Intelligent network library retrieval and management method and system for fishery resource samples

Abstract

The invention relates to the technical field of fishery resource sample management, in particular to an intelligent network library retrieval and management method and system for fishery resource samples. Multidimensional sample information is collected, and a label coding layer is constructed to paste a unique retrieval label for each sample; a category search tree is constructed by combining a cabinet body structure of a sample storage cabinet and a classification system, three-level codes of cabinet number-column number-box number are formed, and standard classified storage of fishery resource samples is achieved. Sample information is input into an Excel table data network library based on inverted indexes, and multi-condition rapid retrieval of places, time, sample numbers and the like is achieved. Before warehousing, three samples are randomly extracted from each box to verify the consistency of the labels and the database, and no less than 5% of samples are regularly backed up and randomly checked, so that long-term accuracy of the data is ensured. According to the method, the fishery resource samples can be classified, stored and taken for rapid positioning and retrieval, the cost is low, the operability is high, and the sample positioning, retrieval and management efficiency is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of fishery resource sample management technology, and in particular to an intelligent online database retrieval and management method and system for fishery resource samples. Background Technology

[0002] In fisheries resource research experiments, it is often necessary to preserve and manage tens of thousands or even more samples for a long period of time. Sample library management is a crucial link in ensuring data accuracy and research traceability. However, traditional sample management methods mostly rely on manual recording and searching, which easily leads to problems such as inconsistencies between labels and electronic data, non-standardized sample information recording, and low efficiency in searching for samples when the number of samples is large. This results in sample loss or misuse, affecting data accuracy. Therefore, there is an urgent need for a low-cost, efficient, accurate, and easy-to-operate intelligent sample library management and retrieval method to achieve rapid sample information entry, precise location and search, and intelligent retrieval. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides a method and system for intelligent online database retrieval and management of fishery resource samples.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of this invention provides a method for intelligent database retrieval and management of fishery resource samples, comprising the following steps: S101: Obtain multidimensional collection data information of fishery resource samples and construct a multidimensional label coding layer. Based on the label dimension, perform semantic-specific label coding on the multidimensional collection data information on the multidimensional label coding layer, and affix a unique coding label to each fishery resource sample to obtain sample retrieval labels for different fishery resource samples. S102: All sample storage cabinets in the sample library are uniformly numbered, and a category search tree architecture is constructed from the root class node of the sample library according to the cabinet structure and the parent-child-sub-level category gradient. The category search tree architecture is used to query the classification and storage of different fishery resource samples and encode them in three levels to obtain the three-level coding sequence of cabinet number-column number-box number for the classification and storage of fishery resource samples. S103: Based on the three-level coding sequence of cabinet number-column number-box number, different fishery resource samples are stored in the corresponding sample boxes, and the sample retrieval tag is bound to the three-level coding sequence of cabinet number-column number-box number to form sample information; S104: Divide the information of different types of samples into different storage workbooks and enter it into an inverted index Excel spreadsheet to establish an intelligent retrieval database of fishery resource samples; S105: Obtain the target fishery resource samples and target retrieval conditions required for fishery research. By inputting the target retrieval conditions into the intelligent retrieval database, the storage location of the target fishery resource samples is searched in reverse order to achieve rapid location and retrieval of the target fishery resource samples. S106: Randomly select three samples from each sample box and search them in the intelligent retrieval database. Verify the consistency of the sample information storage and verify the randomly selected fishery resource samples and sample boxes before storing them in the database. S107: Regularly back up and manage the intelligent retrieval database, and conduct secondary verification management on no less than 5% of the samples already in the database to ensure the accuracy of the data under long-term operation of the sample database.

[0005] Preferably, step S101 specifically includes the following steps: The data collected from fishery resource samples includes multidimensional data such as collection location, collection time, sample name, and sample number. Based on the dimensional characteristics of multidimensional collected data, a multidimensional label coding layer and the granularity limit of each multidimensional label coding layer are constructed. At the same time, a word sense feature extraction algorithm is introduced to create a word sense feature encoder. Based on the granularity limit, a word sense feature encoder is used on each multidimensional label encoding layer to perform average pooling calculation on the word sense vectors contained in the multidimensional acquired data information and map them to integer indices, so as to obtain the independent discrete feature encoding form of each multidimensional label encoding layer for each multidimensional acquired data information. By acquiring multi-dimensional collected data information through big data, a semantic encoding character library is obtained for different label dimensions such as collection location, collection time, sample name, and sample number. The character bit width of the semantic encoding character library is then used to determine the encoding standard for different label dimensions. Based on the character bit width and following the discrete feature encoding form, the continuous bit stream of tag meaning in the multidimensional acquired data information is grouped and segmented to obtain several binary bit groups corresponding to the tag meaning of the multidimensional acquired data information; By querying the coded characters of each binary bit group through the semantic coded character library index, the coded character sequence of each multidimensional data information located in the corresponding label dimension is output. Finally, all the coded character sequences of the label dimension are combined to obtain the exclusive coded string of each multidimensional data information. Each sample is compiled and affixed with a unique coded label based on its collection location, collection time, sample name, and sample number, thus obtaining sample retrieval labels for different fishery resource samples.

