Penaeus monodon cultivation management method and system suitable for industrial aquaculture
By real-time monitoring and multi-dimensional analysis of environmental parameters in the factory farming of tiger prawns, a decision tree prediction model was constructed, which solved the problems of lack of regional density prediction and decision lag in traditional technologies, and achieved efficient and accurate farming regulation and prediction.
Patent Information
- Application Number
- CN202510999929.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-21
- Publication Date
- 2025-11-21
AI Technical Summary
Existing technologies lack effective regional density prediction methods in the industrialized farming of tiger prawns, making it difficult to achieve the expected farming conditions. Furthermore, the reliance on human experience for farming regulation results in problems such as decision-making lag and low prediction accuracy.
By using water quality sensors and camera units to monitor environmental parameters and shrimp activity data in real time, multi-dimensional analysis and clustering are performed to construct a multivariate linear fitting model and a decision tree prediction model, generating an environment-density correlation record table to realize shrimp activity migration prediction and breeding programs.
It improves aquaculture efficiency and solves the problems of single environmental monitoring dimensions and insufficient multi-parameter coupling in traditional technologies, thus achieving efficient and accurate aquaculture regulation and prediction.
Smart Images
Figure CN120996336A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent management of aquaculture, and more specifically, to a method and system for the cultivation and management of tiger prawns suitable for factory farming. Background Technology
[0002] The tiger prawn is an important economic shrimp species in my country, and factory farming has become the mainstream model due to its advantages such as strong controllability and stable yield. However, existing technologies have many problems with factory farming of tiger prawns, especially in the process of high-density shrimp farming, where there is a lack of effective regional density prediction methods, making it difficult to achieve the expected farming conditions.
[0003] In addition, traditional techniques often analyze individual environmental parameters and farming conditions for shrimp farming at a fixed time period (such as a cycle), lacking comprehensive analysis of data changes across dimensions. This makes it difficult to fully explore the relationship between multiple environmental parameters and farming density, resulting in low accuracy in farming predictions under complex farming environments and high-density farming conditions. It also makes it difficult to carry out efficient factory farming and real-time farming control, leading to decision-making lag and reliance on human experience for farming control. Summary of the Invention
[0004] This invention overcomes the shortcomings of the prior art and proposes a cultivation and management method and system for tiger prawns suitable for factory farming.
[0005] The first aspect of this invention provides a method for cultivating and managing tiger prawns suitable for factory farming, comprising:
[0006] S101: During multiple breeding cycles, environmental monitoring and shrimp activity monitoring are carried out in multiple breeding areas of tiger prawns through water quality sensors and camera units, and environmental monitoring data and shrimp activity data are collected.
[0007] S102: Based on environmental monitoring data and shrimp activity data, conduct regionalized multi-dimensional environmental parameter change analysis and farming density statistics. Based on the similarity of environmental parameter changes, cluster the farming areas to form multiple regional clusters.
[0008] S103: Perform multiple linear fitting of environmental parameters and stocking density for each regional cluster, determine the associated regional clusters based on the degree of fitting, analyze the long and short period variation characteristics of the environmental parameters corresponding to the associated regional clusters, and generate a periodic environment-density correlation record table in combination with stocking density.
[0009] S104: Construct a prediction model based on decision trees, perform classification condition transformation based on the environment-density association record table to form condition nodes, and construct a complete prediction model based on heuristic algorithms;
[0010] S105: Import real-time environmental monitoring data into the prediction model to predict the stocking density, and based on the prediction results, predict the shrimp activity migration in different stocking areas and generate a stocking and breeding plan.
[0011] In this solution, S101 specifically refers to:
[0012] Within the overall farming area of the tiger prawn, multiple farming zones were divided, and water quality sensors and camera units were installed in each farming zone;
[0013] Water quality and shrimp activity were monitored in the aquaculture area during multiple aquaculture cycles.
[0014] Environmental monitoring data and shrimp activity data were collected for each aquaculture area during multiple aquaculture cycles.
[0015] Environmental monitoring data include parameters such as dissolved oxygen, water temperature, air temperature, salinity, pH value, and ammonia nitrogen content.
[0016] In this solution, S101 further includes:
[0017] Water quality sensors are used to collect and record multiple environmental parameters of the aquaculture area and store them in a database.
[0018] The camera unit acquires underwater images of the aquaculture area within a unit of time. The YOLOv5 target detection model is used to detect, count, and record target activity status in the underwater images, thus obtaining shrimp activity data.
[0019] In this solution, S102 specifically refers to:
[0020] Based on the analysis of a breeding area, the environmental monitoring data is represented by vectors of multi-dimensional environmental parameters to obtain environmental feature vectors, and multiple environmental feature vectors are formed based on multiple periods.
[0021] Manhattan distance is introduced to calculate the distance between multiple environmental feature vectors based on the time dimension. The obtained distance values are then vectorized to obtain the environmental change features.
[0022] Using the environmental change characteristics of each aquaculture area as cluster sample data, the cluster sample data is clustered based on the k-means clustering algorithm. The elbow rule is introduced in the clustering process to determine the optimal number of clusters K. The Manhattan distance is used to calculate the distance between the cluster sample data and the center point, and the clustering results are obtained.
[0023] The aquaculture areas are grouped according to the clustering results, forming multiple regional clusters.
