure 2 The flowchart illustrating the construction of a benthic organism characteristic database according to the present invention is shown;

[0054] Figure 3 The flowchart illustrating the final identification results obtained by this invention is shown.

[0055] Figure 4A block diagram of a data mining-based benthic organism sample identification and analysis system of the present invention is shown. Detailed Implementation

[0056] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0057] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0058] Figure 1 A flowchart of a data mining-based method for identifying and analyzing benthic organism samples according to the present invention is shown.

[0059] like Figure 1 As shown, the first aspect of this invention provides a data mining-based method for identifying and analyzing benthic organism samples, comprising:

[0060] S102, acquire standard image data of benthic organisms at different growth stages, extract the appearance features of benthic organisms based on the standard image data, and construct a benthic organism feature database based on the appearance features;

[0061] S104, construct a benthic organism identification model based on the benthic organism characteristic database, collect target benthic organism samples using an underwater robot, and obtain three-dimensional structural image data of benthic organism communities within a preset range of the collection point. Identify the appearance characteristics of the target benthic organism samples based on the benthic organism identification model to obtain the initial identification results of the target benthic organism samples.

[0062] S106, Perform a uniqueness test on the initial identification results. If the initial identification results of the target benthic organism sample are not unique, perform data mining on the three-dimensional structure image data of the benthic organism community to determine the interspecific relationships of the benthic organism community.

[0063] S108, each initial identification result is fused with the interspecific relationship, and a relationship network after the fusion of each initial identification result and the interspecific relationship is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks;

[0064] S110, determine the credibility of each initial identification result in the community population relationship network based on the community interspecific relationship network, and perform secondary identification on the target benthic organism sample based on the credibility of existence to obtain the final identification result.

[0065] It should be noted that by acquiring standard image data covering the appearance characteristics of benthic organisms at different growth stages, the comprehensiveness of the identification model is ensured. A benthic organism characteristic database is constructed, linking the characteristics of biological samples with their growth stages to form structured data, facilitating subsequent matching. Based on the characteristic database, an identification model is built. Underwater robots are used to collect three-dimensional structural image data of samples and benthic communities. The identification model is then used to identify the appearance characteristics of the samples to obtain initial identification results. This process enables rapid preliminary identification of target benthic organism samples, improving identification efficiency. The initial identification results are checked for uniqueness. If not unique, the three-dimensional structural image data is mined to determine interspecific relationships, resolving the problem of ambiguous identification results and providing more dimensional information for accurate identification. Integrating the initial identification results with interspecific relationships to construct a relationship network allows for in-depth analysis from an ecological perspective, determining the credibility of existence, and conducting secondary identification to obtain the final result, significantly improving the accuracy and reliability of benthic organism sample identification. The benthic organism samples do not include microorganisms that are not visible to the naked eye.

[0066] Figure 2 The flowchart illustrating the construction of a benthic organism characteristic database according to the present invention is shown.

[0067] According to an embodiment of the present invention, the steps of acquiring standard image data of benthic organisms at different growth stages, extracting appearance features of benthic organisms based on the standard image data, and constructing a benthic organism feature database based on the appearance features are as follows:

[0068] S202, Obtain standard image data of benthic organisms at different growth stages, perform edge detection on the standard image data based on the Canny operator, and extract the contour features of benthic organisms at different growth stages;

[0069] S204. Calculate the perimeter, area, length-to-width ratio, and roundness of benthic organisms at different growth stages based on the outline features to obtain the scale characteristics of benthic organisms.

[0070] S206, Extract the shape features of benthic organisms from the standard image data based on Hu invariant moments, and extract the texture features of benthic organisms from the standard image data based on the gray-level co-occurrence matrix;

[0071] S208, perform histogram analysis on the standard image data to determine the color characteristics of benthic organisms at different growth stages, and label the outline features, scale features, shape features, texture features, and color features as the appearance features of benthic organisms;

[0072] S210, construct a benthic organism characteristic database, import the appearance characteristics into the benthic organism characteristic database for storage, and associate the appearance characteristics in the benthic organism characteristic database with the growth stages of benthic organisms.

[0073] It should be noted that by acquiring standard image data of benthic organisms at different growth stages, edge detection using the Canny operator to extract contour features can accurately delineate the external outlines of benthic organisms, providing a foundation for subsequent analysis. Based on the contour features, perimeter, area, and other scale features can be calculated to quantify the size and shape of benthic organisms. Hu invariant moments and gray-level co-occurrence matrices are used to extract shape and texture features, depicting their appearance details from different angles. Histogram analysis is used to determine color features, enriching the dimensions of appearance features. These features are calibrated and imported into a database for storage, and also correlated with growth stages, constructing a comprehensive and structured benthic organism feature database, laying the data foundation for subsequent benthic organism identification. The standard image data refers to clear image data of benthic organisms.