[0006] Preferably, step S102 specifically includes the following steps: Obtain the classification numbering rules for different types of sample storage and the overall sample library. Based on the classification numbering rules, assign a unified number to all sample storage cabinets in the sample library to obtain the cabinet type number for different sample storage cabinets. Construct the root class node for the storage and query of fishery resource sample types based on the overall sample library. Based on the cabinet structure of the sample storage cabinet, a three-level splitting strategy is set up to classify fishery resource samples according to the root class from parent level, child level to subclass, and the category storage function of the sample storage cabinet is obtained. The cabinet structure has three rows of cabinets for each sample storage cabinet, each row of cabinets has 20 sample boxes, and each sample box contains several sample containers labeled with numbers. Anchor the starting source point with the root class node, and traverse the key input position from the root class node based on the category storage function to generate the storage key node of the sample storage cabinet. Based on the three-order splitting strategy, the storage key node is split to continuously generate category leaf nodes, forming a category search tree architecture of sample storage cabinet-cabinet queue-sample box. The parent category keywords, sub-category keywords, and sub-category keywords of different fishery resource samples are identified and obtained through the knowledge graph of fishery resource samples. The target bytes of the corresponding category keywords of fishery resource samples are scanned and queried in the order of parent, child and sub-level from the root node through the category search tree architecture to obtain the category byte query results. Based on the category byte query results, the parent category byte, child category byte, and sub-category category byte of the fishery resource sample type are found inside the root category node, and the target sample storage cabinet, target cabinet queue, and target sample box for storing different fishery resource sample categories are determined. Based on the target sample storage cabinet, target cabinet queue, and target sample box, samples are classified and stored, and a three-level system coding is performed to obtain the three-level coding sequence of cabinet number-column number-box number for the classification and storage of fishery resource samples.

[0007] Preferably, step S104 specifically includes the following steps: Different fishery resource samples are stored in corresponding sample boxes according to the three-level coding sequence of cabinet number-column number-box number, while ensuring that the sample number recorded on the sample retrieval label is consistent with the label number of the sample container. After storage, the forward index mode of the Excel spreadsheet is called to bind the sample retrieval tag with the cabinet number-column number-box number three-level coding sequence as a prefix index term to form the sample information of the fishery resource sample, and to create a storage workbook for different fishery resource samples. Import the stored workbook into the forward index mode and reorganize it according to the prefix index terms, mapping it backwards to the stored workbook. Generate an Excel spreadsheet containing the inverted index list for the prefix retrieval of corresponding sample information for different fishery resource samples. Based on the online loading of the inverted index list, a data table format that can be quickly retrieved is formed, and an intelligent retrieval database of fishery resource samples is established.

[0008] Preferably, step S105 specifically includes the following steps: Obtain retrieval tasks for fisheries research, and extract the target fisheries resource samples and corresponding target search conditions required for fisheries research through retrieval tasks; A retrieval vector space is constructed, and the k-means clustering algorithm is introduced into the retrieval vector space to perform clustering calculations on the storage index vectors corresponding to all sample information entered into the intelligent retrieval database, thereby obtaining the k retrieval cluster centers of the intelligent retrieval database. Calculate the Mahalanobis distance between the storage index vector of different sample information in the intelligent retrieval database and the center of each retrieval cluster, and extract the autonomous inverted list structure of each retrieval cluster center; If the Mahalanobis distance is less than the preset Mahalanobis distance, then the center of the retrieval cluster is designated as the candidate retrieval bucket, and the stored index vector is assigned to the autonomous inverted list structure corresponding to the candidate retrieval bucket. Repeat the above Mahalanobis distance determination allocation steps until all the storage index vectors of the intelligent retrieval database are allocated, and obtain the inverted index list of the intelligent retrieval database for the retrieval of information from different sample storage locations; Obtain the potential retrieval vector of the target retrieval condition, locate one or more candidate retrieval buckets that are closest to the potential retrieval vector based on Mahalanobis distance, and mark them as the target vector retrieval domain; The correlation between the potential retrieval vector and each stored index vector recorded in the inverted index list corresponding to the retrieval domain of each target vector is calculated by grey relational degree, and several local retrieval correlation degrees are obtained. Only the sample information corresponding to the storage index vector with the maximum local retrieval correlation is extracted, and the cabinet number-column number-box number three-level coding sequence of the sample information is retrieved through the intelligent retrieval database. The target fishery resource sample is quickly located and retrieved based on the cabinet number-column number-box number three-level coding sequence.

[0009] Preferably, step S106 specifically includes the following steps: For each sample box in each sample storage cabinet, three fishery resource samples are randomly selected and marked as random check samples. The corresponding sample information of the random check samples is then imported into the intelligent retrieval database for retrieval. Obtain the storage management requirements for randomly checked samples, and extract one or more main verification factors for the randomly checked samples based on the storage management requirements; Construct a temporary verification structure, extract the library retrieval fields that meet each main verification factor and the corresponding library field value from the intelligent retrieval data network library, and write each library retrieval field into the temporary verification structure based on the library field value to generate a temporary verification blueprint group for the main verification factor. Obtain the actual retrieval field of each random verification sample on each main verification factor and the actual field value corresponding to each actual retrieval field, and construct the actual verification mapping group of the main verification factor based on the actual retrieval field and the actual field value; Project the actual verification mapping group onto the temporary verification blueprint group, and calculate the mapping consistency rate of the actual verification mapping group compared to the temporary verification blueprint group; If the mapping consistency rate is higher than the preset mapping consistency rate threshold, the randomly checked sample will be put back into the sample box and confirmed as being in storage. If there is a discrepancy, the information of the randomly checked sample and the intelligent retrieval database will be corrected and rechecked until they are completely consistent before being confirmed for entry into the database.

[0010] A second aspect of the present invention provides an intelligent database retrieval and management system for fishery resource samples. The system includes: a memory, a processor, and a communication interface. The memory includes a program for an intelligent database retrieval and management method for fishery resource samples. The communication interface is used for data connection and communication between the memory and the processor. When the program for an intelligent database retrieval and management method for fishery resource samples is executed by the processor, it implements any of the steps of the intelligent database retrieval and management method described in the present invention.

