Method for calculating migratory fish sexual maturity cycle through PIT mark monitoring technology

Abstract

The invention provides a method for calculating a migratory fish sexual maturity cycle through a PIT mark monitoring technology, and relates to the technical field of PIT mark monitoring. Comprising the steps of determining starting time and ending time according to electronic tag endurance attributes; based on the starting time and the ending time, installing an electronic tag monitoring device and an environment monitoring sensor at a preset position of the target monitoring area to obtain environment monitoring data and an initial tag monitoring data set of each year in the target monitoring area; determining an initial sexual maturity cycle data set according to the initial label monitoring data set; and correcting the initial sexual maturity cycle data set according to the environmental monitoring data in each year to obtain a final sexual maturity cycle data set. According to the invention, the problem of low fish sexual cycle monitoring efficiency and precision in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of PIT mark monitoring, in particular to a method for estimating the sexual maturity period of migratory fishes by using the PIT mark monitoring technology. Background Art

[0002] The naked carp of Qinghai Lake (Pseudochondrostoma brama) is a unique fish species that primarily inhabits Qinghai Lake and the rivers within its basin. Since the late 20th century, its population has been declining, and it was previously listed as an endangered species. To protect this rare species, numerous conservation measures have been introduced, including tagging and releasing the species and monitoring their sexual cycles. However, existing technologies rely primarily on traditional tagging and manual monitoring methods, which have numerous limitations in identifying and tracking fish behavior and ecological changes.

[0003] First, traditional tagging methods typically use plastic tags or color codes, which are difficult to maintain long-term stability and are prone to falling off or being damaged, resulting in missing data for subsequent monitoring. Furthermore, manual monitoring is inefficient, and the wide monitoring area and dispersed fish populations make real-time tracking and recording extremely difficult. These technical issues directly hinder a detailed understanding of the migration and spawning behavior of naked carp in Qinghai Lake and also restrict comprehensive analysis of population dynamics.

[0004] Secondly, most current methods for analyzing population maturation cycles rely on static data and lack a deep understanding of the dynamics of environmental factors. Existing monitoring technologies are insufficiently adaptable to environmental changes and fail to effectively integrate environmental monitoring data with population data, resulting in reduced prediction accuracy. Summary of the Invention

[0005] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a method for estimating the sexual maturity cycle of migratory fish using PIT mark monitoring technology. The present invention solves the problem of low efficiency and accuracy in fish sexual cycle monitoring in the prior art.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology, comprising:

[0008] Determine the start and end time based on the electronic tag's endurance attributes;

[0009] Based on the start and end times, electronic tag monitoring equipment and environmental monitoring sensors are installed at preset locations in the target monitoring area to obtain environmental monitoring data for each year and the initial tag monitoring data set within the target monitoring area;

[0010] Determining an initial sexual maturity cycle dataset based on the initial tag monitoring dataset;

[0011] The initial sexual maturity cycle data set is corrected according to the environmental monitoring data in each year to obtain a final sexual maturity cycle data set.

[0012] Preferably, the electronic tag is a transistor PIT electronic tag.

[0013] Preferably, determining the initial sexual maturity cycle dataset based on the initial tag monitoring dataset includes:

[0014] A set of recorded times of each fish species in the initial tag monitoring data;

[0015] determining a time interval set based on the recorded time set;

[0016] The data in the time intervals are processed by averaging to obtain an initial sexual maturity cycle data set.

[0017] Preferably, the correction of the initial sexual maturity cycle dataset according to the environmental monitoring data in each year to obtain the final sexual maturity cycle dataset includes:

[0018] Real-time collection and organization of environmental monitoring data for each year and the initial tag monitoring data set for each year, wherein the environmental monitoring data includes: water temperature, dissolved oxygen, conductivity, pH value, and drainage volume;

[0019] Conducting correlation analysis on fish migration behavior based on environmental monitoring data from each year to determine the corrected environmental factors;

[0020] Calibrate the initial label monitoring data set to determine outliers and perform data cleaning and completion to obtain a processed initial label monitoring data set;

[0021] Constructing an environmental correction model based on the processed initial label monitoring data set and the corrected environmental factors;

[0022] Inputting the initial tag monitoring data set of each year into the environmental correction model for correction to obtain a correction value for the sexual maturity cycle;

[0023] A final sexual maturity cycle data set is determined according to the sexual maturity cycle correction amount and the initial sexual maturity cycle data.