[0074] According to an embodiment of the present invention, the step of constructing a benthic organism identification model based on the benthic organism characteristic database, collecting target benthic organism samples using an underwater robot, acquiring three-dimensional structural image data of the benthic organism community within a preset range of the collection point, and identifying the appearance characteristics of the target benthic organism samples based on the benthic organism identification model to obtain the initial identification result of the target benthic organism samples, specifically includes:

[0075] The appearance characteristics of benthic organisms at different growth stages in the benthic organism characteristic database are converted into numerical expression formats, and a benthic organism identification model is constructed based on the appearance characteristics in the data expression formats.

[0076] The underwater robot collects target benthic organism samples, obtains the location information of the target benthic organism sample collection point, and obtains three-dimensional point cloud data within a preset range of the collection point based on the location information of the collection point.

[0077] The benthic community within a preset range of the collection point is modeled based on the three-dimensional point cloud data to obtain three-dimensional structural image data of the benthic community.

[0078] Acquire image data of the target benthic organism sample, extract the appearance features of the target benthic organism sample from the image data, and convert the appearance features of the target benthic organism sample into a numerical expression format;

[0079] The appearance features of the target benthic organism sample in numerical expression format are imported into the benthic organism identification model for data matching, and the Euclidean distance between the appearance features of the target benthic organism sample and the appearance features of benthic organisms at different growth stages in the benthic organism feature database is calculated.

[0080] The similarity between the target benthic organism sample and benthic organisms at each different growth stage in the benthic organism characteristic database is assessed based on the Euclidean distance.

[0081] The benthic organisms whose similarity is greater than the preset value are output along with their growth stage to obtain the initial identification results of the benthic organism identification model for the target benthic organism sample.

[0082] It should be noted that converting the benthic organism characteristic database and the appearance characteristics of the target samples into numerical formats solves the problem of data format differences, achieving data standardization and compatibility. Simultaneously, using Euclidean distance to accurately quantify characteristic differences provides a reliable basis for screening. In the identification process, automated matching identification based on similarity allows the computer to quickly complete data processing, handling a large number of samples in a short time. Furthermore, by setting a preset similarity threshold, interfering items are effectively eliminated, reducing the risk of misjudgment and improving the reliability of identification results.

[0083] According to an embodiment of the present invention, the uniqueness test of the initial identification result is performed. If the initial identification result of the target benthic organism sample is not unique, data mining is performed on the three-dimensional structural image data of the benthic organism community to determine the interspecific relationships of the benthic organism community. Specifically, this involves:

[0084] Obtain the benthic organism identification type data contained in the initial identification results, obtain the similarity of each benthic organism identification type, and construct a similarity distribution curve;

[0085] According to the similarity distribution curve, if the maximum similarity value is greater than the preset similarity value and there are no other data points in the similarity distribution curve that are greater than the preset similarity value, then the benthic organism type and the growth stage of the benthic organism corresponding to the maximum similarity value will be identified as the unique identification result.

[0086] If the initial identification result is the only identification result, the initial identification result shall be taken as the final identification result of the target benthic organism sample;

[0087] If multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, or if the maximum similarity value is less than the preset similarity value, the initial identification result of the target benthic organism sample will be marked as a non-unique identification result.

[0088] If the initial identification result is not unique, the name of each species in the three-dimensional structure image data of the benthic community is labeled, and data mining is performed on each species based on crawler technology to obtain the food web and nutrient relationship information of each species.

[0089] The interspecific relationships within the benthic community of the target benthic organism sample are determined based on the food web and nutrient relationship information.

[0090] It should be noted that different species of benthic organisms may exhibit similar appearance characteristics at certain growth stages. For example, the larvae of some small benthic organisms may have similar morphologies. When matching appearance characteristics with a benthic organism characteristic database, multiple results with high similarity scores can easily occur. When multiple data points in the similarity distribution curve exceed the preset similarity value, it means that several different benthic organism types have reached a high level of similarity with the target benthic organism sample, making it difficult to definitively determine which type the target sample belongs to. Conversely, if the maximum similarity score is less than the preset similarity value, it indicates that the target benthic organism sample does not have sufficient similarity to any type in the benthic organism characteristic database to determine its species, thus making it impossible to provide a unique identification result.

