Animal individual information record tagging system and method for zoo

Abstract

The invention discloses an animal individual information record tagging system and method for a zoo, and belongs to the field of animal data processing. A globally unique ID coding rule and a verification algorithm are adopted to ensure identifier uniqueness and support cross-system data connection; multi-source sensing equipment and a deep learning model are integrated, and automatic feature extraction and high-precision label generation are realized; multi-dimensional correlation analysis and complex retrieval are supported through storage of a hierarchical label system and a graph database; and the data security and the operation traceability are ensured in combination with an authority control and auditing mechanism. According to the scheme, the animal information management efficiency is remarkably improved, individual health monitoring, population optimization management and scientific research decision are supported, and zoo management is promoted to proceed towards the intelligentization and standardization direction.

Description

Technical Field

[0001] The present invention belongs to the technical field of animal data processing, and in particular relates to a system and method for recording and labeling individual animal information in zoos. Background Art

[0002] In zoo animal management, existing technologies mostly rely on scattered paper records or isolated electronic systems, which are subject to problems such as information fragmentation, difficulty in integration, and low retrieval efficiency. Traditional methods usually use basic identification (such as simple numbers) combined with manual records, which makes it difficult to accurately track information on individual animals throughout their life cycle. The data collection dimension is single and lacks the ability to integrate multimodal features, resulting in insufficient individual identification accuracy and an inability to support complex behavioral analysis and health monitoring. In addition, existing systems generally lack a unified identification system, making data difficult to communicate between different parks and hindering cross-system information sharing, which restricts the standardization of animal management and the efficiency of scientific research collaboration.

[0003] Therefore, the core problems of existing technologies include: (1) the animal individual identification system is imperfect and cannot achieve global uniqueness and cross-system compatibility; (2) there is a lack of effective integration and intelligent analysis methods for multi-source heterogeneous data (such as biometrics, behavioral data, and medical records); (3) label generation relies on manual experience, which is inefficient and prone to errors; (4) information retrieval and analysis functions are weak, making it difficult to meet the needs of precise management and scientific research. Summary of the Invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present application provides a system and method for recording and labeling individual animal information in zoos.

[0005] In a first aspect, the present application proposes a system for recording and labeling individual animal information for a zoo, comprising a data collection layer, a data processing layer, an application service layer, a communication network, and a user terminal;

[0006] The data acquisition layer, consisting of image acquisition equipment, biometric acquisition equipment, behavior monitoring equipment and manual input terminals, is used to collect multimodal data of individual animals;

[0007] The data processing layer includes a unique ID generation module, a tag system construction module, a tag extraction module, and a tag management and storage module, which are used to generate a globally unique ID, extract features from raw data, generate tags based on the tag system, and associate and store the tags with the unique ID;

[0008] The unique ID generation module adopts a hierarchical partitioning strategy to generate unique IDs by species, year and serial number, and includes a check code generation algorithm to prevent entry errors;

[0009] The label extraction module performs multi-modal feature fusion through a deep learning model and a rule engine, and realizes automatic label extraction according to the feature fusion result.

[0010] The application service layer includes a retrieval module and a data analysis module, and is configured to provide label retrieval and analysis functions.

[0011] The communication network is configured to connect the data acquisition layer, the data processing layer and the application service layer, support wired and wireless backup, and have a network security protection mechanism.

[0012] The user terminal includes a management center large screen, a workstation terminal and a mobile device, and is configured to display park animal profiles and transmit important information to management personnel.

[0013] In some embodiments, the unique ID generation module is configured to generate and assign animal unique IDs for individual animals according to an ID generation rule, the ID generation rule being that each individual animal is assigned a unique ID in the format SSPP-YYYY-NNNN-C, where SSPP is a combination of species code and subspecies code, YYYY is the birth year, capture year or park entry year, NNNN is a serial number, and C is a check code generated based on a weighted algorithm, the check code being generated by extracting the numerical part of the ID, applying a preset weight factor to calculate a weighted sum, and mapping the result of the modulo 11 operation to the check code.

[0014] In some embodiments, the unique ID generation module further includes a distributed ID assignment unit configured to ensure global uniqueness through pre-assignment of ID segments and a synchronization mechanism.

[0015] In some embodiments, the label system construction module includes a basic information label definition unit and a feature information label definition unit.