[0011] This invention addresses the technical deficiencies in the prior art, and its beneficial technical effects are as follows: Multidimensional data of fishery resource samples were collected and a multidimensional label coding layer was constructed. Based on the label dimensions, semantically specific labels were assigned to the multidimensional collected data on the multidimensional label coding layer, and a unique coded label was affixed to each fishery resource sample to obtain sample retrieval labels for different fishery resource samples. All sample storage cabinets in the sample library were uniformly numbered, and a category search tree architecture was constructed from the root class node of the sample library according to the cabinet structure and the parent-child category gradient. The category search tree architecture was used to query and perform three-level coding for the classification and storage of different fishery resource samples, resulting in a three-level coding sequence of cabinet number-column number-box number for fishery resource sample classification and storage. Based on the cabinet number-column number-box number three-level coding sequence, different fishery resource samples were stored in their corresponding sample boxes, and the sample retrieval labels were associated with the cabinet number-column number-box number. The three-level coding sequence is bundled into sample information; different types of sample information are divided into different storage workbooks and entered into an inverted index Excel spreadsheet to establish an intelligent retrieval database for fishery resource samples; the target fishery resource samples and target retrieval conditions required for fishery research are obtained; by inputting the target retrieval conditions into the intelligent retrieval database, the storage location of the target fishery resource samples is searched in reverse order to achieve rapid location and retrieval of the target fishery resource samples; three samples are randomly selected from each sample box and searched in the intelligent retrieval database to verify the consistency of the sample information storage; the randomly selected fishery resource samples and sample boxes are verified and stored; the intelligent retrieval database is regularly backed up and managed, and at least 5% of the stored samples are randomly checked for secondary verification management to ensure the data accuracy of the sample library under long-term operation. This invention uses a unified coding and storage rule for the sample library, which facilitates standardized sample management and improves the retrieval efficiency of sample storage and retrieval; by utilizing the filtering and search functions of the Excel spreadsheet inverted index, it supports multi-condition combination queries, which can achieve rapid location of any fishery resource sample and significantly improve the sample search speed. Secondly, by verifying the sample label information against the database through a mechanism of "randomly selecting three samples from each box" before warehousing, the problem of data discrepancies caused by human error in data entry is greatly reduced, and the reliability of the sample database is improved. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0013] Figure 1 A flowchart of the first method for intelligent database retrieval and management of fishery resource samples is shown. Figure 2A flowchart of the second method for intelligent database retrieval and management of fishery resource samples is shown; Figure 3 A system framework diagram of an intelligent database retrieval and management system for fishery resource samples is shown. Figure 4 A schematic diagram of the layout structure of different sample storage cabinets is shown; Figure 5 A schematic diagram showing the layout of the cabinet compartments and sample boxes inside the sample storage cabinet is shown. Figure 6 A schematic diagram showing the numbered arrangement of the sample containers stored inside the sample box is shown. Detailed Implementation

[0014] 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.

[0015] 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.

[0016] The first aspect of this invention provides a method for intelligent database retrieval and management of fishery resource samples, such as... Figure 1 As shown, it includes the following steps: S101: Obtain multidimensional collection data information of fishery resource samples and construct a multidimensional label coding layer. Based on the label dimension, perform semantic-specific label coding on the multidimensional collection data information on the multidimensional label coding layer, and affix a unique coding label to each fishery resource sample to obtain sample retrieval labels for different fishery resource samples. S102: All sample storage cabinets in the sample library are uniformly numbered, and a category search tree architecture is constructed from the root class node of the sample library according to the cabinet structure and the parent-to-child category gradient. The category search tree architecture is used to query the classification and storage of different fishery resource samples and encode them in three levels to obtain the three-level coding sequence of cabinet number-column number-box number for the classification and storage of fishery resource samples. S103: Based on the three-level coding sequence of cabinet number-column number-box number, different fishery resource samples are stored in the corresponding sample boxes, and the sample retrieval tag is bound to the three-level coding sequence of cabinet number-column number-box number to form sample information; S104: Divide the information of different types of samples into different storage workbooks and enter it into an inverted index Excel spreadsheet to establish an intelligent retrieval database of fishery resource samples; S105: Obtain the target fishery resource samples and target retrieval conditions required for fishery research. By inputting the target retrieval conditions into the intelligent retrieval database, the storage location of the target fishery resource samples is searched in reverse order to achieve rapid location and retrieval of the target fishery resource samples. S106: Randomly select three samples from each sample box and search them in the intelligent retrieval database. Verify the consistency of the sample information storage and verify the randomly selected fishery resource samples and sample boxes before storing them in the database. S107: Regularly back up and manage the intelligent retrieval database, and conduct secondary verification management on no less than 5% of the samples already in the database to ensure the accuracy of the data under long-term operation of the sample database.

