An environmental regulation method for improving the success rate of the breeding of australian lobster parent stock

Abstract

The application discloses an environmental regulation method for improving the breeding success rate of Australian lobster parent stock, relates to the technical field of intelligent regulation of aquatic product breeding environment, and collects environmental data, behavior data and maintenance record data, performs time alignment, abnormality elimination, stage marking and unit mapping processing on the data, and establishes a stage archive, a device archive and a historical state archive; stage identification, deviation amount, risk level, priority and trigger identification are constructed based on standardized input data, environmental target intervals and control instructions corresponding to control stages are generated, and oxygenation, temperature control, water replacement, illumination, water flow regulation and feeding device linkage execution are controlled; after regulation, updated data is collected again, it is judged whether recovery conditions are met, and switching is performed between a conventional control template and abnormal recovery control, thereby forming a closed-loop control chain. The method is beneficial to improving the environmental regulation stability in the process of breeding Australian lobster parent stock.

Description

Technical Field

[0001] This invention relates to the field of intelligent environmental control technology for aquaculture, specifically an environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock. Background Technology

[0002] In the breeding and preservation of Australian lobster broodstock, factors such as water temperature, dissolved oxygen, pH, salinity, ammonia nitrogen concentration, nitrite concentration, light conditions, water flow, and feeding and activity patterns all contribute to the stability of the breeding stock. This requires not only maintaining the breeding environment within a suitable range but also differentiating environmental control strategies based on various stages, including adaptation to new stocking, stable breeding, intensive recovery, and recovery from abnormal conditions.

[0003] From the perspective of existing technological approaches, one type of approach focuses on environmental management related to the farming or breeding processes of Australian freshwater crayfish. For example, prior art with publication number `CN111493013A` discloses a method for farming adult Australian freshwater crayfish, which monitors dissolved oxygen and water temperature using monitoring devices and adjusts them in conjunction with aeration components, inlet pipes, and outlet pipes. Prior art with publication number `CN109122443B` discloses a recirculating aquaculture system for hatching and raising Australian freshwater crayfish, emphasizing the stabilization of water quality and reduction of stress on broodstock crayfish through recirculating aquaculture systems, constant temperature and water treatment units. This approach can improve environmental conditions in specific breeding or farming stages, but its main technological focus is more on farming processes or recirculating aquaculture system configuration. It typically does not establish phased control logic for the preservation and breeding process around a computer system, nor does it incorporate behavioral data and maintenance record data into a closed-loop control chain along with environmental data.

[0004] Another type of existing technology is intelligent monitoring and automatic control systems for general aquaculture scenarios. For example, the prior art with publication number `CN109006635A` discloses an intelligent aquaculture system that collects data through sensors such as temperature, pH, dissolved oxygen, and conductivity. The host computer or controller compares the real-time data with preset normal values ​​and triggers equipment actions such as aeration, water supply and drainage, and feeding when anomalies occur. This approach has a basic closed loop of sensor acquisition, threshold comparison, and equipment control, but it usually uses general threshold triggering logic, which is difficult to reflect the needs of different environmental target ranges and different control templates at different stages in the breeding and preservation of Australian lobster broodstock. It also lacks the construction of key intermediate quantities such as time alignment, anomaly removal, stage marking, unit mapping, risk level, priority, and trigger identifier for multi-source data.

[0005] In the context of breeding and propagating Australian lobster broodstock, the specific shortcomings of the two existing technologies mentioned above are mainly reflected in the following aspects: the former approach is more inclined towards adult lobster farming processes or management of broodstock hatching systems, while the latter approach is more inclined towards general aquaculture automation. Therefore, existing practices usually either rely on farming process parameters or recirculating aquaculture systems to maintain environmental stability, or rely on sensor data collection and threshold comparison to trigger equipment actions. When relying solely on monitoring a single water quality parameter or using general threshold control, it is difficult to distinguish the environmental control objectives at different stages of breeding, making it difficult for the same set of control logic to take into account states such as pond adaptation, stable breeding, enhanced recovery, and abnormal recovery. Maintenance records such as abnormal behavior, cannibalism, mortality, or abnormal molting are also difficult to enter into the control decision chain in a timely manner, and anomaly identification and control response are prone to lag. After the control action is executed, there is a lack of mechanisms for re-collection, re-judgment, and backtracking around the recovery conditions, making it difficult to form a stable closed loop for abnormal recovery.

[0006] Therefore, the technical problems that need to be solved in this field are: how to construct a phased control mechanism for the breeding and preservation of Australian lobster broodstock, so that different stages can correspond to different environmental target ranges and control templates; how to integrate environmental data, behavioral data and maintenance record data into the computer processing chain to improve the pertinence of anomaly identification and control decisions; and how to perform re-collection, re-judgment and phase rollback based on recovery conditions after regulation to form an anomaly recovery closed loop. Summary of the Invention

[0007] To address the shortcomings of existing environmental control methods for the preservation and breeding of Australian lobster broodstock, such as insufficient phased control capabilities, inadequate utilization of multi-source status information, and incomplete anomaly recovery loops, this invention provides an environmental control method to improve the success rate of Australian lobster broodstock preservation and breeding. This method aims to enable unified processing of environmental data, behavioral data, and maintenance record data within the same computer control chain, and to establish corresponding environmental target ranges, control templates, and recovery criteria around different preservation stages, thereby improving the stability of environmental control, the targeted nature of anomaly identification, and the reliability of recovery control during the preservation process.

[0008] The overall technical concept of this invention lies in using the Australian lobster parent breeding and farming unit as the smallest control object, unifying environmental states, behavioral states, and human maintenance events into a single data processing chain. First, time alignment, anomaly removal, stage marking, and unit mapping are completed. Then, based on this, stage identifiers, deviation quantities, risk levels, priorities, and trigger identifiers are formed to directly support control decisions. Using this set of intermediate state quantities, the control system no longer focuses on responding to the exceeding limits of a single parameter, but rather generates control commands matching the stage state and risk state around the environmental target range of the current control stage, and determines the target equipment and adjustment sequence.