[0016] The basic information label definition unit is configured to define species labels, individual labels, source labels and blood relationship labels, the species labels including species classification and subspecies information, the individual labels including name, unique ID, gender, birth date and age range, the source labels including birthplace, capture location and transfer-in source, and the blood relationship labels including parent information, offspring information and blood purity.

[0017] The feature information label definition unit is configured to define appearance feature labels, biological feature labels and behavior habit labels, the appearance feature labels including body shape, coat color and pattern, the biological feature labels including iris, footprint and sound, and the behavior habit labels including animal state, social relationship and living habit.

[0018] In some embodiments, the label extraction module comprises an image feature extraction unit, a biological feature extraction unit and a behavior feature extraction unit:

[0019] The image feature extraction unit is configured to extract animal appearance features using an improved ResNet-50 or EfficientNet-B3 model, and generate appearance feature labels including body shape, coat color and pattern through target detection and image segmentation;

[0020] The biological feature extraction unit is configured to generate biological feature labels including iris, footprint and voice through iris recognition, footprint recognition and RFID reading;

[0021] The behavior feature extraction unit is configured to use a time series pattern mining algorithm and social network analysis to generate behavior habit labels including animal state, social relationship and living habit.

[0022] In some embodiments, the label management storage module is configured to store label association information in a graph database, record the creation, modification and deletion history of labels, support time dimension query, wherein the label association information comprises taking the animal unique ID as a node, taking the basic information label and the feature information label as an attribute relationship, and storing the animal unique ID in correspondence with the basic information label and the feature information label.

[0023] In some embodiments, the retrieval module is configured to perform multi-dimensional retrieval of package accurate matching, fuzzy query, composite logic retrieval and similarity retrieval on the stored label association information;

[0024] The data analysis module is configured to calculate the target label similarity based on cosine similarity, Jaccard coefficient and weighted fusion algorithm, and detect target label anomaly through modified Z-score method.

[0025] The second aspect of the present application proposes an animal individual information record labeling method for zoos, comprising the following steps:

[0026] A global unique ID is generated for each animal, and its multi-modal data is associated;

[0027] Animal information is collected through sensors and manual input terminals, features are extracted, and corresponding multi-dimensional labels are generated;

[0028] The generated labels are stored in association with the animal unique ID, and are regularly updated to maintain timeliness.

[0029] The third aspect of the present application proposes an electronic device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to realize the steps of the above method.

[0030] In a fourth aspect, the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0031] Beneficial effects of the present invention:

[0032] This solution uses a globally unique ID coding scheme (SSPP-YYYY-NNNN-C) and verification algorithm to ensure unique identification and support cross-system data integration. It integrates multi-source sensor equipment (such as RFID, 3D scanners, and biometric devices) with deep learning models to achieve automated feature extraction and high-precision tag generation. A layered tagging system (basic information, biometrics, behavioral habits, etc.) and graph database storage support multi-dimensional correlation analysis and complex retrieval. Furthermore, combined with permission control and audit mechanisms, it ensures data security and operational traceability. This solution significantly improves the efficiency of animal information management, supports individual health monitoring, population optimization management, and scientific research decision-making, and promotes intelligent and standardized zoo management. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is the overall flow chart of the present invention.

[0034] Figure 2 This is a system principle block diagram of the present invention. DETAILED DESCRIPTION

[0035] The following will describe exemplary embodiments of the present invention in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein; rather, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0036] In the first aspect, the present application proposes a system for recording and labeling individual animal information in a zoo, such as Figure 1 As shown, it includes data acquisition layer, data processing layer, application service layer, communication network and user terminal;

[0037] The data acquisition layer, consisting of image acquisition equipment, biometric acquisition equipment, behavior monitoring equipment and manual input terminals, is used to collect multimodal data of individual animals;

[0038] Among them, image acquisition equipment includes high-definition cameras, infrared cameras and 3D scanners, which are installed in animal activity areas. High-definition cameras are used to collect animal appearance characteristics, infrared cameras are used for night monitoring and thermal characteristics collection, and 3D scanners are used to collect three-dimensional data of animal body shape and surface characteristics. Portable filming equipment is also included: used for close-up photography during veterinary examinations;

[0039] Biometric collection equipment includes RFID readers: for reading RFID chips implanted in animals, biometric identification equipment: for collecting biometrics such as irises and footprints of animals, and DNA sampling tools: for collecting DNA samples from animals;

[0040] Behavioral monitoring equipment includes motion sensors that monitor animal activity patterns, sound collection devices that record animal vocalizations, and pressure sensors installed in specific areas to record animal weight and walking characteristics.