[0017] Preferably, step S101 specifically includes the following steps: The data collected from fishery resource samples includes multidimensional data such as collection location, collection time, sample name, and sample number. Based on the dimensional characteristics of multidimensional collected data, a multidimensional label coding layer and the granularity limit of each multidimensional label coding layer are constructed. At the same time, a word sense feature extraction algorithm is introduced to create a word sense feature encoder. Based on the granularity limit, a word sense feature encoder is used on each multidimensional label encoding layer to perform average pooling calculation on the word sense vectors contained in the multidimensional acquired data information and map them to integer indices, so as to obtain the independent discrete feature encoding form of each multidimensional label encoding layer for each multidimensional acquired data information. By acquiring multi-dimensional collected data information through big data, a semantic encoding character library is obtained for different label dimensions such as collection location, collection time, sample name, and sample number. The character bit width of the semantic encoding character library is then used to determine the encoding standard for different label dimensions. Based on the character bit width and following the discrete feature encoding form, the continuous bit stream of tag meaning in the multidimensional acquired data information is grouped and segmented to obtain several binary bit groups corresponding to the tag meaning of the multidimensional acquired data information; By querying the coded characters of each binary bit group through the semantic coded character library index, the coded character sequence of each multidimensional data information located in the corresponding label dimension is output. Finally, all the coded character sequences of the label dimension are combined to obtain the exclusive coded string of each multidimensional data information. Each sample is compiled and affixed with a unique coded label based on its collection location, collection time, sample name, and sample number, thus obtaining sample retrieval labels for different fishery resource samples.

[0018] It should be noted that this method labels relevant collected data and information onto samples. However, multidimensional collected data has multiple dimensions, including time, location, sample identity, and identification. Traditional methods typically encode sample labels by simply entering data into a table, without specifically labeling the corresponding fishery resource samples. This can easily lead to confusion or loss of sample information, resulting in inaccurate sample database construction and retrieval, and reduced retrieval reliability. To address this, this method constructs a hierarchical coding framework for labels based on different sample data dimensions, namely a multidimensional label coding layer. This captures and analyzes the unique label information for different data dimensions, avoiding coding chaos and information loss during subsequent sample labeling, and ensuring that the sample label coding system comprehensively covers the multi-scale semantics of multidimensional collected data. Simultaneously, a granularity limit is set for each layer, clearly constraining the fine-grained sample information captured at each layer, ensuring multi-level representation of label coding across different label dimensions. Subsequently, the multidimensional collected data is standardized and mapped to integer indices, obtaining discrete feature forms that can be coded from the sample data. This provides semantic feature inputs expressing the corresponding dimensional information for subsequent hierarchical coding of different label dimensions. The semantic encoding character library defines the encoding space for semantic tags in multi-dimensional collected data, determining the completeness and readability of the encoded tag characters. The character bit width defines the fixed encoding length of the tag character index, ensuring the accuracy of the sample tag information.

[0019] It should be noted that this method follows a discrete feature encoding format, grouping and segmenting the continuous bitstream of tag meanings in multidimensional collected data according to character width. This divides the bitstream of tag-dimensional related collected data into fragment units that can be directly mapped to the semantic encoding character library index. This achieves an encoding mechanism for sample tags from the binary world to the character world, and then to printable characters in the intelligent retrieval database. This results in diverse tag encoding characters for different fishery resource samples in the intelligent retrieval database, allowing each sample to be affixed with unique data information tags regarding collection location, collection time, sample name, and sample number, improving the identifiability of fishery resource sample retrieval. This method enables each sample to be affixed with a unique, different-dimensional sample tag encoding, giving the sample a unique and personalized retrieval signature during storage, facilitating sample information management, and effectively improving the accuracy of fishery resource sample retrieval.

[0020] Preferably, step S102 specifically includes the following steps: Obtain the classification numbering rules for different types of sample storage and the overall sample library. Based on the classification numbering rules, assign a unified number to all sample storage cabinets in the sample library to obtain the cabinet type number for different sample storage cabinets. Construct the root class node for the storage and query of fishery resource sample types based on the overall sample library. Based on the cabinet structure of the sample storage cabinet, a three-level splitting strategy is set up to classify fishery resource samples according to the root class from parent level, child level to subclass, and the category storage function of the sample storage cabinet is obtained. The cabinet structure has three rows of cabinets for each sample storage cabinet, each row of cabinets has 20 sample boxes, and each sample box contains several sample containers labeled with numbers. Anchor the starting source point with the root class node, and traverse the key input position from the root class node based on the category storage function to generate the storage key node of the sample storage cabinet. Based on the three-order splitting strategy, the storage key node is split to continuously generate category leaf nodes, forming a category search tree architecture of sample storage cabinet-cabinet queue-sample box. The parent category keywords, sub-category keywords, and sub-category keywords of different fishery resource samples are identified and obtained through the knowledge graph of fishery resource samples. The target bytes of the corresponding category keywords of fishery resource samples are scanned and queried in the order of parent, child and sub-level from the root node through the category search tree architecture to obtain the category byte query results. Based on the category byte query results, the parent category byte, child category byte, and sub-category category byte of the fishery resource sample type are found inside the root category node, and the target sample storage cabinet, target cabinet queue, and target sample box for storing different fishery resource sample categories are determined. Based on the target sample storage cabinet, target cabinet queue, and target sample box, samples are classified and stored, and a three-level system coding is performed to obtain the three-level coding sequence of cabinet number-column number-box number for the classification and storage of fishery resource samples.

[0021] It should be noted that there are several sample storage cabinets, such as... Figure 4 As shown, each sample storage cabinet corresponds to a major species (parent species) of a fishery resource sample, such as fish, crustaceans, echinoderms, or mollusks. Each sample storage cabinet contains compartments numbered A, B, and C, as follows: Figure 5 As shown, each cabinet row corresponds to a coarse species (sub-species) for sample storage, such as shrimp, mud crab, yellow croaker, or grouper. The 20 sample boxes in each cabinet row are also labeled with box numbers from 01 to 20, and each sample box corresponds to a fine species (sub-category), such as muscle, otoliths, gonads, stomach contents, or scales. This forms a hierarchical classification system of sample storage cabinet (parent) - cabinet row (child) - sample box (sub-category), ensuring more standardized, precise, and accurate classification management of fishery resource samples. Furthermore, each sample box contains 88 sample containers, numbered from 01 to 88, such as... Figure 6As shown. However, traditional sample classification and storage relies on manual experience for location and search, which can easily lead to ambiguity in category positioning and incorrect storage when the sample quantity is large, resulting in mixed storage, non-standard practices, and inconsistent information recording. To address this, this method constructs a root class node based on a sample database. The sample database records the basic category information and attributes of all fishery resource samples, and the root class node provides a unified and known starting point for recursive sample classification. Specifically, the three-level splitting strategy includes: setting the cabinet category number as the first-level splitting node, the column number of the cabinet queue as the second-level splitting node, and the box number of the sample box as the third-level splitting node; thus, based on the actual structure of the cabinet, a parent-child-sub-classification three-level splitting model is defined, and a category hierarchy architecture is proposed for the three-level coding system of "cabinet number-column number-box number".