[0009] Furthermore, this invention incorporates the process of collecting and updating data again after regulation and determining whether the recovery conditions are met into the main control chain, forming a closed-loop relationship between the conventional control template and the abnormal recovery control that can be switched back, is sustainable, and can be recorded. The reason for this setup is that stress, cannibalism, abnormal feeding, and abnormal molting during the breeding and preservation of Australian lobster broodstock are not always directly manifested by a single water quality parameter. Only by organizing phased control, multi-source data fusion, state quantity construction, multi-parameter linkage control, and recovery and rollback mechanisms into a continuous chain can the actual regulation needs in the breeding and preservation scenario be more stably adapted.

[0010] To achieve the above objectives, this invention adopts the following technical solution: Environmental data, behavioral data, and maintenance record data are collected for the breeding and rearing unit of Australian lobster broodstock. Environmental data characterizes the state of the aquaculture water and environment; behavioral data characterizes feeding and activity levels; and maintenance record data characterizes the results of manual inspections and events such as mortality, cannibalism, and abnormal molting. The control system performs time alignment, anomaly removal, stage marking, and unit mapping processing on the collected multi-source data. It also establishes stage archives, equipment archives, and historical status archives by combining the stocking time, breeding batch, manual inspection results, and historical event records, thereby obtaining standardized input data corresponding to the breeding unit.

[0011] After obtaining standardized input data, the control system constructs stage identifiers, deviation values, risk levels, priorities, and trigger identifiers. Stage identifiers characterize the current control stage of the aquaculture unit, assigning different environmental target ranges and control templates to different control stages. Deviation values ​​characterize the degree of deviation of real-time environmental values ​​from the environmental target range of the current control stage. Risk levels characterize the overall risk level of the current aquaculture unit based on environmental data, behavioral data, and maintenance record data. Priorities characterize the environmental variables that are prioritized for adjustment at the current moment and the corresponding equipment sequence. Trigger identifiers characterize whether abnormal recovery control is required. Based on these intermediate state variables, the control system generates the environmental target range and control commands for the current aquaculture unit and determines the target equipment and their adjustment sequence among aeration equipment, temperature control equipment, water exchange equipment, lighting equipment, water flow regulation equipment, and feeding equipment according to priority, in order to execute phased environmental adjustments under the corresponding control stage.

[0012] After completing environmental adjustment, the control system re-collects updated environmental and behavioral data and compares the updated results with the environmental target range of the corresponding control stage before adjustment to determine whether the aquaculture unit meets the recovery conditions. When the recovery conditions are met, the control system switches the aquaculture unit back to the normal control template of the current control stage; when the recovery conditions are not met, the control system continues to execute abnormal recovery control, thus forming a closed-loop control chain that connects data acquisition, preprocessing, intermediate quantity construction, status determination, equipment execution, effect recovery, and stage rollback.

[0013] Furthermore, the control phase can be divided into the following stages according to the breeding and preservation process: the adaptation phase, the stable breeding phase, the enhanced recovery phase, the breeding preparation phase, and the anomaly recovery phase. Different control phases can correspond to different environmental target ranges and control templates. This allows the control system to employ adjustment strategies adapted to the current stage under different breeding conditions, avoiding the application of the same control logic to the entire breeding and preservation process.

[0014] Furthermore, behavioral data can be generated from automatic identification results and manual observation records, and maintenance record data can be linked and stored with environmental data. This allows behavioral status, manual inspection results, and environmental status to form consistent data points within the same breeding unit, providing a more stable data foundation for risk level generation and anomaly recovery assessment.

[0015] Furthermore, anomaly removal can identify and process distorted sampling, missing sampling, and abnormal fluctuation sampling; stage identifiers can be determined based on the time of entry into the pond, breeding batch, manual inspection results, and historical event records; risk levels can be determined based on changes in environmental conditions, changes in behavioral conditions, and abnormal event records; and priorities can be determined based on the risk level and the order of impact of each environmental variable on breeding stability under the current control stage. Through these further arrangements, the control system can integrate differences in multi-source data, stage status, and risk sources into the same control logic.

[0016] Furthermore, when the risk level reaches the abnormal recovery trigger condition, control commands can be used to execute coordinated actions on at least some of the equipment involved in oxygenation, temperature control, water exchange, lighting, water flow regulation, and feeding. Recovery conditions can be determined based on the matching results between updated environmental data, behavioral data, and the target range of the current stage. Simultaneously with generating control commands, alarm information can be output to the management terminal, and control logs, event logs, and stage switching records can be output to the storage unit. Through these further solutions, the control system can uniformly manage risk triggering, action execution, recovery switching, and operational logging, thereby providing clear support for subsequent control adjustments and on-site maintenance.

[0017] The beneficial effects of this invention are: By using environmental data, behavioral data, and maintenance record data as inputs to a phased environmental control method, this approach provides a more comprehensive basis for risk identification in breeding and conservation scenarios compared to existing technologies that primarily rely on single water quality parameters or limited sensor data. Current technologies, which rely solely on localized information such as dissolved oxygen, water temperature, or pH, struggle to promptly reflect changes directly related to breeding stability, such as decreased feeding, abnormal activity, increased cannibalism, or abnormal molting. This proposed solution treats environmental changes, behavioral changes, and maintenance events within the same aquaculture unit, allowing control decisions to be constrained by multi-source state information, thus improving the targeting and stability of control measures.

[0018] By performing time alignment, anomaly removal, stage marking, and unit mapping on environmental data, behavioral data, and maintenance record data, this approach improves the consistency of multi-source data before it enters the control chain, compared to existing technologies that allow inconsistent sampling cycles and direct participation of abnormal sampled values ​​in control. In existing technologies, if power-down sampled values, offline missing values, instantaneous jump values, or misaligned timestamps are compared without processing, it can easily lead to distortion of deviations and risk levels, resulting in unnecessary equipment actions. Because this solution preprocesses the data and maps it uniformly to the same aquaculture unit, it reduces misjudgments caused by differences in data sources, thus improving the reliability of subsequent risk identification and equipment linkage control.

[0019] By establishing stage archives, equipment archives, and historical status archives, and constructing stage identifiers, deviation quantities, risk levels, priorities, and trigger identifiers, this approach provides clearer intermediate quantity support for the control process compared to existing technologies that use a single threshold to trigger equipment start-up and shutdown. Existing technologies struggle to distinguish the differences in control objectives across different stages such as pool adaptation, stable seeding, enhanced recovery, and abnormal recovery, making it difficult for the same control logic to accommodate the needs of each stage. Since this solution first uses stage identifiers to determine the current control stage, and then uses deviation quantities, risk levels, and priorities to progressively compress status information, it enables different stages to correspond to different environmental target ranges and control templates, thus improving the adaptability of staged control.

