Fish seed breeding method and system

Abstract

The invention relates to the technical field of aquatic product breeding, in particular to a fingerling breeding method and system, comprising the following steps: collecting fry behavior data, numbering and segmenting to form a time sequence, matching environmental fluctuation parameters, extracting disturbance and offset pairing information, identifying response differences and grouping, and establishing a behavior regulation mapping map. According to the method, the accuracy of fish fry state recognition and environment disturbance perception is improved by extracting the response relation between fish fry behavior characteristics and environment parameters and establishing a time pairing mechanism of disturbance starting and behavior response, and a disturbance response effective node data set comprises behavior mutation points, environment fluctuation boundaries and behavior continuity differences. The method is used for supporting accurate judgment and regulation of dynamic states of fries in a breeding period, improving the problems of fuzzy behaviors and delayed early warning under traditional experience judgment, enhancing the response capability and the adaptation level of the fries to external changes, and realizing collaborative improvement of culture behaviors and environment regulation in a breeding process.

Description

Technical Field

[0001] The present invention relates to the technical field of aquatic breeding, and in particular to a fish breeding method and system. Background Art

[0002] The field of aquatic breeding technology includes core issues such as genetic improvement of aquatic animals, protection of germplasm resources, variety selection and breeding, and reproduction control. It aims to improve the production performance, disease resistance and environmental adaptability of aquatic animals through scientific methods, and promote the sustainable development of the aquaculture industry. The technical field includes specific technical directions such as parent screening and cultivation, artificial breeding technology, embryo development monitoring, seedling cultivation management, and molecular marker-assisted breeding. It is an important support for ensuring the supply of high-quality aquatic seed sources in modern aquaculture.

[0003] A fish breeding method and system refers to a set of operational breeding processes and supporting methods proposed to address the problems of low breeding efficiency, large individual differences, and unstable breeding success rates in fish breeding during aquatic breeding. This includes: identifying and classifying key behavioral states based on continuous behavioral recording data collected from fry during their developmental stages; synchronizing fry development by regulating water temperature, photoperiod, and nutrient composition; selecting the developmental environment based on parameters such as behavioral consistency and response delay differences; and using specific flow rate and temperature control parameters during the incubation stage to ensure uniform embryonic development.

[0004] Existing technologies rely on manual experience to judge the maturity of fish and the optimal breeding period. During operation, it is difficult to accurately grasp the dynamic relationship between individual behavioral changes and environmental factors. In particular, there is a lack of continuous behavioral monitoring methods in the fry stage, resulting in delayed identification of physiological indicators such as feeding activity and swimming abnormalities. In addition, the setting of hatching environment parameters is generally based on empirical assumptions, which cannot timely capture the direct impact of sudden disturbances on fish behavior. In actual operation, the stress response of the fry is often not detected due to the failure to identify environmental changes in a timely manner, thereby affecting the stability of the fry development process and hatching survival rate. A typical example is the failure to adjust the control strategy in time under conditions of light fluctuations or short-term reduction of dissolved oxygen, resulting in problems such as fish behavioral disorders and increased early mortality. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, the present invention provides a method and system for breeding fish. The technical solution is as follows: In order to achieve the above object, the present invention adopts the following technical solution, a fish breeding method, comprising the following steps: S1: Continuous behavioral data is collected during the developmental stages of fry under artificial breeding conditions. By identifying the initiation of swimming, active feeding, and body color formation, the state of each behavior is classified and time-located. The behaviors of each stage are numbered and segmented using chronological order to obtain the time series of behavioral node achievement. S2: Using the behavior nodes to achieve a time series, calling the change records of the water environment in which the fry are located, performing trend identification and fluctuation classification on the water temperature, light, pH and dissolved oxygen data, matching the environmental parameters of the corresponding behavior nodes, and generating an environmental parameter set; S3: Using the environmental parameter set, based on the behavior segments corresponding to the behavior node achievement time, combined with the environmental change process, extract the corresponding observation time interval, identify the disturbance process and behavior deviation manifestation within the interval, analyze the sequential relationship between the disturbance initiation and the initial appearance of the behavioral response, calculate the corresponding time interval and record the pairing information, and obtain the disturbance-behavior response pairing data; S4: Using the disturbance-behavior response pairing data, identify the behavior deviation direction and continuity within the behavior node, analyze the disturbance superposition and response differences, standardize the grouped response behavior, and obtain the disturbance response valid node data set.

[0006] As a further solution of the present invention, the behavior node achievement time series includes the behavior stage number, the node achievement time point, and the behavior classification label; the environmental parameter set includes the water temperature change trend, the light fluctuation range, and the pH stage change value; the disturbance-behavior response pairing data includes the disturbance start time point, the response first time point, and the time interval value; the disturbance response valid node data set includes the directional consistency coefficient, the response delay standard value, and the disturbance difference grouping identifier.

[0007] As a further solution of the present invention, the step S1 is specifically as follows: S101: Continuous behavioral record data of fry development stages under artificial breeding conditions is collected, and continuous frame segments in the time-series image are called. By comparing the pixel edges of the fry morphological posture area in the continuous frames, the frame points of the overall displacement of the fish body and the local morphological mutation are identified. According to the order of occurrence of the morphological mutation, the fry swimming start, feeding activity and body color formation behaviors are extracted to generate a stage behavior recognition sequence; S102: Based on the staged behavior recognition sequence, statistics are collected on the continuous frame segments of each type of behavior in the image sequence, and the start frame and end frame of the behavior are recorded. By analyzing the timestamps corresponding to the frame number intervals, the duration of the behavior is calculated and the time positioning point is calibrated to generate the behavior state duration positioning information; S103: Based on the behavior state duration positioning information, the order of behavior appearance time in the frame sequence is called, the time periods to which each type of behavior belongs are numbered in sequence, and the behaviors are classified into corresponding stage segments according to the numbers, thereby completing the sequence organization of the staged behaviors and obtaining the time sequence of the behavior node achievement.

[0008] As a further solution of the present invention, the step S2 is specifically as follows: S201: Using the behavior nodes to achieve a time sequence, calling the original recorded data of water temperature, light, pH and dissolved oxygen of the fry group at each node, integrating each data item according to the time label, filtering and grouping according to the time labels under multiple behavior nodes of the fry group, marking the node corresponding to the environmental factor data segment, and generating the node environment sequence interval value; S202: calling the node environmental sequence interval value, performing adjacent comparisons on each data item at the time interval within the node based on the increase and decrease change direction of water temperature, light, pH, and dissolved oxygen in the corresponding time period, calibrating the difference interval between adjacent data points, and dividing the time period into a stable segment and a changing segment according to the fluctuation rate threshold, thereby generating the environmental factor fluctuation trend segment value; The fluctuation rate threshold is set by calculating the median of the change rate of each environmental factor in the time series corresponding to the behavior node plus 1.5 times the median absolute deviation; S203: According to the fluctuation trend segment value of the environmental factors, the fluctuation stage data segments of water temperature, light, pH and dissolved oxygen under the behavior node are matched, the mean, range and fluctuation amplitude of the four environmental factors corresponding to each node are calculated, and data classification and integration and time point alignment processing are performed to generate an environmental parameter set.

[0009] As a further solution of the present invention, the step S3 is specifically as follows: S301: Using the environmental parameter set and the time period information corresponding to the behavior node, extract the water temperature, light, pH, and dissolved oxygen data series covered by each behavior segment on the time axis, integrate them into continuous observation intervals in chronological order, set start and end time tags for the intervals, and generate observation time interval values; S302: Calling the observation time interval value, calculating the change rate of the corresponding data point based on the time series change amplitude of water temperature, light, pH and dissolved oxygen in the interval and setting a disturbance identification change amplitude threshold, marking the time segment where the rate exceeds the threshold as a disturbance segment, and combining the fry behavior data in the node to identify the starting point of the offset feature, to obtain a pair of disturbance and offset starting time points; The disturbance identification change amplitude threshold is an adaptive mutation determination benchmark set by extracting the median of the environmental factor change amplitude sequence and adding 2 times the standard deviation; S303: Based on the disturbance and offset start time point pairs, calculate the time interval between each set of disturbance start points and the first occurrence time of the corresponding offset behavior, and associate the corresponding behavior segment identifier and disturbance type for each set of time intervals, integrate them into structured table data, and obtain disturbance-behavior response pairing data.