[0091] According to an embodiment of the present invention, the step of fusing each initial identification result with the interspecific relationship, and constructing a relationship network after fusing each initial identification result with the interspecific relationship based on a graph theory algorithm to obtain multiple community interspecific relationship networks, specifically involves:

[0092] When the initial identification result is not unique and multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, the benthic organism identification type corresponding to the data points with a similarity value greater than the preset similarity value is used as the fusion species element.

[0093] When the initial identification result is not unique and the maximum similarity value in the similarity distribution curve is less than the preset similarity value, the benthic organism identification type corresponding to the preset number of data points with the highest similarity in the similarity distribution curve is used as the fusion species element.

[0094] It should be noted that when multiple data points in the similarity distribution curve exceed the preset similarity value, the benthic organism identification types corresponding to these data points have a high similarity to the target benthic organism sample, and they could all be the true classification type of the target sample. Using these types as fusion species elements significantly reduces the amount of data analysis and avoids further processing of identification results with low similarity. Conversely, when the maximum similarity value is less than the preset similarity value, it indicates that there is no clearly matching type with extremely high similarity to the target sample. In this case, selecting the benthic organism identification types corresponding to the preset number of data points with the highest similarity in the similarity distribution curve as fusion species elements is because although the similarity of these data points does not reach the preset high value, they are relatively the closest to the target sample. By selecting them, the most likely benthic organism types can be screened within a limited range, improving the accuracy of the final identification.

[0095] Obtain food web and trophic relationship information for each fused species element, calculate the similarity between each fused species element and the food web and trophic relationship information of each species in the benthic community, and determine the niche overlap index between each fused species element and each species in the benthic community based on the similarity.

[0096] If the niche overlap index between the fused species element and a certain species in the benthic community is greater than a preset index, the fused species element and that species are identified as competing, and the interspecific relationships between the fused species element and other species in the benthic community are determined based on the interspecific relationships of the benthic community, thus obtaining the first fusion scheme.

[0097] If the niche overlap index between the fused species element and the species in the benthic community is not greater than the preset index, then the interspecific relationship between the fused species element and each species in the benthic community is determined based on the food web and trophic relationship information of the fused species element, and the second fusion scheme is obtained.

[0098] According to the first fusion scheme and the second fusion scheme, the fused species elements are fused with the interspecific relationships of the benthic community, and a relationship network after the fusion of fused species elements and interspecific relationships is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks.

[0099] It should be noted that when the niche overlap index between the fused species element and a species in the benthic community is greater than a preset index, it indicates a significant degree of competition between the two in terms of resource acquisition and living space. For example, two benthic organisms that both rely on the same algae as their primary food source and have highly overlapping habitats will have a high niche overlap index. Based on this, they are labeled as competitors, and the interspecific relationships between the fused species element and other species in the benthic community can be determined according to the interspecific relationships, resulting in the first fusion scheme. If the niche overlap index is not greater than the preset index, it indicates that the competition between the fused species element and the species in the community is not significant. In this case, based on the food web and trophic relationship information of the fused species element itself, its interspecific relationships with each species in the community can be determined more comprehensively, forming the second fusion scheme. The process of fusing fusion species elements with the interspecific relationships of the benthic community involves inserting the fusion species elements into the interspecific relationships of the benthic community. For example, there are fusion species elements A, B, and C, and interspecific relationships containing species D, E, and F. Each fusion only incorporates one element into the interspecific relationships, resulting in multiple interspecific relationships containing elements such as (A, D, E, F), (B, D, E, F), and (C, D, E, F). The purpose of this is to determine the plausibility of the fusion species elements in the interspecific relationships of the benthic community in subsequent invention steps, and then use the most plausible fusion species element as the final identified species. The higher the similarity between the fusion species element and the food web and trophic relationship information of the species in the benthic community, the higher the niche overlap index. The fusion species element then serves as a candidate species for the final species identification type of the target benthic organism sample.

[0100] Figure 3 A flowchart illustrating the final identification results obtained by this invention is shown.