[0020] By generating environmental target ranges and control commands based on stage identifiers, deviations, risk levels, priorities, and trigger identifiers, and determining target equipment and their adjustment sequence according to priority, this approach improves the synergy of multi-parameter linkage control compared to existing technologies that involve independent responses or simple parallel actions by each device. In existing technologies, the lack of sequential constraints between oxygenation, water exchange, lighting, water flow regulation, and feeding equipment can easily lead to fluctuations in control effects, where local parameters are adjusted only to be replaced by new fluctuations introduced by other equipment. Since this scheme directly applies priority to equipment selection and adjustment sequence, it can address environmental variables with a greater impact on the stability of the current stage within the same control cycle, thus reducing control conflicts and improving closed-loop response efficiency.

[0021] By initiating anomaly recovery control when the risk level reaches a preset level, and incorporating the coordinated actions of increasing dissolved oxygen target values, reducing light intensity, reducing feeding frequency, adjusting circulation flow, and initiating water replenishment or replacement operations into a single recovery chain, this approach improves the completeness of anomaly management compared to existing technologies that only adjust individual devices after an anomaly occurs. Existing technologies often only temporarily alleviate a single anomaly, failing to simultaneously suppress continuous fluctuations caused by stress, cannibalism, or abnormal feeding. Because this solution combines anomaly triggering with coordinated recovery actions and uses priority control to manage the order of recovery actions, it enables more targeted recovery adjustments based on the source of the anomaly, which is beneficial for improving environmental stability during the breeding process.

[0022] By re-collecting updated data after performing phased environmental adjustments and switching between conventional control templates and abnormal recovery control based on recovery conditions, this approach creates a verifiable closed loop in the entire control chain, unlike existing technologies that lack data re-collection, verification, and rollback mechanisms after adjustments. In existing technologies, if equipment actions are not re-evaluated based on recovery conditions, it can easily remain in an inappropriate recovery state for an extended period or prematurely exit abnormal control before stabilization. This solution uses updated environmental and behavioral data, along with the current stage's target range, for recovery determination. Furthermore, it enters a safe convergence path when consecutive recovery failures occur. This allows recovery control and stage rollback to work in tandem, improving the sustainability of abnormal recovery control, the stability of stage switching, and the maintainability of the system engineering implementation. Attached Figure Description

[0023] Figure 1 This is the main flowchart of the phased environmental control of the present invention; Figure 2 This is a flowchart of the multi-source data fusion and linkage control of the present invention; Figure 3 This is a closed-loop diagram of anomaly recovery and phase rollback in this invention. Detailed Implementation

[0024] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1 like Figure 1 As shown, in this embodiment, a phased intelligent environmental control method for improving the success rate of breeding and raising Australian lobster broodstock is deployed in an Australian lobster broodstock breeding and raising unit. The raising unit consists of a breeding pond, corresponding data acquisition nodes, a controller, a management terminal, a storage unit, and controlled execution devices. The controlled execution devices include aeration equipment, temperature control equipment, water exchange equipment, lighting equipment, water flow regulation equipment, and feeding equipment. The data acquisition nodes are used to collect environmental data, behavioral data, and maintenance record data. The environmental data includes water temperature, dissolved oxygen, pH value, salinity, conductivity, ammonia nitrogen concentration, nitrite concentration, light intensity, light duration, water flow velocity, and circulation flow rate of the breeding water. The behavioral data includes feeding status and activity status. The maintenance record data includes manual inspection results, mortality records, cannibalism records, and abnormal molting records. To ensure that multi-source data can be aligned and used within the same breeding unit, the controller adds a breeding unit identifier, sampling timestamp, sampling source identifier, and batch identifier to each collection record, and writes the collection results into the stage archive, equipment archive, and historical status archive.

[0026] In this embodiment, the sampling period of the environmental data is set as the parameter to be calibrated. The The initial value range is 60s to 300s. The calibration sample source is the online sampling records of three consecutive production batches within the same breeding workshop. The calibration trigger condition is that the environmental deviation exceeds the preset deviation threshold for two consecutive sampling periods. The update frequency is once after each breeding batch is completed. The sampling period for behavioral data is set to the parameter to be calibrated. The The initial value range is 300s to 900s. Calibration samples are sourced from video recognition records, weighing records, and manual observation records. The calibration trigger condition is an inverse fluctuation between the feeding state and activity state within two consecutive sampling windows. The update frequency is once every seven days. The time alignment window is set to the parameter to be calibrated. The The initial value range is 1 to 3 times the environmental data sampling period. The calibration sample source is the environmental sampling log and manual inspection log of the same breeding unit. The calibration trigger condition is that the timestamps of multiple source data appear in the same time period. The update frequency is updated after the equipment calibration is completed. , , , and These are used to configure environmental sampling frequency, behavior sampling frequency, time alignment window, deviation judgment threshold, and risk trigger threshold, respectively. They can be combined with continuous operation logs, video recognition and manual inspection comparison records, equipment calibration records, and historical recovery records for calibration and updates.

[0027] In this embodiment, the controller receives water temperature, dissolved oxygen, pH value, salinity, conductivity, ammonia nitrogen concentration, nitrite concentration, light intensity, light duration, water flow velocity, and circulation flow rate uploaded by the corresponding sampling node at the end of each environmental data sampling cycle. At the end of each behavioral data sampling cycle, it receives feeding status and activity status. After manual inspection, it receives mortality records, cannibalism records, and abnormal molting records entered by the inspectors. The controller first performs time alignment processing to ensure that time differences exceeding [a certain threshold] are correct. The data is marked as records to be re-collected; then anomaly removal processing is performed, removing or re-collecting abnormal sampling values ​​generated under power failure, missing sampling values ​​formed under offline conditions, instantaneous jump sampling values, and distorted sampling values ​​exceeding the confidence interval; then stage marking and unit mapping processing is performed, uniformly mapping environmental data, behavioral data, and maintenance record data to the same breeding unit identifier to obtain standardized input data. The confidence interval is jointly defined by sensor range, historical archives, and manual inspection verification results; the stage marking determines whether the breeding unit is currently in the breeding adaptation stage, stable breeding stage, enhanced recovery stage, breeding preparation stage, or abnormal recovery stage based on the time of entry into the pond, breeding batch, manual inspection results, and historical event records.