[0010] As a further solution of the present invention, the step S4 is specifically as follows: S401: using the disturbance-behavior response paired data, identifying the direction of the fry position change at each behavior node, extracting the coordinate point displacement values ​​and direction angles in the continuous response segments, performing consistency analysis on the offset directions at consecutive time points, and generating a behavior offset continuous directionality; The continuous response segment refers to the behavioral data period of the fry that continues to deviate in the same direction within a certain period of time after the disturbance occurs; S402: calling the behavior deviation continuous directionality, classifying and counting the number of identified disturbances and the corresponding response time differences in the behavior node, comparing the response amplitude and delay value of each node under the same disturbance condition, and extracting the difference parameters to obtain a disturbance response difference index value; The specific formula for comparing the response amplitude and delay value of each node under the same disturbance conditions is: ; Calculate the disturbance response difference index value; in, For the The disturbance response difference index value of each behavior node under all disturbance conditions, For the The behavior node is in The amplitude deviation normalization parameter under disturbance conditions, For the The behavior node is in The delay deviation normalization parameter under disturbance conditions, For the The perturbation structure coupling factor corresponding to the perturbation condition is: For the The response modulation gain coefficient corresponding to the disturbance condition is, is the total number of perturbation conditions, is the index number of the behavior node, is the index number of the disturbance condition; S403: Grouping and classifying the response amplitude, direction consistency, and response delay combinations according to the disturbance response difference index value, integrating and marking multiple groups of behavior node numbers, screening stable structured combinations, and establishing a disturbance response valid node data set; The screening of stable structured combinations refers to extracting node combinations with parameter centralization and behavioral regularity characteristics based on the joint stability standard of three indicators: response amplitude fluctuation coefficient, direction consistency and delay time standard deviation.

[0011] As a further embodiment of the present invention, the method further comprises: S5: Integrate the disturbance response valid node dataset and the behavior node achievement time series data, identify the normalized behavior offset characteristics and corresponding disturbance types, combine the behavior change sequence with the environmental fluctuation process, analyze the action chain, adjust the water temperature, light cycle, pH and dissolved oxygen regulation parameters, and generate a behavior regulation map; The behavior regulation mapping map includes a regulation parameter matching matrix, a disturbance type action sequence, and a behavior offset response mapping group.

[0012] As a further solution of the present invention, the step S5 is specifically as follows: S501: Integrate the disturbance response valid node dataset and the behavior node achievement time series data, extract the disturbance type, behavior response time and offset direction information corresponding to each node, unify the time label format, establish the corresponding structure of disturbance and behavior performance between nodes, and generate a disturbance behavior association matching table; S502: calling the disturbance behavior association matching table, identifying the association pattern between the change direction of environmental factors and the behavioral response sequence during the disturbance period based on the change trends of water temperature, light, pH, and dissolved oxygen before and after the node disturbance, and normalizing the offset angle distribution and disturbance characteristic value to obtain a standardized response matching parameter value; The standardization process refers to uniformly mapping the disturbance and response parameters to the range of 0 to 1 according to the range formula, which is used to eliminate scale differences and facilitate feature comparison; S503: According to the standardized response matching parameter values, combined with the node behavior change sequence and the fluctuation stage difference of the environmental factors, the response chain in the disturbance path is extracted, and the optimal combination interval of the water temperature regulation value, light cycle ratio, pH regulation amplitude and dissolved oxygen regulation frequency under each disturbance type is calculated, and the corresponding behavioral response nodes are mapped to the regulation matrix to generate a behavioral regulation mapping map.

[0013] As a further solution of the present invention, the specific formula combining the sequence of node behavior changes and the difference in the fluctuation stage of environmental factors is: ; Calculate the disturbance response difference value ΔR; in, is the disturbance response difference value of the i-th type of disturbance in the behavioral response dimension, is the amplitude of the node behavior change under the i-th type of disturbance in the j-th stage, is the environmental factor fluctuation intensity weight corresponding to the i-th type of disturbance in the j-th stage, is the total number of fluctuation stages contained in the perturbation path, is the arithmetic mean of the behavior change amplitude of the nodes in all n stages of the i-th type of disturbance, is the critical disturbance threshold in the behavioral response process corresponding to the i-th type of disturbance, is the average value of the disturbance trend offset intensity of the i-th type disturbance in t observation periods, is the behavioral dimension identifier of the disturbance response type, is the disturbance intensity index of behavior change, is the weight source index of environmental factor fluctuations, is the superscript of the threshold index of the critical condition, is the total number of disturbance trend observation cycles index superscript, is the perturbation type index, It is the fluctuation phase index.

[0014] In another aspect, a fish seed breeding system is provided, which is applied to a fish seed breeding method, and the system comprises: The behavior sequence extraction module obtains the observed data of fry behavior and segments it into a time series. It obtains the behaviors of swimming initiation, feeding activity, and body color formation, forms behavioral parameters, performs interval division according to the duration and occurrence order of the behaviors, analyzes the sequence of behaviors and completes the numbering, extracts the achievement time, and outputs the achievement time series of the behavior nodes, which is then passed to the environmental feature construction module and the path map generation module. The environmental feature construction module obtains water temperature, light, pH, and dissolved oxygen records based on the time series achieved by the behavior nodes, analyzes the execution trend of the node time period, identifies the fluctuation frequency and change direction of each environmental parameter, outputs the environmental parameter set and passes it to the disturbance response matching module; The disturbance-response matching module, based on the environmental parameter set and the behavior segment corresponding to the behavior node reaching time, identifies the disturbance starting point and response time point, extracts the corresponding observation interval, analyzes the disturbance process and offset performance, calculates the time difference and completes the disturbance-response pairing annotation, outputs the disturbance-behavior response pairing data and passes it to the behavior offset classification module; A behavior deviation classification module obtains the disturbance type and deviation performance based on the disturbance-behavior response pairing data, groups them according to the disturbance source and response continuity, analyzes the response differences under the same disturbance, outputs the disturbance response valid node data set and passes it to the path map generation module; The path map generation module, based on the disturbance response valid node data set and the behavior node achievement time series, identifies the offset characteristics and corresponding disturbance types after behavior standardization, combines the behavior change sequence with the environmental fluctuation process, analyzes the action chain, adjusts the regulation parameters of water temperature, light cycle, pH and dissolved oxygen, and generates a behavior regulation mapping map.

[0015] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: By classifying and dividing the continuous behavioral states of fry into time series and stages, and combining the information of water environment changes to construct the response relationship between behavior and environmental parameters, the sensitive nodes of fry to external disturbances during the critical growth period and the trend analysis of their behavioral responses are identified. The interference pairing data between behavior and environment provides a quantitative basis for judging the developmental status of fish and their adaptability to the external environment. Identifying behavioral deviation characteristics and response patterns with the help of behavioral trajectories before and after disturbances helps to clarify the key environmental factors affecting reproductive stability, and extracting effective nodes for standardized grouping of behavioral responses, thus facilitating the accurate identification of the interaction between embryonic development status and environment in the process of aquatic fry cultivation, and thus providing clear data support for environmental regulation and behavioral judgment in the breeding process, and realizing high-precision perception of fry growth behavior and dynamic analysis of disturbance response. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of 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 creative work.

