Environment self-adaptive adjusting method and system for roe incubation

Abstract

The invention discloses an environment self-adaptive adjusting method and system for roe incubation, and relates to the technical field of non-electrical variable control, and the method comprises the steps: deploying an environment sensing device in a roe incubator, obtaining an environment sensing multi-dimensional data flow in real time, carrying out the feature extraction of the environment sensing multi-dimensional data flow, and carrying out the feature extraction of the environment sensing multi-dimensional data flow; and obtaining an environment perception multi-dimensional feature set. And building a fish egg incubation environment prediction network library, calling and obtaining a target fish egg incubation environment prediction network, and performing environment prediction on the environment perception multi-dimensional feature set to obtain an environment prediction trend. And performing incubation condition adjustment analysis on the fish egg incubation environment prediction trend to obtain fish egg incubation environment adjustment strategy parameters. And performing environment closed-loop regulation and control on the fish egg incubator based on a closed-loop feedback optimization mechanism and the fish egg incubation environment regulation strategy parameters. The technical problems that in the prior art, a roe hatching method has the defects of low efficiency, slow response and insufficient accuracy in environmental parameter adjustment, and it is difficult to ensure the optimal state of the hatching environment are solved.

Description

Technical Field

[0001] The present application relates to the technical field related to the control of non-electrical variables, and in particular to an environmental adaptive regulation method and system for fish egg hatching. Background Art

[0002] In modern aquaculture, fish egg hatching is an important step to ensure efficient production and high-quality fry. However, the fish egg hatching process has extremely strict requirements on environmental conditions, including precise control of multiple parameters such as temperature, humidity, pH value and dissolved oxygen. Traditional hatching methods usually rely on manual monitoring and manual adjustment, which is not only inefficient, but also difficult to respond to environmental changes in real time, which can easily lead to hatching failure or deterioration of fry quality.

[0003] Therefore, in the prior art, the fish egg hatching method has the defects of low efficiency, slow response and insufficient accuracy in regulating environmental parameters, and it is difficult to ensure the technical problem of the optimal state of the hatching environment. Summary of the invention

[0004] The present application provides an environmental adaptive adjustment method and system for fish egg hatching, which solves the technical problem that the existing fish egg hatching method has low efficiency, slow response and insufficient accuracy in adjusting environmental parameters, and it is difficult to ensure the optimal state of the hatching environment. It realizes real-time monitoring and optimization of hatching environmental parameters, improves hatching success rate and fry quality, and achieves the technical effect of automation and precision of environmental regulation.

[0005] The present application provides an environment adaptive regulation method for fish egg hatching, the method comprising: deploying an environment sensing device in a fish egg incubator, the environment sensing device comprising a temperature sensor, a humidity sensor, a pH sensor and a dissolved oxygen sensor; acquiring an environment sensing multidimensional data stream in real time through the environment sensing device, performing feature extraction on the environment sensing multidimensional data stream, and obtaining an environment sensing multidimensional feature set; building a fish egg hatching environment prediction network library, matching the target hatching fish egg type of the fish egg incubator with the fish egg hatching environment prediction network library, and calling to obtain a target fish egg hatching environment prediction network; performing environmental prediction on the environment sensing multidimensional feature set based on the target fish egg hatching environment prediction network, and obtaining a predicted trend of the fish egg hatching environment; performing hatching condition regulation analysis on the predicted trend of the fish egg hatching environment, and obtaining fish egg hatching environment regulation strategy parameters; establishing a closed-loop feedback optimization mechanism, and performing environmental closed-loop regulation on the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg hatching environment regulation strategy parameters.

[0006] In the implementation method, the obtaining of the environment-aware multidimensional feature set includes: performing outlier identification and data cleaning processing on the environment-aware multidimensional data stream to obtain a usable environment-aware multidimensional data stream; obtaining environment data source dimension information based on the usable environment-aware multidimensional data stream; determining a multidimensional data application standard based on the environment data source dimension information, and using the multidimensional data application standard to perform standardized conversion on the usable environment-aware multidimensional data stream to obtain a standard environment-aware multidimensional data stream; performing feature selection and feature extraction on the standard environment-aware multidimensional data stream to obtain the environment-aware multidimensional feature set.

[0007] In the implementation method, the method of obtaining the environmental perception multidimensional feature set includes: performing data characteristic analysis on the environmental data source dimensional information to obtain multidimensional data application characteristic information, and determining multidimensional data selection characteristic information based on the multidimensional data application characteristic information; performing feature extraction on the standard environmental perception multidimensional data stream according to the multidimensional data selection characteristic information to obtain an initial multidimensional environmental feature set; standardizing and decentralizing the initial multidimensional environmental feature set to obtain a covariance matrix, and decomposing the covariance matrix to obtain eigenvalues ​​and eigenvectors; selecting feature principal component information based on the eigenvalues ​​and eigenvectors, projecting the initial multidimensional environmental feature set data onto the feature principal component information, and obtaining the environmental perception multidimensional feature set.

[0008] In the implementation method, the construction of the fish egg hatching environment prediction network library includes: mining and obtaining a fish egg hatching environment database, classifying and identifying the fish egg hatching environment database according to the type of hatching fish eggs, and obtaining a fish egg hatching environment classification data set; performing feature extraction on the fish egg hatching environment classification data set to obtain a fish egg hatching multidimensional environment classification feature set; performing environmental trend prediction training on the fish egg hatching multidimensional environment classification feature set to obtain a multidimensional environmental trend prediction branch network set; classifying and serially fusing the multidimensional environmental trend prediction branch network set according to the type of hatching fish eggs to generate a multi-fish egg type environment prediction network set; identifying and integrating the multi-fish egg type environment prediction network set to build the fish egg hatching environment prediction network library.