[0022] It should be noted that the specific meaning of the category storage function is that the sample storage cabinet provides corresponding storage rules and management functions based on the characteristics of different categories of samples in its classification storage design. This is an important premise and clear guide for traversing the keyword nodes of different sample storage cabinets from the root category node, which determines the category matching accuracy of each sample storage cabinet classification system for different fishery resource samples. Subsequently, the category keywords are traversed downwards from the root category node and the leaf nodes are split, thereby completing the automatic classification system mapping from logical classification to physical cabinet storage structure, constructing a multi-level category search tree of sample storage cabinet - cabinet queue - sample box. The search tree structure enables the sample classification storage system to have digital spatial navigation capabilities, providing an index foundation for subsequent sample category keyword queries. Next, the sample category keywords are retrieved level by level in the category search tree according to the internal order of parent-child-subdivision. This recursive method accurately finds the location bytes that match the sample in the tree structure, thereby converting the abstract classification bytes into physical storage coordinates (target sample storage cabinet, target cabinet queue, and target sample box). This enables the rapid determination of the specific storage location of the sample after level-by-level classification, realizing a calculable mapping from category semantics to the physical storage location of the cabinet, and avoiding the ambiguity and confusion of manually classifying and storing samples. This method enables the unified numbering of all sample storage cabinets in the sample library and the construction of a search tree structure for the sample classification and storage system. Based on different semantic categories, the sample storage location can be quickly queried and located, thus storing the samples in the corresponding cabinets. This forms a three-level coding system of "cabinet number - column number - box number". For example, the scale sample of yellow croaker should be stored in "07-A-17", which represents cabinet number 7 (category: fish) - column A - (category: yellow croaker) - box number 17 (category: scale). This achieves precise positioning of sample storage, replacing the tedious steps of traditional manual experience-based positioning, reducing the error rate of storage, and improving the efficiency and standardization of sample management.

[0023] Preferably, S104, as Figure 2As shown, the specific steps include: Different fishery resource samples are stored in corresponding sample boxes according to the three-level coding sequence of cabinet number-column number-box number, while ensuring that the sample number recorded on the sample retrieval label is consistent with the label number of the sample container. After storage, the forward index mode of the Excel spreadsheet is called to bind the sample retrieval tag with the cabinet number-column number-box number three-level coding sequence as a prefix index term to form the sample information of the fishery resource sample, and to create a storage workbook for different fishery resource samples. Import the stored workbook into the forward index mode and reorganize it according to the prefix index terms, mapping it backwards to the stored workbook. Generate an Excel spreadsheet containing the inverted index list for the prefix retrieval of corresponding sample information for different fishery resource samples. Based on the online loading of the inverted index list, a data table format that can be quickly retrieved is formed, and an intelligent retrieval database of fishery resource samples is established.

[0024] It should be noted that by reconstructing the original forward index structure from sample to information through reverse mapping, an inverted index list is generated where "index terms point to sample sets." This supports reverse positioning based on arbitrary prefix index conditions (cabinet number, column number, box number, sample number), enabling the constructed intelligent retrieval database to quickly locate target fishery resource samples or their storage locations based on sample retrieval conditions (type, number, name, etc.). It also allows for tracing or batch filtering of samples and their classification information based on data prefixes, achieving intelligent bidirectional retrieval of fishery resource samples. This provides underlying retrieval mechanism support for high-concurrency, large-scale, multi-condition, and multi-dimensional rapid sample queries, significantly improving sample retrieval hit rate, functionality, and retrieval efficiency. This method utilizes Excel spreadsheets as the database carrier to divide sample information into different workbooks, forming a data table with inverted index retrieval capabilities. It establishes a fishery resource sample database, supporting multi-condition combined queries to quickly locate any sample. It is low-cost, suitable for large-scale sample management, and significantly improves sample search speed and accuracy.

[0025] Preferably, step S105 specifically includes the following steps: Obtain retrieval tasks for fisheries research, and extract the target fisheries resource samples and corresponding target search conditions required for fisheries research through retrieval tasks; A retrieval vector space is constructed, and the k-means clustering algorithm is introduced into the retrieval vector space to perform clustering calculations on the storage index vectors corresponding to all sample information entered into the intelligent retrieval database, thereby obtaining the k retrieval cluster centers of the intelligent retrieval database. Calculate the Mahalanobis distance between the storage index vector of different sample information in the intelligent retrieval database and the center of each retrieval cluster, and extract the autonomous inverted list structure of each retrieval cluster center; If the Mahalanobis distance is less than the preset Mahalanobis distance, then the center of the retrieval cluster is designated as the candidate retrieval bucket, and the stored index vector is assigned to the autonomous inverted list structure corresponding to the candidate retrieval bucket. Repeat the above Mahalanobis distance determination allocation steps until all the storage index vectors of the intelligent retrieval database are allocated, and obtain the inverted index list of the intelligent retrieval database for the retrieval of information from different sample storage locations; Obtain the potential retrieval vector of the target retrieval condition, locate one or more candidate retrieval buckets that are closest to the potential retrieval vector based on Mahalanobis distance, and mark them as the target vector retrieval domain; The correlation between the potential retrieval vector and each stored index vector recorded in the inverted index list corresponding to the retrieval domain of each target vector is calculated by grey relational degree, and several local retrieval correlation degrees are obtained. Only the sample information corresponding to the storage index vector with the maximum local retrieval correlation is extracted, and the cabinet number-column number-box number three-level coding sequence of the sample information is retrieved through the intelligent retrieval database. The target fishery resource sample is quickly located and retrieved based on the cabinet number-column number-box number three-level coding sequence.