[0017] Figure 1 It is a schematic diagram of the workflow of the present invention; Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0019] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0020] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0021] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0022] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0023] See also Figure 1 The present invention provides a technical solution, a fish breeding method, comprising the following steps: S1: Continuous behavioral data is collected during the developmental stages of fry under artificial breeding conditions. By identifying the initiation of swimming, active feeding, and body color formation, the state of each behavior is classified and time-located. The behaviors of each stage are numbered and segmented using chronological order to obtain the time series of behavioral node achievement. S2: Use behavior nodes to achieve time series, call the change records of the water environment in which the fry are located, identify trends and divide fluctuations in water temperature, light, pH and dissolved oxygen data, match the environmental parameters of the corresponding behavior nodes, and generate an environmental parameter set; S3: Using the environmental parameter set, based on the behavioral segments corresponding to the behavioral node achievement time, combined with the environmental change process, the corresponding observation time interval is extracted, the disturbance process and behavioral deviation manifestation within the interval are identified, the sequential relationship between the onset of the disturbance and the initial appearance of the behavioral response is analyzed, the corresponding time interval is calculated and the pairing information is recorded to obtain the disturbance-behavior response pairing data; S4: Using the disturbance-behavior response pairing data, identify the behavior deviation direction and continuity within the behavior node, analyze the disturbance superposition and response differences, standardize the grouped response behavior, and obtain the disturbance response valid node data set; S5: Integrate the disturbance response effective node data set and the behavior node achievement time series data, identify the offset characteristics and corresponding disturbance types after behavior standardization, combine the behavior change sequence with the environmental fluctuation process, analyze the action chain, adjust the regulation parameters of water temperature, light cycle, pH and dissolved oxygen, and generate a behavior regulation mapping map.

[0024] The behavioral node achievement time series includes the behavioral stage number, the node achievement time point, and the behavioral classification label. The environmental parameter set includes the water temperature change trend, the light fluctuation range, and the pH stage change value. The disturbance-behavior response pairing data includes the disturbance start time point, the response first time point, and the time interval value. The disturbance response valid node data set includes the direction consistency coefficient, the response delay standard value, and the disturbance difference grouping identifier. The behavioral regulation mapping map includes the adjustment parameter matching matrix, the disturbance type action order, and the behavioral offset response mapping group.

[0025] See also Figure 1 , the steps of S1 are as follows: S101: Continuous behavioral record data of fry development stages under artificial breeding conditions is collected, and continuous frame segments in the time-series image are called. By comparing the pixel edges of the fry morphological posture area in the continuous frames, the frame points of the overall displacement of the fish body and the local morphological mutation are identified. According to the order of occurrence of the morphological mutation, the fry swimming start, feeding activity and body color formation behaviors are extracted to generate a stage behavior recognition sequence; Continuous behavioral recording data of fry development stages under artificial breeding conditions were collected. First, a high-definition underwater camera was arranged in the fry hatching pond, and the shooting parameters were set to 25 frames per second. The entire process of the fry from hatching to body color formation was continuously filmed. After obtaining the image sequence, each frame of the image was numbered in chronological order and accompanied by a shooting timestamp, such as the 0th second was the 0th frame and the 1st second was the 25th frame. A total of 45,000 frames were recorded for 30 minutes. Subsequently, continuous frame segments in the image sequence, such as from the 300th frame to the 1800th frame, were called. All image frames in this segment were processed, and the fry pixel area was extracted in each frame. By setting the initial background model and combining the contour detection method, the pixels of the fry boundary area were extracted. Then, the center of mass calculation method was used to determine the coordinate position of the center of mass of the fish body in each frame. , calculate the displacement between consecutive frames , as in Frame and Displacement of the fry's center of mass between frames Pixels, exceeding the set threshold of 5 pixels, it is determined to be an overall displacement frame. On this basis, the local grayscale structure changes of the fish body are calculated, and local areas of each frame of the fish body image, such as the tail and abdomen, are extracted. A 16-dimensional hash value is generated for each area using the mean hashing algorithm. The Hamming distance is compared between consecutive frames. If the hash distance of the tail area is 8 between frames 420 and 421, which is greater than the set mutation threshold of 6, it is determined to be a local morphological mutation frame. Each frame mutation point is added to the behavior judgment candidate list and arranged in ascending order by frame number. Behavior judgment is performed on the segments with consecutive mutation frames in the sequence. For example, if frames 312 to 320 are continuously determined to be displacement frames, they are defined as swimming start behavior. Combined with the obvious opening and closing changes of the mouth contour between frames 420 and 426, it is defined as feeding active behavior. From frame 880, a pigment deposition area with a grayscale value of 90 to 130 appears, which lasts until frame 890, and is determined to be body color formation behavior. The final behavior recognition sequence is .

[0026] S102: Based on the staged behavior recognition sequence, count the consecutive frame segments of each behavior in the image sequence, record the start and end frames of the behavior, calculate the duration of the behavior by analyzing the timestamps corresponding to the frame number intervals, calibrate the time positioning points, and generate behavior state duration positioning information; Recognition sequence based on staged behavior , respectively, expand ±10 frames before and after each behavior frame point to establish a continuous frame segment. For example, the behavior segment corresponding to the 312th frame is the 302nd to 322nd frame. The system performs statistical processing on the image in each frame, reads the fish body area contour and center of mass coordinates in each frame image, and determines whether the change range of the center of mass coordinates between consecutive frames is within the set range If there are 8 consecutive frames that meet this condition within the pixel, it is marked as the valid starting frame of the behavior. They are [5.4, 6.1, 6.3, 5.7, 5.1, 5.6, 5.0, 5.9], all of which meet the requirements. Then the 302th frame is marked as the starting frame, and the change range is determined to be maintained. If the change range is maintained until the 322nd frame, If it gradually drops to 2.1 and is smaller than the threshold for three consecutive frames, the 319th frame is marked as the termination frame, and the continuous frame length of the behavior is calculated as frame, multiplied by the frame interval 0.04 seconds to get the duration The rest of the behaviors, such as the feeding behavior segment from frame 420, range from frame 410 to frame 430. The analysis shows that the width of the fish mouth shrinks from 12 pixels to 9 pixels in consecutive frames, with a change of more than 25%, which is judged to be an effective feeding action. The start frame is marked as 410 and the end frame is marked as 428, so the continuous frame length is 19 frames, corresponding to a duration of 0.76 seconds. The body color formation behavior starts from frame 880, and the average RGB brightness value of the body surface increases from (90, 85, 82) to (120, 115, 110). The standard deviation of the grayscale value in consecutive frames increases from 5 to 15, which lasts until frame 896, marked as start 880 and end 896, with a continuous frame length of 17 frames, corresponding to a duration of 0.68 seconds. Finally, the system generates the start and end frames, continuous frame length, and duration of each behavior and organizes them into the following sets: Swimming start: (302, 319, 18 frames, 0.72 seconds), feeding activity: (410, 428, 19 frames, 0.76 seconds), body color formation: (880, 896, 17 frames, 0.68 seconds) .

[0027] S103: Based on the behavior state duration location information, the order of behavior occurrence time in the frame sequence is called, the time periods of each type of behavior are numbered in sequence, and the behaviors are classified into corresponding stage segments according to the numbers, thus completing the sequence organization of the staged behaviors and obtaining the time sequence of the behavior node achievement; Extract the starting frame number of each behavior based on the behavior state duration positioning information And arrange them in ascending order, and set the frame sequence of the entire image to to Frame, first of all, the starting frame of the behavior is used as the segmentation point to generate stage time segments, the first stage is 0 to 301 frames, the second stage is 302 to 409 frames, the third stage is 410 to 879 frames, and the fourth stage is 880 to 1200 frames. Each segment is numbered as stage 0, stage 1, stage 2, and stage 3. Then, the frame interval of each behavior is mapped to its stage number. The swimming start behavior is from 302 to 319 frames and is classified into stage 1. The feeding active behavior is from 410 to 428 frames and is classified into stage 2. The body color formation behavior is from 880 to 896 frames and is classified into stage 3. The behavioral organization structure is stage 1: swimming start, stage 2: feeding active, stage 3: body color formation, and the start and end time of each stage is further converted into a time positioning point through the frame number. For example, the start and end time of stage 1 is Seconds to seconds, and stage 2 corresponds to Seconds to seconds, and stage 3 corresponds to Seconds to seconds, and the final behavior node stage sequence is Stage 1: (12.08s, 16.36s, swimming starts), Stage 2: (16.4s, 35.16s, feeding is active), Stage 3: (35.2s, 48.0s, body color formation) .