[0009] In the implementation method, the method of obtaining the fish egg hatching environment adjustment strategy parameters includes: determining the applicable conditions of the target hatching environment according to the target hatching fish egg type; using the applicable conditions of the target hatching environment to compare and judge the predicted trend of the fish egg hatching environment to obtain the fish egg environment parameters to be optimized; performing adjustment strategy analysis on the fish egg environment parameters to be optimized to obtain the fish egg environment adjustment adaptive strategy, and performing data search according to the fish egg environment adjustment adaptive strategy to construct an environment adjustment strategy space; performing strategy parameter optimization on the fish egg environment parameters to be optimized based on the environment adjustment strategy space to obtain the fish egg hatching environment adjustment strategy parameters.

[0010] In the implementation, the obtaining of fish egg hatching environment adjustment strategy parameters includes: performing strategy parameter analysis on the fish egg environment parameters to be optimized based on the environment adjustment strategy space to obtain available adjustment strategy parameter thresholds; performing indicator decomposition and effect fitting on the fish egg hatching environment adjustment target to construct a fish egg environment effect evaluation fitness function; initializing particle swarm parameters according to the available adjustment strategy parameter thresholds, the particle swarm parameters including particle position and particle velocity; using the fish egg environment effect evaluation fitness function to evaluate and iteratively update the particle swarm parameters until a preset termination condition is met, and optimizing and determining the particle with the largest fitness as the fish egg hatching environment adjustment strategy parameter.

[0011] In the implementation method, the environmental closed-loop regulation of the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg incubation environment regulation strategy parameters includes: performing environmental regulation monitoring on the fish egg incubator based on the closed-loop feedback optimization mechanism to obtain fish egg incubation environment feedback parameters; obtaining expected environmental regulation effects, performing iterative feedback statistical fitting on the deviation values ​​of the fish egg incubation environment feedback parameters and the expected environmental regulation effects based on the closed-loop feedback optimization mechanism to obtain environmental regulation compensation factors; performing compensation optimization on the fish egg incubation environment regulation strategy parameters based on the environmental regulation compensation factors, and performing environmental closed-loop regulation on the fish egg incubator through the compensated and optimized fish egg incubation environment regulation strategy parameters.

[0012] The present application also provides an environment adaptive regulation system for fish egg hatching, comprising: an environment perception module, which is used to deploy an environment perception device in a fish egg incubator, wherein the environment perception device comprises a temperature sensor, a humidity sensor, a pH sensor and a dissolved oxygen sensor; an environment feature extraction module, which is used to obtain an environment perception multidimensional data stream in real time through the environment perception device, perform feature extraction on the environment perception multidimensional data stream, and obtain an environment perception multidimensional feature set; an environment prediction network construction module, which is used to build a fish egg hatching environment prediction network library, match the target hatching fish egg type of the fish egg incubator with the fish egg hatching environment prediction network library, and call to obtain a target fish egg hatching environment prediction network; an environment prediction module, which is used to perform environmental prediction on the environment perception multidimensional feature set based on the target fish egg hatching environment prediction network, and obtain a predicted trend of the fish egg hatching environment; an adjustment strategy parameter acquisition module, which is used to perform hatching condition adjustment analysis on the predicted trend of the fish egg hatching environment, and obtain fish egg hatching environment adjustment strategy parameters; a closed-loop control module, which is used to establish a closed-loop feedback optimization mechanism, and perform environmental closed-loop control on the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg hatching environment adjustment strategy parameters.

[0013] The environmental adaptive regulation method and system for fish egg hatching proposed in the present application are proposed. The environmental sensing device is deployed in the fish egg incubator. The environmental sensing device includes a temperature sensor, a humidity sensor, a pH sensor and a dissolved oxygen sensor. The environmental sensing device obtains the environmental sensing multidimensional data stream in real time, extracts the features of the environmental sensing multidimensional data stream, and obtains the environmental sensing multidimensional feature set. The fish egg hatching environment prediction network library is built, and the target hatching fish egg type of the fish egg incubator is matched with the fish egg hatching environment prediction network library, and the target fish egg hatching environment prediction network is called to obtain the target fish egg hatching environment prediction network. The environmental sensing multidimensional feature set is predicted based on the target fish egg hatching environment prediction network to obtain the predicted trend of the fish egg hatching environment. The predicted trend of the fish egg hatching environment is analyzed for hatching condition regulation to obtain the parameters of the fish egg hatching environment regulation strategy. A closed-loop feedback optimization mechanism is established, and the fish egg incubator is regulated in an environmental closed-loop based on the closed-loop feedback optimization mechanism and the parameters of the fish egg hatching environment regulation strategy. The technical problem that the fish egg hatching method in the prior art has the defects of low efficiency, slow response and insufficient accuracy in regulating environmental parameters, and it is difficult to ensure the optimal state of the hatching environment is solved. It realizes real-time monitoring and optimization of hatching environment parameters, improves hatching success rate and fry quality, and achieves the technical effect of automated and precise environmental control. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0015] Figure 1 A schematic flow chart of an environmental adaptive adjustment method for fish egg hatching provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of an environmental adaptive regulation system for fish egg hatching provided in an embodiment of the present application.