[0041] Manual input terminals include mobile terminals: tablet computers or mobile phones used by breeders, fixed workstations: computer terminals used by veterinarians and managers, and voice input devices: support voice recording of information.

[0042] The data processing layer includes a unique ID generation module, a tag system construction module, a tag extraction module, and a tag management and storage module, which are used to generate a globally unique ID, extract features from raw data, generate tags based on the tag system, and associate and store the tags with the unique ID;

[0043] The unique ID generation module adopts a hierarchical partitioning strategy to generate unique IDs by species, year and serial number, and includes a check code generation algorithm to prevent entry errors;

[0044] In some embodiments, the unique ID generation module is used to generate and assign an animal unique ID for an individual animal through an ID generation rule, and the ID generation rule is: assign a unique ID to each individual animal in the format: SSPP-YYYY-NNNN-C, where SSPP is a combination of the species code and the subspecies code, YYYY is the year of birth, capture year or admission year, NNNN is the serial number, and C is a check code generated based on a weighted algorithm, and the check code is generated by the following steps: extracting the digital part of the ID, applying a preset weight factor to calculate the weighted sum, and mapping it to the check code after taking the modulus of 11.

[0045] The ID allocation method is as follows:

[0046] Format: SSPP-YYYY-NNNN-C;

[0047] SS: species code (2 letters, such as TI for tiger);

[0048] PP: subspecies code (2 letters, such as BE for Bengal tiger);

[0049] YYYY: year of birth / enrollment (4 digits);

[0050] NNNN: serial number (4 digits);

[0051] C: Check code (1 letter or number);

[0052] For example: TIBE-2023-0001-A means the first Bengal tiger in 2023;

[0053] The weighted algorithm for generating the check code includes:

[0054] Assign different weights to each digit of the ID number and calculate the modulo 11 remainder of the weighted sum;

[0055] Weight design basis: using a decreasing weight sequence [7, 6, 5, 4, 3, 2, 1, 1];

[0056] Higher weights are assigned to higher positions: Year information (YYYY) uses higher weights of 7, 6, 5, and 4 to strengthen the protection of important identification information.

[0057] Moderate median weight: The high digits of the sequence number use a medium weight of 3,2 to balance the weight distribution;

[0058] The lower bits of the sequence number have a lower weight of 1,1 to keep the algorithm simple.

[0059] Verification code mapping rules:

[0060] Remainders 0-9 are mapped to characters "0"-"9"

[0061] The remainder 10 is mapped to the character "A"

[0062] Error detection capabilities:

[0063] Can detect 100% of single-character errors;

[0064] Can detect 98.9% of adjacent digit swap errors;

[0065] Detects over 90% of common multi-character errors.

[0066] Special ID Rules:

[0067] Wild-caught animals: The year represents the year of capture, with a W prefix (e.g., WTIBE-2023-0001-A);

[0068] Animals born in the zoo: The year represents the birth year, and the prefix is added with a B marker (e.g., BTIBE-2023-0001-A);

[0069] Animals transferred in: The original ID is kept, a transfer record is added, and it is internally mapped to a new ID by the system;

[0070] Specifically, the ID check code calculation formula is:

[0071]

[0072] where, represents the i-th digit of the ID number part, represents the weight factor of the i-th digit, represents the number of digits of the ID number part, represents the modulo operation, which calculates the remainder when divided by 11, represents the check code, ranging from 0 to 10, where 10 is mapped to the character "A".