[0024] In this solution, S103 specifically refers to:
[0025] Within a regional cluster, periodic environmental parameter data of the aquaculture area within the cluster are extracted, the periodic environmental parameter data are cleaned, the aquaculture density of the aquaculture area within the cluster in multiple aquaculture cycles is analyzed, and the density data is extracted and time-series converted to obtain aquaculture density parameter data. Parameter data and density data from multiple time periods are randomly selected as observation data.
[0026] Periodic environmental parameter data and stocking density parameter data were used as fitting data, with each environmental parameter as the independent variable and stocking density as the dependent variable for multiple linear fitting.
[0027] The goodness-of-fit R² is calculated based on the fitting results and the observation data. If the goodness-of-fit R² is greater than the preset goodness-of-fit, the region cluster is marked as an associated region cluster.
[0028] Filter out all related region clusters.
[0029] In this solution, S103 further includes:
[0030] Obtain periodic environmental parameter data corresponding to the aquaculture areas in the associated regional cluster;
[0031] Set long and short period time windows T1 and T2, and move the window with a preset step size based on the periodic environmental parameter data in multiple breeding cycles based on windows T1 and T2. During each movement, calculate the environmental change characteristics of multiple environmental parameters from window T, and associate the environmental change characteristics with the breeding density corresponding to window T to form an environment-density association record.
[0032] The environmental parameters of all associated regional clusters are analyzed for long- and short-period variation characteristics. The associated records are stored and a periodic analysis environmental-density association record table is generated.
[0033] In this solution, S104 specifically refers to:
[0034] Construct a prediction model based on decision trees;
[0035] Based on the environment-density association record table, each environment-density association record is transformed by conditions. The feature data part is used as the feature classification condition, the breeding density data is used as the classification result, and a condition node is formed for each record.
[0036] Calculate the information gain of all conditional nodes, determine the root node and leaf nodes, determine the remaining internal nodes based on the CART heuristic algorithm, and obtain the complete prediction model;
[0037] Based on the test data, the prediction model is trained and its nodes are optimized and adjusted.
[0038] In this solution, S105 specifically refers to:
[0039] Within a real-time cycle, collect real-time environmental monitoring data of the target aquaculture area;
[0040] By analyzing the environmental parameter change characteristics of real-time environmental monitoring data, the real-time change characteristics can be obtained.
[0041] Real-time change characteristics are imported into a prediction model to perform stocking density prediction analysis and obtain prediction results.
[0042] Based on the comparison between the predicted results and the real-time stocking density of the target aquaculture area, the analysis is conducted to determine whether the shrimp stocking density trend meets expectations, and early warning information is set.
[0043] Based on the prediction results of different aquaculture areas, shrimp migration activity is predicted, and aquaculture and breeding programs are set in combination with the target aquaculture density range of different aquaculture areas.
[0044] A second aspect of the present invention also provides a Penaeus monodon cultivation and management system suitable for factory farming. The system includes a memory and a processor. The memory includes a Penaeus monodon cultivation and management program suitable for factory farming. When the processor executes the Penaeus monodon cultivation and management program suitable for factory farming, it performs the following steps:
[0045] S101: During multiple breeding cycles, environmental monitoring and shrimp activity monitoring are carried out in multiple breeding areas of tiger prawns through water quality sensors and camera units, and environmental monitoring data and shrimp activity data are collected.
[0046] S102: Based on environmental monitoring data and shrimp activity data, conduct regionalized multi-dimensional environmental parameter change analysis and farming density statistics. Based on the similarity of environmental parameter changes, cluster the farming areas to form multiple regional clusters.
[0047] S103: Perform multiple linear fitting of environmental parameters and stocking density for each regional cluster, determine the associated regional clusters based on the degree of fitting, analyze the long and short period variation characteristics of the environmental parameters corresponding to the associated regional clusters, and generate a periodic environment-density correlation record table in combination with stocking density.
[0048] S104: Construct a prediction model based on decision trees, perform classification condition transformation based on the environment-density association record table to form condition nodes, and construct a complete prediction model based on heuristic algorithms;
[0049] S105: Import real-time environmental monitoring data into the prediction model to predict the stocking density, and based on the prediction results, predict the shrimp activity migration in different stocking areas and generate a stocking and breeding plan.
[0050] A third aspect of the present invention also provides a computer-readable storage medium comprising a suitable program for the cultivation and management of Penaeus monodon for industrialized farming. When executed by a processor, the suitable program for the cultivation and management of Penaeus monodon for industrialized farming implements the steps of the method for cultivating and managing Penaeus monodon for industrialized farming as described in any of the preceding claims.
[0051] This invention discloses a method and system for cultivating and managing Penaeus monodon suitable for factory farming, belonging to the field of intelligent aquaculture management. The specific steps include: collecting environmental parameters and shrimp activity data in real time using water quality sensors and camera units throughout multiple Penaeus monodon farming cycles; performing regional clustering and grouping based on regional environmental parameter change analysis and farming density statistics to form multiple regional clusters; performing multiple linear fitting and long- and short-cycle characteristic analysis on the farming environmental parameters and density of the regional clusters to generate an environment-density correlation record table; constructing a decision tree prediction model based on the correlation table; importing real-time environmental data into the model to predict farming density and shrimp activity migration, thereby regulating the farming environment. This invention solves the problems of single-dimensional environmental monitoring, lack of multi-parameter coupling and farming density prediction analysis in traditional shrimp farming, addresses the problem of decision lag, and effectively improves farming efficiency. Attached Figure Description
[0052] Figure 1 A flowchart of a method for cultivating and managing tiger prawns suitable for factory farming according to the present invention is shown;
[0053] Figure 2 A flowchart of the aquaculture area clustering process of the present invention is shown;
[0054] Figure 3 A block diagram of a tiger prawn cultivation and management system suitable for factory farming according to the present invention is shown. Detailed Implementation
[0055] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.