[0016] Explanation of the reference numerals: 11 environmental perception module, 12 environmental feature extraction module, 13 environmental prediction network construction module, 14 environmental prediction module, 15 adjustment strategy parameter acquisition module, 16 closed-loop control module. DETAILED DESCRIPTION

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0018] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0019] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0020] The present application provides an environment adaptive adjustment method and system for fish egg hatching, such as Figure 1 As shown, the method includes: An environment sensing device is deployed in a fish egg incubator, wherein the environment sensing device includes a temperature sensor, a humidity sensor, a pH sensor, and a dissolved oxygen sensor; an environment sensing multidimensional data stream is acquired in real time through the environment sensing device, and features are extracted from the environment sensing multidimensional data stream to obtain an environment sensing multidimensional feature set; a fish egg hatching environment prediction network library is built, and the target hatching fish egg type of the fish egg incubator is matched with the fish egg hatching environment prediction network library, and a target fish egg hatching environment prediction network is called to obtain; An environmental sensing device is deployed in a fish egg incubator to monitor the multidimensional environmental parameters in the fish egg incubator in real time to ensure the stability and suitability of the incubation environment. The environmental sensing device includes a temperature sensor, a humidity sensor, a pH sensor and a dissolved oxygen sensor. The temperature sensor is used to measure the temperature in the incubator, the humidity sensor is used to monitor the humidity in the incubator, the pH sensor is used to measure the pH of the water body, and the dissolved oxygen sensor is used to detect the dissolved oxygen content in the water. Subsequently, the collected data of each sensor is obtained in real time through the environmental sensing device to obtain an environmental sensing multidimensional data stream. The environmental sensing multidimensional data stream contains monitoring data collected by the sensor in a time series. Feature extraction is performed on the environmental sensing multidimensional data stream to obtain an environmental sensing multidimensional feature set. Further, a fish egg hatching environment prediction network library is built, and the fish egg hatching environment prediction network library stores environmental prediction networks corresponding to multiple fish egg categories, which is used to predict the trend of future environmental parameters based on known environmental data. According to the target hatching fish egg type of the fish egg incubator, the fish egg hatching environment prediction network library is matched, and the target fish egg hatching environment prediction network is called to obtain the target fish egg hatching environment prediction network.

[0021] The method provided in the embodiment of the present application also includes: performing outlier identification and data cleaning processing on the environmental perception multidimensional data stream to obtain a usable environmental perception multidimensional data stream; obtaining environmental data source dimension information based on the usable environmental perception multidimensional data stream; determining a multidimensional data application standard based on the environmental data source dimension information, and using the multidimensional data application standard to perform standardized conversion on the usable environmental perception multidimensional data stream to obtain a standard environmental perception multidimensional data stream; performing feature selection and feature extraction on the standard environmental perception multidimensional data stream to obtain the environmental perception multidimensional feature set.

[0022] The method of obtaining the multidimensional feature set of environmental perception includes: performing outlier identification and data cleaning on the multidimensional data stream of environmental perception to obtain an available multidimensional data stream of environmental perception. When performing outlier identification, data points in the data set that are significantly different from other data are identified to obtain an outlier identification result. These data points may be generated due to sensor failure, environmental interference or other reasons. Data cleaning is performed based on the outlier identification result to remove outliers and other inaccurate and incomplete data to obtain a high-quality available data set. According to the available multidimensional data stream of environmental perception, the dimension information of the source of environmental data is obtained. The dimension information of the source of environmental data is the data after the data stream corresponds to the corresponding monitoring dimension. The monitoring dimension includes the specific dimensions of temperature, humidity, pH value and dissolved oxygen. Based on the dimension information of the source of environmental data, the multidimensional data application standard is determined. The multidimensional data application standard is the environmental parameter range during the hatching of fish eggs, including the environmental parameter range of the specific dimensions of temperature, humidity, pH value and dissolved oxygen. The available multidimensional data stream of environmental perception is standardized and converted using the multidimensional data application standard, and the multidimensional data stream of environmental perception is converted to the range of 0-1 to obtain a standard multidimensional data stream of environmental perception. Feature selection and feature extraction are performed on the standard environment-aware multidimensional data stream to obtain the environment-aware multidimensional feature set.

[0023] The method provided in the embodiment of the present application also includes: performing data characteristic analysis on the source dimensional information of the environmental data to obtain multidimensional data application characteristic information, and determining multidimensional data selection feature information based on the multidimensional data application characteristic information; performing feature extraction on the standard environment perception multidimensional data stream according to the multidimensional data selection feature information to obtain an initial multidimensional environment feature set; standardizing and decentralizing the initial multidimensional environment feature set to obtain a covariance matrix, and decomposing the covariance matrix to obtain eigenvalues ​​and eigenvectors; selecting feature principal component information based on the eigenvalues ​​and eigenvectors, projecting the initial multidimensional environment feature set data onto the feature principal component information, and obtaining the environment perception multidimensional feature set.

[0024] Perform data characteristic analysis on the environmental data source dimension information, obtain commonly used application characteristic information of the corresponding environmental dimension during the data characteristic analysis, such as spatial characteristics, time series characteristics, etc., and obtain multidimensional data application characteristic information, which is commonly used analysis characteristics of different environmental dimension data, such as spatial characteristics, time series characteristics, etc. of temperature data. Determine multidimensional data selection characteristic information based on the multidimensional data application characteristic information, that is, obtain specific analysis characteristics of different environmental data dimensions.