[0073] Example: Generate a unique ID check code for a new Bengal tiger entering the zoo

[0074] Basic information: Species code: TI (Tiger), Subspecies code: BE (Bengal), Year of entry: 2023, Serial number: 0001 (the first of the year);

[0075] Step 1: Combine the ID basic part;

[0076] ID basic part: TIBE-2023-0001;

[0077] Step 2: Extract the number part for check code calculation;

[0078] Number part: 2023-0001 → "20230001";

[0079] Step 3: Apply the check code weight factor;

[0080] Weight factor: [7, 6, 5, 4, 3, 2, 1, 1] (high weight for high digits, low weight for low digits);

[0081] Step 4: Calculate the weighted sum;

[0082] Check sum = 2 × 7 + 0 × 6 + 2 × 5 + 3 × 4 + 0 × 3 + 0 × 2 + 0 × 1 + 1 × 1 = 37;

[0083] Step 5: Calculate the check code;

[0084]

[0085] Step 6: Final ID;

[0086] Final ID: TIBE-2023-0001-4;

[0087] Conclusion: The unique ID for this Bengal tiger is TIBE-2023-0001-4, where the check code "4" can detect single-character errors and most adjacent digit transposition errors during ID entry.

[0088] In some embodiments, the unique ID generation module further includes a distributed ID allocation unit for ensuring global uniqueness through pre-allocated ID segments and synchronization mechanisms.

[0089] The distributed ID allocation unit includes an ID conflict prevention mechanism, an ID backup and recovery mechanism, and an ID association mechanism.

[0090] The ID conflict prevention mechanism includes multiple verifications: database duplication confirmation before new ID generation, pre-generation rules: important species ID can be planned and reserved in advance, cross-system coordination: regular exchange of ID information with international animal databases to avoid cross-system conflicts, historical ID reuse rules: deceased / transferred animal IDs are not reused for at least 10 years to prevent historical record confusion.

[0091] The ID backup and recovery mechanism includes multi-level backup: daily incremental backup and weekly full backup of the ID database, distributed storage: core ID mapping table is backed up in multiple geographically isolated data centers, emergency recovery: ID service can be restored from backup systems within 30 minutes in case of main system failure.

[0092] The ID association mechanism includes physical association: association with RFID chips, ear tags, collars, etc., biological association: mapping with biological characteristics (such as stripe patterns), system association: interface with international animal database systems.

[0093] In some embodiments, the tag system construction module includes a basic information tag definition unit and a feature information tag definition unit.

[0094] The basic information tag definition unit defines species tags, individual tags, source tags, and blood relationship tags. The species tag includes species classification and subspecies information. The individual tag includes name, unique ID, gender, birth date, and age range. The source tag includes birthplace, capture location, and transfer source. The blood relationship tag includes parent information, offspring information, and blood purity.

[0095] The feature information label definition unit is configured to define appearance feature labels, biological feature labels, and behavior habit labels, wherein the appearance feature labels include body shape, fur color, and pattern, the biological feature labels include iris, footprint, and sound, and the behavior habit labels include animal state, social relationship, and living habit.

[0096] The label extraction module performs multi-modal feature fusion through a deep learning model and a rule engine, and automatically extracts labels according to a feature fusion result.

[0097] In some embodiments, the label extraction module includes an image feature extraction unit, a biological feature extraction unit, and a behavior feature extraction unit.

[0098] The image feature extraction unit is configured to extract animal appearance features using an improved ResNet-50 or EfficientNet-B3 model, and generate appearance feature labels including body shape, fur color, and pattern through target detection and image segmentation.

[0099] The biological feature extraction unit is configured to generate biological feature labels including iris, footprint, and sound through iris recognition, footprint recognition, and RFID reading.

[0100] The behavior feature extraction unit is configured to use a time series pattern mining algorithm and social network analysis to include animal state, social relationship, and living habit in the behavior habit labels.

[0101] The target detection and label generation are associated, specifically, animal posture detection is performed using an improved YOLOv5 model, key points (such as head, limbs, and torso) obtained by detection are used to generate a skeleton structure, and skeleton features are automatically mapped to body shape related labels: for example, body length ratio → body shape label ("tall", "stout", "thin", etc.), limb angle → animal state label ("standing", "lying", "alert", etc.), head position → living habit label ("foraging", "resting", "observing", etc.);

[0102] The image segmentation to label conversion mechanism is also included, specifically, fine segmentation is performed using DeepLabv3+, hair, skin, and other regions are identified at the pixel level, and segmented region features are mapped to labels: color histogram feature → fur color label (such as "mainly brown", "black and white", etc.), texture feature → pattern label (such as "stripes", "spots", "rings", etc.), region proportion → feature distribution label (such as "dark back", "light belly", etc.), individual unique marker mapping: using SIFT / ORB algorithm to extract feature points, identifying unique markers such as scars and missing parts through feature point anomaly detection, and automatically generating descriptive labels (such as "left ear notch", "right front limb scar", etc.);

