Aquaculture environment monitoring method and system

Abstract

The invention provides an aquaculture environment monitoring method and system. The method comprises the steps that an aquaculture environment is divided into a plurality of space areas which are the same in size and connected with one another, sensor groups are arranged in the space areas which are connected with one another, and aquaculture environment parameters are obtained; the aquaculture environment parameters are transmitted to a central processing system through LoRa, and a prediction model built in the central processing system processes the aquaculture environment parameters; the central processing system processes the aquaculture environment index, obtains an aquaculture environment prediction value after processing, calculates whether the aquaculture environment prediction value exceeds a preset threshold value through a comprehensive aquaculture environment index model, and sends alarm information to equipment of aquaculture personnel if the aquaculture environment prediction value exceeds the preset threshold value. Through the method and the corresponding system, the intelligence of aquaculture is improved, manpower is greatly liberated to monitor various environmental parameters, and the abnormality of the aquaculture environmental parameters can be predicted in time.

Description

Technical Field

[0001] The invention proposes an aquaculture environment monitoring method and system, belonging to the technical field of aquaculture supervision. Background Art

[0002] Aquaculture is an important part of ensuring global food security and nutrition supply. In recent decades, the depletion of wild fishery resources, the rising global population and the growing demand for food and protein have jointly promoted the rapid development of aquaculture.

[0003] With the development of Internet of Things technology, the application of Internet of Things technology in the field of aquaculture has also become one of the current research hotspots. However, in traditional aquaculture environmental monitoring, it is usually necessary to deploy multiple sensors in different areas of the aquaculture environment, and the user-side page will display the values ​​of these areas at the same time. For large aquaculture ponds, the process of manually checking a large number of environmental parameters is cumbersome and time-consuming. In addition, even if an alarm is issued when a certain environmental parameter exceeds the threshold, it is often too late to take remedial measures. Therefore, the traditional monitoring method has many shortcomings and urgently needs to be improved and optimized. Summary of the invention

[0004] The present invention provides an aquaculture environment monitoring method and system. Considering the accuracy and reliability of sensors, different sensors may have different accuracy and stability, so the quality of data they provide may be different. In this case, weighted fusion is used to obtain various aquaculture environment parameters of the entire aquaculture environment, and input into the prediction model to check whether the predicted environment value exceeds the threshold, so as to solve the above-mentioned problems:

[0005] The present invention proposes a method for monitoring an aquaculture environment, the method comprising:

[0006] The aquaculture environment is divided into a number of spatial areas of equal volume and interconnected, and sensor groups are respectively deployed in the interconnected spatial areas to obtain aquaculture environment parameters;

[0007] The aquaculture environment parameters are transmitted to a central processing system via LoRa, and a built-in prediction model of the central processing system processes the aquaculture environment parameters;

[0008] After processing, the predicted value of the aquaculture environment is obtained, and the comprehensive aquaculture environment index model is used to calculate whether the predicted value of the aquaculture environment exceeds a preset threshold. If it exceeds the preset threshold, the central processing system sends an alarm message to the equipment of the aquaculture personnel.

[0009] Furthermore, a method for monitoring an aquaculture environment is provided, wherein the aquaculture environment is divided into a number of spatial regions of equal volume and interconnected with each other, and sensor groups are respectively deployed in the interconnected spatial regions to obtain aquaculture environment parameters, including:

[0010] The aquaculture environment is divided into a number of spatial areas of equal volume and interconnected with each other, and sensor groups are respectively deployed in the interconnected spatial areas, wherein the sensor groups include a dissolved oxygen sensor, a temperature sensor, a pH sensor, an ammonia nitrogen sensor, a total phosphorus sensor, a total nitrogen sensor and a turbidity sensor;

[0011] Collecting environmental factor data measured by the sensor group, and preprocessing the environmental factor data, wherein the preprocessing includes filtering, denoising and smoothing;

[0012] After preprocessing, the weight of each sensor is dynamically adjusted based on the sensor performance evaluation results and environmental difference analysis;

[0013] The sensor data from each area are weightedly fused using the adjusted weights to obtain various aquaculture environmental parameters.

[0014] Furthermore, an aquaculture environment monitoring method dynamically adjusts the weight of each sensor according to the sensor performance evaluation results and environmental difference analysis, including:

[0015] The sensor is placed in an environment with constant environmental parameters for testing, and the stability of the sensor is obtained by calculating the standard deviation of the sensor readings over a period of time;

[0016] The mean absolute error is used to calculate the accuracy evaluation value of the sensor, and the sensor performance evaluation result is obtained according to the stability and accuracy evaluation value of the sensor;

[0017] For each of the environmental factor data, the difference between the data and the standard value is calculated, and the differences are standardized to the same scale, and the environmental difference coefficient is calculated by combining multiple environmental factors;

[0018] The weight of the sensor is obtained according to the sensor performance evaluation results and the environmental difference coefficient.

[0019] Furthermore, an aquaculture environment monitoring method is provided, wherein the aquaculture environment parameters are transmitted to a central processing system via LoRa, and a built-in prediction model of the central processing system processes the aquaculture environment parameters, including:

[0020] The central processing system acquires historical data of the sensor group, and performs preliminary processing on the historical data of the sensor group, wherein the preliminary processing includes data cleaning, data transformation and feature extraction, wherein the feature extraction includes extracting statistical features and time series features from the original data, and after the preliminary processing, the data is divided into a training set and a test set;

[0021] Building an aquaculture environment prediction model through a long short-term memory network (LSTM), and training the aquaculture environment prediction model using a training set in the preliminarily processed data;

[0022] Using the test set in the initially processed data to evaluate and verify the trained aquaculture environment prediction model;

[0023] The aquaculture environment parameters are input into the aquaculture environment prediction model to obtain an aquaculture environment prediction value.

[0024] Further, an aquaculture environment monitoring method is provided, wherein after processing, an aquaculture environment prediction value is obtained, and a comprehensive aquaculture environment index model is used to calculate whether the aquaculture environment prediction value exceeds a preset threshold value. If the preset threshold value is exceeded, a central processing system sends an alarm message to a device of an aquaculture personnel, including:

[0025] The aquaculture environment prediction value is input into the aquaculture environment index model, and the aquaculture environment model calculates whether the aquaculture environment prediction value exceeds a preset threshold. Specifically, the aquaculture environment model is

[0026]

[0027] Where CEI is the comprehensive environmental index, n is the total number of key parameters, and w i is the weight of the i-th parameter, indicating the importance of this parameter to the suitability of the fish living environment, P i is the normalized value of the ith parameter, indicating the relative position between the current measured value of the parameter and its ideal range, e i is the exponential coefficient of the i-th parameter, which is used to adjust the influence of the parameter on the comprehensive index;

[0028] If the predicted value of the aquaculture environment exceeds a preset threshold, the central processing system sends an alarm message to the mobile device of the aquaculture personnel.