[0024] Preferably, the performing of correlation analysis on fish migration behavior based on the environmental monitoring data of each year to determine the modified environmental factors includes:

[0025] Extracting spatiotemporal features from the environmental monitoring data to obtain dynamic derivative indicators;

[0026] Construct spatiotemporal weight matrix;

[0027] Determining key environmental factors based on the spatiotemporal weight matrix and the dynamic derivative indicators to obtain final modified environmental factors;

[0028] Among them, the expression of the spatiotemporal weight matrix is:

[0029]

[0030] The expression of the modified environmental factor is:

[0031]

[0032] Among them, W ij is the spatiotemporal weight between the first monitoring point i and the second monitoring point j, d ij is the river distance between the first monitoring point i and the second monitoring point j, σ is the weight attenuation factor, F is the significant correction environmental factor, X k is the potential correction environmental factor, including: dynamic derivative indicators and environmental monitoring data, W is the spatiotemporal weight matrix, P(Y t ∣Y t-1 ,X k ,W) is the migration time Y at a given time t-1 t-1 and environmental factor X k In the case of the spatiotemporal weight matrix W, the migration time Y t The probability of P(Y t ∣Y t-1 ) is the migration time Y at a given time t-1 t-1 In the case of t The probability of p is the significance level.

[0033] Preferably, the dynamic derivative indicators include:

[0034] Annual effective accumulated temperature, dissolved oxygen fluctuation intensity, daily average change slope and proportion of flow during the spawning season;

[0035] The expression of the annual effective accumulated temperature is:

[0036]

[0037] The expression of the dissolved oxygen fluctuation intensity is:

[0038]

[0039] The expression of the daily average change slope of conductivity is:

[0040]

[0041] The expression of the flow ratio in the spawning season is:

[0042]

[0043] Among them, T cum is the annual effective accumulated temperature, T t is the average daily water temperature on day t, T th is the metabolic activation threshold, n is the total number of days, DO var is the dissolved oxygen fluctuation intensity, is the annual average dissolved oxygen, DO t is the average daily dissolved oxygen on day t, EC slope is the daily average change slope of conductivity, EC max 、EC min are the maximum and minimum values ​​of daily average conductivity, Δt is the time interval between the maximum and minimum values ​​of daily average conductivity, is the average flow rate during the spawning season, The average flow rate for the whole year.

[0044] Preferably, the initial label monitoring data set is calibrated to determine outliers and data cleaning and completion is performed to obtain a processed initial label monitoring data set, including:

[0045] Conduct clock traceability calibration on electronic tag monitoring equipment installed in each year to establish a unified time base;

[0046] The theoretical maximum migration distance was calculated based on the maximum physiological swimming speed threshold of fish and the time interval between two consecutive detections;

[0047] When the spatial distance between two adjacent detections is greater than the maximum migration distance, the corresponding record will be marked as an outlier;

[0048] Based on the determined outliers, an adaptive density clustering algorithm is used to cluster the high-frequency repeated readings generated at the same monitoring point in a very short period of time, retaining the cluster center readings and eliminating the redundant readings within the cluster;

[0049] In order to solve the problem of missing labels, the Kalman smoothing-linear interpolation hybrid strategy is used to complete the data by combining the synchronous records of adjacent monitoring points to obtain the processed initial label monitoring data set.

[0050] Preferably, the constructing of the environmental correction model based on the processed initial tag monitoring data set and the corrected environmental factors includes:

[0051] Build an LSTM-Transformer hybrid model architecture;

[0052] Constructing a dynamic correction equation based on the LSTM-Transformer hybrid model architecture and the initial labeled monitoring dataset;

[0053] The expression of the dynamic correction equation is:

[0054]

[0055] Among them, α is the environmental contribution coefficient; β k is the environmental factor weight; γ is the individual difference adjustment coefficient; PIT seq is the individual migration time series data, obtained through the initial tag monitoring data; ΔT is the correction value of the sexual maturity cycle, F k is the kth modified environmental factor, k is the index number of the environmental factor, and tanh() is the hyperbolic tangent function.

[0056] Preferably, the LSTM-Transformer hybrid model architecture includes:

[0057] Spatiotemporal attention module and environmental response module;

[0058] The spatiotemporal attention module is used to calculate the spatiotemporal association weights between environmental factors and migratory paths, and the environmental response module is used to input and modify environmental factors;

[0059] The expression of the spatiotemporal correlation weight between the environmental factors and the migratory path is:

[0060]

[0061] Among them, Q i ,K j The query vector and key vector are generated through the migratory trajectory features.