[0103] Among them, taking stripe recognition as an example, the stripe pattern recognition algorithm includes stripe feature extraction and stripe matching similarity;

[0104] Stripe feature extraction includes the Gabor filter formula:

[0105]

[0106]

[0107]

[0108] in, and represents the pixel coordinates in the image, represents the wavelength, which determines the spatial frequency of the filter, Indicates the direction, the edge direction detected by the filter, represents the phase, determines the symmetry of the filter, represents the standard deviation, which determines the size of the Gaussian envelope. represents the aspect ratio, which determines the shape of the filter, and represents the rotated coordinate system, Indicates that the Gabor filter is at point ( , ) response value.

[0109] Example: Extracting tiger stripe features

[0110] Input: TIBE-2023-0003 tiger side view high-definition image (2048×1536 pixels)

[0111] Step 1: Preprocess the image: including grayscale, contrast enhancement, and size normalization to 512×384 pixels;

[0112] Step 2: Set the Gabor filter parameter group: wavelength ( ):[4,8,16](corresponding to stripes of different scales), direction( ):[0°,45°,90°,135°](covering stripes in different directions), phase( ):0, standard deviation( ): / 2 (adaptively adjusted according to wavelength), aspect ratio ( ):0.5;

[0113] Step 3: Apply Gabor filter bank: Apply 12 filter combinations (3 wavelengths × 4 directions) to the image, and each pixel position ( , )calculate 12 response maps are generated;

[0114] Step 4: Generate feature vector, divide each response map into an 8x8 grid, calculate the average response value of each grid area, combine all grid values to form a 768-dimensional feature vector (12x8x8), normalize the feature vector to obtain the final stripe feature vector;

[0115] Output: 768-dimensional stripe feature vector, describing the unique stripe pattern of the tiger.

[0116] Further, the stripe matching similarity is:

[0117]

[0118]

[0119]

[0120] where, represents the first stripe feature vector , represents the second stripe feature vector , represents the dot product of the vectors, represents Euclidean norm, represents Euclidean norm, represents the vector similarity, ranging from [-1, 1], the closer the value to 1, the more similar.

[0121] Example: Compare the stripe feature similarity of two tigers;

[0122] For simplicity, use a 3-dimensional simplified feature vector:

[0123] Tiger A (TIBE-2022-0007) feature vector: [0.8, 0.6, 0.4];

[0124] Tiger B (TIBE-2023-0008) feature vector: [0.7, 0.5, 0.3]

[0125] Step 1: Calculate the dot product of the vectors:

[0126]

[0127] Step 2: Calculate the vector length:

[0128]

[0129] Step 3: Calculate the cosine similarity:

[0130]

[0131] Conclusion: The stripe patterns of the two tigers are almost identical, with a similarity of 99.9%.

[0132] Further comprising a multi-modal feature fusion label generation process, specifically multi-sensor time series data fusion: using long short-term memory network (LSTM) to process time series data, extract behavior patterns, generate habit-related labels, detect behavior anomalies, and trigger animal state label updates.

[0133] In some embodiments, the label management storage module is used to store label association information in a graph database, record the creation, modification and deletion history of labels, and support time dimension queries, wherein the label association information includes taking the animal unique ID as a node, taking the basic information label and the feature information label as attribute relationships, and corresponding associating and storing the animal unique ID with the basic information label and the feature information label.