[0056] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.
[0057] Figure 1 A flowchart of a method for cultivating and managing tiger prawns suitable for factory farming according to the present invention is shown.
[0058] like Figure 1As shown, the first aspect of the present invention provides a method for cultivating and managing tiger prawns suitable for factory farming, comprising:
[0059] S101: During multiple breeding cycles, environmental monitoring and shrimp activity monitoring are carried out in multiple breeding areas of tiger prawns through water quality sensors and camera units, and environmental monitoring data and shrimp activity data are collected.
[0060] S102: Based on environmental monitoring data and shrimp activity data, conduct regionalized multi-dimensional environmental parameter change analysis and farming density statistics. Based on the similarity of environmental parameter changes, cluster the farming areas to form multiple regional clusters.
[0061] S103: Perform multiple linear fitting of environmental parameters and stocking density for each regional cluster, determine the associated regional clusters based on the degree of fitting, analyze the long and short period variation characteristics of the environmental parameters corresponding to the associated regional clusters, and generate a periodic environment-density correlation record table in combination with stocking density.
[0062] S104: Construct a prediction model based on decision trees, perform classification condition transformation based on the environment-density association record table to form condition nodes, and construct a complete prediction model based on heuristic algorithms;
[0063] S105: Import real-time environmental monitoring data into the prediction model to predict the stocking density, and based on the prediction results, predict the shrimp activity migration in different stocking areas and generate a stocking and breeding plan.
[0064] It should be noted that the aforementioned aquaculture areas include multiple areas, divided from the overall aquaculture area (or overall aquaculture range). Because shrimp stocking density and swimming patterns often exhibit regional differences in large and complex aquaculture environments, regional environmental characteristic analysis of the overall aquaculture area helps in the precise prediction of shrimp distribution and the achievement of efficient aquaculture environment control. Furthermore, for different aquaculture areas, regional density differences will vary to varying degrees due to environmental characteristics, feed programs, and other factors, and these differences often do not conform to the expected breeding outcomes.
[0065] According to an embodiment of the present invention, S101 specifically includes:
[0066] Within the overall farming area of the tiger prawn, multiple farming zones were divided, and water quality sensors and camera units were installed in each farming zone;
[0067] Water quality and shrimp activity were monitored in the aquaculture area during multiple aquaculture cycles.
[0068] Environmental monitoring data and shrimp activity data were collected for each aquaculture area during multiple aquaculture cycles.
[0069] Environmental monitoring data include parameters such as dissolved oxygen, water temperature, air temperature, salinity, pH value, and ammonia nitrogen content.
[0070] It should be noted that environmental monitoring data may include environmental parameters such as dissolved oxygen, water temperature, air temperature, salinity, pH value, and ammonia nitrogen content. Based on research needs, additional environmental parameters can be reasonably added to comprehensively evaluate aquaculture targets.
[0071] According to an embodiment of the present invention, S101 further includes:
[0072] Water quality sensors are used to collect and record multiple environmental parameters of the aquaculture area and store them in a database.
[0073] The camera unit acquires underwater images of the aquaculture area within a unit of time. The YOLOv5 target detection model is used to detect, count, and record target activity status in the underwater images, thus obtaining shrimp activity data.
[0074] It should be noted that the shrimp activity data specifically refers to information such as the number of shrimp in the aquaculture area, the range of shrimp activity, and the range of shrimp movement trajectories at different time periods, which is used to calculate the aquaculture density in subsequent statistics.
[0075] Figure 2 A flowchart of the aquaculture area clustering process of the present invention is shown.
[0076] According to an embodiment of the present invention, step S102 specifically includes:
[0077] S201, based on an aquaculture area, analyzes environmental monitoring data and performs multi-dimensional vector representation of environmental parameters to obtain environmental feature vectors, and forms multiple environmental feature vectors based on multiple periods;
[0078] S202 introduces Manhattan distance, calculates the distance between multiple environmental feature vectors based on the time dimension, and vectorizes the obtained multiple distance values to obtain environmental change features;
[0079] S203. Using the environmental change characteristics of each aquaculture area as clustering sample data, the clustering sample data is clustered based on the k-means clustering algorithm. The elbow rule is introduced in the clustering process to determine the optimal number of clusters K. The Manhattan distance is used to calculate the distance between the clustering sample data and the center point, and the clustering results are obtained.
[0080] S204: Based on the clustering results, the aquaculture areas are grouped into multiple regional clusters.
[0081] It should be noted that the environmental feature vector includes multiple dimensional values, each corresponding to an environmental parameter. Specifically, in the process of calculating the distance between multiple environmental feature vectors based on the time dimension and vectorizing these distance values, the distance between environmental feature vectors within adjacent periods is calculated, resulting in multiple distance values. These multiple distance values are then vectorized in multiple dimensions to obtain the environmental change feature. This feature is specifically a multi-dimensional change feature vector used to represent the state of environmental change.