[0026] It should be noted that when retrieving a specific fishery resource sample, the corresponding search conditions are entered into the intelligent retrieval database. However, traditional database retrieval methods lack vector retrieval engines that support indexing by time, location, cabinet number, column number, box number, or sample number. This makes it difficult for the intelligent retrieval database to accurately locate the specific sample, reducing the accuracy and reliability of intelligent sample retrieval. To address this, this method encodes the storage information of each sample in the intelligent retrieval database into a storage index vector. Then, k-means clustering is performed on all sample vectors in the retrieval vector space to generate k retrieval cluster centers. The retrieval cluster centers define and plan the search interval for the search conditions, thus clarifying the retrieval distribution structure of the storage index vectors. Next, the storage index vectors of different sample information are assigned to the autonomous inverted list structure of the nearest retrieval cluster center according to the minimum Mahalanobis distance, i.e., candidate retrieval bucketing. This further buckets all vectors, ensuring that retrieval of similar samples falls into the same or similar clusters. This allows subsequent sample retrieval to be conducted only within a few search ranges, eliminating the need for a global search across all vectors and improving the orientation of the intelligent database for different target retrieval conditions. The autonomous inverted index structure shifts the retrieval of index vectors from a global to a local approach, enabling rapid matching of sample information to vectors within clusters (buckets), thus improving retrieval hit rate and correspondence rate. The target retrieval conditions are then converted into potential retrieval vectors. Based on the previously obtained Mahalanobis distance, the target retrieval domain closest to the center of each retrieval cluster is found, thereby locking in the retrieval range most likely containing the target sample based on the sample information distribution. Local retrieval correlation measures the relevance and matching degree between the input target retrieval conditions and the corresponding index vectors of each sample in the intelligent database. The index vector with the highest local retrieval correlation is the sample information that meets the target retrieval conditions. This method enables the intelligent retrieval database to perform high-precision, directional matching retrieval of input retrieval conditions using stored vectors, thereby accurately locating the location (cabinet number, column number, box number) of the target fishery resource sample, achieving rapid location and retrieval, and solving problems such as inaccurate information retrieval, incorrect sample retrieval and location, low management efficiency, and difficulty in searching in traditional methods.

[0027] Preferably, step S106 specifically includes the following steps: For each sample box in each sample storage cabinet, three fishery resource samples are randomly selected and marked as random check samples. The corresponding sample information of the random check samples is then imported into the intelligent retrieval database for retrieval. Obtain the storage management requirements for randomly checked samples, and extract one or more main verification factors for the randomly checked samples based on the storage management requirements; Construct a temporary verification structure, extract the library retrieval fields that meet each main verification factor and the corresponding library field value from the intelligent retrieval data network library, and write each library retrieval field into the temporary verification structure based on the library field value to generate a temporary verification blueprint group for the main verification factor. Obtain the actual retrieval field of each random verification sample on each main verification factor and the actual field value corresponding to each actual retrieval field, and construct the actual verification mapping group of the main verification factor based on the actual retrieval field and the actual field value; Project the actual verification mapping group onto the temporary verification blueprint group, and calculate the mapping consistency rate of the actual verification mapping group compared to the temporary verification blueprint group; If the mapping consistency rate is higher than the preset mapping consistency rate threshold, the randomly checked sample will be put back into the sample box and confirmed as being in storage. If there is a discrepancy, the information of the randomly checked sample and the intelligent retrieval database will be corrected and rechecked until they are completely consistent before being confirmed for entry into the database.

[0028] It should be noted that traditional sample management methods often rely on manual recording and searching, which can easily lead to inconsistencies between sample labels and electronic data, as well as non-standard sample information recording. Therefore, regular verification of samples upon entry into the warehouse is necessary. However, traditional sample entry verification typically involves checking each of the main verification factors one by one. These main verification factors include: whether the sample number matches, whether the collection location and time are consistent, and whether the storage cabinet, column, and box information correspond. This undoubtedly increases the workload of manual sample management verification, especially when the number of samples is large, reducing the efficiency of sample information management and increasing the likelihood of verification errors and incorrect entry. To address this, this method employs a quantitative verification approach for each sample box, randomly selecting three fishery resource samples and retrieving them from an intelligent retrieval database. A temporary verification structure is then constructed. This temporary verification structure is a range-based, coordinated space that responds to verification data corresponding to different main verification factors. Through field distribution, verification data belonging to the same entity are logically aggregated, directly presenting the comparison relationship and ensuring the comparability of sample and database data verification.