[0134] The label management storage module stores examples as follows:

[0135] The application service layer includes a retrieval module and a data analysis module, used to provide label retrieval and analysis functions;

[0136] Step 1: Create an animal node:

[0137] / / Create a node representing tiger "Zhuangzhuang" in the graph database

[0138] CREATE (:Animal {

[0139] uid: "TGRR-2020-0123-9", / / Global unique ID

[0140] name: "Zhuangzhuang",

[0141] species: "Northeast Tiger"

[0142] });

[0143] Step 2: Add basic information labels (relationship edges):

[0144] / / Add blood relationship labels (connect parent nodes)

[0145] MATCH (zhuang:Animal {uid:"TGRR-2020-0123-9"})

[0146] MATCH (father:Animal {uid:"TGRR-2015-0456-2"})

[0147] MATCH (mother:Animal {uid:"TGRR-2016-0789-5"})

[0148] CREATE (zhuang)-[:FATHER]->(father),

[0149] (zhuang)-[:MOTHER]->(mother);

[0150] / / Add source tag

[0151] CREATE (zhuang)-[r:SOURCE]->(:Label {

[0152] type: "origin",

[0153] value: "Changbai Mountain Nature Reserve",

[0154] create_time: "2020-05-01"

[0155] });

[0156] Step 3: Add feature tags (dynamic update):

[0157] / / Add biometrics after the 2023 physical examination

[0158] MATCH (zhuang:Animal {uid:"TGRT-2020-0123-9"})

[0159] CREATE (zhuang)-[r:HAS_FEATURE]->(:Label {

[0160] type: "iris features",

[0161] data: "iriscode_8D7F...",

[0162] update_time: "2023-08-15"

[0163] });

[0164] / / Updated fur color changes in 2024

[0165] MATCH (zhuang)-[old:HAS_FEATURE]->(feature)

[0166] WHERE feature.type = "Coat color"

[0167] DELETE old

[0168] CREATE (zhuang)-[:HAS_FEATURE]->(:Label {

[0169] type: "hair color",

[0170] value: "Orange and yellow stripes deepen",

[0171] update_time: "2024-03-10"

[0172] });

[0173] Time dimension query example:

[0174] / / Query all features of "Zhuangzhuang" in 2023

[0175] MATCH (a:Animal {uid:"TGRR-2020-0123-9"})-[r]->(l)

[0176] WHERE r.create_time >= "2023-01-01" AND r.create_time <= "2023-12-31";

[0177] RETURN l.type AS feature type, l.value AS detailed information.

[0178] In some embodiments, the retrieval module is used to perform a multi-dimensional search including exact matching, fuzzy query, compound logic search and similarity search on the stored tag association information;

[0179] Precise search supports direct search by unique ID and precise matching query by tag, for example, searching for animals with "male" + "over 5 years old" + "Bengal tiger";

[0180] Fuzzy search supports range queries and fuzzy matching of tag values, for example, searching for animals weighing between 150kg and 200kg.

[0181] Compound search: supports multi-label combination query, uses AND, OR, NOT logic, and supports nested query conditions;

[0182] Similarity search is to find similar animals based on tag similarity, for example, to find other animals (of the same species) with "appearance feature similarity > 80%" to a specific animal (of the same species).

[0183] The data analysis module is configured to calculate target label similarity based on cosine similarity, Jaccard coefficient and weighted fusion algorithm, and detect target label anomaly by correcting Z-score.

[0184] The label similarity calculation algorithm includes a numerical label similarity calculation.

[0185] The normalization formula is:

[0186]

[0187] represents the numerical value of the label to be analyzed, represents the minimum reference value of the attribute, represents the maximum reference value of the attribute, represents the normalized value, ranging from 0 to 1;

[0188] Z-score standardization:

[0189]

[0190] wherein, represents the mean of the label data set, represents the standard deviation of the label data set, represents the standardized value, representing the standard deviation multiple of the deviation from the mean;

[0191] Numerical similarity calculation:

[0192]

[0193] represents the normalized value of the first label value, represents the normalized value of the second label value, represents the numerical similarity, ranging from 0 to 1, and the larger the value, the more similar.

[0194] Example: Calculate the weight similarity of two Bengal tigers:

[0195] Tiger A (TIBE-2023-0001) weighs 180 kg;

[0196] Tiger B (TIBE-2023-0002) weighs 160 kg;

[0197] Reference range: The weight range of adult Bengal tigers is 100-200 kg

[0198] Step 1: Normalize the weight value:

[0199]

[0200] Step 2: Calculate similarity:

[0201]

[0202] Conclusion: The two Bengal tigers have 80% similarity in weight characteristics;

[0203] Furthermore, it also includes category label similarity. Specifically, the Jaccard similarity coefficient is calculated as:

[0204]

[0205] in, Represents the first tag set, Represents the second tag set, Represents the number of elements in the intersection of two sets. Represents the number of elements in the union of two sets. Indicates label similarity, ranging from [0,1];

[0206] The weighted Jaccard similarity coefficient is calculated as

[0207]

[0208] in, represents the weight of label i in set A, represents the weight of label i in set B, represents the smaller of the two weights, represents the larger of the two weights, Indicates the similarity taking weight into account, ranging from [0,1].