[0062] The present invention discloses the following technical effects:

[0063] The present invention provides a method for estimating the sexual maturity cycle of migratory fish using PIT tag monitoring technology, comprising: determining a start time and an end time according to an electronic tag endurance attribute; installing electronic tag monitoring equipment and environmental monitoring sensors at preset locations in a target monitoring area based on the start time and the end time to obtain environmental monitoring data and an initial tag monitoring data set for each year in the target monitoring area; determining an initial sexual maturity cycle data set based on the initial tag monitoring data set; and correcting the initial sexual maturity cycle data set based on the environmental monitoring data in each year to obtain a final sexual maturity cycle data set. The present invention utilizes the endurance properties of PIT electronic tags to determine unified monitoring start and end times. PIT tag monitoring equipment and environmental monitoring sensors are deployed in target river sections according to a unified clock, synchronously collecting years of continuous tag data and environmental data such as water temperature and dissolved oxygen. The initial sexual maturity cycle of fish at both the individual and population levels is calculated based on tag timestamps. An annual environmental monitoring curve is introduced to dynamically correct the initial cycle to obtain a final sexual maturity cycle dataset. This method solves the problems of traditional plastic tags that are easily detached and have a short lifespan, enabling uninterrupted individual identification and behavioral recording for months to years. High-frequency, non-invasive electronic reading and writing replace manual fishing, allowing large-sample, long-sequence data to be obtained without disturbing the fish, significantly improving monitoring efficiency. Real-time environmental factors and fish behavior data are integrated into a single data link, overcoming the shortcomings of existing methods that "only use static samples and do not consider interannual climate differences," making sexual maturity cycle predictions more accurate to actual ecological contexts. A standardized and automated monitoring-correction-prediction process is formed that can be quickly transferred to other river sections or similar endangered fish species, enhancing scientific decision-making support for the protection and management of aquatic organisms. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0065] Figure 1 A flow chart of a method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0067] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0068] like Figure 1 As shown, the present invention provides a method for estimating the sexual maturity period of migratory fish using PIT mark monitoring technology, comprising:

[0069] Step 100: Determine the start time and end time based on the battery life attribute of the electronic tag;

[0070] Step 200: Based on the start time and the end time, electronic tag monitoring equipment and environmental monitoring sensors are installed at preset locations in the target monitoring area to obtain environmental monitoring data for each year in the target monitoring area and an initial tag monitoring data set;

[0071] Specifically, river section zoning and survey:

[0072] Using RTK-GPS and drone aerial surveying, we conducted 1:2000 horizontal precision topographic mapping of the main channels of the Buha River, Quanji River, Shaliu River, and Heima River;

[0073] Based on river width, average flow velocity and bottom conditions, each river is divided into four sections: "lake entrance-downstream slow flow section-midstream mainstream section-upstream tributary section".

[0074] Site selection for deployment:

[0075] Select one representative section at the beginning and end of each level section;

[0076] Section selection principles: river width <15m, stable riverbed, convenient for laying ground cage antennas;

[0077] Finally, 8 PIT monitoring points and 4 environmental sensor points (temperature / dissolved oxygen / conductivity / flow composite probes) were confirmed for each river.

[0078] PIT Antenna Installation:

[0079] ①Use weather-resistant HDPE tube to encapsulate the loop antenna, cross the river, and bury it at a depth of 30cm;

[0080] ② The antenna is connected to the Reader host through a waterproof interface, and the Reader is installed in a steel cabinet 1m above the flood level;

[0081] ③ A 200W solar panel + lithium battery pack is installed on the top of the cabinet to achieve 24h off-grid power supply.

[0082] Environmental sensor installation:

[0083] ①The probe is hung at 60% depth in the center of the cross section;

[0084] ②Use RS-485 bus to connect to low-power data logger, with a recording frequency of 5 minutes per time;

[0085] ③The sensor is manually calibrated once a quarter and new coefficients are written after calibration.

[0086] Unified clock and network:

[0087] ① The full-site reader and recorder have a built-in GPS timing module, which completes the timing within 10 minutes after startup;

[0088] ②The LoRa-WAN private gateway regularly transmits key heartbeat packets, and automatically sends SMS alarms if the packet loss rate is greater than 5%.