[0028] The stage identifier is generated using formula (1): in, Indicates the target control phase The corresponding stage matching degree; Indicates the time and stage of pool entry Time matching quantity; Indicates the current breeding batch and stage Batch matching quantity; Indicates the results and stages of manual inspection. The state matching quantity; Representing historical event records and stages The number of events matched; , , and These represent the stage weights of the four types of matching quantities mentioned above, and satisfy the following conditions: The variables are sourced from pond entry registration records, aquaculture batch records, manual inspection results, and historical event records. The controller calculates the corresponding values ​​for each control stage. and will maximize The corresponding stage is determined as the current stage identifier. Formula (1) compresses time information, batch information, manual inspection information and historical event information into a single stage matching degree, thereby avoiding switching control stages based solely on a single water quality indicator.

[0029] After determining the current stage identifier, the controller calculates the real-time environmental deviation based on the environmental target interval corresponding to the current stage. The environmental deviation is generated using formula (2): in, Indicates the amount of environmental deviation; Indicates the first Real-time environmental values ​​of several environment variables; Indicates the first control phase under the current control phase. Target values ​​for several environmental variables; Indicates the first The deviation weights of each environmental variable under the current control phase; This indicates the number of environmental variables. The variables are sourced from online environmental sampling records and the current stage target range configuration. The controller uses... It characterizes the degree of deviation of real-time environmental values ​​from the target environmental range of the current control phase, and... Write it into the historical status archive to provide input for risk level determination and recovery condition determination.

[0030] In this embodiment, the risk level is generated using a combination of environmental data, behavioral data, and maintenance record data, rather than solely based on a single environmental indicator. The risk level is generated using formula (3): in, Indicates the basic quantity of risk level; Indicates the amount of environmental deviation; This indicates the decrease in food intake, reflecting the extent to which the food intake status within the current sampling window has decreased relative to the historical stable state; This indicates an abnormal activity level, reflecting the deviation of the current activity state from the historical stable state within the sampling window. This indicates the number of abnormal events, reflecting the combined intensity of events such as death records, mutual injury records, and molting abnormality records; , , and These represent the risk weights for environmental deviations, decreased food intake, abnormal activity, and abnormal events, respectively. The variables are sourced from standardized input data, historical status archives, and maintenance record data. The controller will... With preset risk threshold When comparing, Reaching or exceeding A trigger flag for entering abnormal recovery control is generated at that time. The threshold is derived from the successful recovery sample cutoff value in the historical recovery record, and is recalculated and updated once after each preservation batch ends.

[0031] In this embodiment, priority is used to determine which environmental variables should be adjusted first under the current control phase and current risk level, and to further determine the target equipment and adjustment sequence. Priority is generated using formula (4): in, Indicates the first The adjustment priority of each environment variable; Indicates the basic quantity of risk level; Indicates the first The impact of each environmental variable on seed conservation stability under the current control phase; and These represent the risk impact weight and the stage impact weight, respectively. The variables are sourced from the risk level assessment results, the current stage control template, and historical recovery records. The controller calculates the corresponding values ​​for all environmental variables. , and according to Arrange them from largest to smallest to obtain the adjustment order of the target equipment. When the oxygenation variable corresponds to... At its highest level, the oxygenation equipment is triggered first; when the light variable corresponds to When the water flow variable is higher than the light flow variable, the lighting equipment adjustment is executed first, followed by the water flow adjustment; when the water exchange variable and the temperature control variable both reach high priority, the water exchange equipment is executed first to reduce water quality fluctuations, followed by the temperature control equipment to converge the target range.

[0032] The working process is as follows: The controller first receives multi-source data uploaded by each sampling node, behavior recognition node and manual inspection terminal, and then obtains standardized input data in the order of time alignment, anomaly removal, stage marking and unit mapping; then determines the current stage identifier according to formula (1), obtains the environmental deviation amount according to formula (2), obtains the risk level according to formula (3), and obtains the adjustment priority of each environmental variable according to formula (4); then the controller reads the environmental target range and control template corresponding to the current stage, determines the target equipment and adjustment sequence according to the priority, and outputs the corresponding oxygenation, temperature control, water exchange, light, water flow adjustment and feeding control instructions; after the control instructions are executed, the controller recalculates the environmental deviation amount and risk level with the updated environmental data and behavior data. When the updated environmental data returns to the environmental target range of the current control stage, and the feeding state and activity state are restored to the preset stable state, the controller determines that the recovery conditions are met and maintains the current stage regular control template. When the updated environmental data still does not return or the behavior state is still abnormal, the controller keeps the abnormal recovery control template running and writes the recovery results, target equipment, adjustment sequence and triggering reason of this round into the stage switching record.

[0033] In this embodiment, the anomaly recovery control is embedded in the feedback loop of the main control chain. Specifically, when Reaching or exceeding When the controller sets the trigger flag to the abnormal recovery state, it switches the current control stage to the abnormal recovery control template. The abnormal recovery control template includes at least two linkage actions among increasing the dissolved oxygen target value, reducing light intensity, reducing feeding frequency, adjusting circulation flow, and initiating water replenishment and water exchange operations. The priority order of abnormal recovery control is determined by formula (4), and different risk sources correspond to different linkage action combinations. If hypoxia and abnormal activity coexist, the action intensity of oxygenation equipment and water flow regulation equipment is increased first; if light stimulation and decreased feeding coexist, the action intensity of light equipment and feeding equipment is reduced first; if ammonia nitrogen concentration and nitrite concentration exceed the target range of the current stage, the water exchange equipment and water replenishment operation are executed first. The threshold source is determined by successful recovery samples in the historical recovery record, and the recovery conditions are jointly determined by the regression results of the updated environmental data and the target range of the current stage, the stable feeding state results, and the stable activity state results. The priority always participates in the ranking of target equipment in the abnormal recovery chain.