[0025] Furthermore, feature extraction is performed on the standard environment perception multidimensional data stream according to the multidimensional data selection feature information to obtain data features corresponding to each dimensional data in the multidimensional data selection feature information. Furthermore, the initial multidimensional environment feature set is standardized and decentralized to obtain a covariance matrix, and the covariance matrix is ​​decomposed to obtain eigenvalues ​​and eigenvectors. Finally, the main components selected by principal component analysis (PCA) are used to obtain the characteristic principal component information of the eigenvalues ​​and eigenvectors, and the initial multidimensional environment feature set data is projected onto the characteristic principal component information to obtain the environment perception multidimensional feature set. Through the above steps, the hatchery can extract high-quality and representative feature sets from the original multidimensional data stream, providing a reliable data basis for subsequent environmental prediction and adaptive adjustment.

[0026] The method provided in the embodiment of the present application also includes: mining and obtaining a fish egg hatching environment database, classifying and identifying the fish egg hatching environment database according to the type of hatching fish eggs, and obtaining a fish egg hatching environment classification data set; performing feature extraction on the fish egg hatching environment classification data set to obtain a fish egg hatching multi-dimensional environment classification feature set; performing environmental trend prediction training on the fish egg hatching multi-dimensional environment classification feature set to obtain a multi-dimensional environmental trend prediction branch network set; classifying and serially fusing the multi-dimensional environmental trend prediction branch network set according to the type of hatching fish eggs to generate a multi-fish egg type environment prediction network set; identifying and integrating the multi-fish egg type environment prediction network set to build the fish egg hatching environment prediction network library.

[0027] The method of building a fish egg hatching environment prediction network library comprises: obtaining various environmental parameter data related to the fish egg hatching process, such as temperature, humidity, pH value, dissolved oxygen, etc., and result data such as hatching success rate, etc., by mining fish egg hatching data. Based on the above data, a fish egg hatching environment database is constructed, and the fish egg hatching environment database is classified and identified according to the type of hatching fish eggs, so as to obtain a fish egg hatching environment classification data set, and the fish egg hatching environment classification data set contains the hatching environment corresponding to various types of fish eggs and the result data such as hatching success rate, etc. Feature extraction is performed on the fish egg hatching environment classification data set respectively, and data corresponding to the multidimensional features of environmental perception of multiple dimensions of fish egg hatching are obtained to obtain a fish egg hatching multidimensional environment classification feature set. Environmental trend prediction training is performed on the fish egg hatching multidimensional environment classification feature set respectively, that is, the neural network model is trained respectively by each classification feature and the corresponding data trend in the fish egg hatching multidimensional environment classification feature set, until the model training is completed to obtain a multidimensional environmental trend prediction branch network set. The multidimensional environmental trend prediction branch network set is used to predict future environmental parameter trends based on known environmental data. The multi-dimensional environmental trend prediction branch network set is classified and connected in series according to the hatching fish egg type, that is, according to the environmental characteristics corresponding to the hatching fish egg type, multiple multi-dimensional environmental trend prediction branches are connected in series to generate a multi-fish egg type environmental prediction network set. The multi-fish egg type environmental prediction network set is identified and integrated, that is, each fish egg type environmental prediction network is identified with the corresponding fish egg category, and the fish egg hatching environment prediction network library is built.

[0028] Based on the target fish egg hatching environment prediction network, the environment perception multidimensional feature set is predicted to obtain a fish egg hatching environment prediction trend; the predicted trend of the fish egg hatching environment is analyzed for hatching condition adjustment to obtain fish egg hatching environment adjustment strategy parameters; a closed-loop feedback optimization mechanism is established, and the fish egg incubator is closed-loop regulated based on the closed-loop feedback optimization mechanism and the fish egg hatching environment adjustment strategy parameters.

[0029] Based on the target fish egg hatching environment prediction network, the environment perception multidimensional feature set is predicted, that is, the environment perception multidimensional feature set is input into the target fish egg hatching environment prediction network to obtain the fish egg hatching environment prediction trend. The predicted trend of the fish egg hatching environment is analyzed for hatching condition adjustment to obtain the fish egg hatching environment adjustment strategy parameters. Finally, a closed-loop feedback optimization mechanism is established, and the fish egg incubator is regulated in an environmental closed-loop based on the closed-loop feedback optimization mechanism and the fish egg hatching environment adjustment strategy parameters to ensure that the environmental parameters continue to remain within the expected target range. The technical problem that the fish egg hatching method in the prior art has the defects of low efficiency, slow response and insufficient accuracy for environmental parameter adjustment, and it is difficult to ensure the optimal state of the hatching environment is solved. The hatching environment parameters are monitored and optimized in real time, the hatching success rate and the quality of fry are improved, and the technical effect of automation and precision of environmental regulation is achieved.

[0030] The method provided in the embodiment of the present application also includes: determining the applicable conditions of the target hatching environment according to the target hatching fish egg type; using the applicable conditions of the target hatching environment to compare and judge the predicted trend of the fish egg hatching environment to obtain the fish egg environment parameters to be optimized; performing adjustment strategy analysis on the fish egg environment parameters to be optimized to obtain a fish egg environment adjustment adaptive strategy, and performing data search based on the fish egg environment adjustment adaptive strategy to construct an environment adjustment strategy space; performing strategy parameter optimization on the fish egg environment parameters to be optimized based on the environment adjustment strategy space to obtain fish egg hatching environment adjustment strategy parameters.