[0029] The present invention provides an aquaculture environment monitoring system, the system comprising:

[0030] The aquaculture environment monitoring module is used to divide the aquaculture environment into a number of spatial areas of equal volume and interconnected, and to deploy sensor groups in the interconnected spatial areas to obtain aquaculture environment parameters;

[0031] A module for predicting aquaculture environmental parameters, used for transmitting the aquaculture environmental parameters to a central processing system via LoRa, and a prediction model built into the central processing system processes the aquaculture environmental parameters;

[0032] The alarm module is used to obtain the predicted value of the aquaculture environment after processing, and calculate whether the predicted value of the aquaculture environment exceeds a preset threshold through a comprehensive aquaculture environment index model. If it exceeds the preset threshold, the central processing system sends an alarm message to the equipment of the aquaculture personnel.

[0033] Further, an aquaculture environment monitoring system, the aquaculture environment monitoring module comprises:

[0034] A sensor group deployment module is used to divide the aquaculture environment into a number of spatial areas of equal volume and interconnected, and respectively deploy sensor groups in the interconnected spatial areas, wherein the sensor groups include a dissolved oxygen sensor, a temperature sensor, a pH sensor, an ammonia nitrogen sensor, a total phosphorus sensor, a total nitrogen sensor, and a turbidity sensor;

[0035] A preprocessing module, used for collecting the environmental factor data measured by the sensor group, and preprocessing the environmental factor data, wherein the preprocessing includes filtering, denoising and smoothing;

[0036] Dynamically adjust sensor weight module, which is used to dynamically adjust the weight of each sensor according to sensor performance evaluation results and environmental difference analysis after preprocessing;

[0037] The module for obtaining aquaculture environmental parameters is used to perform weighted fusion on the sensor data of each area using the adjusted weights to obtain various aquaculture environmental parameters.

[0038] Furthermore, in an aquaculture environment monitoring system, the module for dynamically adjusting sensor weights includes:

[0039] The sensor stability acquisition module is used to test the sensor in an environment with constant environmental parameters and obtain the sensor stability by calculating the standard deviation of the sensor readings over a period of time;

[0040] A module for obtaining sensor performance evaluation results is used to calculate the accuracy evaluation value of the sensor using the mean absolute error, and obtain the sensor performance evaluation result according to the stability and accuracy evaluation value of the sensor;

[0041] An environment difference coefficient acquisition module is used to calculate the difference between each environmental factor data and the standard value, standardize these differences to the same scale, and combine multiple environmental factors to calculate the environment difference coefficient;

[0042] The sensor weight obtaining module is used to obtain the sensor weight according to the sensor performance evaluation result and the environmental difference coefficient.

[0043] Furthermore, in an aquaculture environment monitoring system, the module for predicting aquaculture environment parameters comprises:

[0044] A preliminary processing module, used for the central processing system to obtain historical data of the sensor group, and to perform preliminary processing on the historical data of the sensor group, wherein the preliminary processing includes data cleaning, data transformation and feature extraction, wherein the feature extraction includes extracting statistical features and time series features from the original data, and after the preliminary processing, the data is divided into a training set and a test set;

[0045] A training model module is used to construct an aquaculture environment prediction model through a long short-term memory network (LSTM), and train the aquaculture environment prediction model using a training set in the preliminarily processed data;

[0046] A testing module, used to evaluate and verify the trained aquaculture environment prediction model using a test set in the preliminarily processed data;

[0047] The module for obtaining the predicted value of the aquaculture environment is used to input the aquaculture environment parameters into the aquaculture environment prediction model to obtain the predicted value of the aquaculture environment.

[0048] Further, in an aquaculture environment monitoring system, the alarm module comprises:

[0049] The module for calculating whether the value exceeds a preset threshold is used to input the predicted value of the aquaculture environment into the aquaculture environment index model, and the aquaculture environment model calculates whether the predicted value of the aquaculture environment exceeds a preset threshold.

[0050] The module for sending alarm information is used for the central processing system to send alarm information to the mobile device of the aquaculture personnel if the predicted value of the aquaculture environment exceeds a preset threshold.

[0051] Beneficial effects of the present invention: through real-time monitoring and sensor data collection, the system can obtain water quality parameters in real time, and use the prediction model to predict and evaluate the aquaculture environment; the central processing system determines whether the predicted value of the aquaculture environment exceeds the preset threshold value according to the preset aquaculture environment index model. Once the predicted value exceeds the threshold value, the system will trigger the alarm mechanism; with the help of the analysis of the intelligent prediction model, the system can timely discover abnormal water quality conditions, help aquaculture personnel to respond quickly, and avoid aquaculture losses caused by abnormal water quality; the intelligent monitoring system can provide sufficient water quality data support to help aquaculture personnel better understand the real-time situation of the aquaculture environment; with the help of the intelligent prediction model and the alarm mechanism, the system can accurately warn of water quality problems, which helps to avoid unexpected losses and improve aquaculture efficiency; improving monitoring accuracy and system response speed, helps to finely manage the aquaculture environment and improve aquaculture benefits; the intelligent aquaculture environment monitoring system can realize real-time monitoring and prediction of water quality environment changes, and can issue alarm information in time, thereby helping aquaculture personnel to adjust management measures in time, ensure the stability and health of the aquaculture environment, and improve aquaculture efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a schematic diagram of an aquaculture environment monitoring method according to the present invention. DETAILED DESCRIPTION

[0053] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0054] One embodiment of the present invention provides an aquaculture environment monitoring method, the method comprising:

[0055] The aquaculture environment is divided into a number of spatial areas of equal volume and interconnected, and sensor groups are respectively deployed in the interconnected spatial areas to obtain aquaculture environment parameters;

[0056] The aquaculture environment parameters are transmitted to a central processing system via LoRa, and a built-in prediction model of the central processing system processes the aquaculture environment parameters;

[0057] After processing, the predicted value of the aquaculture environment is obtained, and the comprehensive aquaculture environment index model is used to calculate whether the predicted value of the aquaculture environment exceeds a preset threshold. If it exceeds the preset threshold, the central processing system sends an alarm message to the equipment of the aquaculture personnel.