[0034] Environmental fluctuations during the breeding and propagation of Australian lobsters are not only manifested as exceeding limits for a single water quality parameter, but are also reflected in behaviors and maintenance records such as decreased feeding, abnormal activity, increased cannibalism, and abnormal molting. Therefore, it is difficult to reliably identify the true risks using only single parameter thresholds such as water temperature or dissolved oxygen. By performing time alignment, anomaly removal, stage marking, and unit mapping after data acquisition, environmental data, behavioral data, and maintenance record data can be made comparable within the same breeding unit. By compressing time, batch, manual inspection, and historical events into stage matching degree using formula (1), the controller can perform evidence-based stage switching between pond adaptation, stable breeding, enhanced recovery, breeding preparation, and anomaly recovery. By unifying the deviations of different environmental variables from the target range into environmental deviation amount using formula (2), the fluctuations of multiple indicators can be transformed into sortable control inputs. By unifying environmental deviation, decreased feeding, abnormal activity, and abnormal events into risk level using formula (3), behavioral changes and maintenance events can be directly introduced into risk judgment. By coupling risk level and stage impact amount into priority using formula (4), the selection and adjustment sequence of target equipment can be transformed from single-point triggering into multi-parameter linkage control. Therefore, after each round of regulation, the controller re-collects updated environmental and behavioral data and determines the recovery conditions, enabling closed-loop switching between the conventional control template and the abnormal recovery control template, thereby reducing the risk of stress-induced death, mutual injury, and delayed recovery caused by drastic environmental fluctuations.

[0035] Example 2 like Figure 1 and Figure 2 As shown, in this embodiment, the behavioral data fusion chain, maintenance record constraint chain, and multi-parameter linkage action generation chain are enhanced based on Embodiment 1 to further improve the robustness of risk identification and the targeted nature of equipment coordinated adjustment in the breeding and preservation scenario. Except for the new content in this embodiment, the basic structure of the breeding unit, environmental data sampling nodes, controller, management terminal, storage unit, aeration equipment, temperature control equipment, water exchange equipment, lighting equipment, water flow regulation equipment, and feeding equipment is the same as in Embodiment 1. Environmental data sampling frequency... Behavioral data sampling frequency Time alignment window Preset deviation threshold and risk threshold Same as Example 1, and continue to use the field structure, collection criteria and update time rules of the stage archive, equipment archive and historical status archive in Example 1.

[0036] In this embodiment, behavioral data no longer uses only single-path feeding and activity states. Instead, behavioral quantities are extracted from video recognition results, weighing results, and manual observation records, and consistency processing is performed on behavioral quantities from different sources. Video recognition results provide the frequency of feeding actions, feeding pause duration, activity trajectory dispersion, and avoidance / stay ratio within a unit time window; weighing results provide the amount of feed consumed before and after feeding and the amount of uneaten feed; manual observation records provide the group aggregation state, mutual attack state, and individual recovery state after molting. The controller first aggregates behavioral quantities from different sources according to the same breeding unit identifier, the same time window, and the same batch identifier, and then performs dimension normalization and direction unification on each behavioral quantity, so that an increase in value corresponds to an increase in risk or a decrease in stability. Parameters to be calibrated This indicates the length of the video recognition statistical window, with an initial value range of 300s to 1200s. The sample source is continuous video frame recordings and manual observation comparison records. The calibration trigger condition is that the activity recognition result and the manual observation result are inconsistent for two consecutive times. The update frequency is once every seven days. Parameters to be calibrated. This indicates the length of the weighing statistics window. The initial value range is 0.5 to 1 times that of a feeding cycle. The sample source is the weighing records before and after feeding and the residual feed verification records. The calibration trigger condition is that the residual feed amount and the frequency of feeding actions in the video show an inverse fluctuation. The update frequency is once after each feeding cycle is completed. , , , , , , , , , and These are used to configure the behavior fusion window, weighing window, behavior fusion weight, maintenance constraint weight, and enhanced risk level coefficient, respectively. They can be calibrated and updated by combining video recognition and manual observation comparison records, feeding and weighing records, abnormal event ledgers, and historical recovery sample recalculation results.

[0037] In this embodiment, after completing time alignment, anomaly removal, stage marking, and unit mapping in Embodiment 1, the controller further performs multi-source behavior fusion processing. The behavior fusion quantity is generated using formula (5): in, Indicates the amount of behavior fusion; This represents the behavioral quantity obtained from the video recognition results; This represents the behavioral quantity obtained from the weighing result; This represents the amount of behavior obtained through manual observation and recording; , and Let represent the fusion weights of video recognition, weighing, and human observation, respectively, and satisfy . The variables originate from video recognition nodes, weighing recording nodes, and manual observation terminals, respectively. The controller will... Write it to the historical status archive and put As a direct input for enhancing risk level calculation, Formula (5) unifies the behavioral quantities from different sources into a single behavioral fusion quantity, thereby avoiding risk identification bias caused by distortion of a single behavioral source.

[0038] In this embodiment, to ensure that manual maintenance records can provide clear constraints for risk assessment, the controller extracts the number of death records, mutual injury records, and molting anomaly records from the maintenance record data to form maintenance constraint quantities. The maintenance constraint quantities are generated using formula (6): in, Indicates the maintenance constraint quantity; This indicates the number of deaths recorded in the current statistics window; This indicates the number of mutual damage records within the current statistics window; This indicates the number of abnormal molting records within the current statistics window; , and These represent the constraint weights of the three types of records mentioned above. The variables are sourced from manual inspection results, abnormal event logs, and recovery control event records. The controller... The quantification of maintenance records directly impacts risk identification, enabling risk assessment to no longer rely solely on environmental deviations and behavioral changes.

[0039] In obtaining behavioral fusion quantity and maintenance constraints Subsequently, the controller operates at the basic risk level of Example 1. The enhanced risk level is generated above. The enhanced risk level is generated using formula (7): in, This indicates an increased risk level; This indicates the basic risk level obtained in Example 1; Indicates the amount of behavior fusion; Indicates the maintenance constraint quantity; , and These represent the enhanced weights of basic risk, behavior fusion, and maintenance constraints, respectively. The variables are sourced from the risk level output of Example 1, the multi-source behavior fusion processing results, and the maintenance record processing results. The controller will... This serves as the core input for prioritizing and generating the intensity of linked actions at the current stage. If higher than , it indicates that relying solely on environmental deviations is insufficient to reflect the actual germplasm conservation risk, and the controller accordingly increases the response intensity to abnormal behaviors and maintenance events.