[0028] See also Figure 1 , the steps of S2 are as follows: S201: Using behavior nodes to achieve a time sequence, calling the original recorded data of water temperature, light, pH and dissolved oxygen of the fry group at each node, integrating each data item by time label, filtering and grouping according to the time labels under multiple behavior nodes of the fry group, marking the node corresponding to the environmental factor data segment, and generating the node environment sequence interval value; To achieve time series using behavior nodes, we first need to set multiple behavior nodes based on the behavioral characteristics of the fry, such as eating, aggregating, and dispersing. Each behavior node records the corresponding time tag, and then calls the original recorded data of the water environment in which the fry are located for the time tag, including water temperature, light intensity, pH value, and dissolved oxygen concentration. During the call process, the accuracy matching window needs to be determined according to the recording frequency of the recording device. For example, if the sampling frequency of environmental parameters is once every 10 minutes, it is necessary to filter the environmental data segment of ±5 minutes before and after the behavior node time to complete the node docking. Next, the above environmental parameters are synchronized and integrated according to the time tag to ensure that the four types of environmental factor data in the time dimension of each behavior node can be aligned. For example, if the behavior node T1 = 2024-06-01 10:00, then 2024-06-01 is extracted from the original data set. The sampling values ​​of water temperature, light, pH and dissolved oxygen from 09:55 to 10:05 are indexed and stored as segment data for all data points in this time period. If the water temperature collection value is 22.3, 22.5, 22.4℃, the light is 300, 310, 295lx, the pH value is 7.1, 7.2, 7.1, and the dissolved oxygen is 6.5, 6.6, 6.5mg / L, then this data segment is marked as the environmental data segment of node T1. In the screening and grouping process, all data points in this time period are indexed and stored as segment data. The behavior nodes are traversed in order of their timestamps, and the time ranges corresponding to different nodes are segmented and numbered for archiving, ultimately forming a behavior node environmental sequence interval value set. Its structure can be represented by key-value pairs, such as T1: [22.3, 22.5, 22.4], [300, 310, 295], [7.1, 7.2, 7.1], [6.5, 6.6, 6.5], T2: […]. Each data segment serves as the basis for subsequent environmental factor trend analysis.

[0029] S202: Calling the node environmental sequence interval value, based on the increase and decrease change direction of water temperature, light, pH and dissolved oxygen in the corresponding time period, performing adjacent comparisons on each data item at the time interval within the node, calibrating the difference interval between adjacent data points, and dividing the time period into a stable segment and a changing segment according to the fluctuation rate threshold, to generate the environmental factor fluctuation trend segment value; Call the node environment sequence interval value. First, take each node in the time series as the analysis unit, and calculate the difference between adjacent sampling points of the four parameters of water temperature, light, pH and dissolved oxygen within the node. For example, the water temperature in node T1 is 22.3, 22.5, and 22.4℃, with adjacent differences of 0.2 and -0.1℃, and the light intensity is 300, 310, and 295lx, with adjacent differences of 10 and -15lx. The direction of each group of differences is determined, that is, the positive or negative sign is judged by the sign to obtain an increasing or decreasing trend, forming a fluctuation trend sequence classified as "increasing", "decreasing", and "stable". Then calculate the fluctuation rate of each environmental factor in the time segment of the corresponding behavior node, and set the threshold by combining the median fluctuation rate and median absolute deviation of the same factor of all nodes, such as water temperature fluctuation. The median rate is set to 0.2°C / 10min, and the MAD is 0.05. The fluctuation rate threshold is 0.2 + 1.5 × 0.05 = 0.275°C / 10min. This threshold is used to determine whether a certain period of time is a changing segment. When the water temperature rate at two consecutive sampling points is 0.3°C / 10min, it is greater than the threshold and is judged to be a changing segment. Otherwise, it is a stable segment. By traversing the adjacent time periods of each factor and comparing its rate with the corresponding threshold, the parameter time period of each node is segmented according to whether it exceeds the threshold. Finally, the environmental factor fluctuation trend segment value marked as stable or changing is generated, such as the water temperature changing segment of node T1: [09:55–10:00], the stable segment: [10:00–10:05], and the light changing segment: [09:55–10:05]. The fluctuation rate threshold is set by calculating the median of the change rate of each environmental factor in the time series corresponding to the behavior node plus 1.5 times the median absolute deviation.

[0030] S203: Based on the fluctuation trend segment values ​​of the environmental factors, the fluctuation phase data segments of water temperature, light, pH, and dissolved oxygen under the behavior nodes are matched, the mean, range, and fluctuation amplitude of the four environmental factors corresponding to each node are calculated, and data classification and integration and time point alignment are performed to generate an environmental parameter set; According to the fluctuation trend segment value of the environmental factor, the environmental data segment corresponding to each behavior node is matched and extracted. For each parameter, its mean, range and fluctuation amplitude are calculated in the fluctuation trend segment. The mean is obtained by summing all the values ​​in the segment and dividing it by the number of data points. For example, the water temperature value of node T1 is 22.3, 22.5, and 22.4, then the mean is (22.3+22.5+22.4) / 3=22.4℃, the range is the maximum value minus the minimum value 22.5–22.3=0.2℃, and the fluctuation amplitude is the difference between each adjacent value. The average absolute value of (0.2 + 0.1) / 2 = 0.15 °C is calculated. The same process is applied to light, pH, and dissolved oxygen. Subsequently, the four types of parameters at different nodes are aligned and summarized at a unified time point. In the data classification and integration, a structured approach is used to generate an environmental parameter set. Each parameter set should include five items: node number, time period, mean, range, and fluctuation amplitude. To ensure data consistency, samples that do not have the same behavior node need to be eliminated or interpolated to form a standardized behavior node environmental parameter set. The structure is shown in the following table. Table 1 Example of environmental parameter set

[0031] As shown in Table 1, the environmental parameter set of node T1 can be used to show the detailed characteristic values ​​of the four environmental factors in the corresponding time period, which is convenient for subsequent analysis or modeling operations.

[0032] See also Figure 1 , the specific steps of S3 are: S301: Using the environmental parameter set and the time period information corresponding to the behavior node, extract the water temperature, light, pH, and dissolved oxygen data series covered by each behavior segment on the time axis, integrate them into continuous observation intervals in chronological order, set start and end time tags for the intervals, and generate observation time interval values; Using the environmental parameter set and combining the time period information corresponding to the behavior node, we first obtain the behavior node time period recorded by the fry behavior recognition system, including the start and end time of each behavior, corresponding to behaviors such as feeding, swimming, and resting, and extract the time series data of four types of environmental parameters, namely water temperature, light, pH, and dissolved oxygen, covered by the behavior segment on the time axis. The above four types of parameters are sliced ​​according to the start and end time of the behavior. For example, if the start time of a feeding behavior segment is 08:30 on May 1, 2024, and the end time is 08:45, it is necessary to extract the observed values ​​of water temperature, light, pH, and dissolved oxygen in this time period from the records collected by the environmental monitoring equipment. Assuming that the water temperature is collected once per minute, there are 16 water temperature data points in this behavior segment, which are recorded as , light, pH, and dissolved oxygen also use the same frequency to extract data, which are recorded as ,, ,, Then, the various environmental parameter data of each behavior segment are merged and integrated into a set of continuous observation interval sequences in chronological order to form a matrix data structure. Each row of the matrix is ​​a time point, and the columns are five pieces of information: time, water temperature, light, pH, and dissolved oxygen. During the integration process, the start time and end time corresponding to each behavior segment need to be set as the label value, that is, the time label of the observation time interval, which is used for subsequent event calibration and response matching processing. For example, the start and end times of behavior segment 1 are 08:30 and 08:45 respectively, and the start and end times of behavior segment 2 are 09:00 and 09:10, then the final observation time interval value is , forming a standard observation unit.