[0058] The working principle of the above technical solution is as follows: the aquaculture environment is divided into multiple spatial areas of equal volume and interconnected with each other. This division can ensure detailed and comprehensive monitoring of the entire aquaculture environment. In each divided spatial area, a sensor group is deployed to obtain the aquaculture environmental parameters of the area in real time, including temperature, dissolved oxygen, ammonia nitrogen content, etc.; through LoRa (long-range radio) communication technology, the sensor group of each spatial area transmits the collected aquaculture environmental parameters to the central processing system. LoRa technology is known for its low power consumption and long-distance communication capabilities, which is suitable for the data transmission requirements of this distributed monitoring system; after receiving the data, the central processing system uses its built-in prediction model to process these aquaculture environmental parameters. The model predicts the changing trend of the aquaculture environment based on machine learning. Through the comprehensive aquaculture environmental index model, the central processing system will calculate the predicted value of the aquaculture environment and compare it with the preset threshold. If the predicted value exceeds the preset threshold, it means that some environmental parameters may reach or exceed the dangerous level, which may have an adverse impact on aquaculture. In this case, the central processing system will immediately send an alarm message to the equipment of the aquaculture personnel so that they can take quick measures to respond.

[0059] The effects of the above technical solution are as follows: through real-time monitoring and sensor data collection, the system can obtain water quality parameters in real time, and use the prediction model to predict and evaluate the aquaculture environment; the central processing system determines whether the predicted value of the aquaculture environment exceeds the preset threshold according to the preset aquaculture environment index model. Once the predicted value exceeds the threshold, the system will trigger the alarm mechanism; with the help of the analysis of the intelligent prediction model, the system can timely detect abnormal water quality conditions, help aquaculture personnel to respond quickly, and avoid aquaculture losses caused by abnormal water quality; the intelligent monitoring system can provide sufficient water quality data support to help aquaculture personnel better understand the real-time situation of the aquaculture environment; with the help of intelligent prediction models and alarm mechanisms, the system can accurately warn of water quality problems, help avoid unexpected losses and improve aquaculture efficiency; improve monitoring accuracy and system response speed, help to finely manage the aquaculture environment and improve aquaculture benefits; the intelligent aquaculture environment monitoring system can realize real-time monitoring and prediction of water quality environment changes, and can issue alarm information in time, so as to help aquaculture personnel adjust management measures in time, ensure the stability and health of the aquaculture environment, and improve aquaculture efficiency and quality.

[0060] One embodiment of the present invention is a method for monitoring an aquaculture environment, which divides an aquaculture environment into a number of spatial regions of equal volume and interconnected with each other, and respectively deploys sensor groups in the interconnected spatial regions to obtain aquaculture environment parameters, including:

[0061] The aquaculture environment is divided into a number of spatial areas of equal volume and interconnected with each other, and sensor groups are respectively deployed in the interconnected spatial areas, wherein the sensor groups include a dissolved oxygen sensor, a temperature sensor, a pH sensor, an ammonia nitrogen sensor, a total phosphorus sensor, a total nitrogen sensor and a turbidity sensor;

[0062] Collecting environmental factor data measured by the sensor group, and preprocessing the environmental factor data, wherein the preprocessing includes filtering, denoising and smoothing;

[0063] After preprocessing, the weight of each sensor is dynamically adjusted based on the sensor performance evaluation results and environmental difference analysis;

[0064] The sensor data from each area are weightedly fused using the adjusted weights to obtain various aquaculture environmental parameters.

[0065] The working principle of the above technical solution is as follows: the aquaculture environment is first finely divided into multiple spatial areas of equal volume and interconnected. In each such spatial area, a set of sensor groups are deployed, including dissolved oxygen sensors, temperature sensors, pH sensors, ammonia nitrogen sensors, total phosphorus sensors, total nitrogen sensors and turbidity sensors. These sensors can comprehensively and real-time monitor the key environmental factors of the area; the sensor group will continuously collect the corresponding environmental factor data, such as dissolved oxygen content, temperature, pH value, etc. Before being transmitted to the central processing system, these data will be pre-processed, including filtering, denoising and smoothing. Filtering can remove high-frequency noise in the data, denoising can further reduce random errors, and smoothing can make the data more stable and continuous; after pre-processing, the system will dynamically adjust the weight of each sensor according to the sensor performance evaluation results and environmental difference analysis. Through dynamic weight adjustment, the system can more accurately reflect the actual environmental conditions of each area; using the adjusted weights, the system will weightedly fuse the sensor data of each area to obtain the comprehensive value of various aquaculture environmental parameters. These comprehensive values ​​not only take into account the raw data of each sensor, but also fully consider the impact of sensor performance and environmental differences on the final results.

[0066] The effects of the above technical solutions are as follows: through preprocessing and dynamic weight adjustment, the quality and accuracy of sensor data can be improved to ensure the reliability of collected data; by adopting weighted fusion algorithm and combining data from different sensors, more comprehensive and comprehensive aquaculture environmental parameters can be obtained to provide more complete information for aquaculture management; according to sensor performance evaluation and environmental difference analysis, the weight of the sensor is dynamically adjusted to improve the adaptability and flexibility of the system; by integrating data from multiple sensors, the system can provide more accurate and complete aquaculture environmental parameters to help aquaculture personnel understand the water quality more comprehensively; dynamic weight adjustment and data fusion improve the intelligence level of the system, better adapt to the characteristics of different areas of the aquaculture environment, and improve the intelligence level of the system; by providing high-quality environmental parameter data and intelligent data processing methods, it can help aquaculture personnel optimize management decisions and improve aquaculture efficiency and production quality; the intelligent aquaculture environment monitoring system realizes more accurate and comprehensive monitoring and evaluation of the aquaculture environment by comprehensively utilizing data from multiple environmental factors, performing data preprocessing and dynamic weight adjustment, which helps to improve the quality of aquaculture water and management efficiency.

[0067] One embodiment of the present invention is an aquaculture environment monitoring method, which dynamically adjusts the weight of each sensor according to sensor performance evaluation results and environmental difference analysis, including:

[0068] The sensor is placed in an environment with constant environmental parameters for testing. The stability of the sensor is obtained by calculating the standard deviation of the sensor readings over a period of time. Specifically, the stability of the sensor can be calculated using the following model:

[0069] P s,i (t) = 1-β i (t) / β max

[0070] Among them, P s,i (t) represents the stability of the i-th sensor in time period t, β i (t) is the standard deviation of the i-th sensor reading in time period t, and βmax is the maximum standard deviation among all sensors;

[0071] The mean absolute error is used to calculate the accuracy evaluation value of the sensor, and the sensor performance evaluation result is obtained according to the stability and accuracy evaluation value of the sensor. Specifically, the accuracy evaluation value of the sensor can be calculated by the following model:

[0072] P a,i (t) = 1-M i (t) / M max

[0073] Among them, M i(t) is the mean absolute error of the ith sensor in time period t, M max is the preset maximum mean absolute error;

[0074] Specifically, the sensor performance evaluation results are obtained according to the stability and accuracy evaluation values ​​of the sensor through the following model:

[0075] P i (t) = W S *P s,i (t)+W a *P a,i (t)

[0076] Among them, W S and W a is the weight of sensor stability assessment and accuracy assessment, W S +W a =1.