[0089] Functional acceptance:

[0090] ① Place 10 test-tagged fish at each monitoring point, and those that pass the test three times in a row with a reading success rate of ≥98% are considered qualified;

[0091] ②Compared with the handheld YSI multi-parameter instrument, the four sensor errors are all less than ±2%.

[0092] More specifically, multi-level segmentation + unified GPS timing ensures that PIT data corresponds one-to-one with environmental curves; the buried loop antenna and off-grid power supply ensure uninterrupted monitoring throughout the year, avoiding data gaps caused by winter ice or power outages during flood season; and the standardized process of survey-site selection-calibration-acceptance facilitates rapid replication in other rivers flowing into the lake.

[0093] Step 300: determining an initial sexual maturity cycle dataset based on the initial tag monitoring dataset;

[0094] Step 400: Modify the initial sexual maturity cycle dataset based on the environmental monitoring data in each year to obtain a final sexual maturity cycle dataset.

[0095] Furthermore, the electronic tag is a transistor PIT electronic tag.

[0096] Furthermore, determining an initial sexual maturity cycle dataset based on the initial tag monitoring dataset includes:

[0097] A set of recorded times of each fish species in the initial tag monitoring data;

[0098] determining a time interval set based on the recorded time set;

[0099] The data in the time intervals are processed by averaging to obtain an initial sexual maturity cycle data set.

[0100] Specifically, since 2021, the Qinghai Lake Naked Carp Rescue Center has tagged 150,000 one-year-old fingerlings (Naked Carp of Qinghai Lake) and broodstock with transistor PIT electronic tags. These tagged fish are then released back into Qinghai Lake for breeding. Naked carp of Qinghai Lake migrate annually to spawn, primarily in the rivers that feed Qinghai Lake, including the Buha, Quanji, Shaliu, and Heima rivers.

[0101] During spawning season, PIT monitoring equipment is deployed along the Buha, Quanji, Shaliu, and Heima rivers. The equipment records the PIT number, time, location, and number of times the broodstock pass through in real time.

[0102] By analyzing and organizing the data collected annually, we can determine whether a fish has reached sexual maturity and is migrating for spawning. The first migration data collected is recorded, and if the same fish is monitored again after a few years, it can be considered as a second migration for sexual maturity and spawning. The time interval between the first and second migrations can be considered the sexual maturity cycle (initial sexual maturity cycle).

[0103] Furthermore, the initial sexual maturity cycle dataset is corrected according to the environmental monitoring data in each year to obtain a final sexual maturity cycle dataset, including:

[0104] Real-time collection and organization of environmental monitoring data for each year and the initial tag monitoring data set for each year, wherein the environmental monitoring data includes: water temperature, dissolved oxygen, conductivity, pH value, and drainage volume;

[0105] Conducting correlation analysis on fish migration behavior based on environmental monitoring data from each year to determine the corrected environmental factors;

[0106] Calibrate the initial label monitoring data set to determine outliers and perform data cleaning and completion to obtain a processed initial label monitoring data set;

[0107] Constructing an environmental correction model based on the processed initial label monitoring data set and the corrected environmental factors;

[0108] Inputting the initial tag monitoring data set of each year into the environmental correction model for correction to obtain a correction value for the sexual maturity cycle;

[0109] A final sexual maturity cycle data set is determined according to the sexual maturity cycle correction amount and the initial sexual maturity cycle data.

[0110] Furthermore, the correlation analysis of fish migration behavior is performed based on the environmental monitoring data of each year to determine the modified environmental factors, including:

[0111] Extracting spatiotemporal features from the environmental monitoring data to obtain dynamic derivative indicators;

[0112] Construct spatiotemporal weight matrix;

[0113] Determining key environmental factors based on the spatiotemporal weight matrix and the dynamic derivative indicators to obtain final modified environmental factors;

[0114] Among them, the expression of the spatiotemporal weight matrix is:

[0115]

[0116] The expression of the modified environmental factor is:

[0117]

[0118] in,

[0119] W ij is the spatiotemporal weight between the first monitoring point i and the second monitoring point j, d ij is the river distance between the first monitoring point i and the second monitoring point j, σ is the weight attenuation factor, F is the significant correction environmental factor, X k is the potential correction environmental factor, including: dynamic derivative indicators and environmental monitoring data, W is the spatiotemporal weight matrix, P(Y t ∣Y t-1 ,X k ,W) is the migration time Y at a given time t-1 t-1 and environmental factor X k In the case of the spatiotemporal weight matrix W, the migration time Y t The probability of P(Y t ∣Y t-1 ) is the migration time Y at a given time t-1 t-1 In the case of t The probability of p is the significance level.