[0031] The method of obtaining the parameters of the regulation strategy for the hatching environment of fish eggs comprises: determining the applicable conditions of the target hatching environment according to the target hatching fish eggs type, wherein the applicable conditions of the target hatching environment are the optimal environmental parameter conditions when hatching the target fish eggs. Subsequently, the predicted trend of the hatching environment of fish eggs is compared and judged by using the applicable conditions of the target hatching environment, and the predicted trend of the hatching environment of fish eggs that is not in the applicable conditions of the target hatching environment is obtained, and the environmental parameters of fish eggs to be optimized are obtained. Further, the environmental parameters of fish eggs to be optimized are analyzed by the regulation strategy, and the deviation relationship between the environmental parameters of fish eggs to be optimized and the applicable conditions of the corresponding target hatching environment is determined, such as the data is too large or the data is too small. According to the deviation relationship, the adaptive strategy for regulating the environment of fish eggs is obtained, and the adaptive strategy for regulating the environment of fish eggs is the direction of parameter optimization. Data search is performed according to the adaptive strategy for regulating the environment of fish eggs, and the range of optimizable parameters for the direction of parameter optimization corresponding to the environmental dimension is searched and obtained, and the environment regulation strategy space is constructed. Based on the environment regulation strategy space, the policy parameters of the environmental parameters of fish eggs to be optimized are optimized, and the parameters of the regulation strategy for the hatching environment of fish eggs are obtained.

[0032] The method provided in the embodiment of the present application also includes: performing strategy parameter analysis on the fish egg environmental parameters to be optimized based on the environmental adjustment strategy space to obtain available adjustment strategy parameter thresholds; performing indicator decomposition and effect fitting on the fish egg hatching environment adjustment target to construct a fish egg environment effect evaluation fitness function; initializing particle swarm parameters according to the available adjustment strategy parameter thresholds, the particle swarm parameters including particle position and particle velocity; using the fish egg environment effect evaluation fitness function to evaluate and iteratively update the particle swarm parameters until a preset termination condition is met, and optimizing and determining the particle with the largest fitness as the fish egg hatching environment adjustment strategy parameter.

[0033] Based on the environmental adjustment strategy space, the strategy parameter of the fish egg environment parameter to be optimized is analyzed to obtain the threshold value of the available adjustment strategy parameter. The threshold value of the available adjustment strategy parameter is the allowable range or boundary of each adjustment strategy parameter, which is used to limit the value of the parameter to avoid exceeding the actual operable range. The index disassembly and effect fitting are performed on the adjustment target of the fish egg hatching environment, that is, the relationship between the hatching rate and the adjustment target of the fish egg hatching environment is fitted by multivariate linear regression for historical hatching data, and the fitness function for evaluating the effect of the fish egg environment is constructed. The particle position and speed in the particle swarm optimization algorithm are initialized using the threshold value of the available adjustment strategy parameter to prepare for subsequent optimization iterations. The particle swarm parameters include particle position and particle speed. Particle swarm optimization (PSO) is an optimization algorithm based on swarm intelligence, which finds the optimal solution by simulating the foraging behavior of bird flocks. Among them, the particle is the basic unit in the PSO algorithm, and each particle represents a possible solution, that is, a set of adjustment strategy parameters. The particle position is the current solution of the particle in the strategy space, which represents a set of specific strategy parameter values. The particle speed is the moving direction and step length of the particle in the strategy space, which is used to update the particle position. The particle swarm parameters are evaluated and iteratively updated using the fitness function for evaluating the fish egg environment effect until a preset termination condition is met, and the particle with the largest fitness is optimally determined as the parameter of the fish egg hatching environment adjustment strategy. That is, through iterative optimization, the position and speed of the particle swarm are continuously evaluated and updated using the fitness function, and the optimal adjustment strategy parameters are gradually approached until a preset termination condition is met, such as reaching a maximum number of iterations or a fitness change is lower than a threshold, and the particle with the largest fitness is used as the parameter of the fish egg hatching environment adjustment strategy.

[0034] The method provided in the embodiment of the present application also includes: performing environmental control monitoring on the fish egg incubator based on the closed-loop feedback optimization mechanism to obtain fish egg incubation environment feedback parameters; obtaining expected environmental control effects, and performing iterative feedback statistical fitting on the deviation values ​​of the fish egg incubation environment feedback parameters and the expected environmental control effects based on the closed-loop feedback optimization mechanism to obtain environmental control compensation factors; performing compensation optimization on the fish egg incubation environment control strategy parameters based on the environmental control compensation factors, and performing environmental closed-loop control on the fish egg incubator through the compensated and optimized fish egg incubation environment control strategy parameters.