[0209] Example: Calculating the similarity of behavioral labels of two snow leopards

[0210] Snow Leopard A (SNLE-2022-0005) behavior tag set: {eating (0.8), resting (0.6), running (0.9)}

[0211] Snow Leopard B (SNLE-2022-0008) behavior tag set: {eating (0.7), resting (0.8), climbing (0.9)}

[0212] Step 1: Identify shared and unique tags;

[0213] Total tags: {eating, resting};

[0214] All tags: {eating, resting, running, climbing};

[0215] Step 2: Calculate the simple Jaccard similarity coefficient:

[0216]

[0217] Step 3: Calculate the weighted Jaccard similarity coefficient;

[0218]

[0219] Conclusion: Considering the weights, the similarity of the two snow leopards in behavior characteristics is 38.2%.

[0220] Further, the target label anomaly detection by modifying the Z-score method includes:

[0221]

[0222] where, represents the target label data set, represents the median of the target label data set, represents the median absolute deviation, a more robust measure of dispersion for outliers, represents the label value to be detected, is a standardization constant, which makes the MAD correspond to the standard deviation of the standard normal distribution, represents the modified Z-score, which is used to judge the degree of anomaly.

[0223] Example: Detect the weight label value anomaly of Bengal tiger TIBE-2020-0003;

[0224] Historical weight records (kg): [165, 170, 168, 172, 169, 171, 190];

[0225] Latest measured weight: 190 kg

[0226] Step 1: Calculate the median:

[0227]

[0228] Step 2: Calculate the absolute value of the deviation of each measurement from the median:

[0229] ;

[0230] Step 3: Calculate the MAD (Median Absolute Deviation):

[0231]

[0232] Step 4: Calculate the modified Z-score of the latest weight:

[0233]

[0234] Step 5: Determine if there is an anomaly:

[0235] The threshold is set to 3.0 (based on statistical theory, more than 3.0 is considered to be a significant anomaly), since 6.745>3.0, it is determined that the weight is abnormal.

[0236] Conclusion: The system automatically generates a "weight abnormality" label and triggers a health monitoring alert to remind the veterinarian and the keeper to pay attention to the changes in the health of the tiger, and updates the weight label.

[0237] The communication network is used to connect the data acquisition layer, data processing layer and application service layer, supports wired and wireless backup, and has a network security protection mechanism;

[0238] The user terminal includes a management center large screen, a workstation terminal and a mobile device, and is used to display the general situation of animals in the park and transmit important information to management personnel.

[0239] The second aspect of the present application provides an animal individual information record labeling method for a zoo, as shown in Figure 2 The method comprises the following steps:

[0240] S100: Generate a globally unique ID for each animal and associate its multi-modal data;

[0241] S200: Collect animal information through sensors and manual input terminals, extract features and generate corresponding multi-dimensional labels;

[0242] S300: Store the generated label in association with the unique ID of the animal, and update it regularly to maintain timeliness.

[0243] The third aspect of the present application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to realize the steps of the above method.

[0244] The fourth aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the above method.

[0245] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0246] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0247] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0248] In the embodiments provided in the present disclosure, it should be understood that the disclosed apparatus / computer equipment and methods can be implemented in other ways. For example, the apparatus / computer equipment embodiments described above are merely schematic. For example, the division of modules or units is merely a logical function division. In actual implementation, there may be other division methods. Multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection of the apparatus or unit, which may be electrical, mechanical or other forms.

[0249] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0250] In addition, each function unit in each embodiment of the present disclosure can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0251] If the integrated module / unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment described above can be implemented. The computer program can include computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0252] The above is only the preferred embodiment of the present application, and it should be noted that the technical solutions of several modifications and improvements made by those skilled in the art without departing from the present technical solution should also be considered to fall within the scope of the present application.