[0082] It is worth noting that this invention utilizes multi-dimensional environmental feature data for vector representation, and periodically analyzes parameter changes and collects data on these changes. Based on clustering, it analyzes and groups the similarities in changes across different aquaculture areas. Furthermore, it couples the changing features of multi-dimensional parameters with the similarities in changes across multiple regions, thereby extracting groups of regions with similar environment-density relationships. This allows for subsequent analysis of long- and short-term change characteristics and the construction of decision trees. Traditional techniques, on the other hand, often focus on analyzing single environmental values and aquaculture conditions within a specific timeframe (such as a cycle), lacking comprehensive analysis of data changes across dimensions. This makes it difficult to fully explore the relationship between multiple environmental parameters and aquaculture density, often resulting in low prediction accuracy for complex aquaculture environments and high-density aquaculture conditions, hindering efficient industrialized aquaculture.
[0083] According to an embodiment of the present invention, step S103 specifically includes:
[0084] Within a regional cluster, periodic environmental parameter data of the aquaculture area within the cluster are extracted, the periodic environmental parameter data are cleaned, the aquaculture density of the aquaculture area within the cluster in multiple aquaculture cycles is analyzed, and the density data is extracted and time-series converted to obtain aquaculture density parameter data. Parameter data and density data from multiple time periods are randomly selected as observation data.
[0085] Periodic environmental parameter data and stocking density parameter data were used as fitting data, with each environmental parameter as the independent variable and stocking density as the dependent variable for multiple linear fitting.
[0086] The goodness-of-fit R-value was determined based on the fitting results and the observed data. 2 If the goodness of fit R is greater than the preset goodness of fit, then the region cluster is marked as an associated region cluster.
[0087] Filter out all related region clusters.
[0088] It should be noted that the observed data must meet a certain parameter data volume requirement. Periodic environmental parameter data refers to parameter data extracted from environmental monitoring data, sorted based on both time and parameter dimensions, for subsequent multiple linear evaluation. Parameter data for multiple time periods specifically involves extracting environmental parameter data and stocking density data for specific time periods across multiple aquaculture cycles. The preset goodness-of-fact ratio is set to 0.8.
[0089] In multiple linear fitting, the fitting equation is as follows:
[0090] ρ=β0+β1·X1+β2·X2+β3·X3+β4·X4+β5·X5+ε;
[0091] Where X1~X5 are preset environmental parameters, β0~β5 are regression coefficients, ε is the error term, and ρ is the stocking density.
[0092] Depending on the research needs, the fitting of environmental parameters can be increased or decreased.
[0093] According to an embodiment of the present invention, S103 further includes:
[0094] Obtain periodic environmental parameter data corresponding to the aquaculture areas in the associated regional cluster;
[0095] Set long and short period time windows T1 and T2, and move the window with a preset step size based on the periodic environmental parameter data in multiple breeding cycles based on windows T1 and T2. During each movement, calculate the environmental change characteristics of multiple environmental parameters from window T, and associate the environmental change characteristics with the breeding density corresponding to window T to form an environment-density association record.
[0096] The environmental parameters of all associated regional clusters are analyzed for long- and short-period variation characteristics. The associated records are stored and a periodic analysis environmental-density association record table is generated.
[0097] It should be noted that the window T represents both long and short time periods, including T1 and T2. The environmental change characteristic analysis process is described in S102. Multiple aquaculture cycles are specifically continuous cycles, with the overall cycle time denoted by T. P It means that in T P In the middle, set the moving time window T1, T2 and the preset step size, from T... P The time window is shifted during the analysis, and T1 and T2 can be set to long period (≥7 days) and short period (<3 days) respectively. The calculation process of the environment-density correlation record includes long period change characteristic analysis and short period change characteristic analysis.
[0098] This invention mines clusters of related regions for initial screening of regional groups, enabling the fusion analysis of environmental parameter data in regions with similar changes. Furthermore, it utilizes long- and short-cycle variation characteristics for parameter correlation assessment, sets classification condition nodes based on a decision tree model, and constructs a predictive model. This predictive model allows for efficient and rapid aquaculture prediction based on real-time environmental parameters and their variation characteristics, providing data model support for rapid environmental optimization and control, and adjustment of target stocking densities. This effectively solves the problem of decision-making lag in traditional factory farming.
[0099] According to an embodiment of the present invention, step S104 specifically includes:
[0100] Construct a prediction model based on decision trees;
[0101] Based on the environment-density association record table, each environment-density association record is transformed by conditions. The feature data part is used as the feature classification condition, the breeding density data is used as the classification result, and a condition node is formed for each record.
[0102] Calculate the information gain of all conditional nodes, determine the root node and leaf nodes, determine the remaining internal nodes based on the CART heuristic algorithm, and obtain the complete prediction model;
[0103] Based on the test data, the prediction model is trained and its nodes are optimized and adjusted.
[0104] It should be noted that the test data can be obtained based on the periodic environmental parameter data and breeding density data corresponding to the breeding areas in the associated regional cluster, and the prediction optimization of the decision tree can be achieved.
[0105] According to an embodiment of the present invention, step S105 specifically includes:
[0106] Within a real-time cycle, collect real-time environmental monitoring data of the target aquaculture area;
[0107] By analyzing the environmental parameter change characteristics of real-time environmental monitoring data, the real-time change characteristics can be obtained.
[0108] Real-time change characteristics are imported into a prediction model to perform stocking density prediction analysis and obtain prediction results.
[0109] Based on the comparison between the predicted results and the real-time stocking density of the target aquaculture area, the analysis is conducted to determine whether the shrimp stocking density trend meets expectations, and early warning information is set.
[0110] Based on the prediction results of different aquaculture areas, shrimp migration activity is predicted, and aquaculture and breeding programs are set in combination with the target aquaculture density range of different aquaculture areas.