[0120] Specifically, spatiotemporal causal analysis: breaking through the limitations of traditional statistical methods, the spatial topological relationship of the river channel was introduced for the first time in fish ecology research, and it was found that the impact of water temperature in the upper reaches of the Buha River on the sexual maturity cycle was 1.8 times that of the lower reaches.

[0121] Furthermore, the dynamic derivative indicators include:

[0122] Annual effective accumulated temperature, dissolved oxygen fluctuation intensity, daily average change slope and proportion of flow during the spawning season;

[0123] The expression of the annual effective accumulated temperature is:

[0124]

[0125] The expression of the dissolved oxygen fluctuation intensity is:

[0126]

[0127] The expression of the daily average change slope of conductivity is:

[0128]

[0129] The expression of the flow ratio in the spawning season is:

[0130]

[0131] Among them, T cum is the annual effective accumulated temperature, T t is the average daily water temperature on day t, T th is the metabolic activation threshold, n is the total number of days, DO var is the dissolved oxygen fluctuation intensity, is the annual average dissolved oxygen, DO t is the average daily dissolved oxygen on day t, EC slope is the daily average change slope of conductivity, EC max 、EC min are the maximum and minimum values ​​of daily average conductivity, Δt is the time interval between the maximum and minimum values ​​of daily average conductivity, is the average flow rate during the spawning season, The average flow rate for the whole year.

[0132] Specifically, Annual Effective Temperature (AET):

[0133] Objective: To measure the "available energy" (AET) accumulated over a year for gonadal development in naked carp of Qinghai Lake. A high AET indicates sufficient "heat surplus" in the water, leading to earlier energy allocation to the fish's gonads and a shorter maturation cycle.

[0134] Dissolved-Oxygen Fluctuation (DOF):

[0135] Purpose: To quantify dissolved oxygen stability throughout the year and reflect the impact of water purification and algal metabolic fluctuations on fish stress. High DOF indicates dramatic fluctuations in dissolved oxygen, requiring frequent adjustments in osmotic pressure and respiratory rate, which can inhibit maturation. High DOF indicates dramatic fluctuations in dissolved oxygen, requiring frequent adjustments in osmotic pressure and respiratory rate, which can inhibit maturation.

[0136] Daily average change slope of electrical conductivity (Electrical-Conductivity Slope, ECS):

[0137] Purpose: To capture the gradual shift of conductivity from high salinity to low salinity or vice versa throughout the year, indirectly reflecting changes in rainfall runoff and groundwater recharge. A large absolute slope indicates rapid changes in salinity gradients throughout the year, which may burden fish with osmotic regulation and delay or disrupt migratory rhythms.

[0138] Flow-QuotainSpawningseason (FQS):

[0139] Assess the intensity of river recharge during the key spawning months (April-July) relative to the annual average. Wet periods provide greater surface area and bottom scouring, facilitating migration and spawning; dry periods may delay maturation or reduce migration success.

[0140] More specifically, AET is closely positively correlated with spawning hormone synthesis, DOF and ECS mainly describe the negative modulation of "environmental stress", and FQS is the positive drive of "hydrodynamic channel"; the four together characterize the comprehensive impact of "heat-gas-salinity-hydrodynamics" on the annual maturation process of naked carp in Qinghai Lake.

[0141] Furthermore, the initial label monitoring data set is calibrated to determine outliers and data cleaning and completion are performed to obtain a processed initial label monitoring data set, including:

[0142] Conduct clock traceability calibration on electronic tag monitoring equipment installed in each year to establish a unified time base;

[0143] The theoretical maximum migration distance was calculated based on the maximum physiological swimming speed threshold of fish and the time interval between two consecutive detections;

[0144] When the spatial distance between two adjacent detections is greater than the maximum migration distance, the corresponding record will be marked as an outlier;

[0145] Based on the determined outliers, an adaptive density clustering algorithm is used to cluster the high-frequency repeated readings generated at the same monitoring point in a very short period of time, retaining the cluster center readings and eliminating the redundant readings within the cluster;

[0146] In order to solve the problem of missing labels, the Kalman smoothing-linear interpolation hybrid strategy is used to complete the data by combining the synchronous records of adjacent monitoring points to obtain the processed initial label monitoring data set.