[0033] S302: Calling the observation time interval value, calculating the change rate of the corresponding data point based on the time series change amplitude of water temperature, light, pH and dissolved oxygen in the interval and setting the disturbance identification change amplitude threshold, marking the time segment where the rate exceeds the threshold as a disturbance segment, and combining the fry behavior data in the node to identify the starting point of the offset feature, and obtaining the disturbance and offset starting time point pair; Call the above observation time interval value, calculate the time series change rate of water temperature, light, pH and dissolved oxygen in each interval respectively, and process it by using the change amount of adjacent data points and time interval calculation method. For example, if the water temperature rises from 20.5°C to 21.2°C in one minute, the minute change rate is , repeat this action for all adjacent data points in the entire behavior interval to obtain the rate sequence , the change rate sequences of the other three parameters are obtained in the same way. When setting the change amplitude threshold for disturbance identification, the median and standard deviation of the change amplitude sequence of a parameter in the behavior segment must be extracted first. For example, the pH value change amplitude sequence is , the median is 0.035, and the standard deviation is , then the disturbance judgment threshold is When the pH change rate exceeds 0.095 in any time period, it is determined to be a disturbance time segment. This determination method is performed independently in all parameter dimensions. Then, combined with the behavioral node data corresponding to the time segment, it is observed whether the fry activity has an offset feature. For example, the fry that was originally in a static state begins to swim violently. The time point when the offset behavior first appears and the corresponding disturbance time point form a pair of time points, which are recorded as the disturbance start time. With offset start time For example, if the disturbance time period is 08:35 and the offset behavior starts at 08:36, then the time point pair is .

[0034] Table 2 Example of environmental parameter change amplitude sequence and disturbance threshold setting

[0035] As shown in Table 2 , the disturbance thresholds of different environmental factors in different behavioral segments are different, which are determined by the actual measured data in each segment and are adaptively set by adding twice the standard deviation to the median of the variation range; The disturbance identification change amplitude threshold is an adaptive mutation judgment benchmark set by extracting the median of the environmental factor change amplitude sequence and adding 2 times the standard deviation.

[0036] S303: Based on the disturbance and offset start time point pairs, calculate the time interval between each set of disturbance start points and the first occurrence time of the corresponding offset behavior, associate the corresponding behavior segment identifier and disturbance type for each set of time intervals, integrate them into structured table data, and obtain disturbance-behavior response pairing data; Based on the disturbance and offset start time point pairs obtained above, the time interval of each time pair is calculated. For example, if the disturbance time is 08:35 and the offset time is 08:36:30, the time interval is 90 seconds. This processing is performed on all time point pairs to obtain an array of disturbance response interval values. Then, each group of interval values ​​is associated with the behavior segment identifier and disturbance source type. For example, the response interval generated by the water temperature disturbance in behavior segment 001 is 60 seconds, and the response interval generated by the pH disturbance in behavior segment 002 is 120 seconds. This forms structured table data, in which each row contains five items: behavior segment number, disturbance parameter type, disturbance start time, offset start time, and time interval. This table serves as the input source of disturbance-behavior response pairing data for subsequent analysis.

[0037] See also Figure 1 , the steps of S4 are as follows: S401: Using the disturbance-behavior response paired data, identify the direction of the fry position change at each behavior node, extract the coordinate point displacement value and direction angle in the continuous response segment, perform consistency analysis on the offset direction of consecutive time points, and generate the behavior offset continuous direction degree; Using the disturbance-behavior response paired data, we first extracted the position coordinate information of the fry in the water body before and after the disturbance for each behavior node after the disturbance event, and recorded the position points in the form of two-dimensional coordinates. For example, point A is the position before the disturbance (100, 200), and point B is the response point 2 seconds after the disturbance (104, 207). By calculating the displacement value between the coordinate points ≈8.06mm to obtain the spatial displacement, and then calculate the offset direction angle ((207−200) / (104−100))≈ (1.75)≈60.25°. The above process requires traversing the behavior nodes under each disturbance event, extracting the continuous position points within the response time period in the time series, and obtaining the fry trajectory points with a sampling unit of 0.5 seconds. In the continuous response segment, if the angular variation of the 6 groups of coordinate points within 3 seconds is less than ±15°, it is considered to be consistent in direction. In order to judge the degree of directional consistency, the angular difference between adjacent time points is counted and the directional consistency is calculated. The calculation process is to count the standard deviation of the continuous directional difference σθ. If σθ≤10°, it is defined as a high consistency segment. The specific value is obtained from the directional sequence such as [60°, 62°, 63°, 59°, 61°]. The average value is calculated to be 61°. The difference between each directional angle and the average value is [-1, 1, 2, -2, 0]. The sum of squares is 10, the variance = 10 / 5 = 2, and the standard deviation ≈1.41°, judging that σθ=1.41° is far below the set judgment standard of 15°, and the behavior deviation is determined to be continuous and consistent; the continuous direction consistency segment of each node is extracted and archived as "behavior deviation continuous direction degree" data, which can be further used to classify the behavior response type; A continuous response segment refers to a period of behavioral data in which the fry continuously deviate in the same direction within a certain period of time after the disturbance occurs.

[0038] S402: Invoke the behavior deviation continuity direction, classify and count the number of identified disturbances in the behavior node and the corresponding response time difference, compare the response amplitude and delay value of each node under the same disturbance condition, and extract the difference parameters to obtain the disturbance response difference index value; The specific formula for comparing the response amplitude and delay value of each node under the same disturbance conditions is: ; Calculate the disturbance response difference index value; in, For the The disturbance response difference index value of each behavior node under all disturbance conditions, For the The behavior node is in The amplitude deviation normalization parameter under disturbance conditions, For the The behavior node is in The delay deviation normalization parameter under disturbance conditions, For the The perturbation structure coupling factor corresponding to the perturbation condition is: For the The response modulation gain coefficient corresponding to the disturbance condition is, is the total number of perturbation conditions, is the index number of the behavior node, is the index number of the disturbance condition; formula: ; Detailed explanation of the formula and the process of formula calculation and derivation: The formula is used to calculate the response difference index value of the behavior node under multiple disturbance conditions, and the results are used to evaluate the sensitivity and response consistency of each node to disturbance; Parameter meaning and setting value: : No. The behavior node is in The amplitude deviation normalization parameter under disturbance conditions is calculated as: ,in, For the The behavior node is in The response amplitude value under the disturbance condition is For the The average value of the response amplitude of all behavior nodes under the disturbance condition; : No. The behavior node is in The delay deviation normalization parameter under disturbance conditions is calculated as: ,in, For the The behavior node is in The response delay time under disturbance conditions is For the The average response delay time of all behavior nodes under disturbance conditions; : No. The disturbance structure coupling factor corresponding to each disturbance condition reflects the degree of coupling of the disturbance in the system structure. By analyzing the propagation path and impact range of the disturbance in the system structure, the structural impact factor method is used to quantify it, with a value range of 0.1 to 0.9; : No. The response modulation gain coefficient corresponding to each disturbance condition represents the modulation intensity of the disturbance on the node response. The amplitude of the node response change caused by the disturbance is measured experimentally and fitted using the least squares method, with a value range of 0.5 to 1.5. : The total number of perturbation conditions, set to 3; : The index number of the behavior node, set to 1; : The index number of the disturbance condition, ranging from 1 to ; Set the following parameter values: , ; , ; , ; , ; , ; , ; , ; , ; , ; Substitute the parameters into the formula for calculation: Calculate each item: ; ; ; ; ; ; calculate : ; The result of 0.1121 indicates that the response of the first behavior node under the three disturbance conditions is relatively low and has good response consistency. This result can be used to prioritize behavior nodes, prioritizing nodes with large response differences to improve the overall stability and response efficiency of the system.