[0111] It should be noted that the aquaculture and cultivation plan includes environmental control plans for different aquaculture areas, shrimp farming plans, etc. The environmental control plan includes real-time control of aerator equipment, setting water exchange cycle, and monitoring and improving the bottom sediment of the aquaculture area. The shrimp farming plan includes setting the feed feeding cycle, setting the feeding area and frequency, stocking shrimp larvae, and administering medications.
[0112] Based on the prediction of breeding density and the analysis of regional density changes, it is possible to analyze the migration situation in different regions. For breeding areas with low breeding density and migration that does not meet expectations, real-time early warning investigation and analysis can be carried out to achieve precise breeding control.
[0113] It is worth mentioning that for high-density shrimp farming and intensive aquaculture, traditional techniques often suffer from delayed decision-making and reliance on human experience, lacking methods for analyzing density change trends across multiple regions. The lack of methods to integrate the correlation between cyclical environmental characteristics and density change characteristics makes it difficult for existing technologies to conduct precise analysis of shrimp farming density trends and predictive analysis of behavioral models. Consequently, it is difficult to achieve multi-regional shrimp farming regulation and realize intensive, intelligent aquaculture.
[0114] Based on this, this invention utilizes periodic analysis of environmental parameters and stocking density variations in multiple aquaculture areas to cluster different aquaculture areas. Multiple environmental parameters are used to perform linear fitting on these clusters, and based on the fitting results, clusters showing correlation between environmental parameters and stocking density are selected. Furthermore, long- and short-period variation characteristic analysis is conducted on these correlated clusters to uncover the environment-density correlation information under complex environmental conditions and high-density shrimp farming. This correlation information is then transformed to construct a highly efficient and rapid classification prediction model. This model enables efficient and rapid aquaculture prediction based on real-time environmental monitoring data. Based on the predicted stocking density, shrimp migration prediction analysis is performed. Combined with the target stocking density in different aquaculture areas, regulation and early warning are implemented under the predicted conditions. This effectively manages shrimp growth and density control in different areas, fully utilizes aquaculture resources, and sets optimal local aquaculture plans to improve aquaculture efficiency.
[0115] Figure 3 A block diagram of a tiger prawn cultivation and management system suitable for factory farming according to the present invention is shown.
[0116] A second aspect of the present invention also provides a Penaeus monodon cultivation and management system 3 suitable for factory farming. The system includes a memory 31 and a processor 32. The memory includes a Penaeus monodon cultivation and management program suitable for factory farming. When the processor executes the Penaeus monodon cultivation and management program suitable for factory farming, it performs the following steps:
[0117] S101: During multiple breeding cycles, environmental monitoring and shrimp activity monitoring are carried out in multiple breeding areas of tiger prawns through water quality sensors and camera units, and environmental monitoring data and shrimp activity data are collected.
[0118] S102: Based on environmental monitoring data and shrimp activity data, conduct regionalized multi-dimensional environmental parameter change analysis and farming density statistics. Based on the similarity of environmental parameter changes, cluster the farming areas to form multiple regional clusters.
[0119] S103: Perform multiple linear fitting of environmental parameters and stocking density for each regional cluster, determine the associated regional clusters based on the degree of fitting, analyze the long and short period variation characteristics of the environmental parameters corresponding to the associated regional clusters, and generate a periodic environment-density correlation record table in combination with stocking density.
[0120] S104: Construct a prediction model based on decision trees, perform classification condition transformation based on the environment-density association record table to form condition nodes, and construct a complete prediction model based on heuristic algorithms;
[0121] S105: Import real-time environmental monitoring data into the prediction model to predict the stocking density, and based on the prediction results, predict the shrimp activity migration in different stocking areas and generate a stocking and breeding plan.
[0122] It should be noted that the aforementioned aquaculture areas include multiple areas, divided from the overall aquaculture area (or overall aquaculture range). Because shrimp stocking density and swimming patterns often exhibit regional differences in large and complex aquaculture environments, regional environmental characteristic analysis of the overall aquaculture area helps in the precise prediction of shrimp distribution and the achievement of efficient aquaculture environment control. Furthermore, for different aquaculture areas, regional density differences will vary to varying degrees due to environmental characteristics, feed programs, and other factors, and these differences often do not conform to the expected breeding outcomes.
[0123] According to an embodiment of the present invention, S101 specifically includes:
[0124] Within the overall farming area of the tiger prawn, multiple farming zones were divided, and water quality sensors and camera units were installed in each farming zone;
[0125] Water quality and shrimp activity were monitored in the aquaculture area during multiple aquaculture cycles.
[0126] Environmental monitoring data and shrimp activity data were collected for each aquaculture area during multiple aquaculture cycles.
[0127] Environmental monitoring data include parameters such as dissolved oxygen, water temperature, air temperature, salinity, pH value, and ammonia nitrogen content.
[0128] It should be noted that environmental monitoring data may include environmental parameters such as dissolved oxygen, water temperature, air temperature, salinity, pH value, and ammonia nitrogen content. Based on research needs, additional environmental parameters can be reasonably added to comprehensively evaluate aquaculture targets.
[0129] According to an embodiment of the present invention, S101 further includes:
[0130] Water quality sensors are used to collect and record multiple environmental parameters of the aquaculture area and store them in a database.
[0131] The camera unit acquires underwater images of the aquaculture area within a unit of time. The YOLOv5 target detection model is used to detect, count, and record target activity status in the underwater images, thus obtaining shrimp activity data.