[0040] In this embodiment, the linkage action intensity is used to determine the action amplitude of each target device in the current stage, rather than merely determining whether the device acts. The linkage action intensity is generated using Equation (8): wherein, denotes the linkage action intensity corresponding to the target device; denotes the environmental variable priority generated in Embodiment 1; denotes the enhanced risk level; and respectively denote the priority action coefficient and the enhanced risk action coefficient. The variables are sourced from the priority output by Embodiment 1 and the enhanced risk level generated in this embodiment. The controller determines the final action combination according to the magnitude rule of all : when the corresponding to the aeration device is the maximum, the power or working duty cycle of the aeration device is increased first; when the corresponding to the lighting device is greater than the corresponding to the feeding device, the lighting intensity is reduced first, and then the feeding frequency is reduced; when the corresponding to the water changing device and the water flow regulating device are both at high levels, the controller starts the water changing device first, and then adjusts the water flow regulating device, so that the introduction of new water and the redistribution of the flow field are performed in coordination. As a result, the linkage action is subjected to graded control with the participation of the enhanced risk level.

[0041] The working process is as follows: the controller first obtains standardized input data according to the sampling and preprocessing procedures of Embodiment 1, and obtains the basic risk level and the priority ; then the controller extracts the behavior quantity from the video recognition result, the weighing result and the manual observation record, and generates the behavior fusion quantity according to Equation (5) , then extracts the maintenance constraint quantity from the death record, the mutual injury record and the ecdysis abnormality record and calculates it according to Equation (6); after that, the controller generates the enhanced risk level according to Equation (7) , and substitutes together with the priority of Embodiment 1 into Equation (8) to calculate the linkage action intensity corresponding to each target device ; the controller then according to Determine the target equipment action combination and adjustment sequence, and output enhanced oxygenation, temperature control, water exchange, lighting, water flow adjustment and feeding control commands; after the control commands are executed, the controller synchronously outputs low oxygen risk alarm, water quality fluctuation alarm, abnormal feeding alarm, abnormal activity alarm or control status recovery alarm, and writes the control command generation time, target equipment identification, adjustment sequence, current control stage, enhanced risk level and triggering reason into the control log, event log and stage switching record.

[0042] In this embodiment, the enhancement of the anomaly / recovery path is mainly reflected in two aspects. First, when the video recognition result is inconsistent with the manual observation record, the controller adjusts according to the historical consistency record. , and The value of gives higher weight to data sources with higher credibility in formula (5); secondly, when maintaining constraints Even when the environmental deviation continues to rise, Even if the risk level does not increase significantly, the controller can still enhance the risk level using formula (7). Raising the risk level to near the risk trigger threshold allows for the early initiation of coordinated action sequences to mitigate the subsequent spread of deaths caused by mutual harm, group stress, or molting vulnerability. The abnormal trigger criterion is: when the risk level is increased... Greater than or equal to the preset risk threshold Or maintain constraint quantities Greater than or equal to the preset constraint threshold When the controller determines that the enhanced abnormal chain has been triggered, the priority is still based on the linkage action strength generated by formula (8). The sorting is determined. The threshold source is the same as in Example 1, wherein... and Prioritize determining the boundary between successful and failed recovery batches in historical recovery samples, and enhance the weighting. , and The determination is primarily based on the recalculation results of successfully recovered batches from historical recovery samples. If historical samples are insufficient, the parameters to be calibrated are used, and their version field, source batch, and update time are recorded in the log. Recovery conditions are: when the risk level is enhanced. The value has fallen back below the preset recovery threshold, and the behavior fusion amount... With maintenance constraints When all actions fall back to the allowable range of the current control stage, the controller ends the enhanced linkage action combination of this embodiment and returns the subsequent stage rollback to the recovery determination chain of embodiment 3.

[0043] In the context of breeding and propagating Australian lobsters, environmental deviations often initially manifest as increased risk through abnormal behavior and maintenance records, before further escalating into significant environmental instability or mass mortality. If only the basic risk level in Example 1 is relied upon... While the controller can identify environmental fluctuations, its ability to quantify abnormal feeding, abnormal activity, increased cannibalism, and molting vulnerability in advance is insufficient. This embodiment integrates behavioral information from video recognition, weighing, and manual observation into a unified behavioral fusion quantity using formula (5). The death, mutual destruction, and molting anomalies are transformed into maintenance constraints using formula (6). Then, the basic risk, behavioral integration, and maintenance constraints are coupled into an enhanced risk level through formula (7). This allows risk identification to more closely reflect the actual state of the species; further, formula (8) is used to prioritize... and enhance risk level Combined into coordinated action intensity This allows the sequence and magnitude of actions of each target device to be simultaneously constrained by environmental risks, behavioral risks, and maintenance events, thereby improving the targeting and robustness of multi-parameter linkage adjustment. Since the enhanced risk level and action intensity are both written to the log and participate in subsequent recovery control, this embodiment can provide a more stable input basis for the recovery criteria and phase rollback in Embodiment 3.

[0044] Example 3 like Figures 1 to 3 As shown, in this embodiment, based on Embodiments 1 and 2, abnormal recovery control, recovery condition determination, stage rollback, and safety convergence path are further developed to ensure that the phased intelligent environmental control method maintains an executable and rewritable closed-loop control chain even in scenarios of continuous abnormalities, repeated fluctuations, or recovery failures. Except for the new content in this embodiment, the basic structure of the aquaculture unit, environmental data sampling nodes, controller, management terminal, storage unit, aeration equipment, temperature control equipment, water exchange equipment, lighting equipment, water flow regulation equipment, and feeding equipment is the same as in Embodiment 1. The behavior fusion quantity... Maintenance constraints Enhance risk level and the intensity of the linkage action The generation caliber is the same as in Example 2, with the same stage identifier and deviation. Risk threshold Deviation threshold The definitions of the environmental target range and control template for each control stage are the same as in Example 1. The updated deviation is denoted as... This represents the deviation amount recalculated based on updated environmental data after the execution of anomaly recovery control; the updated stable food intake is denoted as... , used to represent the degree of regression of the updated feeding state relative to the current control phase steady state; the updated activity steady state is denoted as . This indicates the degree of regression of the updated activity state relative to the current stable state in the control phase; the number of consecutive recovery failure rounds is denoted as... This indicates the number of rounds in which abnormal recovery control has been continuously executed under the current control phase but the recovery conditions have not yet been met.