[0039] S403: Grouping and classifying the response amplitude, direction consistency, and response delay combinations according to the disturbance response difference index value, integrating and labeling multiple groups of behavior node numbers, screening stable structured combinations, and establishing a disturbance response valid node data set; According to the disturbance response difference index value, the three parameters of response amplitude, direction consistency and response delay are extracted and classified. First, the amplitude of the disturbance response corresponding to each behavior node is defined as the maximum displacement value after the disturbance minus the displacement value corresponding to the initial position before the disturbance. For example, if the maximum displacement reaches 12.3 within 3 seconds after the disturbance, mm, the static displacement before disturbance is 0.5 mm, then the response amplitude is 11.8 mm; the directional consistency is calculated with reference to the standard deviation σθ in S401, and the response delay is defined as the time difference between the disturbance time point and the time point where the direction-consistent offset is first detected. If the disturbance time is 10:00:00 and the time of the first offset point is 10:00:01.8, then the response delay is 1.8 seconds; after quantifying and summarizing the above three indicators for all behavior node numbers, perform grouping and classification operations. First, set the upper threshold of the response amplitude fluctuation coefficient (CV) to 0.25, that is, CV = standard deviation / mean. If the response amplitude value within a certain combination group is [11.8, 11.7, 12.0, 11.9] mm, then the mean is 1 1.85mm, standard deviation ≈ 0.1118mm, CV = 0.1118 / 11.85 ≈ 0.0094, which is much smaller than the threshold of 0.25; the directional consistency requires that the mean σθ within the group does not exceed 8°, and the standard deviation of the response delay must be controlled within 0.5 seconds. If the delay time of a group is [1.8, 1.6, 1.9, 1.7] seconds, the mean is 1.75 seconds, and the standard deviation is ≈ 0.1118 seconds, it meets the set standards; the behavior node combination that meets the stability requirements of the three indicators is marked as a "stable structured combination", and these combinations are included in the disturbance response valid node data set. In the final data structure, each record contains three fields: node number, amplitude value, directional consistency, and response delay, as shown in Table 3; Table 3 Disturbance response effective node data table

[0040] As shown in Table 3, the response amplitudes of the four nodes differ slightly, the directional consistency is less than 2°, and the standard deviation of the response delay is only 0.1118 seconds, meeting the criteria for a stable structured combination. This result indicates that the above node combination has regular characteristics in the disturbance response and can be included in the disturbance response effective node dataset; Screening stable structured combinations refers to extracting node combinations with parameter concentration and behavioral regularity characteristics based on the joint stability standard of three indicators: response amplitude fluctuation coefficient, directional consistency and delay time standard deviation.

[0041] See also Figure 1 , the specific steps of S5 are: S501: Integrate the disturbance response effective node data set and the behavior node achievement time series data, extract the disturbance type, behavior response time and offset direction information corresponding to each node, unify the time label format, establish the corresponding structure between the disturbance and behavior performance between nodes, and generate the disturbance behavior association matching table; Integrate the disturbance response valid node dataset and the behavior node achievement time series data. First, extract the disturbance type field recorded in each valid node. Its value comes from the disturbance event classification identified in the previous step, such as water temperature mutation, pH mutation, light reduction or dissolved oxygen drop, etc. Its value is set to a string type enumeration item. Then call the response time information of the corresponding behavior node. This time field records the first achievement time of behaviors such as fry displacement, activity enhancement, and position migration. It needs to be converted into a standard timestamp format such as "2024-05-0108:34:00" and then combined with the offset direction information output by the behavior recognition module. This information is usually calculated by the camera tracking system and its value is in the form of a two-dimensional vector. For example, if the fry swims from the initial position to the upper right, the offset direction vector is On this basis, the time information involved in all nodes is unified into the Unix timestamp format (i.e., the number of seconds from January 1, 1970) to facilitate subsequent sorting and processing. At the same time, a mapping relationship structure is established for different disturbance types and behavioral response events. Each row records the disturbance type, disturbance occurrence time, response time, offset direction vector, behavior node number and other fields of a node. Finally, a disturbance behavior association matching table is constructed. The matching representation is shown in the following table: Table 4. Perturbation behavior correlation matching table

[0042] As shown in Table 4, the structured matching data is constructed by extracting the disturbance type, disturbance and response time points node by node and calculating their timestamps. The offset direction vector is then uniformly represented.

[0043] S502: Calling the disturbance behavior association matching table, based on the changing trends of water temperature, light, pH, and dissolved oxygen before and after the node disturbance, identifying the correlation pattern between the change direction of environmental factors and the order of behavioral responses during the disturbance period, normalizing the offset angle distribution and disturbance characteristic values, and obtaining the standardized response matching parameter value; Call the disturbance type, response time point and offset direction vector corresponding to each node in the disturbance behavior association matching table, extract the data sequence of water temperature, light, pH and dissolved oxygen in the corresponding time period before and after the disturbance, and segment the data before and after the disturbance. Take the data 30 seconds before the disturbance time before the disturbance, and take the data 30 seconds after the response time after the disturbance. Calculate the mean change trend of the four environmental factors in the time period. For example, if the mean water temperature before the disturbance is 20.8°C and the mean water temperature after the disturbance is 21.6°C, the change direction is warming, and the direction is assigned to +1. If the dissolved oxygen before the disturbance is 7.5mg / L and after the disturbance is 7.2 mg / L, the direction is assigned to -1, and the change direction vector is constructed for all nodes accordingly. Then, the behavioral response and the order of the disturbance event are analyzed to determine whether the response lags behind the disturbance. The judgment is made based on whether the time difference is greater than 5 seconds. If the result of subtracting the disturbance time from the response time is greater than 5 seconds, it is marked as "delayed response", otherwise it is "synchronous response". Then, the offset direction angle of each node is extracted and an angle set is constructed. For example, the offset direction is 45° to the southeast, 90° to the south, and other angle data. Then, the disturbance eigenvalues ​​such as the environmental change amplitude and the response angle are uniformly processed by the range normalization method. For each eigenvalue According to the formula: , for calculation, for example, the maximum value of the dissolved oxygen variation amplitude in each node of a group is 0.8, the minimum value is 0.1, and the dissolved oxygen disturbance value of a node is 0.6, then the normalized value of the disturbance is: For example, the offset angle is extracted from each node, the maximum angle is 180°, and the minimum angle is 0°. If the angle of a node is 90°, its normalized value is: This mapping is performed on the above disturbance values ​​and response angle values ​​to obtain standardized parameter value pairs for all nodes, which facilitates the construction of a matching feature set at a unified scale. This result shows that the matching parameter structure between the disturbance feature and the behavioral response direction at each node has been completed through a standardized scale. Standardization refers to mapping the disturbance and response parameters uniformly to the range of 0 to 1 according to the range formula, which is used to eliminate scale differences and facilitate feature comparison.