[0132] It should be noted that the shrimp activity data specifically refers to information such as the number of shrimp in the aquaculture area, the range of shrimp activity, and the range of shrimp movement trajectories at different time periods, which is used to calculate the aquaculture density in subsequent statistics.
[0133] According to an embodiment of the present invention, step S102 specifically includes:
[0134] Based on the analysis of a breeding area, the environmental monitoring data is represented by vectors of multi-dimensional environmental parameters to obtain environmental feature vectors, and multiple environmental feature vectors are formed based on multiple periods.
[0135] Manhattan distance is introduced to calculate the distance between multiple environmental feature vectors based on the time dimension. The obtained distance values are then vectorized to obtain the environmental change features.
[0136] Using the environmental change characteristics of each aquaculture area as cluster sample data, the cluster sample data is clustered based on the k-means clustering algorithm. The elbow rule is introduced in the clustering process to determine the optimal number of clusters K. The Manhattan distance is used to calculate the distance between the cluster sample data and the center point, and the clustering results are obtained.
[0137] The aquaculture areas are grouped according to the clustering results, forming multiple regional clusters.
[0138] It should be noted that the environmental feature vector includes multiple dimensional values, each corresponding to an environmental parameter. Specifically, in the process of calculating the distance between multiple environmental feature vectors based on the time dimension and vectorizing these distance values, the distance between environmental feature vectors within adjacent periods is calculated, resulting in multiple distance values. These multiple distance values are then vectorized in multiple dimensions to obtain the environmental change feature. This feature is specifically a multi-dimensional change feature vector used to represent the state of environmental change.
[0139] It is worth noting that this invention utilizes multi-dimensional environmental feature data for vector representation, and periodically analyzes parameter changes and collects data on these changes. Based on clustering, it analyzes and groups the similarities in changes across different aquaculture areas. Furthermore, it couples the changing features of multi-dimensional parameters with the similarities in changes across multiple regions, thereby extracting groups of regions with similar environment-density relationships. This allows for subsequent analysis of long- and short-term change characteristics and the construction of decision trees. Traditional techniques, on the other hand, often focus on analyzing single environmental values and aquaculture conditions within a specific timeframe (such as a cycle), lacking comprehensive analysis of data changes across dimensions. This makes it difficult to fully explore the relationship between multiple environmental parameters and aquaculture density, often resulting in low prediction accuracy for complex aquaculture environments and high-density aquaculture conditions, hindering efficient industrialized aquaculture.
[0140] According to an embodiment of the present invention, step S103 specifically includes:
[0141] Within a regional cluster, periodic environmental parameter data of the aquaculture area within the cluster are extracted, the periodic environmental parameter data are cleaned, the aquaculture density of the aquaculture area within the cluster in multiple aquaculture cycles is analyzed, and the density data is extracted and time-series converted to obtain aquaculture density parameter data. Parameter data and density data from multiple time periods are randomly selected as observation data.
[0142] Periodic environmental parameter data and stocking density parameter data were used as fitting data, with each environmental parameter as the independent variable and stocking density as the dependent variable for multiple linear fitting.
[0143] The goodness-of-fit R-value was determined based on the fitting results and the observed data. 2 If the goodness of fit R is greater than the preset goodness of fit, then the region cluster is marked as an associated region cluster.
[0144] Filter out all related region clusters.
[0145] It should be noted that the observed data must meet a certain parameter data volume requirement. Periodic environmental parameter data refers to parameter data extracted from environmental monitoring data, sorted based on both time and parameter dimensions, for subsequent multiple linear evaluation. Parameter data for multiple time periods specifically involves extracting environmental parameter data and stocking density data for specific time periods across multiple aquaculture cycles. The preset goodness-of-fact ratio is set to 0.8.
[0146] In multiple linear fitting, the fitting equation is as follows:
[0147] ρ=β0+β1·X1+β2·X2+β3·X3+β4·X4+β5·X5+ε;
[0148] Where X1~X5 are preset environmental parameters, β0~β5 are regression coefficients, ε is the error term, and ρ is the stocking density.
[0149] Depending on the research needs, the fitting of environmental parameters can be increased or decreased.
[0150] According to an embodiment of the present invention, S103 further includes:
[0151] Obtain periodic environmental parameter data corresponding to the aquaculture areas in the associated regional cluster;
[0152] Set long and short period time windows T1 and T2, and move the window with a preset step size based on the periodic environmental parameter data in multiple breeding cycles based on windows T1 and T2. During each movement, calculate the environmental change characteristics of multiple environmental parameters from window T, and associate the environmental change characteristics with the breeding density corresponding to window T to form an environment-density association record.
[0153] The environmental parameters of all associated regional clusters are analyzed for long- and short-period variation characteristics. The associated records are stored and a periodic analysis environmental-density association record table is generated.
[0154] It should be noted that the window T represents both long and short time periods, including T1 and T2. The environmental change characteristic analysis process is described in S102. Multiple aquaculture cycles are specifically continuous cycles, with the overall cycle time denoted by T. P It means that in T P In the middle, set the moving time window T1, T2 and the preset step size, from T... P The time window is shifted during the analysis, and T1 and T2 can be set to long period (≥7 days) and short period (<3 days) respectively. The calculation process of the environment-density correlation record includes long period change characteristic analysis and short period change characteristic analysis.