[0045] In this embodiment, after each round of linkage action is completed, the controller re-collects environmental and behavioral data according to the same sampling period as in Embodiment 1, and uses the updated environmental and behavioral data, the target range of the current control stage, and historical recovery control results as inputs for recovery determination. If the enhanced risk level in Embodiment 2... Still higher than the preset risk threshold, or the updated deviation If the deviation remains above the preset threshold, the controller will maintain the abnormal recovery control template; if the risk level is increased... Fallback and updated deviation Reduce, and update the stable intake amount and the stable amount of activity after the update If the current control phase enters the allowed range, the controller will roll back the execution phase; if there are consecutive failed recovery cycles... If the preset failure round threshold is exceeded, the controller will no longer increase the intensity of the regular linkage action, but will enter the safe convergence path, perform high-frequency sampling, conservative oxygenation, conservative illumination and conservative feeding control, and freeze the phase switching until the new recovery judgment meets the recovery conditions.

[0046] In this embodiment, the trigger identifier is generated using formula (9): in, Indicates the trigger identifier; This indicates the enhanced risk level obtained in Example 2; Indicates the risk threshold; This represents the deviation obtained in Example 1; This represents the deviation threshold. The variable sources are the enhanced risk level output, deviation output, and threshold configuration table. The controller calculates this at least once per control cycle. ,when A value of 1 indicates entry into abnormal recovery control. A value of 0 indicates that the normal control template of the current control phase is maintained. Formula (9) compresses the enhanced risk level and deviation into a single abnormal trigger criterion, so that the entry condition of abnormal recovery control has a directly executable gating form.

[0047] In this embodiment, to avoid judging the completion of recovery based solely on a single recovery threshold, the controller comprehensively evaluates the degree of recovery through the recovery convergence value. The recovery convergence value is generated using formula (10): in, Indicates the recovery of convergence; This represents the normalized value of the deviation after the update; This indicates the stable amount of food intake after the update; This indicates the stable activity level after the update; , and Let represent the contribution weights of deviation reduction, feeding recovery, and activity recovery to the recovery convergence, respectively, and satisfy . The variables are sourced from updated environmental data, updated behavioral data, and the current stable state configuration for the control phase. The controller will... Write to the historical status archive, and with Evaluate the overall convergence degree after anomaly recovery control.

[0048] In this embodiment, the recovery criterion is generated using formula (11): in, This indicates the result of the recovery condition determination; Indicates the recovery of convergence; Indicates the recovery threshold; This indicates the deviation after the update; This indicates an update to the deviation threshold. The variables are derived from the recovered convergence output, the updated deviation output, and the recovered decision threshold configuration. The controller... The value is 1 and is calculated after one round of anomaly recovery control is completed. ,when When the value is 1, the recovery condition is considered met; when A value of 0 indicates that the recovery condition is not met and the abnormal recovery control template is maintained. The threshold is derived from the dividing point of successfully recovered samples in the historical recovery control results and is updated based on the recalculation records after the batch ends. The minimum action chain for the recovery condition is: when When the value is 1, first stop any additional abnormal recovery actions added in the current round, then retain the basic oxygenation and monitoring actions of the current control phase, then write the recovery results to the phase switching record and historical status archive, and finally switch back to the normal control template of the current control phase; when When the value is 0, the current abnormal recovery control template is maintained and the process proceeds to the next round of high-frequency sampling and re-judgment.

[0049] In this embodiment, to handle the scenario of continuous recovery failure, the controller sets a safe convergence value and determines whether to enter the safe convergence path accordingly. The safe convergence value is generated using formula (12): in, Indicates the safe convergence amount; Indicates the number of consecutive recovery failures; This indicates an increased risk level; This indicates the deviation after the update; , and These represent the convergence weights for the recovery failure round, the enhanced risk level, and the updated deviation, respectively. The variables are sourced from the recovery control event log, the enhanced risk level output, and the updated deviation output. The controller will... With preset safe convergence threshold When comparing, Greater than or equal to When, enter the safe convergence path; when Less than At that time, continue to perform the normal recovery judgment under the abnormal recovery control template. The minimum executable action chain of the safe convergence path is: increase the environmental data sampling frequency to high frequency sampling mode, maintain the oxygenation equipment at a conservative high level, reduce the lighting equipment to a conservative low level, freeze the new feeding action of the feeding equipment, limit the single water exchange range of the water exchange equipment, keep the current control stage identifier from switching, and continue to perform the recovery judgment of formula (10) and formula (11) after each round of high frequency sampling.

[0050] The working process is as follows: After the linkage action in Example 2 is executed, the controller receives updated environmental and behavioral data, and combines this data with the target range of the current control phase, historical recovery control results, and enhanced risk level. First, generate the trigger identifier using formula (9). ;when When the value is 0, the controller maintains the normal control template of the current control phase and records the result of this round as no recovery control triggered; when When the value is 1, the controller executes the abnormal recovery control template and, after completing one round of recovery actions, calculates the recovery convergence amount based on the updated environmental data, updated behavioral data, and the current stable state of the control phase using formula (10). Then, the recovery condition determination result is obtained through formula (11). ;when When the value is 1, the controller sequentially stops the abnormal additional actions, writes back the successful recovery record, switches back to the normal control template of the current control stage, and resets the consecutive recovery failure rounds. ;when When the value is 0, the controller causes the number of consecutive recovery failures to be zero. Add one, and with , and The safe convergence amount is calculated using formula (12) for the input. ;when Reaching or exceeding the preset safe convergence threshold At that time, the controller enters the safe convergence path and freezes the phase transition. Not achieved At this point, the controller continues to execute the next round of anomaly recovery control. Throughout the process, the controller writes the trigger cause, recovery judgment result, number of consecutive recovery failure rounds, time of entering or exiting the safe convergence path, and corresponding target device actions into the phase switching record and historical status archive.

[0051] In this embodiment, the abnormal trigger quantification criteria, threshold sources, priorities, and recovery conditions are as follows: the abnormal trigger criterion is given by formula (9), and the threshold source is the boundary value between successful recovery samples and failed recovery samples in the historical recovery control results; the recovery condition criterion is given by formula (11), and the threshold source is the representative regression interval of the successful recovery batch under the current control stage; the priority follows the method in embodiment 2, which is based on the intensity of the linkage action. The determined equipment adjustment sequence continues to be effective in the abnormal recovery control template; the action chain to be executed after the recovery conditions are met and the action chain to be executed after the recovery conditions are not met are both clearly given in this embodiment. , and These are used to configure the convergence threshold, the deviation recovery threshold, and the safe convergence threshold, respectively. They can be calibrated and updated by combining historical recovery sample recalculation results, continuous monitoring records, recovery event ledgers, and batch verification records.