[0077] For each of the environmental factor data, the difference between the data and the standard value is calculated, and the differences are standardized to the same scale, and the environmental difference coefficient is calculated by combining multiple environmental factors;

[0078] Specifically, the environmental difference coefficient is calculated by the following model:

[0079]

[0080] Among them, e tmp,i (t) is the water temperature difference coefficient of the i-th sensor at time t, T i (t) is the water temperature in the area where the i-th sensor is located at time t, Ts is the standard water temperature, T r is the optimal operating temperature range of the sensor. If the optimal operating temperature range of the sensor is 18℃~20℃, then T r =12℃, here, only the calculation model of water temperature difference coefficient is given, the calculation model of difference coefficients such as dissolved oxygen and pH is the same as the calculation model of water temperature difference coefficient.

[0081] Because there are multiple environmental factors to consider, they are combined using a weighted average approach:

[0082] e i (t) = W tmp *e tmp,i (t)+W dio *e dio,i (t)+……

[0083] Among them, e i (t) is the environmental difference coefficient of the i sensors at time t, e dio,i (t) is the dissolved oxygen difference coefficient of the ith sensor at time t, Wtmp is the weight of the environmental factor water temperature, W dio is the weight of the environmental factor dissolved oxygen.

[0084] The weight of the sensor is obtained according to the sensor performance evaluation results and the environmental difference coefficient.

[0085]

[0086] Among them, W i (t) is the weight of the i-th sensor at time t, P i (t) is the performance evaluation result of the i-th sensor at time t, N is the total number of sensors of this type in the entire aquaculture environment, and e i (t) is the environmental difference coefficient of the ith sensor at time t.

[0087] The working principle and effect of the above technical solution are as follows: the sensor is placed in an environment where the environmental parameters are constant for testing. The stability of the sensor is evaluated by recording the sensor readings over a period of time and calculating the standard deviation of these readings. A small standard deviation means that the output of the sensor is relatively stable and has small fluctuations. The mean absolute error is used to calculate the accuracy evaluation value of the sensor. By comparing the difference between the sensor reading and the true value or standard value, a numerical value representing the accuracy of the sensor can be obtained. A smaller mean absolute error value indicates that the sensor has higher accuracy. For each environmental factor data (such as temperature, dissolved oxygen, etc.), the difference between it and the standard value or ideal value is calculated. Then, these differences are standardized to the same scale so that multiple environmental factors can be compared and combined. Finally, multiple environmental factors are combined to calculate a comprehensive environmental difference coefficient, which is used to represent the degree of difference between the current environment and the standard or ideal environment. According to the stability evaluation results and accuracy evaluation values ​​of the sensor, as well as the environmental difference coefficient, the weight of the sensor can be comprehensively obtained. The weight reflects the importance and reliability of the sensor in the current environment and can be used in subsequent sensor data fusion or decision-making processes. By comprehensively considering the stability and accuracy of the sensor and the impact of environmental factors on the performance of the sensor, the performance of the sensor can be evaluated more accurately. This evaluation method is more objective and comprehensive, avoiding the one-sidedness of single indicator evaluation; based on the performance evaluation results of the sensor and the environmental difference coefficient, the sensor can be selected and used more specifically. In a complex and changing environment, the weight of the sensor can be adjusted according to actual needs to improve the performance and reliability of the entire system; by calculating the environmental difference coefficient, the degree of influence of different environmental factors on sensor performance can be quantified. This enables the sensor to better adapt to different working environments and improve its accuracy and stability under various conditions.

[0088] The formulas for performance evaluation, accuracy evaluation, and comprehensive performance evaluation are designed to quantify the performance of sensors, so as to facilitate the comparison and selection of sensors. These formulas provide an objective and comparable method to measure the performance of sensors, avoiding subjective assumptions and uncertainties. Specifically, the design idea of ​​the stability evaluation and accuracy evaluation formulas is to measure the performance of sensors based on the difference between the sensor reading and the ideal value or reference value. By standardizing the difference (using the mean absolute error) and comparing the result with the maximum possible difference, an evaluation value between 0 and 1 can be obtained. This design makes the evaluation result more intuitive and easy to understand, and also facilitates the comparison of the performance of different sensors; the comprehensive performance evaluation formula is designed to combine multiple performance indicators to form a comprehensive evaluation result. By using methods such as weighted average, different performance indicators can be given different weights according to actual needs, so as to obtain a comprehensive evaluation value that is more in line with actual needs. Such a design can more comprehensively reflect the performance of the sensor and avoid the one-sidedness of a single indicator evaluation. The calculation of the water temperature environmental difference coefficient and the comprehensive environmental difference coefficient is to quantify the difference between the environmental conditions in the area where the sensor is located and the standard conditions or ideal conditions. This difference may affect the performance of the sensor, so it needs to be included in the performance evaluation. By calculating the coefficient of difference, a relatively objective measure can be obtained to compare the performance of different sensors in different environments. Such calculations can more accurately reflect the performance of sensors in actual applications and provide a more reliable basis for the selection and use of sensors. These formulas provide a standardized and comparable method to measure the performance of sensors, avoiding subjective assumptions and uncertainties, making the evaluation results more accurate and reliable. The function for calculating sensor weights ensures that the sum of all weights is 1, and the output can be interpreted as a probability distribution, which is very suitable for multi-classification tasks; through the exponential function, the function amplifies the difference between different input values, so that relatively large input values ​​correspond to larger weights, thereby highlighting important features; the function is continuous and differentiable, which makes it possible to use it in gradient-based optimization algorithms; P i (t) as the numerator, which means that sensors with better performance will receive a greater weight; This part uses the environmental difference coefficient ei(t) to adjust the weight, but avoids the problems that may be caused by directly using its reciprocal (such as the weight becoming too large when ei(t) is close to zero). By adding 1 and then taking the reciprocal, it ensures that when the environmental difference increases, the weight will decrease smoothly instead of suddenly becoming very large or very small.