[0035] The method of performing closed-loop environmental regulation on the egg incubator based on the closed-loop feedback optimization mechanism and the egg incubation environment regulation strategy parameters includes: using the closed-loop feedback optimization mechanism to continuously monitor the environmental parameters in the egg incubator and obtain real-time environmental feedback data. The expected environmental regulation effect is obtained, and the expected environmental regulation effect is the expected regulation target of the incubation environment. Based on the closed-loop feedback optimization mechanism, the deviation values ​​of the egg incubation environment feedback parameters and the expected environmental regulation effect are subjected to iterative feedback statistical fitting to obtain an environmental regulation compensation factor, and the environmental regulation compensation factor is used to correct and optimize the coefficient or adjustment value of the regulation strategy parameters to ensure that the environmental parameters approach the expected target. Based on the environmental regulation compensation factor, the egg incubation environment regulation strategy parameters are compensated and optimized to eliminate or reduce the deviation of the environmental parameters, and to achieve accurate control of the environmental parameters, and the egg incubator is subjected to closed-loop environmental regulation by the compensated and optimized egg incubation environment regulation strategy parameters. That is, based on the real-time feedback data and the optimized strategy parameters, the environment in the incubator is dynamically adjusted to ensure that the environmental parameters continue to remain within the expected target range.

[0036] In the above, refer to Figure 1 The environmental adaptive adjustment method for fish egg hatching according to an embodiment of the present invention is described in detail. Figure 2 An environmental adaptive regulation system for fish egg hatching according to an embodiment of the present invention is described.

[0037] The environmental adaptive regulation system for fish egg hatching according to the embodiment of the present invention solves the technical problem that the fish egg hatching method in the prior art has the defects of low efficiency, slow response and insufficient accuracy in regulating environmental parameters, and it is difficult to ensure the optimal state of the hatching environment. It realizes real-time monitoring and optimization of hatching environmental parameters, improves hatching success rate and fry quality, and realizes the technical effect of automation and precision of environmental regulation. The environmental adaptive regulation system for fish egg hatching includes: an environmental perception module 11, an environmental feature extraction module 12, an environmental prediction network construction module 13, an environmental prediction module 14, an adjustment strategy parameter acquisition module 15, and a closed-loop control module 16.

[0038] The environment sensing module 11 is used to deploy an environment sensing device in the fish egg incubator, wherein the environment sensing device includes a temperature sensor, a humidity sensor, a pH sensor and a dissolved oxygen sensor.

[0039] The environment feature extraction module 12 is used to obtain the environment perception multidimensional data stream in real time through the environment perception device, perform feature extraction on the environment perception multidimensional data stream, and obtain an environment perception multidimensional feature set.

[0040] The environment prediction network construction module 13 is used to build a fish egg hatching environment prediction network library, match the target hatching fish egg type of the fish egg incubator with the fish egg hatching environment prediction network library, and call to obtain the target fish egg hatching environment prediction network.

[0041] The environment prediction module 14 is used to perform environment prediction on the environment perception multidimensional feature set based on the target fish egg hatching environment prediction network to obtain a predicted trend of the fish egg hatching environment.

[0042] The adjustment strategy parameter acquisition module 15 is used to perform an incubation condition adjustment analysis on the predicted trend of the fish egg incubation environment to obtain the fish egg incubation environment adjustment strategy parameters.

[0043] The closed-loop control module 16 is used to establish a closed-loop feedback optimization mechanism, and perform closed-loop control on the environment of the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg incubation environment adjustment strategy parameters.

[0044] The specific configuration of the environmental feature extraction module 12 will be described in detail below. The environmental feature extraction module 12 may further include: the obtaining of the environmental perception multidimensional feature set includes: performing outlier identification and data cleaning processing on the environmental perception multidimensional data stream to obtain an available environmental perception multidimensional data stream. According to the available environmental perception multidimensional data stream, the environmental data source dimension information is obtained. Based on the environmental data source dimension information, the multidimensional data application standard is determined, and the available environmental perception multidimensional data stream is standardized and converted using the multidimensional data application standard to obtain a standard environmental perception multidimensional data stream. Feature selection and feature extraction are performed on the standard environmental perception multidimensional data stream to obtain the environmental perception multidimensional feature set.

[0045] The specific configuration of the environmental feature extraction module 12 will be described in detail below. The environmental feature extraction module 12 further includes: obtaining the environmental perception multidimensional feature set, including: performing data characteristic analysis on the environmental data source dimension information to obtain multidimensional data application characteristic information, and determining multidimensional data selection feature information based on the multidimensional data application characteristic information. Feature extraction is performed on the standard environmental perception multidimensional data stream according to the multidimensional data selection feature information to obtain an initial multidimensional environmental feature set. The initial multidimensional environmental feature set is standardized and decentralized to obtain a covariance matrix, and the covariance matrix is ​​decomposed to obtain eigenvalues ​​and eigenvectors. Feature principal component information is selected based on the eigenvalues ​​and eigenvectors, and the initial multidimensional environmental feature set data is projected onto the feature principal component information to obtain the environmental perception multidimensional feature set.

[0046] The specific configuration of the environment prediction network construction module 13 will be described in detail below. The environment prediction network construction module 13 may further include: the construction of the fish egg hatching environment prediction network library includes: mining and obtaining a fish egg hatching environment database, classifying and identifying the fish egg hatching environment database according to the type of hatched fish eggs, and obtaining a fish egg hatching environment classification data set. Feature extraction is performed on the fish egg hatching environment classification data set respectively to obtain a fish egg hatching multi-dimensional environment classification feature set. Environmental trend prediction training is performed on the fish egg hatching multi-dimensional environment classification feature set respectively to obtain a multi-dimensional environment trend prediction branch network set. The multi-dimensional environment trend prediction branch network set is classified and serially integrated according to the hatching fish egg type to generate a multi-fish egg type environment prediction network set. The multi-fish egg type environment prediction network set is identified and integrated to build the fish egg hatching environment prediction network library.