[0147] Specifically, clock traceability correction:

[0148] Enable satellite timing function in each fixed reader and complete time synchronization within ten minutes of powering on;

[0149] Compare the device's local log with the satellite standard time. If the cumulative error exceeds one second, the device's system clock will be forcibly adjusted through a remote script.

[0150] Retroactively correct the timestamps of the original readings saved in the last 24 hours and overwrite the original entries in the database;

[0151] The server collects heartbeat packets from each monitoring point every fifteen minutes and checks whether the timestamp variance exceeds 0.3 seconds. If it exceeds the standard, it will immediately send a text message alarm to the maintenance personnel.

[0152] Determination of the theoretical maximum migration distance:

[0153] Based on laboratory flume tests, the upper limit of the sustainable swimming speed of naked carp of Qinghai Lake was determined.

[0154] For the same tagged fish, the time interval between two consecutive readings can be directly obtained by subtracting the timestamps;

[0155] Multiply the maximum sustainable swimming speed by the interval time to get the maximum distance the fish can theoretically move within the interval.

[0156] If the actual measured river channel distance between two readings exceeds this maximum distance, the latter record will be marked as "abnormal swimming speed".

[0157] High frequency read clusters are merged:

[0158] When the fish stays near the antenna, the same reader will generate many duplicate numbers in a very short time;

[0159] First, calculate the mean time interval and fluctuation range of all two adjacent readings of the same reader within a single day;

[0160] The mean plus three times the fluctuation amplitude is used as the initial distance threshold, and the improved density clustering algorithm is called to cluster the readings with a time interval less than the threshold;

[0161] If the average time interval within a cluster is still too large after clustering, the threshold is reduced by 80% and clustering is performed again until convergence;

[0162] For each time cluster, only the central reading is retained and other redundant records in the same cluster are deleted.

[0163] Abnormal swimming speed and redundant reading elimination:

[0164] Aggregate the "abnormal swimming speed" label and the "high-frequency redundant" label into the same temporary list;

[0165] Remove these invalid records at one time through database batch deletion instructions;

[0166] After the deletion is complete, the number of remaining records can be checked in real time to ensure that the percentage of accidental deletion is less than 0.02%.

[0167] The deletion log is encrypted and archived for easy tracing in the future.

[0168] Missing test completion:

[0169] Based on the successive readings of adjacent monitoring points upstream and downstream, the recursive smoothing method is used to estimate the fish body position at the missing moment;

[0170] If the spatial difference between two consecutive estimated values ​​and the true readings does not exceed thirty meters, an interpolated record is directly generated at the missing position;

[0171] When the missing time is long and interpolation alone cannot meet the required accuracy, the Kalman smoothing process is started:

[0172] Taking the upstream readings as the initial state, the prediction step is used to extrapolate to the missing interval;

[0173] Once downstream reads are captured, all predictions are corrected using an update step;

[0174] The final completed record is marked with a "virtual" field in the database to distinguish it from the real reading;

[0175] After the statistics are completed, if the overall missed measurement rate is still higher than one percent, an order will be automatically sent to the equipment maintenance team for investigation.

[0176] Furthermore, constructing an environmental correction model based on the processed initial label monitoring data set and the corrected environmental factors includes:

[0177] Build an LSTM-Transformer hybrid model architecture;

[0178] Constructing a dynamic correction equation based on the LSTM-Transformer hybrid model architecture and the initial labeled monitoring dataset;

[0179] The expression of the dynamic correction equation is:

[0180]

[0181] Among them, α is the environmental contribution coefficient; β k is the environmental factor weight; γ is the individual difference adjustment coefficient; PIT seq is the individual migration time series data, obtained through the initial tag monitoring data; ΔT is the correction value of the sexual maturity cycle, F k is the kth modified environmental factor, k is the index number of the environmental factor, and tanh() is the hyperbolic tangent function.

[0182] Specifically, the hyperbolic tangent function compresses the linear weighted sum of environmental factors into a reasonable range (-1.5 to +1.5 years) to avoid extreme correction values; it also reflects the progressive saturation effect of the environment on the sexual maturity cycle (for example, after the accumulated temperature exceeds the threshold, the promoting effect no longer grows linearly).