[0089] One embodiment of the present invention is a method for monitoring an aquaculture environment, wherein the aquaculture environment parameters are transmitted to a central processing system via LoRa, and a prediction model built into the central processing system processes the aquaculture environment parameters, including:

[0090] The central processing system acquires historical data of the sensor group, and performs preliminary processing on the historical data of the sensor group, wherein the preliminary processing includes data cleaning, data transformation and feature extraction, wherein the feature extraction includes extracting statistical features and time series features from the original data, and after the preliminary processing, the data is divided into a training set and a test set;

[0091] Building an aquaculture environment prediction model through a long short-term memory network (LSTM), and training the aquaculture environment prediction model using a training set in the preliminarily processed data;

[0092] Using the test set in the initially processed data to evaluate and verify the trained aquaculture environment prediction model;

[0093] The aquaculture environment parameters are input into the aquaculture environment prediction model to obtain an aquaculture environment prediction value.

[0094] The working principle and effect of the above technical solution are as follows: the central processing system first obtains historical data from the sensor group, which may contain noise and outliers, so preliminary processing is required. Preliminary processing includes data cleaning, data transformation and feature extraction. Data cleaning aims to remove noise and outliers, and data transformation converts data into a format suitable for model processing, such as standardization or normalization. Feature extraction is to extract meaningful statistical features and time series features from the original data, which will be used for model training and evaluation; the data after preliminary processing is divided into training set and test set. The training set is used for model training, while the test set is used to evaluate the performance of the model; a long short-term memory network (LSTM) is used to build an aquaculture environment prediction model. LSTM is a recurrent neural network suitable for processing time series data. The model is trained using the training set, and the model parameters are adjusted through the optimization algorithm to minimize the difference between the predicted value and the actual value; the trained model is evaluated and verified using the test set. Evaluation indicators may include precision, recall rate, F1 value, etc., which are used to measure the performance of the model on unseen data. The validation process ensures that the model has good generalization ability and can provide accurate predictions in practical applications; the actual aquaculture environmental parameters are input into the trained model to obtain the predicted values ​​of the aquaculture environment. These predicted values ​​can provide farmers with information about future environmental conditions and help them make reasonable farming decisions. Through preliminary processing and data set division, the quality of data input to the model can be ensured to be high, thereby improving the prediction accuracy of the model. The LSTM model can capture long-term dependencies in time series data, further improving the accuracy of the prediction; accurate environmental predictions can help farmers understand future changes in environmental conditions in advance, thereby optimizing farming strategies, reducing risks and increasing production. This is of great significance to the sustainable development of the aquaculture industry; promoting intelligent farming, by combining sensor technology with machine learning models, intelligent monitoring and prediction of the aquaculture environment are realized, which helps to promote the modernization and intelligentization of the aquaculture industry and improve production efficiency and management level.

[0095] An embodiment of the present invention is a method for monitoring an aquaculture environment, wherein after processing, a predicted value of an aquaculture environment is obtained, and a comprehensive aquaculture environment index model is used to calculate whether the predicted value of the aquaculture environment exceeds a preset threshold value. If the predicted value exceeds the preset threshold value, a central processing system sends an alarm message to a device of an aquaculture personnel, including:

[0096] The aquaculture environment prediction value is input into the aquaculture environment index model, and the aquaculture environment model calculates whether the aquaculture environment prediction value exceeds a preset threshold. Specifically, the aquaculture environment model is

[0097]

[0098] Where CEI is the comprehensive environmental index, n is the total number of environmental parameters, and w j is the weight of the jth parameter, indicating the importance of this parameter to the suitability of the aquaculture living environment, x j is the predicted value of the jth parameter, x min ≤x j ≤x max , x min and x max are the minimum and maximum possible values ​​of the parameter suitable for fish life, respectively; ej is the exponential coefficient of the jth parameter, which is used to adjust the degree of influence of the parameter on the comprehensive index;

[0099] If the predicted value of the aquaculture environment exceeds a preset comprehensive environmental index threshold, the central processing system sends an alarm message to the mobile device of the aquaculture personnel.

[0100] It should be noted that the predicted values ​​of various environmental parameters need to be compared with the environmental parameter thresholds stored in the central processing system in advance. If a predicted value of an environmental parameter exceeds the threshold, the information that the environmental parameter will exceed the threshold will be sent directly to the aquaculture personnel to guide the aquaculture personnel to make adjustments. If none of the predicted environmental parameters exceeds the threshold, the predicted values ​​of each environment will be input into the aquaculture environmental index model.

[0101] In aquaculture environmental monitoring, a single parameter not exceeding the threshold does not mean that the overall environmental conditions are safe. The environment is a complex system, and there may be interactions and influences between various parameters. Sometimes, even if each individual environmental parameter is within the normal range, when they are considered together, unexpected effects may occur. The comprehensive environmental index model is designed to solve this problem. It takes into account multiple environmental parameters and combines these parameters into an index value through a certain algorithm. This index value can more comprehensively reflect the overall condition of the environment, thereby providing a more accurate basis for judgment. Therefore, even if none of the parameters exceed the threshold, after entering the comprehensive environmental index model, it is still possible to find that the overall environmental conditions exceed the preset threshold. In this case, timely measures need to be taken to adjust and improve to ensure the sustainability and safety of the aquaculture environment.

[0102] The working principle and effect of the above technical solution are as follows: the predicted aquatic environment value obtained by the LSTM model is input into the aquaculture environment index model. This environmental index model is a mathematical model that integrates multiple environmental parameters and is used to evaluate the overall condition of the aquaculture environment. The aquaculture environment model calculates whether the predicted aquatic environment value exceeds these thresholds based on the preset environmental parameter thresholds. These thresholds are usually set based on the growth requirements, environmental tolerance and industry standards of aquaculture organisms. If the predicted aquatic environment value exceeds the preset threshold, it means that the current or upcoming environmental conditions may have an adverse effect on aquaculture organisms. At this time, the central processing system triggers an alarm mechanism. The central processing system sends a message to the connected mobile network or other communication methods. The mobile device of the aquaculture personnel sends an alarm message; by inputting the predicted value of the aquatic environment into the environmental index model in real time, the system can continuously monitor the status of the aquaculture environment and issue a timely warning when unfavorable conditions are predicted; since the alarm information is sent directly to the mobile device of the aquaculture personnel, this ensures that the aquaculture personnel can be quickly notified of environmental abnormalities, so that they can take necessary measures as soon as possible to reduce or avoid potential losses; through timely environmental warnings and rapid responses, aquaculture personnel can better manage the aquaculture environment, improve the survival rate and growth efficiency of aquaculture organisms, and reduce aquaculture risks caused by environmental mutations; the entire process is highly automated, which reduces the frequency and intensity of manual inspections, improves the level of intelligence in aquaculture, and helps promote the modernization of the aquaculture industry. The model is used to calculate the Comprehensive Environmental Index (CEI) to evaluate the suitability of the aquaculture living environment. The first part is the standardization of the jth environmental parameter. This step converts the parameter value to the interval [0,1] to facilitate the comparison and weighting between different parameters. j represents the weight of the jth parameter, reflecting the importance of the parameter to the suitability of the aquaculture environment. The larger the weight, the greater the impact of the parameter on the comprehensive index. j is the exponential coefficient of the jth parameter, which is used to adjust the nonlinear effect of the parameter on the comprehensive index. j The value of can magnify or reduce the impact of parameter changes on the comprehensive index. By adjusting the weights and index coefficients, it can flexibly adapt to different aquaculture environments and fish needs. The formula takes into account multiple environmental parameters and provides a comprehensive assessment framework that can more accurately reflect the overall suitability of the aquaculture environment. The formula allows multiple environmental parameters to be integrated into a single index, which helps to provide a comprehensive and concise environmental status assessment indicator, making it easier for decision makers to quickly understand the overall status of the current environment.