[0101] According to an embodiment of the present invention, the step of determining the credibility of each initial identification result in the community population relationship network based on the community interspecific relationship network, and performing a secondary identification of the target benthic organism sample based on the credibility of the existence to obtain the final identification result, specifically involves:

[0102] S302, determine the number of connections between each initial identification result and other network nodes in the community population relationship network based on the community interspecific relationship network, and determine the degree of association of each initial identification result in the community interspecific relationship network based on the number of connections;

[0103] S304, Determine the niche rationality of each initial identification result in the interspecific relationship network of the community based on the niche index and interspecific relationship of each initial identification result;

[0104] S306, Based on the correlation degree and the ecological niche rationality, a comprehensive score is performed to determine the credibility of the existence of each initial identification result in the interspecific relationship network of the community;

[0105] S308, Based on the aforementioned credibility, a secondary identification of the target benthic organism sample is performed, and the initial identification result with the highest credibility is taken as the final identification result.

[0106] It is important to note that the determination of existence credibility integrates correlation and niche rationality. Correlation reflects the number of connections between the initial identification result and other nodes in the community's population relationship network. A higher number of connections indicates a closer relationship between the identification result and other organisms in the community, and a more universal and reasonable existence within the community. Niche rationality, based on the niche index and interspecific relationships, reflects the rationality of the identification result in niche allocation and interspecific relationships. This comprehensive scoring of these two dimensions fully considers the position and role of benthic organisms in the community's ecological structure, avoiding the one-sidedness of judgments based on a single factor. By determining existence credibility, the result that best fits the community's ecological relationships can be selected from multiple candidate results. This process further verifies and optimizes the initial identification results, effectively eliminating unreasonable results and significantly improving the accuracy and reliability of the final identification results. When the identification results are not unique, it is difficult to determine the species of the target benthic organism sample based solely on appearance characteristics. Analyzing the community's interspecific relationship network can provide more identification clues from an ecological perspective. For example, if two similar-looking benthic organisms are located in different positions in the food web, one feeding on algae and the other on small zooplankton, they can be accurately distinguished based on the food web information, thereby effectively improving the accuracy of identification and reducing misjudgment.

[0107] Figure 4 A block diagram of a data mining-based benthic organism sample identification and analysis system of the present invention is shown.

[0108] A second aspect of the present invention also provides a data mining-based benthic organism sample identification and analysis system 4, the system comprising: a memory 41 and a processor 42, wherein the memory includes a data mining-based benthic organism sample identification and analysis method program, and when the data mining-based benthic organism sample identification and analysis method program is executed by the processor, the following steps are implemented:

[0109] Obtain standard image data of benthic organisms at different growth stages, extract the appearance features of benthic organisms based on the standard image data, and construct a benthic organism feature database based on the appearance features;

[0110] A benthic organism identification model is constructed based on the benthic organism characteristic database. The target benthic organism sample is collected by an underwater robot, and three-dimensional structural image data of the benthic organism community within a preset range of the collection point is obtained. The appearance characteristics of the target benthic organism sample are identified based on the benthic organism identification model to obtain the initial identification result of the target benthic organism sample.

[0111] The initial identification results are tested for uniqueness. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community.

[0112] Each initial identification result is fused with the interspecific relationship, and a relationship network after the fusion of each initial identification result and the interspecific relationship is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks;

[0113] The credibility of each initial identification result in the community population relationship network is determined based on the community interspecific relationship network. The target benthic organism sample is then re-identified based on the credibility of the existence to obtain the final identification result.

[0114] This invention discloses a method and system for identifying and analyzing benthic organism samples based on data mining. The method first acquires standard image data of benthic organisms at different growth stages, extracts appearance features, and constructs a feature database. Based on this database, an identification model is established, and an underwater robot is used to collect target samples and three-dimensional structural image data of the community for initial identification. If the results are not unique, interspecific relationships are mined, and a community relationship network is constructed using graph theory algorithms. The reliability of the identification results is calculated, and a second identification is performed to obtain the final result. This invention improves the accuracy and reliability of identification and is applicable to marine ecological monitoring and biodiversity research.