[0183] Furthermore, the LSTM-Transformer hybrid model architecture includes:

[0184] Spatiotemporal attention module and environmental response module;

[0185] The spatiotemporal attention module is used to calculate the spatiotemporal association weights between environmental factors and migratory paths, and the environmental response module is used to input and modify environmental factors;

[0186] The expression of the spatiotemporal correlation weight between the environmental factors and the migratory path is:

[0187]

[0188] Among them, Q i ,K j The query vector and key vector are generated through the migratory trajectory features.

[0189] Specifically, migration trajectory feature extraction:

[0190] The migration trajectory characteristics recorded at each monitoring point include the following core information:

[0191] Time series: timestamp of each individual passing through a monitoring point (accurate to the minute):

[0192] Spatial coordinates: GPS location (longitude, latitude) of the monitoring point and its relative position to the river (kilometers from the river mouth):

[0193] Movement characteristics: migration speed (m / s) and direction (upstream or downstream of the river) between adjacent monitoring points;

[0194] Duration of stay: the cumulative time (hours) that an individual stays near the monitoring point;

[0195] Temporal feature encoding:

[0196] Divide the consecutive timestamps into 3-hour windows and extract the following within each window:

[0197] Arrival frequency: the number of times an individual passes through a monitoring point within the window;

[0198] Temporal distribution entropy: Shannon entropy is calculated through timestamps to quantify the temporal regularity of individual activities;

[0199] Circadian rhythm markers: distinguishing the activity intensity ratio between daytime (6:00-18:00) and nighttime;

[0200] Spatial feature embedding:

[0201] Map the monitoring point position to a 128-dimensional vector, including:

[0202] Normalized value of distance to the estuary;

[0203] Upstream and downstream direction (upstream is 1, downstream is 0);

[0204] River channel curvature (calculated by the curvature radius of the 5 km river section before and after);

[0205] Spatial attention mask: Based on historical migration paths, a migration probability matrix between monitoring points is pre-generated to weight spatial correlation;

[0206] Trajectory feature fusion:

[0207] LSTM time series modeling: Input the time window features into the bidirectional LSTM to obtain a 256-dimensional time series hidden state input dimension: time features (3 dimensions) + motion features (2 dimensions);

[0208] Output: Temporal context representation of each time window

[0209] Spatial-temporal feature concatenation: concatenate the LSTM output with the spatial embedding vector to form a 320-dimensional joint feature vector;

[0210] Query / key vector generation

[0211] Linear projection transformation: The joint feature vector is mapped to a low-dimensional space through a trainable weight matrix for query / key vector generation.

[0212] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0213] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology, characterized in that: include: Determine the start and end time based on the electronic tag's endurance attributes; Based on the start and end times, electronic tag monitoring equipment and environmental monitoring sensors are installed at preset locations in the target monitoring area to obtain environmental monitoring data for each year and the initial tag monitoring data set within the target monitoring area; Determining an initial sexual maturity cycle dataset based on the initial tag monitoring dataset; The initial sexual maturity cycle data set is corrected according to the environmental monitoring data in each year to obtain a final sexual maturity cycle data set.

2. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 1, characterized in that: The electronic tag is a transistor PIT electronic tag.

3. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 1, characterized in that: Determining an initial sexual maturity cycle dataset based on the initial tag monitoring dataset includes: A set of recorded times of each fish species in the initial tag monitoring data; determining a time interval set based on the recorded time set; The data in the time intervals are processed by averaging to obtain an initial sexual maturity cycle data set.

4. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 1, characterized in that: The initial sexual maturity cycle dataset is corrected according to the environmental monitoring data in each year to obtain a final sexual maturity cycle dataset, including: Real-time collection and organization of environmental monitoring data for each year and the initial tag monitoring data set for each year, wherein the environmental monitoring data includes: water temperature, dissolved oxygen, conductivity, pH value, and drainage volume; Conducting correlation analysis on fish migration behavior based on environmental monitoring data from each year to determine the corrected environmental factors; Calibrate the initial label monitoring data set to determine outliers and perform data cleaning and completion to obtain a processed initial label monitoring data set; Constructing an environmental correction model based on the processed initial label monitoring data set and the corrected environmental factors; Inputting the initial tag monitoring data set of each year into the environmental correction model for correction to obtain a correction value for the sexual maturity cycle; A final sexual maturity cycle data set is determined according to the sexual maturity cycle correction amount and the initial sexual maturity cycle data.

5. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 4, characterized in that: The correlation analysis of fish migration behavior based on the environmental monitoring data of each year to determine the modified environmental factors includes: Extracting spatiotemporal features from the environmental monitoring data to obtain dynamic derivative indicators; Construct spatiotemporal weight matrix; Determining key environmental factors based on the spatiotemporal weight matrix and the dynamic derivative indicators to obtain final modified environmental factors; Among them, the expression of the spatiotemporal weight matrix is: The expression of the modified environmental factor is: Among them, W ij is the spatiotemporal weight between the first monitoring point i and the second monitoring point j, d ij is the river distance between the first monitoring point i and the second monitoring point j, σ is the weight attenuation factor, F is the significant correction environmental factor, X k is the potential correction environmental factor, including: dynamic derivative indicators and environmental monitoring data, W is the spatiotemporal weight matrix, P(Y t ∣Y t-1 ,X k ,W) is the migration time Y at a given time t-1 t-1 and environmental factor X k In the case of the spatiotemporal weight matrix W, the migration time Y t The probability of P(Y t ∣Y t-1 ) is the migration time Y at a given time t-1 t-1 The probability of migration time Y under the condition of , p is the significance level.

6. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 5, characterized in that: The dynamic derivative indicators include: Annual effective accumulated temperature, dissolved oxygen fluctuation intensity, daily average change slope and proportion of flow during the spawning season; The expression of the annual effective accumulated temperature is: The expression of the dissolved oxygen fluctuation intensity is: The expression of the daily average change slope of conductivity is: The expression of the flow ratio in the spawning season is: Among them, T cum is the annual effective accumulated temperature, T t is the average daily water temperature on day t, T th is the metabolic activation threshold, n is the total number of days, DO var is the dissolved oxygen fluctuation intensity, is the annual average dissolved oxygen, DO t is the average daily dissolved oxygen on day t, EC slope is the daily average change slope of conductivity, EC max 、EC min are the maximum and minimum values ​​of daily average conductivity, Δt is the time interval between the maximum and minimum values ​​of daily average conductivity, is the average flow rate during the spawning season, The average flow rate for the whole year.

7. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 4, characterized in that: The initial label monitoring data set is calibrated to determine outliers and data cleaning and completion is performed to obtain a processed initial label monitoring data set, including: Conduct clock traceability calibration on electronic tag monitoring equipment installed in each year to establish a unified time base; The theoretical maximum migration distance was calculated based on the maximum physiological swimming speed threshold of fish and the time interval between two consecutive detections; When the spatial distance between two adjacent detections is greater than the maximum migration distance, the corresponding record will be marked as an outlier; Based on the determined outliers, an adaptive density clustering algorithm is used to cluster the high-frequency repeated readings generated at the same monitoring point in a very short period of time, retaining the cluster center readings and eliminating the redundant readings within the cluster; In order to solve the problem of missing labels, the Kalman smoothing-linear interpolation hybrid strategy is used to complete the data by combining the synchronous records of adjacent monitoring points to obtain the processed initial label monitoring data set.

8. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 4, characterized in that: The step of constructing an environmental correction model based on the processed initial label monitoring data set and the corrected environmental factors includes: Build an LSTM-Transformer hybrid model architecture; Constructing a dynamic correction equation based on the LSTM-Transformer hybrid model architecture and the initial labeled monitoring dataset; The expression of the dynamic correction equation is: Among them, α is the environmental contribution coefficient; β k is the environmental factor weight; γ is the individual difference adjustment coefficient; PIT seq is the individual migration time series data, obtained through the initial tag monitoring data; ΔT is the correction value of the sexual maturity cycle, F k is the kth modified environmental factor, k is the index number of the environmental factor, and tanh() is the hyperbolic tangent function.

9. The method for estimating the sexual maturity period of migratory fish using PIT marker monitoring technology according to claim 8, characterized in that: The LSTM-Transformer hybrid model architecture includes: Spatiotemporal attention module and environmental response module; The spatiotemporal attention module is used to calculate the spatiotemporal association weights between environmental factors and migratory paths, and the environmental response module is used to input and modify environmental factors; The expression of the spatiotemporal correlation weight between the environmental factors and the migratory path is: Among them, Q i ,K j The query vector and key vector are generated through the migratory trajectory features.

Images