[0103] One embodiment of the present invention provides an aquaculture environment monitoring system, the system comprising:

[0104] The aquaculture environment monitoring module is used to divide the aquaculture environment into a number of spatial areas of equal volume and interconnected, and to deploy sensor groups in the interconnected spatial areas to obtain aquaculture environment parameters;

[0105] A module for predicting aquaculture environmental parameters, which transmits the aquaculture environmental parameters to a central processing system via LoRa, and a built-in prediction model of the central processing system processes the aquaculture environmental parameters;

[0106] The alarm module is used to calculate whether the aquaculture environment prediction value exceeds a preset threshold value through a comprehensive aquaculture environment index model when the aquaculture environment prediction value is obtained after processing. If the preset threshold value is exceeded, the central processing system sends an alarm message to the equipment of the aquaculture personnel.

[0107] The working principle of the above technical solution is as follows: the aquaculture environment is divided into multiple spatial areas of equal volume and interconnected with each other. This division can ensure detailed and comprehensive monitoring of the entire aquaculture environment. In each divided spatial area, a sensor group is deployed to obtain the aquaculture environmental parameters of the area in real time, including temperature, dissolved oxygen, ammonia nitrogen content, etc.; through LoRa (long-range radio) communication technology, the sensor group of each spatial area transmits the collected aquaculture environmental parameters to the central processing system. LoRa technology is known for its low power consumption and long-distance communication capabilities, which is suitable for the data transmission requirements of this distributed monitoring system; after receiving the data, the central processing system uses its built-in prediction model to process these aquaculture environmental parameters. The model predicts the changing trend of the aquaculture environment based on machine learning. Through the comprehensive aquaculture environmental index model, the central processing system will calculate the predicted value of the aquaculture environment and compare it with the preset threshold. If the predicted value exceeds the preset threshold, it means that some environmental parameters may reach or exceed the dangerous level, which may have an adverse impact on aquaculture. In this case, the central processing system will immediately send an alarm message to the equipment of the aquaculture personnel so that they can take quick measures to respond.

[0108] The effects of the above technical solution are as follows: through real-time monitoring and sensor data collection, the system can obtain water quality parameters in real time, and use the prediction model to predict and evaluate the aquaculture environment; the central processing system determines whether the predicted value of the aquaculture environment exceeds the preset threshold according to the preset aquaculture environment index model. Once the predicted value exceeds the threshold, the system will trigger the alarm mechanism; with the help of the analysis of the intelligent prediction model, the system can timely detect abnormal water quality conditions, help aquaculture personnel to respond quickly, and avoid aquaculture losses caused by abnormal water quality; the intelligent monitoring system can provide sufficient water quality data support to help aquaculture personnel better understand the real-time situation of the aquaculture environment; with the help of intelligent prediction models and alarm mechanisms, the system can accurately warn of water quality problems, help avoid unexpected losses and improve aquaculture efficiency; improve monitoring accuracy and system response speed, help to finely manage the aquaculture environment and improve aquaculture benefits; the intelligent aquaculture environment monitoring system can realize real-time monitoring and prediction of water quality environment changes, and can issue alarm information in time, so as to help aquaculture personnel adjust management measures in time, ensure the stability and health of the aquaculture environment, and improve aquaculture efficiency and quality.

[0109] One embodiment of the present invention is a method for monitoring an aquaculture environment, which divides an aquaculture environment into a number of spatial regions of equal volume and interconnected with each other, and respectively deploys sensor groups in the interconnected spatial regions to obtain aquaculture environment parameters, including:

[0110] A sensor group deployment module is used to divide the aquaculture environment into a number of spatial areas of equal volume and interconnected, and respectively deploy sensor groups in the interconnected spatial areas, wherein the sensor groups include a dissolved oxygen sensor, a temperature sensor, a pH sensor, an ammonia nitrogen sensor, a total phosphorus sensor, a total nitrogen sensor, and a turbidity sensor;

[0111] A preprocessing module, used for collecting the environmental factor data measured by the sensor group, and preprocessing the environmental factor data, wherein the preprocessing includes filtering, denoising and smoothing;

[0112] Dynamically adjust sensor weight module, which is used to dynamically adjust the weight of each sensor according to sensor performance evaluation results and environmental difference analysis after preprocessing;

[0113] The module for obtaining aquaculture environmental parameters is used to perform weighted fusion on the sensor data of each area using the adjusted weights to obtain various aquaculture environmental parameters.

[0114] The working principle of the above technical solution is as follows: the aquaculture environment is first finely divided into multiple spatial areas of equal volume and interconnected. In each such spatial area, a set of sensor groups are deployed, including dissolved oxygen sensors, temperature sensors, pH sensors, ammonia nitrogen sensors, total phosphorus sensors, total nitrogen sensors and turbidity sensors. These sensors can comprehensively and real-time monitor the key environmental factors of the area; the sensor group will continuously collect the corresponding environmental factor data, such as dissolved oxygen content, temperature, pH value, etc. Before being transmitted to the central processing system, these data will be pre-processed, including filtering, denoising and smoothing. Filtering can remove high-frequency noise in the data, denoising can further reduce random errors, and smoothing can make the data more stable and continuous; after pre-processing, the system will dynamically adjust the weight of each sensor according to the sensor performance evaluation results and environmental difference analysis. Through dynamic weight adjustment, the system can more accurately reflect the actual environmental conditions of each area; using the adjusted weights, the system will weightedly fuse the sensor data of each area to obtain the comprehensive value of various aquaculture environmental parameters. These comprehensive values ​​not only take into account the raw data of each sensor, but also fully consider the impact of sensor performance and environmental differences on the final results.