Claims

1. A system for recording and labeling individual animal information in a zoo, characterized by: Includes data acquisition layer, data processing layer, application service layer, communication network and user terminal; The data acquisition layer consists of image acquisition equipment, biometric acquisition equipment, behavior monitoring equipment and manual input terminals, and is used to collect multimodal data of individual animals; The data processing layer includes a unique ID generation module, a tag system construction module, a tag extraction module, and a tag management and storage module, which are used to generate a globally unique ID, extract features from raw data, generate tags based on the tag system, and associate and store the tags with the unique ID; The unique ID generation module adopts a hierarchical partitioning strategy to generate unique IDs by species, year and serial number, and includes a check code generation algorithm to prevent entry errors; The label extraction module performs multimodal feature fusion through a deep learning model and a rule engine, and realizes automatic label extraction based on the feature fusion results; The application service layer includes a retrieval module and a data analysis module, which are used to provide tag retrieval and analysis functions; The communication network is used to connect the data acquisition layer, data processing layer and application service layer, supports wired and wireless backup, and has a network security protection mechanism; The user terminals include a large screen in the management center, workstation terminals and mobile devices, which are used to display an overview of the animals in the park and transmit important information to management personnel.

2. The method according to claim 1, wherein: The unique ID generation module is used to generate and assign an animal unique ID for an individual animal through an ID generation rule. The ID generation rule is: each individual animal is assigned a unique ID in the format of SSPP-YYYY-NNNN-C, where SSPP is a combination of a species code and a subspecies code, YYYY is the year of birth, capture year, or admission year, NNNN is a serial number, and C is a check code generated based on a weighted algorithm. The check code is generated by extracting the digital part of the ID, calculating a weighted sum using a preset weight factor, and mapping the result modulo 11 to the check code.

3. The system according to claim 2, characterized in that: The unique ID generation module further comprises a distributed ID allocation unit, which is configured to ensure global uniqueness through pre-allocated ID segments and a synchronization mechanism.

4. The system according to claim 3, wherein: The label system building module includes a basic information label definition unit and a feature information label definition unit; The basic information tag definition unit is used to define species tags, individual tags, source tags, and lineage tags. The species tag includes genus classification and subspecies information. The individual tag includes name, unique ID, gender, date of birth, and age range. The source tag includes birthplace, capture place, and transfer source. The lineage tag includes parent information, offspring information, and bloodline purity. The feature information tag definition unit is used to define appearance feature tags, biometric feature tags and behavioral habit tags. The appearance feature tags include body shape, fur color and pattern; the biometric feature tags include iris, footprints and sound; and the behavioral habit tags include animal status, social relationships and living habits.

5. The method according to claim 4, characterized in that: The tag extraction module includes an image feature extraction unit, a biometric feature extraction unit, and a behavior feature extraction unit: The image feature extraction unit is used to extract animal appearance features using an improved ResNet-50 or EfficientNet-B3 model, and generate appearance feature labels including body shape, fur color and pattern through target detection and image segmentation; The biometric feature extraction unit is used to generate a biometric feature tag including iris, footprint and voice through iris recognition, footprint recognition and RFID reading; The behavior feature extraction unit is used to use a temporal pattern mining algorithm and a social network to analyze behavior habit labels including animal status, social relationships, and living habits.

6. The method according to claim 5, characterized in that: The tag management storage module is used to use a graph database to store tag association information, record the creation, modification and deletion history of tags, and support time dimension query. The tag association information includes the animal's unique ID as a node, the basic information tag and the feature information tag as an attribute relationship, and the animal's unique ID is stored in a corresponding association with the basic information tag and the feature information tag.

7. The method according to claim 6, characterized in that: The retrieval module is used to perform multi-dimensional retrieval including exact matching, fuzzy query, compound logic retrieval and similarity retrieval on the stored tag association information; The data analysis module is used to calculate the target label similarity based on cosine similarity, Jaccard coefficient and weighted fusion algorithm, and detect target label anomalies through modified Z score method.

8. A method for recording and labeling individual animal information in a zoo, characterized by: The following steps are involved: Generate a globally unique ID for each animal and associate its multimodal data; Collect animal information through sensors and manual input terminals, extract features and generate corresponding multi-dimensional labels; The generated tags are associated with the animal's unique ID and stored, and updated regularly to maintain timeliness.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to claim 8 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to claim 8 are implemented.

Images