[0115] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0116] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0117] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0118] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0119] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0120] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for identifying and analyzing benthic organism samples based on data mining, characterized in that, Includes the following steps: Obtain standard image data of benthic organisms at different growth stages, extract the appearance features of benthic organisms based on the standard image data, and construct a benthic organism feature database based on the appearance features; A benthic organism identification model is constructed based on the benthic organism characteristic database. The target benthic organism sample is collected by an underwater robot, and three-dimensional structural image data of the benthic organism community within a preset range of the collection point is obtained. The appearance characteristics of the target benthic organism sample are identified based on the benthic organism identification model to obtain the initial identification result of the target benthic organism sample. The initial identification results are tested for uniqueness. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community. Each initial identification result is fused with the interspecific relationships, and a relationship network is constructed based on graph theory algorithms to obtain multiple interspecific relationship networks of communities, specifically: When the initial identification result is not unique and multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, the benthic organism identification type corresponding to the data points with a similarity value greater than the preset similarity value is used as the fusion species element. When the initial identification result is not unique and the maximum similarity value in the similarity distribution curve is less than the preset similarity value, the benthic organism identification type corresponding to the preset number of data points with the highest similarity in the similarity distribution curve is used as the fusion species element. Obtain food web and trophic relationship information for each fused species element, calculate the similarity between each fused species element and the food web and trophic relationship information of each species in the benthic community, and determine the niche overlap index between each fused species element and each species in the benthic community based on the similarity. The species elements are fused with the interspecific relationships of the benthic community, and a relationship network after the fusion of species elements and interspecific relationships is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks. The credibility of each initial identification result in the community population relationship network is determined based on the community interspecific relationship network. The target benthic organism sample is then re-identified based on the credibility of the existence to obtain the final identification result.

2. The method for identifying and analyzing benthic organism samples based on data mining according to claim 1, characterized in that, The process of acquiring standard image data of benthic organisms at different growth stages, extracting appearance features of benthic organisms based on the standard image data, and constructing a benthic organism feature database based on the appearance features specifically involves: Standard image data of benthic organisms at different growth stages are obtained, and edge detection is performed on the standard image data based on the Canny operator to extract the contour features of benthic organisms at different growth stages. Based on the outline features, the perimeter, area, length-to-width ratio, and roundness of benthic organisms at different growth stages are calculated to obtain the scale characteristics of benthic organisms. Shape features of benthic organisms in the standard image data are extracted based on Hu invariant moments, and texture features of benthic organisms in the standard image data are extracted based on gray-level co-occurrence matrix. Histogram analysis was performed on the standard image data to determine the color characteristics of benthic organisms at different growth stages, and the outline features, scale features, shape features, texture features, and color features were labeled as the appearance features of benthic organisms. A benthic organism characteristic database is constructed, the appearance characteristics are imported into the benthic organism characteristic database for storage, and the appearance characteristics in the benthic organism characteristic database are associated with the growth stages of the benthic organisms.

3. The method for identifying and analyzing benthic organism samples based on data mining according to claim 1, characterized in that, The process involves constructing a benthic organism identification model based on the benthic organism characteristic database, collecting target benthic organism samples using an underwater robot, acquiring three-dimensional structural image data of the benthic organism community within a preset range of the collection point, and identifying the appearance characteristics of the target benthic organism samples based on the benthic organism identification model to obtain the initial identification results of the target benthic organism samples. Specifically: The appearance characteristics of benthic organisms at different growth stages in the benthic organism characteristic database are converted into numerical expression formats, and a benthic organism identification model is constructed based on the appearance characteristics in the data expression formats. The underwater robot collects target benthic organism samples, obtains the location information of the target benthic organism sample collection point, and obtains three-dimensional point cloud data within a preset range of the collection point based on the location information of the collection point. The benthic community within a preset range of the collection point is modeled based on the three-dimensional point cloud data to obtain three-dimensional structural image data of the benthic community. Acquire image data of the target benthic organism sample, extract the appearance features of the target benthic organism sample from the image data, and convert the appearance features of the target benthic organism sample into a numerical expression format; The appearance features of the target benthic organism sample in numerical expression format are imported into the benthic organism identification model for data matching, and the Euclidean distance between the appearance features of the target benthic organism sample and the appearance features of benthic organisms at different growth stages in the benthic organism feature database is calculated. The similarity between the target benthic organism sample and benthic organisms at each different growth stage in the benthic organism characteristic database is assessed based on the Euclidean distance. The benthic organisms whose similarity is greater than the preset value are output along with their growth stage to obtain the initial identification results of the benthic organism identification model for the target benthic organism sample.

4. The method for identifying and analyzing benthic organism samples based on data mining according to claim 1, characterized in that, The initial identification results are subjected to a uniqueness test. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community. Specifically: Obtain the benthic organism identification type data contained in the initial identification results, obtain the similarity of each benthic organism identification type, and construct a similarity distribution curve; According to the similarity distribution curve, if the maximum similarity value is greater than the preset similarity value and there are no other data points in the similarity distribution curve that are greater than the preset similarity value, then the benthic organism type and the growth stage of the benthic organism corresponding to the maximum similarity value will be identified as the unique identification result. If the initial identification result is the only identification result, the initial identification result shall be taken as the final identification result of the target benthic organism sample; If multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, or if the maximum similarity value is less than the preset similarity value, the initial identification result of the target benthic organism sample will be marked as a non-unique identification result. If the initial identification result is not unique, the name of each species in the three-dimensional structure image data of the benthic community is labeled, and data mining is performed on each species based on crawler technology to obtain the food web and nutrient relationship information of each species. The interspecific relationships within the benthic community of the target benthic organism sample are determined based on the food web and nutrient relationship information.