[0115] The effects of the above technical solutions are as follows: through preprocessing and dynamic weight adjustment, the quality and accuracy of sensor data can be improved to ensure the reliability of collected data; by adopting weighted fusion algorithm and combining data from different sensors, more comprehensive and comprehensive aquaculture environmental parameters can be obtained to provide more complete information for aquaculture management; according to sensor performance evaluation and environmental difference analysis, the weight of the sensor is dynamically adjusted to improve the adaptability and flexibility of the system; by integrating data from multiple sensors, the system can provide more accurate and complete aquaculture environmental parameters to help aquaculture personnel understand the water quality more comprehensively; dynamic weight adjustment and data fusion improve the intelligence level of the system, better adapt to the characteristics of different areas of the aquaculture environment, and improve the intelligence level of the system; by providing high-quality environmental parameter data and intelligent data processing methods, it can help aquaculture personnel optimize management decisions and improve aquaculture efficiency and production quality; the intelligent aquaculture environment monitoring system realizes more accurate and comprehensive monitoring and evaluation of the aquaculture environment by comprehensively utilizing data from multiple environmental factors, performing data preprocessing and dynamic weight adjustment, which helps to improve the quality of aquaculture water and management efficiency.

[0116] In one embodiment of the present invention, an aquaculture environment monitoring system, the module for dynamically adjusting sensor weights includes:

[0117] The sensor stability acquisition module is used to test the sensor in an environment with constant environmental parameters. The sensor stability is obtained by calculating the standard deviation of the sensor readings over a period of time. Specifically, the sensor stability can be calculated using the following model:

[0118] P s,i (t) = 1-β i (t) / β max

[0119] Among them, P s,i (t) represents the stability of the i-th sensor in time period t, β i (t) is the standard deviation of the i-th sensor reading in time period t, and βmax is the maximum standard deviation among all sensors;

[0120] The module for obtaining sensor performance evaluation results is used to calculate the accuracy evaluation value of the sensor using the mean absolute error, and obtain the sensor performance evaluation result according to the stability and accuracy evaluation value of the sensor. Specifically, the accuracy evaluation value of the sensor can be calculated by the following model:

[0121] P a,i (t) = 1-M i (t) / M max

[0122] Among them, M i (t) is the mean absolute error of the ith sensor in time period t, M max is the preset maximum mean absolute error;

[0123] Specifically, the sensor performance evaluation results are obtained according to the stability and accuracy evaluation values ​​of the sensor through the following model:

[0124] P i (t) = W S *P s,i (t)+W a *P a,i (t)

[0125] Among them, W S and W a is the weight of sensor stability assessment and accuracy assessment, W S +W a =1.

[0126] An environment difference coefficient acquisition module is used to calculate the difference between each environmental factor data and the standard value, standardize the differences to the same scale, and combine multiple environmental factors to calculate the environment difference coefficient;

[0127] Specifically, the environmental difference coefficient is calculated by the following model:

[0128]

[0129] Among them, e tmp,i (t) is the water temperature difference coefficient of the i-th sensor at time t, T i (t) is the water temperature in the area where the i-th sensor is located at time t, Ts is the standard water temperature, T r is the optimal operating temperature range of the sensor. If the optimal operating temperature range of the sensor is 18℃~20℃, then T r =12℃, here, only the calculation model of water temperature difference coefficient is given, the calculation model of difference coefficients such as dissolved oxygen and pH is the same as the calculation model of water temperature difference coefficient.

[0130] Because there are multiple environmental factors to consider, they are combined using a weighted average approach:

[0131] e i (t) = W tmp *e tmp,i (t)+W dio *e dio,i (t)+……

[0132] Among them, e i (t) is the environmental difference coefficient of the i sensors at time t, e dio,i (t) is the dissolved oxygen difference coefficient of the ith sensor at time t, W tmp is the weight of the environmental factor water temperature, W dio is the weight of the environmental factor dissolved oxygen.

[0133] The sensor weight module is used to obtain the sensor weight according to the sensor performance evaluation result and the environmental difference coefficient. Specifically, the weight calculation model is as follows:

[0134]

[0135] Among them, W i (t) is the weight of the i-th sensor at time t, P i (t) is the performance evaluation result of the i-th sensor at time t, N is the total number of sensors of this type in the entire aquaculture environment, and e i (t) is the environmental difference coefficient of the ith sensor at time t.

[0136] The working principle and effect of the above technical solution are as follows: the sensor is placed in an environment where the environmental parameters are constant for testing. The stability of the sensor is evaluated by recording the sensor readings over a period of time and calculating the standard deviation of these readings. A small standard deviation means that the output of the sensor is relatively stable and has small fluctuations. The mean absolute error is used to calculate the accuracy evaluation value of the sensor. By comparing the difference between the sensor reading and the true value or standard value, a numerical value representing the accuracy of the sensor can be obtained. A smaller mean absolute error value indicates that the sensor has higher accuracy. For each environmental factor data (such as temperature, dissolved oxygen, etc.), the difference between it and the standard value or ideal value is calculated. Then, these differences are standardized to the same scale so that multiple environmental factors can be compared and combined. Finally, multiple environmental factors are combined to calculate a comprehensive environmental difference coefficient, which is used to represent the degree of difference between the current environment and the standard or ideal environment. According to the stability evaluation results and accuracy evaluation values ​​of the sensor, as well as the environmental difference coefficient, the weight of the sensor can be comprehensively obtained. The weight reflects the importance and reliability of the sensor in the current environment and can be used in subsequent sensor data fusion or decision-making processes. By comprehensively considering the stability and accuracy of the sensor and the impact of environmental factors on the performance of the sensor, the performance of the sensor can be evaluated more accurately. This evaluation method is more objective and comprehensive, avoiding the one-sidedness of single indicator evaluation; based on the performance evaluation results of the sensor and the environmental difference coefficient, the sensor can be selected and used more specifically. In a complex and changing environment, the weight of the sensor can be adjusted according to actual needs to improve the performance and reliability of the entire system; by calculating the environmental difference coefficient, the degree of influence of different environmental factors on sensor performance can be quantified. This enables the sensor to better adapt to different working environments and improve its accuracy and stability under various conditions.