[0044] S503: Based on the standardized response matching parameter values, combined with the node behavior change sequence and the fluctuation stage difference of the environmental factors, the response chain in the disturbance path is extracted, and the optimal combination interval of water temperature regulation value, light cycle ratio, pH regulation amplitude and dissolved oxygen regulation frequency under each disturbance type is calculated. The corresponding behavior response nodes are mapped to the regulation matrix to generate a behavior regulation mapping map; The specific formula combining the order of node behavior changes and the difference in the fluctuation stages of environmental factors is: ; Calculate the disturbance response difference value ΔR; in, is the disturbance response difference value of the i-th type of disturbance in the behavioral response dimension, is the amplitude of the node behavior change under the i-th type of disturbance in the j-th stage, is the environmental factor fluctuation intensity weight corresponding to the i-th type of disturbance in the j-th stage, is the total number of fluctuation stages contained in the perturbation path, is the arithmetic mean of the behavior change amplitude of the nodes in all n stages of the i-th type of disturbance, is the critical disturbance threshold in the behavioral response process corresponding to the i-th type of disturbance, is the average value of the disturbance trend offset intensity of the i-th type disturbance in t observation periods, is the behavioral dimension identifier of the disturbance response type, is the disturbance intensity index of behavior change, is the weight source index of environmental factor fluctuations, is the superscript of the threshold index of the critical condition, is the total number of disturbance trend observation cycles index superscript, is the perturbation type index, is the volatility phase index; formula: ; Detailed explanation of the formula and the process of formula calculation and derivation: The formula is used to calculate the disturbance response difference value of the i-th type of disturbance in the behavioral response dimension, and the obtained result is used to evaluate the impact of the disturbance on the system behavior; Parameter meaning and setting value: : The amplitude of the node behavior change under the i-th type of disturbance in the j-th stage, set to 0.6; : The weight of the environmental factor fluctuation intensity corresponding to the i-th type of disturbance in the j-th stage, set to 0.8; : The total number of fluctuation stages contained in the perturbation path, the set value is 5; : The arithmetic mean of the behavior change amplitude of the node of the i-th type disturbance in all n stages, through The average calculation was obtained in all stages, and the set value was 0.5; : The critical disturbance threshold in the behavioral response process corresponding to the i-th type of disturbance. Through historical data analysis, the threshold at which the system behavior changes significantly is determined, and the set value is 0.4.

[0045] : The average value of the disturbance trend offset intensity of the i-th type disturbance in t observation periods, with a set value of 0.1; Substitute the parameters into the formula for calculation: Calculate the weighted average behavioral change: ; Compute the square root term: ; Calculate the disturbance response difference value: ; The result of 0.033 indicates that the difference in the disturbance response for type i under the behavioral response dimension is small, indicating that the disturbance has a low impact on system behavior. This result can be used to further assess the impact of disturbances on system stability and guide appropriate management measures.

[0046] See also Figure 2 A fish seed breeding system is provided, which is used to implement the above-mentioned fish seed breeding method, and the system comprises: The behavior sequence extraction module obtains the observed data of fry behavior and segments it into a time series. It obtains the behaviors of swimming initiation, feeding activity, and body color formation, forms behavioral parameters, performs interval division according to the duration and occurrence order of the behaviors, analyzes the sequence of behaviors and completes the numbering, extracts the achievement time, and outputs the achievement time series of the behavior nodes, which is then passed to the environmental feature construction module and the path map generation module. The environmental feature construction module obtains water temperature, light, pH, and dissolved oxygen records based on the time series of behavior nodes, analyzes the execution trend of the node time period, identifies the fluctuation frequency and change direction of each environmental parameter, outputs the environmental parameter set and passes it to the disturbance response matching module; The disturbance-response matching module, based on the environmental parameter set and the behavior segment corresponding to the behavior node reaching time, identifies the disturbance starting point and response time point, extracts the corresponding observation interval, analyzes the disturbance process and offset performance, calculates the time difference and completes the disturbance-response pairing annotation, outputs the disturbance-behavior response pairing data and passes it to the behavior offset classification module; The behavior deviation classification module obtains the disturbance type and deviation performance based on the disturbance-behavior response pairing data, groups them according to the disturbance source and response continuity, analyzes the response differences under the same disturbance, outputs the disturbance response valid node data set and passes it to the path map generation module; The path map generation module, based on the disturbance response effective node data set and the behavior node achievement time series, identifies the offset characteristics and corresponding disturbance types after behavior standardization, combines the behavior change sequence with the environmental fluctuation process, analyzes the action chain, adjusts the regulation parameters of water temperature, light cycle, pH and dissolved oxygen, and generates a behavior regulation mapping map.

[0047] The above embodiments can be implemented in whole or in part via software, hardware (e.g., circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0048] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0049] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0050] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0051] 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 the present invention.

[0052] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0053] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0054] The units described as separate components may or may not be physically separate, and the 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.

[0055] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0056] If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical disks.

[0057] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for breeding fish species, characterized in that: The method comprises: S1: Continuous behavioral data is collected during the developmental stages of fry under artificial breeding conditions. By identifying the initiation of swimming, active feeding, and body color formation, the state of each behavior is classified and time-located. The behaviors of each stage are numbered and segmented using chronological order to obtain the time series of behavioral node achievement. S2: Using the behavior nodes to achieve a time series, calling the change records of the water environment in which the fry are located, performing trend identification and fluctuation classification on the water temperature, light, pH and dissolved oxygen data, matching the environmental parameters of the corresponding behavior nodes, and generating an environmental parameter set; S3: Using the environmental parameter set, based on the behavior segments corresponding to the behavior node achievement time, combined with the environmental change process, extract the corresponding observation time interval, identify the disturbance process and behavior deviation manifestation within the interval, analyze the sequential relationship between the disturbance initiation and the initial appearance of the behavioral response, calculate the corresponding time interval and record the pairing information, and obtain the disturbance-behavior response pairing data; S4: Using the disturbance-behavior response pairing data, identify the behavior deviation direction and continuity within the behavior node, analyze the disturbance superposition and response differences, standardize the grouped response behavior, and obtain the disturbance response valid node data set.

2. The fish breeding method according to claim 1, wherein: The behavior node achievement time series includes the behavior stage number, the node achievement time point, and the behavior classification label; the environmental parameter set includes the water temperature change trend, the light fluctuation range, and the pH stage change value; the disturbance-behavior response paired data includes the disturbance start time point, the response first time point, and the time interval value; the disturbance response valid node data set includes the direction consistency coefficient, the response delay standard value, and the disturbance difference grouping identifier.

3. The fish breeding method according to claim 1, wherein: The steps of S1 are as follows: S101: Continuous behavioral record data of fry development stages under artificial breeding conditions is collected, and continuous frame segments in the time-series image are called. By comparing the pixel edges of the fry morphological posture area in the continuous frames, the frame points of the overall displacement of the fish body and the local morphological mutation are identified. According to the order of occurrence of the morphological mutation, the fry swimming start, feeding activity and body color formation behaviors are extracted to generate a stage behavior recognition sequence; S102: Based on the staged behavior recognition sequence, statistics are collected on the continuous frame segments of each type of behavior in the image sequence, and the start frame and end frame of the behavior are recorded. By analyzing the timestamps corresponding to the frame number intervals, the duration of the behavior is calculated and the time positioning point is calibrated to generate the behavior state duration positioning information; S103: Based on the behavior state duration positioning information, the order of behavior appearance time in the frame sequence is called, the time periods to which each type of behavior belongs are numbered in sequence, and the behaviors are classified into corresponding stage segments according to the numbers, thereby completing the sequence organization of the staged behaviors and obtaining the time sequence of the behavior node achievement.

4. The fish breeding method according to claim 3, characterized in that: The specific steps of S2 are: S201: Using the behavior nodes to achieve a time sequence, calling the original recorded data of water temperature, light, pH and dissolved oxygen of the fry group at each node, integrating each data item according to the time label, filtering and grouping according to the time labels under multiple behavior nodes of the fry group, marking the node corresponding to the environmental factor data segment, and generating the node environment sequence interval value; S202: calling the node environmental sequence interval value, performing adjacent comparisons on each data item at the time interval within the node based on the increase and decrease change direction of water temperature, light, pH, and dissolved oxygen in the corresponding time period, calibrating the difference interval between adjacent data points, and dividing the time period into a stable segment and a changing segment according to the fluctuation rate threshold, thereby generating the environmental factor fluctuation trend segment value; The fluctuation rate threshold is set by calculating the median of the change rate of each environmental factor in the time series corresponding to the behavior node plus 1.5 times the median absolute deviation; S203: According to the fluctuation trend segment value of the environmental factors, the fluctuation stage data segments of water temperature, light, pH and dissolved oxygen under the behavior node are matched, the mean, range and fluctuation amplitude of the four environmental factors corresponding to each node are calculated, and data classification and integration and time point alignment processing are performed to generate an environmental parameter set.