[0047] The specific configuration of the adjustment strategy parameter acquisition module 15 will be described in detail below. The adjustment strategy parameter acquisition module 15 further includes: the acquisition of the fish egg hatching environment adjustment strategy parameters includes: determining the applicable conditions of the target hatching environment according to the target hatching fish egg type. The predicted trend of the fish egg hatching environment is compared and judged using the applicable conditions of the target hatching environment to obtain the fish egg environment parameters to be optimized. The adjustment strategy analysis is performed on the fish egg environment parameters to be optimized to obtain the fish egg environment adjustment adaptive strategy, and data search is performed based on the fish egg environment adjustment adaptive strategy to construct an environment adjustment strategy space. Based on the environment adjustment strategy space, the strategy parameters of the fish egg environment parameters to be optimized are optimized to obtain the fish egg hatching environment adjustment strategy parameters.

[0048] Below, the specific configuration of the adjustment strategy parameter acquisition module 15 will be described in detail. The adjustment strategy parameter acquisition module 15 further includes: the acquisition of the fish egg hatching environment adjustment strategy parameters includes: performing strategy parameter analysis on the fish egg environment parameters to be optimized based on the environmental adjustment strategy space to obtain the available adjustment strategy parameter thresholds. The fish egg hatching environment adjustment target is indexed and effect fitted to construct a fish egg environment effect evaluation fitness function. The particle swarm parameters are initialized according to the available adjustment strategy parameter thresholds, and the particle swarm parameters include particle position and particle velocity. The particle swarm parameters are evaluated and iteratively updated using the fish egg environment effect evaluation fitness function until the preset termination condition is met, and the particle with the largest fitness is determined by optimization as the fish egg hatching environment adjustment strategy parameter.

[0049] Below, the specific configuration of the closed-loop control module 16 will be described in detail. The closed-loop control module 16 further includes: the closed-loop control of the environment of the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg incubation environment adjustment strategy parameters, including: environmental control monitoring of the fish egg incubator based on the closed-loop feedback optimization mechanism to obtain fish egg incubation environment feedback parameters. Obtain the expected environmental adjustment effect, perform iterative feedback statistical fitting on the deviation value of the fish egg incubation environment feedback parameters and the expected environmental adjustment effect based on the closed-loop feedback optimization mechanism, and obtain the environmental adjustment compensation factor. Compensate and optimize the fish egg incubation environment adjustment strategy parameters based on the environmental adjustment compensation factor, and perform environmental closed-loop control of the fish egg incubator through the compensated and optimized fish egg incubation environment adjustment strategy parameters.

[0050] The environmental adaptive adjustment system for fish egg hatching provided by the embodiment of the present invention can execute the environmental adaptive adjustment method for fish egg hatching provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0051] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0052] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for adaptively regulating the environment for hatching fish eggs, characterized in that: The method comprises: Deploy an environmental sensing device in the fish egg incubator, wherein the environmental sensing device includes a temperature sensor, a humidity sensor, a pH sensor, and a dissolved oxygen sensor; Acquire an environment perception multidimensional data stream in real time through the environment perception device, perform feature extraction on the environment perception multidimensional data stream, and obtain an environment perception multidimensional feature set; Building a fish egg hatching environment prediction network library, matching the target hatching fish egg type of the fish egg incubator with the fish egg hatching environment prediction network library, and calling to obtain a target fish egg hatching environment prediction network; Performing environmental prediction on the environmental perception multidimensional feature set based on the target fish egg hatching environment prediction network to obtain a fish egg hatching environment prediction trend; Performing hatching condition adjustment analysis on the predicted trend of the fish egg hatching environment to obtain fish egg hatching environment adjustment strategy parameters; Establishing a closed-loop feedback optimization mechanism, and performing closed-loop regulation of the environment of the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg incubation environment regulation strategy parameters; The method of building a fish egg hatching environment prediction network library comprises: Mining and acquiring a fish egg hatching environment database, classifying and identifying the fish egg hatching environment database according to the types of hatched fish eggs, and obtaining a fish egg hatching environment classification data set; Extracting features from the fish egg hatching environment classification data sets respectively to obtain a fish egg hatching multidimensional environment classification feature set; Performing environmental trend prediction training on the multidimensional environmental classification feature set for fish egg hatching respectively to obtain a multidimensional environmental trend prediction branch network set; Classifying and serially fusing the multidimensional environmental trend prediction branch network set according to the hatching fish egg type to generate a multi-fish egg type environmental prediction network set; The multiple fish egg type environment prediction network sets are labeled and integrated to build the fish egg hatching environment prediction network library.

2. The method for environmental self-adaptation for fish egg hatching according to claim 1, characterized in that: The step of obtaining a multi-dimensional feature set of environmental perception includes: Performing outlier identification and data cleaning processing on the environment perception multidimensional data stream to obtain a usable environment perception multidimensional data stream; Obtaining environmental data source dimension information according to the available environmental perception multidimensional data stream; Based on the source dimension information of the environmental data, determine a multidimensional data application standard, and use the multidimensional data application standard to perform a standardized conversion on the available environmental perception multidimensional data stream to obtain a standard environmental perception multidimensional data stream; Feature selection and feature extraction are performed on the standard environment-aware multidimensional data stream to obtain the environment-aware multidimensional feature set.