5. The method for identifying and analyzing benthic organism samples based on data mining according to claim 4, characterized in that, The step of fusing each initial identification result with the interspecific relationship, constructing a relationship network after fusing each initial identification result with the interspecific relationship based on a graph theory algorithm, and obtaining multiple community interspecific relationship networks, further includes: If the niche overlap index between the fused species element and a certain species in the benthic community is greater than a preset index, the fused species element and that species are identified as competing, and the interspecific relationships between the fused species element and other species in the benthic community are determined based on the interspecific relationships of the benthic community, thus obtaining the first fusion scheme. If the niche overlap index between the fused species element and the species in the benthic community is not greater than a preset index, then the interspecific relationship between the fused species element and each species in the benthic community is determined based on the food web and trophic relationship information of the fused species element, thus obtaining the second fusion scheme.

6. The method for identifying and analyzing benthic organism samples based on data mining according to claim 1, characterized in that, The process involves determining the credibility of each initial identification result within the community population relationship network based on the interspecific relationship network, and then performing a secondary identification of the target benthic organism sample based on the credibility of the existence to obtain the final identification result. Specifically: The number of connections between each initial identification result and other network nodes in the community population relationship network is determined based on the community interspecific relationship network, and the degree of association of each initial identification result in the community interspecific relationship network is determined based on the number of connections. The niche rationality of each initial identification result in the interspecific relationship network of the community is determined based on the niche index and interspecific relationship of each initial identification result in the interspecific relationship network of the community; Based on the correlation and niche rationality, a comprehensive score is calculated to determine the credibility of each initial identification result in the interspecific relationship network of the community. The target benthic organism sample was re-identified based on the credibility of the existence, and the initial identification result with the highest credibility was taken as the final identification result.

7. A data mining-based benthic organism sample identification and analysis system, characterized in that, The data mining-based benthic organism sample identification and analysis system includes a storage device and a processor. The storage device includes a data mining-based benthic organism sample identification and analysis method program. When the data mining-based benthic organism sample identification and analysis method program is executed by the processor, the following steps are implemented: Obtain standard image data of benthic organisms at different growth stages, extract the appearance features of benthic organisms based on the standard image data, and construct a benthic organism feature database based on the appearance features; A benthic organism identification model is constructed based on the benthic organism characteristic database. The target benthic organism sample is collected by an underwater robot, and three-dimensional structural image data of the benthic organism community within a preset range of the collection point is obtained. The appearance characteristics of the target benthic organism sample are identified based on the benthic organism identification model to obtain the initial identification result of the target benthic organism sample. The initial identification results are tested for uniqueness. If the initial identification results of the target benthic organism sample are not unique, data mining is performed on the three-dimensional structural image data of the benthic community to determine the interspecific relationships of the benthic community. Each initial identification result is fused with the interspecific relationships, and a relationship network is constructed based on graph theory algorithms to obtain multiple interspecific relationship networks of communities, specifically: When the initial identification result is not unique and multiple data points with a similarity value greater than the preset similarity value appear in the similarity distribution curve, the benthic organism identification type corresponding to the data points with a similarity value greater than the preset similarity value is used as the fusion species element. When the initial identification result is not unique and the maximum similarity value in the similarity distribution curve is less than the preset similarity value, the benthic organism identification type corresponding to the preset number of data points with the highest similarity in the similarity distribution curve is used as the fusion species element. Obtain food web and trophic relationship information for each fused species element, calculate the similarity between each fused species element and the food web and trophic relationship information of each species in the benthic community, and determine the niche overlap index between each fused species element and each species in the benthic community based on the similarity. The species elements are fused with the interspecific relationships of the benthic community, and a relationship network after the fusion of species elements and interspecific relationships is constructed based on graph theory algorithm to obtain multiple community interspecific relationship networks. The credibility of each initial identification result in the community population relationship network is determined based on the community interspecific relationship network. The target benthic organism sample is then re-identified based on the credibility of the existence to obtain the final identification result.

Images