5. The fish breeding method according to claim 4, characterized in that: The specific steps of S3 are: S301: Using the environmental parameter set and the time period information corresponding to the behavior node, extract the water temperature, light, pH, and dissolved oxygen data series covered by each behavior segment on the time axis, integrate them into continuous observation intervals in chronological order, set start and end time tags for the intervals, and generate observation time interval values; S302: Calling the observation time interval value, calculating the change rate of the corresponding data point based on the time series change amplitude of water temperature, light, pH and dissolved oxygen in the interval and setting a disturbance identification change amplitude threshold, marking the time segment where the rate exceeds the threshold as a disturbance segment, and combining the fry behavior data in the node to identify the starting point of the offset feature, to obtain a pair of disturbance and offset starting time points; The disturbance identification change amplitude threshold is an adaptive mutation determination benchmark set by extracting the median of the environmental factor change amplitude sequence and adding 2 times the standard deviation; S303: Based on the disturbance and offset start time point pairs, calculate the time interval between each set of disturbance start points and the first occurrence time of the corresponding offset behavior, and associate the corresponding behavior segment identifier and disturbance type for each set of time intervals, integrate them into structured table data, and obtain disturbance-behavior response pairing data.

6. The fish breeding method according to claim 5, characterized in that: The specific steps of S4 are: S401: using the disturbance-behavior response paired data, identifying the direction of the fry position change at each behavior node, extracting the coordinate point displacement values ​​and direction angles in the continuous response segments, performing consistency analysis on the offset directions at consecutive time points, and generating a behavior offset continuous directionality; The continuous response segment refers to the behavioral data period of the fry that continues to deviate in the same direction within a certain period of time after the disturbance occurs; S402: calling the behavior deviation continuous directionality, classifying and counting the number of identified disturbances and the corresponding response time differences in the behavior node, comparing the response amplitude and delay value of each node under the same disturbance condition, and extracting the difference parameters to obtain a disturbance response difference index value; The specific formula for comparing the response amplitude and delay value of each node under the same disturbance conditions is: ； Calculate the disturbance response difference index value; in, For the The disturbance response difference index value of each behavior node under all disturbance conditions, For the The behavior node is in The amplitude deviation normalization parameter under disturbance conditions, For the The behavior node is in The delay deviation normalization parameter under disturbance conditions, For the The perturbation structure coupling factor corresponding to the perturbation condition is: For the The response modulation gain coefficient corresponding to the disturbance condition is, is the total number of perturbation conditions, is the index number of the behavior node, is the index number of the disturbance condition; S403: Grouping and classifying the response amplitude, direction consistency, and response delay combinations according to the disturbance response difference index value, integrating and marking multiple groups of behavior node numbers, screening stable structured combinations, and establishing a disturbance response valid node data set; The screening of stable structured combinations refers to extracting node combinations with parameter centralization and behavioral regularity characteristics based on the joint stability standard of three indicators: response amplitude fluctuation coefficient, direction consistency and delay time standard deviation.

7. The fish breeding method according to claim 1, characterized in that: The method further comprises: S5: Integrate the disturbance response valid node dataset and the behavior node achievement time series data, identify the normalized behavior offset characteristics and corresponding disturbance types, combine the behavior change sequence with the environmental fluctuation process, analyze the action chain, adjust the water temperature, light cycle, pH and dissolved oxygen regulation parameters, and generate a behavior regulation map; The behavior regulation mapping map includes a regulation parameter matching matrix, a disturbance type action sequence, and a behavior offset response mapping group.

8. The fish breeding method according to claim 7, characterized in that: The specific steps of S5 are: S501: Integrate the disturbance response valid node dataset and the behavior node achievement time series data, extract the disturbance type, behavior response time and offset direction information corresponding to each node, unify the time label format, establish the corresponding structure of disturbance and behavior performance between nodes, and generate a disturbance behavior association matching table; S502: calling the disturbance behavior association matching table, identifying the association pattern between the change direction of environmental factors and the behavioral response sequence during the disturbance period based on the change trends of water temperature, light, pH, and dissolved oxygen before and after the node disturbance, and normalizing the offset angle distribution and disturbance characteristic value to obtain a standardized response matching parameter value; The standardization process refers to uniformly mapping the disturbance and response parameters to the range of 0 to 1 according to the range formula, which is used to eliminate scale differences and facilitate feature comparison; S503: According to the standardized response matching parameter values, combined with the node behavior change sequence and the fluctuation stage difference of the environmental factors, the response chain in the disturbance path is extracted, and the optimal combination interval of the water temperature regulation value, light cycle ratio, pH regulation amplitude and dissolved oxygen regulation frequency under each disturbance type is calculated, and the corresponding behavioral response nodes are mapped to the regulation matrix to generate a behavioral regulation mapping map.

9. The fish breeding method according to claim 1, characterized in that: The specific formula combining the order of node behavior changes and the difference in the fluctuation stages of environmental factors is: ； Calculate the disturbance response difference value ΔR; in, is the disturbance response difference value of the i-th type of disturbance in the behavioral response dimension, is the amplitude of the node behavior change under the i-th type of disturbance in the j-th stage, is the environmental factor fluctuation intensity weight corresponding to the i-th type of disturbance in the j-th stage, is the total number of fluctuation stages contained in the perturbation path, is the arithmetic mean of the behavior change amplitude of the nodes in all n stages of the i-th type of disturbance, is the critical disturbance threshold in the behavioral response process corresponding to the i-th type of disturbance, is the average value of the disturbance trend offset intensity of the i-th type disturbance in t observation periods, is the behavioral dimension identifier of the disturbance response type, is the disturbance intensity index of behavior change, is the weight source index of environmental factor fluctuations, is the superscript of the threshold index of the critical condition, is the total number of disturbance trend observation cycles index superscript, is the perturbation type index, It is the fluctuation phase index.

10. A fish breeding system, characterized in that: The system is used to implement the fish breeding method according to any one of claims 1 to 9, and the system comprises: The behavior sequence extraction module obtains the observed data of fry behavior and segments it into a time series. It obtains the behaviors of swimming initiation, feeding activity, and body color formation, forms behavioral parameters, performs interval division according to the duration and occurrence order of the behaviors, analyzes the sequence of behaviors and completes the numbering, extracts the achievement time, and outputs the achievement time series of the behavior nodes, which is then passed to the environmental feature construction module and the path map generation module. The environmental feature construction module obtains water temperature, light, pH, and dissolved oxygen records based on the time series achieved by the behavior nodes, analyzes the execution trend of the node time period, identifies the fluctuation frequency and change direction of each environmental parameter, outputs the environmental parameter set and passes it to the disturbance response matching module; The disturbance-response matching module, based on the environmental parameter set and the behavior segment corresponding to the behavior node reaching time, identifies the disturbance starting point and response time point, extracts the corresponding observation interval, analyzes the disturbance process and offset performance, calculates the time difference and completes the disturbance-response pairing annotation, outputs the disturbance-behavior response pairing data and passes it to the behavior offset classification module; A behavior deviation classification module obtains the disturbance type and deviation performance based on the disturbance-behavior response pairing data, groups them according to the disturbance source and response continuity, analyzes the response differences under the same disturbance, outputs the disturbance response valid node data set and passes it to the path map generation module; The path map generation module, based on the disturbance response valid node data set and the behavior node achievement time series, identifies the offset characteristics and corresponding disturbance types after behavior standardization, combines the behavior change sequence with the environmental fluctuation process, analyzes the action chain, adjusts the regulation parameters of water temperature, light cycle, pH and dissolved oxygen, and generates a behavior regulation mapping map.

Images