[0052] In the context of breeding and propagation of Australian lobster broodstock, anomaly recovery control requires continuous quantitative assessment of the degree of anomaly mitigation, the degree of recovery completion, and when the stock enters a conservative convergence state. Formula (9) enhances the risk level. and deviation Jointly determine the trigger identifier This makes the entry conditions for abnormal recovery control subject to both environmental risk and behavioral / maintenance risk; Formula (10) unifies the updated deviation, feeding recovery degree, and activity recovery degree into a recovery convergence quantity. This makes the recovery process no longer dependent on a single threshold; Formula (11) incorporates the recovery convergence factor. Deviation from the updated value Coupling is the result of the recovery condition determination. This enables the controller to distinguish between incomplete recovery and complete recovery states; Formula (12) counts consecutive recovery failure rounds. Enhance risk level and updated deviation Unified conversion to safe convergence quantity This enables the controller to enter a conservative and stable path and limit the continuous amplification of adjustment actions when recovery repeatedly fails.

[0053] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be modified within the scope of the concept described herein by means of the above teachings or the technology or knowledge in related fields.

Claims

1. An environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock, characterized in that, Includes the following steps: Environmental data, behavioral data, and maintenance records were collected for the breeding and rearing unit of Australian lobster broodstock. The environmental data included water temperature, dissolved oxygen, pH, salinity, conductivity, ammonia nitrogen concentration, nitrite concentration, light intensity, light duration, water flow velocity, and circulation flow rate. The behavioral data included feeding and activity status. The maintenance records included manual inspection results, mortality records, cannibalism records, and records of abnormal molting. The environmental data, behavioral data, and maintenance records underwent time alignment, anomaly removal, stage marking, and unit mapping processing, and were further refined based on the time of stocking, breeding batch, and manual inspection results. Based on the results and historical event records, stage files, equipment files, and historical status files corresponding to the aquaculture unit are established to obtain standardized input data corresponding to the aquaculture unit. Based on the standardized input data, stage identifiers, deviation values, risk levels, priorities, and trigger identifiers are constructed. The stage identifier represents the current control stage of the aquaculture unit, and different control stages correspond to different environmental target ranges and control templates. The deviation value represents the degree of deviation of real-time environmental values ​​from the environmental target range of the current control stage. The risk level, based on the environmental data, behavioral data, and maintenance record data, jointly represents low oxygen levels and water quality. The comprehensive risk level caused by fluctuations, light stimulation, decreased feeding, abnormal activity, mortality, cannibalism, and abnormal molting is defined. The priority is based on the risk level and the order of influence of each environmental variable on the stability of the breeding stock under the current control stage, representing the environmental variable that should be adjusted first at the current moment. The trigger flag indicates whether abnormal recovery control has been initiated. Based on the stage flag, deviation, risk level, priority, and trigger flag, the environmental target range and control instructions for the current breeding unit are generated. The target equipment and their adjustment order among aeration equipment, temperature control equipment, water exchange equipment, lighting equipment, water flow regulation equipment, and feeding equipment are determined according to the priority. The control command is sent to the target device to execute the phased environmental adjustment of the corresponding control stage. After the phased environmental adjustment is executed, the updated environmental data and behavioral data are collected again, and the updated environmental data is compared with the environmental target range of the corresponding control stage before the adjustment to determine whether the breeding unit meets the recovery conditions. If the recovery conditions are met, the breeding unit is switched back to the normal control template of the current control stage. If the recovery conditions are not met, the abnormal recovery control continues to be executed, thereby forming a closed-loop control chain of data acquisition, preprocessing, intermediate quantity construction, status determination, equipment execution, effect recovery and stage rollback.

2. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, The control phases include the adaptation phase, the stable breeding phase, the enhanced recovery phase, the breeding preparation phase, and the abnormal recovery phase. Each control phase corresponds to a different environmental target range, sampling frequency, control priority, and alarm threshold.

3. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, The behavioral data consists of video recognition results, weighing results, and corresponding results from manual observation records. The maintenance record data is stored in association with the environmental data according to the breeding batch, breeding unit identifier, and recording time.

4. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, The anomaly removal process includes removing abnormal sampled values ​​generated under power failure conditions, missing sampled values ​​generated under offline conditions, instantaneous jump sampled values, and distorted sampled values ​​that exceed the preset confidence interval.

5. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, The stage identifier is determined jointly based on the time of entry into the pond, the breeding batch, the results of manual inspection, and historical event records. The historical event records include records of abnormal environmental events, abnormal feeding events, abnormal activity events, and recovery and control events.

6. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, The risk level is determined jointly based on the deviation amount, the change in feeding status, the change in activity status, the mortality record, the cannibalism record, and the abnormal molting record. The priority is determined based on the risk level and the order of the impact of each environmental variable on the stability of the species under the current control stage.

7. The phased intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, When the risk level reaches the preset risk level, the control command is used to increase the dissolved oxygen target value, reduce the light intensity, reduce the feeding frequency, adjust the circulation flow rate, and initiate water replenishment and water exchange operations corresponding to the current abnormal state.

8. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, The recovery conditions are determined based on the regression results between the updated environmental data, behavioral data, and the target interval of the current stage, and are determined to be satisfied when the deviation falls back to within the preset deviation threshold and the feeding state and activity state recover to the preset stable state.

9. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 1, characterized in that, While generating the control command, alarm information is output to the management terminal, and control logs, event logs, and stage switching records are output to the storage unit. The stage switching records include at least the control stage switching time, triggering reason, executing device, and recovery result.

10. The staged intelligent environmental control method for improving the success rate of breeding and propagation of Australian lobster broodstock as described in claim 9, characterized in that, The alarm information includes low oxygen risk alarm, water quality fluctuation alarm, abnormal feeding alarm, abnormal activity alarm, and recovery control status alarm. The control log includes at least the control command generation time, target equipment identifier, adjustment sequence, and corresponding control stage.

Images