3. The environmental adaptive adjustment method for fish egg hatching according to claim 2, characterized in that: The step of obtaining the environment perception multidimensional feature set comprises: Performing data characteristic analysis on the source dimension information of the environmental data to obtain multidimensional data application characteristic information, and determining multidimensional data selection feature information based on the multidimensional data application characteristic information; Extract features from the standard environment perception multidimensional data stream according to the multidimensional data selection feature information to obtain an initial multidimensional environment feature set; Standardizing and decentralizing the initial multi-dimensional environmental feature set to obtain a covariance matrix, and decomposing the covariance matrix to obtain eigenvalues ​​and eigenvectors; Feature principal component information is selected according to the feature value and the feature vector, and the initial multi-dimensional environmental feature set data is projected onto the feature principal component information to obtain the environmental perception multi-dimensional feature set.

4. The method for environmental self-adaptation for fish egg hatching according to claim 1, characterized in that: The method of obtaining the parameters of the fish egg hatching environment adjustment strategy comprises: Determining applicable conditions of a target hatching environment according to the target hatching fish egg type; The target hatching environment applicable conditions are used to compare and judge the predicted trend of the fish egg hatching environment to obtain the fish egg environment parameters to be optimized; Performing adjustment strategy analysis on the fish egg environmental parameters to be optimized to obtain a fish egg environmental adjustment adaptive strategy, and performing data search according to the fish egg environmental adjustment adaptive strategy to construct an environmental adjustment strategy space; Based on the environmental regulation strategy space, the fish egg environment parameters to be optimized are optimized to obtain fish egg hatching environment regulation strategy parameters.

5. The method for environmental self-adaptation for fish egg hatching according to claim 4, characterized in that: The method of obtaining the parameters of the fish egg hatching environment adjustment strategy comprises: Based on the environmental adjustment strategy space, the fish egg environmental parameters to be optimized are analyzed for strategy parameters to obtain available adjustment strategy parameter thresholds; The indicators of the regulation target of the fish egg hatching environment are decomposed and the effect is fitted, and the fitness function for evaluating the effect of the fish egg environment is constructed; Initialize particle swarm parameters according to the available adjustment strategy parameter threshold, wherein the particle swarm parameters include particle position and particle velocity; The fish egg environment effect evaluation fitness function is used to evaluate and iteratively update the particle swarm parameters until a preset termination condition is met, and the particle with the largest fitness is optimally determined as the fish egg hatching environment adjustment strategy parameter.

6. The method for environmental self-adaptation for fish egg hatching according to claim 1, characterized in that: The closed-loop control of the environment of the fish egg incubator based on the closed-loop feedback optimization mechanism and the fish egg incubation environment adjustment strategy parameters includes: Based on the closed-loop feedback optimization mechanism, the fish egg incubator is subjected to environmental control monitoring to obtain fish egg incubation environmental feedback parameters; Obtaining the expected environmental regulation effect, performing iterative feedback statistical fitting on the deviation value of the fish egg hatching environment feedback parameter and the expected environmental regulation effect based on the closed-loop feedback optimization mechanism, and obtaining the environmental regulation compensation factor; The fish egg hatching environment adjustment strategy parameters are compensated and optimized based on the environmental adjustment compensation factor, and the fish egg hatching environment adjustment strategy parameters after compensation and optimization are used to perform environmental closed-loop control on the fish egg incubator.

7. An environmental adaptive regulation system for fish egg hatching, characterized in that: The system comprises: An environmental sensing module is used to deploy an environmental sensing device in the fish egg incubator, wherein the environmental sensing device includes a temperature sensor, a humidity sensor, a pH sensor, and a dissolved oxygen sensor; An environmental feature extraction module, used to obtain an environmental perception multidimensional data stream in real time through the environmental perception device, perform feature extraction on the environmental perception multidimensional data stream, and obtain an environmental perception multidimensional feature set; An environment prediction network construction module is used to build a fish egg hatching environment prediction network library, match the target hatching fish egg type of the fish egg incubator with the fish egg hatching environment prediction network library, and call to obtain the target fish egg hatching environment prediction network; An environment prediction module, used to perform environment prediction on the environment perception multidimensional feature set based on the target fish egg hatching environment prediction network to obtain a fish egg hatching environment prediction trend; An adjustment strategy parameter acquisition module is used to perform an incubation condition adjustment analysis on the predicted trend of the fish egg incubation environment to obtain the fish egg incubation environment adjustment strategy parameters; A closed-loop control module, used for establishing a closed-loop feedback optimization mechanism, and performing closed-loop control of the environment of the fish egg incubator based on the closed-loop feedback optimization mechanism and the parameters of the fish egg incubation environment adjustment strategy; The environment prediction network building module is used to: Mining and acquiring a fish egg hatching environment database, classifying and identifying the fish egg hatching environment database according to the types of hatched fish eggs, and obtaining a fish egg hatching environment classification data set; Extracting features from the fish egg hatching environment classification data sets respectively to obtain a fish egg hatching multidimensional environment classification feature set; Performing environmental trend prediction training on the multidimensional environmental classification feature set for fish egg hatching respectively to obtain a multidimensional environmental trend prediction branch network set; Classifying and serially fusing the multidimensional environmental trend prediction branch network set according to the hatching fish egg type to generate a multi-fish egg type environmental prediction network set; The multiple fish egg type environment prediction network sets are labeled and integrated to build the fish egg hatching environment prediction network library.

Images