[0029] The following table shows an example of temporary check structure data:

[0030] Based on the prescribed primary verification factors, the corresponding database retrieval fields and their values ​​are extracted from the intelligent retrieval database and written into a temporary verification structure, forming a temporary verification blueprint group for the primary verification factors. This temporary verification blueprint group reflects the sample storage information currently recorded in the intelligent retrieval database; while the actual verification mapping group reflects the actual storage information of the samples. If the mapping consistency rate is higher than the preset mapping consistency rate threshold, it indicates that the mapping of all verification data groups in the mapping group is completely consistent. Therefore, the randomly checked samples are returned to the sample box and confirmed as being in the database. Conversely, if the mapping consistency rate is lower than the threshold, it indicates that at least one or more verification data groups have a verification inconsistency, such as the sample number not matching the sample number recorded in the intelligent database. Therefore, information correction and re-verification until complete consistency are required before confirmation of database entry. This method enables the design of a pre-entry sampling comparison verification process, establishing a sampling verification mechanism to randomly select a quantitative sample from each sample box and retrieve and verify the sample information in the intelligent retrieval database. This effectively prevents data entry errors, ensures the consistency and accuracy of data and samples, and improves the accuracy of the sample database.

[0031] A second aspect of this invention provides an intelligent database retrieval and management system for fishery resource samples, such as... Figure 3 As shown, the system includes: a memory 301, a processor 302, and a communication interface 303. The memory 301 includes a program for intelligent database retrieval and management of fishery resource samples. The communication interface 303 is used for data connection and communication between the memory 301 and the processor 302. When the program for intelligent database retrieval and management of fishery resource samples is executed by the processor 302, it implements any of the steps of the intelligent database retrieval and management method described above.

[0032] The above are merely specific embodiments 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 intelligent online database retrieval and management of fishery resource samples, characterized in that, Includes the following steps: S101: Obtain multidimensional collection data information of fishery resource samples and construct a multidimensional label coding layer. Based on the label dimension, perform semantic-specific label coding on the multidimensional collection data information on the multidimensional label coding layer, and affix a unique coding label to each fishery resource sample to obtain sample retrieval labels for different fishery resource samples. S102: All sample storage cabinets in the sample library are uniformly numbered, and a category search tree architecture is constructed from the root class node of the sample library according to the cabinet structure and the parent-child-sub-level category gradient. The category search tree architecture is used to query the classification and storage of different fishery resource samples and encode them in three levels to obtain the three-level coding sequence of cabinet number-column number-box number for the classification and storage of fishery resource samples. S103: Based on the three-level coding sequence of cabinet number-column number-box number, different fishery resource samples are stored in the corresponding sample boxes, and the sample retrieval tag is bound to the three-level coding sequence of cabinet number-column number-box number to form sample information; S104: Divide the information of different types of samples into different storage workbooks and enter it into an inverted index Excel spreadsheet to establish an intelligent retrieval database of fishery resource samples; S105: Obtain the target fishery resource samples and target retrieval conditions required for fishery research. By inputting the target retrieval conditions into the intelligent retrieval database, the storage location of the target fishery resource samples is searched in reverse order to achieve rapid location and retrieval of the target fishery resource samples. S106: Randomly select three samples from each sample box and search them in the intelligent retrieval database. Verify the consistency of the sample information storage and verify the randomly selected fishery resource samples and sample boxes before storing them in the database. S107: Regularly back up and manage the intelligent retrieval database, and conduct secondary verification management on no less than 5% of the samples already in the database to ensure the accuracy of the data under long-term operation of the sample database.

2. The intelligent database retrieval and management method for fishery resource samples according to claim 1, characterized in that, S101 specifically includes the following steps: The data collected from fishery resource samples includes multidimensional data such as collection location, collection time, sample name, and sample number. Based on the dimensional characteristics of multidimensional collected data, a multidimensional label coding layer and the granularity limit of each multidimensional label coding layer are constructed. At the same time, a word sense feature extraction algorithm is introduced to create a word sense feature encoder. Based on the granularity limit, a word sense feature encoder is used on each multidimensional label encoding layer to perform average pooling calculation on the word sense vectors contained in the multidimensional acquired data information and map them to integer indices, so as to obtain the independent discrete feature encoding form of each multidimensional label encoding layer for each multidimensional acquired data information. By acquiring multi-dimensional collected data information through big data, a semantic encoding character library is obtained for different label dimensions such as collection location, collection time, sample name, and sample number. The character bit width of the semantic encoding character library is then used to determine the encoding standard for different label dimensions. Based on the character bit width and following the discrete feature encoding form, the continuous bit stream of tag meaning in the multidimensional acquired data information is grouped and segmented to obtain several binary bit groups corresponding to the tag meaning of the multidimensional acquired data information; By querying the coded characters of each binary bit group through the semantic coded character library index, the coded character sequence of each multidimensional data information located in the corresponding label dimension is output. Finally, all the coded character sequences of the label dimension are combined to obtain the exclusive coded string of each multidimensional data information. Each sample is compiled and affixed with a unique coded label based on its collection location, collection time, sample name, and sample number, thus obtaining sample retrieval labels for different fishery resource samples.