[0155] This invention mines clusters of related regions for initial screening of regional groups, enabling the fusion analysis of environmental parameter data in regions with similar changes. Furthermore, it utilizes long- and short-cycle variation characteristics for parameter correlation assessment, sets classification condition nodes based on a decision tree model, and constructs a predictive model. This predictive model allows for efficient and rapid aquaculture prediction based on real-time environmental parameters and their variation characteristics, providing data model support for rapid environmental optimization and control, and adjustment of target stocking densities. This effectively solves the problem of decision-making lag in traditional factory farming.
[0156] According to an embodiment of the present invention, step S104 specifically includes:
[0157] Construct a prediction model based on decision trees;
[0158] Based on the environment-density association record table, each environment-density association record is transformed by conditions. The feature data part is used as the feature classification condition, the breeding density data is used as the classification result, and a condition node is formed for each record.
[0159] Calculate the information gain of all conditional nodes, determine the root node and leaf nodes, determine the remaining internal nodes based on the CART heuristic algorithm, and obtain the complete prediction model;
[0160] Based on the test data, the prediction model is trained and its nodes are optimized and adjusted.
[0161] It should be noted that the test data can be obtained based on the periodic environmental parameter data and breeding density data corresponding to the breeding areas in the associated regional cluster, and the prediction optimization of the decision tree can be achieved.
[0162] According to an embodiment of the present invention, step S105 specifically includes:
[0163] Within a real-time cycle, collect real-time environmental monitoring data of the target aquaculture area;
[0164] By analyzing the environmental parameter change characteristics of real-time environmental monitoring data, the real-time change characteristics can be obtained.
[0165] Real-time change characteristics are imported into a prediction model to perform stocking density prediction analysis and obtain prediction results.
[0166] Based on the comparison between the predicted results and the real-time stocking density of the target aquaculture area, the analysis is conducted to determine whether the shrimp stocking density trend meets expectations, and early warning information is set.
[0167] Based on the prediction results of different aquaculture areas, shrimp migration activity is predicted, and aquaculture and breeding programs are set in combination with the target aquaculture density range of different aquaculture areas.
[0168] It should be noted that the aquaculture and cultivation plan includes environmental control plans for different aquaculture areas, shrimp farming plans, etc. The environmental control plan includes real-time control of aerator equipment, setting water exchange cycle, and monitoring and improving the bottom sediment of the aquaculture area. The shrimp farming plan includes setting the feed feeding cycle, setting the feeding area and frequency, stocking shrimp larvae, and administering medications.
[0169] A third aspect of the present invention also provides a computer-readable storage medium comprising a suitable program for the cultivation and management of Penaeus monodon for industrialized farming. When executed by a processor, the suitable program for the cultivation and management of Penaeus monodon for industrialized farming implements the steps of the method for cultivating and managing Penaeus monodon for industrialized farming as described in any of the preceding claims.
[0170] This invention discloses a method and system for cultivating and managing Penaeus monodon suitable for factory farming, belonging to the field of intelligent aquaculture management. The specific steps include: collecting environmental parameters and shrimp activity data in real time using water quality sensors and camera units throughout multiple Penaeus monodon farming cycles; performing regional clustering and grouping based on regional environmental parameter change analysis and farming density statistics to form multiple regional clusters; performing multiple linear fitting and long- and short-cycle characteristic analysis on the farming environmental parameters and density of the regional clusters to generate an environment-density correlation record table; constructing a decision tree prediction model based on the correlation table; importing real-time environmental data into the model to predict farming density and shrimp activity migration, thereby regulating the farming environment. This invention solves the problems of single-dimensional environmental monitoring, lack of multi-parameter coupling and farming density prediction analysis in traditional shrimp farming, addresses the problem of decision lag, and effectively improves farming efficiency.
[0171] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.
[0172] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.
[0173] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.
[0174] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0175] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.
[0176] 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 changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for cultivating and managing tiger prawns suitable for factory farming, characterized in that, include: S101: During multiple breeding cycles, environmental monitoring and shrimp activity monitoring are carried out in multiple breeding areas of tiger prawns through water quality sensors and camera units, and environmental monitoring data and shrimp activity data are collected. S102: Based on environmental monitoring data and shrimp activity data, conduct regionalized multi-dimensional environmental parameter change analysis and farming density statistics. Based on the similarity of environmental parameter changes, cluster the farming areas to form multiple regional clusters. S103: Perform multiple linear fitting of environmental parameters and stocking density for each regional cluster, determine the associated regional clusters based on the degree of fitting, analyze the long and short period variation characteristics of the environmental parameters corresponding to the associated regional clusters, and generate a periodic environment-density correlation record table in combination with stocking density. S104: Construct a prediction model based on decision trees, perform classification condition transformation based on the environment-density association record table to form condition nodes, and construct a complete prediction model based on heuristic algorithms; S105: Import real-time environmental monitoring data into the prediction model to predict the stocking density, and based on the prediction results, predict the shrimp activity migration in different stocking areas and generate a stocking and breeding plan.
2. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 1, characterized in that, Specifically, S101 is as follows: Within the overall farming area of the tiger prawn, multiple farming zones were divided, and water quality sensors and camera units were installed in each farming zone; Water quality and shrimp activity were monitored in the aquaculture area during multiple aquaculture cycles. Environmental monitoring data and shrimp activity data were collected for each aquaculture area during multiple aquaculture cycles. Environmental monitoring data include parameters such as dissolved oxygen, water temperature, air temperature, salinity, pH value, and ammonia nitrogen content.
3. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 1, characterized in that, S101 further includes: Water quality sensors are used to collect and record multiple environmental parameters of the aquaculture area and store them in a database. The camera unit acquires underwater images of the aquaculture area within a unit of time. The YOLOv5 target detection model is used to detect, count, and record target activity status in the underwater images, thus obtaining shrimp activity data.
4. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 1, characterized in that, Specifically, S102 is as follows: Based on the analysis of a breeding area, the environmental monitoring data is represented by vectors of multi-dimensional environmental parameters to obtain environmental feature vectors, and multiple environmental feature vectors are formed based on multiple periods. Manhattan distance is introduced to calculate the distance between multiple environmental feature vectors based on the time dimension. The obtained distance values are then vectorized to obtain the environmental change features. Using the environmental change characteristics of each aquaculture area as cluster sample data, the cluster sample data is clustered based on the k-means clustering algorithm. The elbow rule is introduced in the clustering process to determine the optimal number of clusters K. The Manhattan distance is used to calculate the distance between the cluster sample data and the center point, and the clustering results are obtained. The aquaculture areas are grouped according to the clustering results, forming multiple regional clusters.
5. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 1, characterized in that, Specifically, S103 is as follows: Within a regional cluster, periodic environmental parameter data of the aquaculture area within the cluster are extracted, the periodic environmental parameter data are cleaned, the aquaculture density of the aquaculture area within the cluster in multiple aquaculture cycles is analyzed, and the density data is extracted and time-series converted to obtain aquaculture density parameter data. Parameter data and density data from multiple time periods are randomly selected as observation data. Periodic environmental parameter data and stocking density parameter data were used as fitting data, with each environmental parameter as the independent variable and stocking density as the dependent variable for multiple linear fitting. The goodness-of-fit R-value was determined based on the fitting results and the observed data. 2 If the goodness of fit R² is greater than the preset goodness of fit, then the region cluster is marked as an associated region cluster. Filter out all related region clusters.
6. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 5, characterized in that, S103 further includes: Obtain periodic environmental parameter data corresponding to the aquaculture areas in the associated regional cluster; Set long and short period time windows T1 and T2, and move the window with a preset step size based on the periodic environmental parameter data in multiple breeding cycles based on windows T1 and T2. During each movement, calculate the environmental change characteristics of multiple environmental parameters from window T, and associate the environmental change characteristics with the breeding density corresponding to window T to form an environment-density association record. The environmental parameters of all associated regional clusters are analyzed for long- and short-period variation characteristics. The associated records are stored and a periodic analysis environmental-density association record table is generated.
7. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 1, characterized in that, Specifically, S104 is as follows: Construct a prediction model based on decision trees; Based on the environment-density association record table, each environment-density association record is transformed by conditions. The feature data part is used as the feature classification condition, the breeding density data is used as the classification result, and a condition node is formed for each record. Calculate the information gain of all conditional nodes, determine the root node and leaf nodes, determine the remaining internal nodes based on the CART heuristic algorithm, and obtain the complete prediction model; Based on the test data, the prediction model is trained and its nodes are optimized and adjusted.
8. The method for cultivating and managing tiger prawns suitable for factory farming according to claim 1, characterized in that, Specifically, S105 is as follows: Within a real-time cycle, collect real-time environmental monitoring data of the target aquaculture area; By analyzing the environmental parameter change characteristics of real-time environmental monitoring data, the real-time change characteristics can be obtained. Real-time change characteristics are imported into a prediction model to perform stocking density prediction analysis and obtain prediction results. Based on the comparison between the predicted results and the real-time stocking density of the target aquaculture area, the analysis is conducted to determine whether the shrimp stocking density trend meets expectations, and early warning information is set. Based on the prediction results of different aquaculture areas, shrimp migration activity is predicted, and aquaculture and breeding programs are set in combination with the target aquaculture density range of different aquaculture areas.
9. A cultivation and management system for tiger prawns suitable for factory farming, characterized in that, The system includes: a memory and a processor. The memory contains a cultivation and management program for Penaeus monodon suitable for factory farming. When the processor executes the Penaeus monodon cultivation and management program suitable for factory farming, it performs the following steps: S101: During multiple breeding cycles, environmental monitoring and shrimp activity monitoring are carried out in multiple breeding areas of tiger prawns through water quality sensors and camera units, and environmental monitoring data and shrimp activity data are collected. S102: Based on environmental monitoring data and shrimp activity data, conduct regionalized multi-dimensional environmental parameter change analysis and farming density statistics. Based on the similarity of environmental parameter changes, cluster the farming areas to form multiple regional clusters. S103: Perform multiple linear fitting of environmental parameters and stocking density for each regional cluster, determine the associated regional clusters based on the degree of fitting, analyze the long and short period variation characteristics of the environmental parameters corresponding to the associated regional clusters, and generate a periodic environment-density correlation record table in combination with stocking density. S104: Construct a prediction model based on decision trees, perform classification condition transformation based on the environment-density association record table to form condition nodes, and construct a complete prediction model based on heuristic algorithms; S105: Import real-time environmental monitoring data into the prediction model to predict the stocking density, and based on the prediction results, predict the shrimp activity migration in different stocking areas and generate a stocking and breeding plan.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a suitable program for the cultivation and management of Penaeus monodon for industrialized farming. When the suitable program for the cultivation and management of Penaeus monodon for industrialized farming is executed by a processor, it implements the steps of the method for the cultivation and management of Penaeus monodon for industrialized farming as described in any one of claims 1 to 8.