3. The intelligent database retrieval and management method for fishery resource samples according to claim 1, characterized in that, S102 specifically includes the following steps: Obtain the classification numbering rules for different types of sample storage and the overall sample library. Based on the classification numbering rules, assign a unified number to all sample storage cabinets in the sample library to obtain the cabinet type number for different sample storage cabinets. Construct the root class node for the storage and query of fishery resource sample types based on the overall sample library. Based on the cabinet structure of the sample storage cabinet, a three-level splitting strategy is set up to classify fishery resource samples according to the root class from parent level, child level to subclass, and the category storage function of the sample storage cabinet is obtained. The cabinet structure has three rows of cabinets for each sample storage cabinet, each row of cabinets has 20 sample boxes, and each sample box contains several sample containers labeled with numbers. Anchor the starting source point with the root class node, and traverse the key input position from the root class node based on the category storage function to generate the storage key node of the sample storage cabinet. Based on the three-order splitting strategy, the storage key node is split to continuously generate category leaf nodes, forming a category search tree architecture of sample storage cabinet-cabinet queue-sample box. The parent category keywords, sub-category keywords, and sub-category keywords of different fishery resource samples are identified and obtained through the knowledge graph of fishery resource samples. The target bytes of the corresponding category keywords of fishery resource samples are scanned and queried in the order of parent, child and sub-level from the root node through the category search tree architecture to obtain the category byte query results. Based on the category byte query results, the parent category byte, child category byte, and sub-category category byte of the fishery resource sample type are found inside the root category node, and the target sample storage cabinet, target cabinet queue, and target sample box for storing different fishery resource sample categories are determined. Based on the target sample storage cabinet, target cabinet queue, and target sample box, samples are classified and stored, and a three-level system coding is performed to obtain the three-level coding sequence of cabinet number-column number-box number for the classification and storage of fishery resource samples.

4. The intelligent database retrieval and management method for fishery resource samples according to claim 1, characterized in that, S104 specifically includes the following steps: Different fishery resource samples are stored in corresponding sample boxes according to the three-level coding sequence of cabinet number-column number-box number, while ensuring that the sample number recorded on the sample retrieval label is consistent with the label number of the sample container. After storage, the forward index mode of the Excel spreadsheet is called to bind the sample retrieval tag with the cabinet number-column number-box number three-level coding sequence as a prefix index term to form the sample information of the fishery resource sample, and to create a storage workbook for different fishery resource samples. Import the stored workbook into the forward index mode and reorganize it according to the prefix index terms, mapping it backwards to the stored workbook. Generate an Excel spreadsheet containing the inverted index list for the prefix retrieval of corresponding sample information for different fishery resource samples. Based on the online loading of the inverted index list, a data table format that can be quickly retrieved is formed, and an intelligent retrieval database of fishery resource samples is established.

5. The intelligent database retrieval and management method for fishery resource samples according to claim 1, characterized in that, S105 specifically includes the following steps: Obtain retrieval tasks for fisheries research, and extract the target fisheries resource samples and corresponding target search conditions required for fisheries research through retrieval tasks; A retrieval vector space is constructed, and the k-means clustering algorithm is introduced into the retrieval vector space to perform clustering calculations on the storage index vectors corresponding to all sample information entered into the intelligent retrieval database, thereby obtaining the k retrieval cluster centers of the intelligent retrieval database. Calculate the Mahalanobis distance between the storage index vector of different sample information in the intelligent retrieval database and the center of each retrieval cluster, and extract the autonomous inverted list structure of each retrieval cluster center; If the Mahalanobis distance is less than the preset Mahalanobis distance, then the center of the retrieval cluster is designated as the candidate retrieval bucket, and the stored index vector is assigned to the autonomous inverted list structure corresponding to the candidate retrieval bucket; Repeat the above Mahalanobis distance determination allocation steps until all the storage index vectors of the intelligent retrieval database are allocated, and obtain the inverted index list of the intelligent retrieval database for the retrieval of information from different sample storage locations; Obtain the potential retrieval vector of the target retrieval condition, locate one or more candidate retrieval buckets that are closest to the potential retrieval vector based on Mahalanobis distance, and mark them as the target vector retrieval domain; The correlation between the potential retrieval vector and each stored index vector recorded in the inverted index list corresponding to the retrieval domain of each target vector is calculated by grey relational degree, and several local retrieval correlation degrees are obtained. Only the sample information corresponding to the storage index vector with the maximum local retrieval correlation is extracted, and the cabinet number-column number-box number three-level coding sequence of the sample information is retrieved through the intelligent retrieval database. The target fishery resource sample is quickly located and retrieved based on the cabinet number-column number-box number three-level coding sequence.

6. The intelligent database retrieval and management method for fishery resource samples according to claim 1, characterized in that, S106 specifically includes the following steps: For each sample box in each sample storage cabinet, three fishery resource samples are randomly selected and marked as random check samples. The corresponding sample information of the random check samples is then imported into the intelligent retrieval database for retrieval. Obtain the storage management requirements for randomly checked samples, and extract one or more main verification factors for the randomly checked samples based on the storage management requirements; Construct a temporary verification structure, extract the library retrieval fields that meet each main verification factor and the corresponding library field value from the intelligent retrieval data network, and write each library retrieval field into the temporary verification structure based on the library field value to generate a temporary verification blueprint group for the main verification factor. Obtain the actual retrieval field of each random verification sample on each main verification factor and the actual field value corresponding to each actual retrieval field. Construct the actual verification mapping group of the main verification factor based on the actual retrieval field and the actual field value. Project the actual verification mapping group onto the temporary verification blueprint group, and calculate the mapping consistency rate of the actual verification mapping group compared to the temporary verification blueprint group; If the mapping consistency rate is higher than the preset mapping consistency rate threshold, the randomly checked sample will be put back into the sample box and confirmed as being in storage. If there is a discrepancy, the information of the randomly checked sample and the intelligent retrieval database will be corrected and rechecked until they are completely consistent before being confirmed for entry into the database.

7. A smart database retrieval and management system for fishery resource samples, characterized in that, The system includes: a memory, a processor, and a communication interface. The memory includes a program for intelligent database retrieval and management of fishery resource samples. The communication interface is used for data connection and communication between the memory and the processor. When the program for intelligent database retrieval and management of fishery resource samples is executed by the processor, it implements the steps of the intelligent database retrieval and management method as described in any one of claims 1-6.

Images