[0137] The formulas for performance evaluation, accuracy evaluation, and comprehensive performance evaluation are designed to quantify the performance of sensors, so as to facilitate the comparison and selection of sensors. These formulas provide an objective and comparable method to measure the performance of sensors, avoiding subjective assumptions and uncertainties. Specifically, the design idea of ​​the stability evaluation and accuracy evaluation formulas is to measure the performance of sensors based on the difference between the sensor reading and the ideal value or reference value. By standardizing the difference (using the mean absolute error) and comparing the result with the maximum possible difference, an evaluation value between 0 and 1 can be obtained. This design makes the evaluation result more intuitive and easy to understand, and also facilitates the comparison of the performance of different sensors; the comprehensive performance evaluation formula is designed to combine multiple performance indicators to form a comprehensive evaluation result. By using methods such as weighted average, different performance indicators can be given different weights according to actual needs, so as to obtain a comprehensive evaluation value that is more in line with actual needs. Such a design can more comprehensively reflect the performance of the sensor and avoid the one-sidedness of a single indicator evaluation. The calculation of the water temperature environmental difference coefficient and the comprehensive environmental difference coefficient is to quantify the difference between the environmental conditions in the area where the sensor is located and the standard conditions or ideal conditions. This difference may affect the performance of the sensor, so it needs to be included in the performance evaluation. By calculating the coefficient of difference, a relatively objective measure can be obtained to compare the performance of different sensors in different environments. Such calculations can more accurately reflect the performance of sensors in actual applications and provide a more reliable basis for the selection and use of sensors. These formulas provide a standardized and comparable method to measure the performance of sensors, avoiding subjective assumptions and uncertainties, making the evaluation results more accurate and reliable. The function for calculating sensor weights ensures that the sum of all weights is 1, and the output can be interpreted as a probability distribution, which is very suitable for multi-classification tasks; through the exponential function, the function amplifies the difference between different input values, so that relatively large input values ​​correspond to larger weights, thereby highlighting important features; the function is continuous and differentiable, which makes it possible to use it in gradient-based optimization algorithms; P i (t) as the numerator, which means that sensors with better performance will receive a greater weight; This part uses the environmental difference coefficient ei(t) to adjust the weight, but avoids the problems that may be caused by directly using its reciprocal (such as the weight becoming too large when ei(t) is close to zero). By adding 1 and then taking the reciprocal, it ensures that when the environmental difference increases, the weight will decrease smoothly instead of suddenly becoming very large or very small.

[0138] In one embodiment of the present invention, an aquaculture environment monitoring system, a module for predicting aquaculture environment parameters includes:

[0139] A preliminary processing module, used for the central processing system to obtain historical data of the sensor group, and to perform preliminary processing on the historical data of the sensor group, wherein the preliminary processing includes data cleaning, data transformation and feature extraction, wherein the feature extraction includes extracting statistical features and time series features from the original data, and after the preliminary processing, the data is divided into a training set and a test set;

[0140] A training model module is used to construct an aquaculture environment prediction model through a long short-term memory network (LSTM), and train the aquaculture environment prediction model using a training set in the preliminarily processed data;

[0141] A testing module, used to evaluate and verify the trained aquaculture environment prediction model using a test set in the preliminarily processed data;

[0142] The module for obtaining the predicted value of the aquaculture environment is used to input the aquaculture environment parameters into the aquaculture environment prediction model to obtain the predicted value of the aquaculture environment.

[0143] The working principle and effect of the above technical solution are as follows: the central processing system first obtains historical data from the sensor group, which may contain noise and outliers, so preliminary processing is required. Preliminary processing includes data cleaning, data transformation and feature extraction. Data cleaning aims to remove noise and outliers, and data transformation converts data into a format suitable for model processing, such as standardization or normalization. Feature extraction is to extract meaningful statistical features and time series features from the original data, which will be used for model training and evaluation; the data after preliminary processing is divided into training set and test set. The training set is used for model training, while the test set is used to evaluate the performance of the model; a long short-term memory network (LSTM) is used to build an aquaculture environment prediction model. LSTM is a recurrent neural network suitable for processing time series data. The model is trained using the training set, and the model parameters are adjusted through the optimization algorithm to minimize the difference between the predicted value and the actual value; the trained model is evaluated and verified using the test set. Evaluation indicators may include precision, recall rate, F1 value, etc., which are used to measure the performance of the model on unseen data. The validation process ensures that the model has good generalization ability and can provide accurate predictions in practical applications; the actual aquaculture environmental parameters are input into the trained model to obtain the predicted values ​​of the aquaculture environment. These predicted values ​​can provide farmers with information about future environmental conditions and help them make reasonable farming decisions. Through preliminary processing and data set division, the quality of data input to the model can be ensured to be high, thereby improving the prediction accuracy of the model. The LSTM model can capture long-term dependencies in time series data, further improving the accuracy of the prediction; accurate environmental predictions can help farmers understand future changes in environmental conditions in advance, thereby optimizing farming strategies, reducing risks and increasing production. This is of great significance to the sustainable development of the aquaculture industry; promoting intelligent farming, by combining sensor technology with machine learning models, intelligent monitoring and prediction of the aquaculture environment are realized, which helps to promote the modernization and intelligentization of the aquaculture industry and improve production efficiency and management level.

[0144] In one embodiment of the present invention, an aquaculture environment monitoring system, the alarm module comprises:

[0145] The module for calculating whether the preset threshold is exceeded is used to input the predicted value of the aquatic environment into the aquaculture environment index model, and the aquaculture environment model calculates whether the predicted value of the aquatic environment exceeds the preset threshold. Specifically, the aquaculture environment model is

[0146]

[0147] Where CEI is the comprehensive environmental index, n is the total number of environmental parameters, and w jis the weight of the jth parameter, indicating the importance of this parameter to the suitability of the aquaculture living environment, x j is the predicted value of the jth parameter, x min ≤x j ≤x max , x min and x max are respectively the minimum and maximum possible values ​​of the parameter suitable for fish life, and ej is the exponential coefficient of the jth parameter, which is used to adjust the influence of the parameter on the comprehensive index.

[0148] The module for sending alarm information is used for the central processing system to send an alarm information to the mobile device of the aquaculture personnel if the predicted value of the aquaculture environment exceeds the preset comprehensive environmental index threshold.

[0149] The working principle and effect of the above technical solution are as follows: the predicted aquatic environment value obtained by the LSTM model is input into the aquaculture environment index model. This environmental index model is a mathematical model that integrates multiple environmental parameters and is used to evaluate the overall condition of the aquaculture environment. The aquaculture environment model calculates whether the predicted aquatic environment value exceeds these thresholds based on the preset environmental parameter thresholds. These thresholds are usually set based on the growth requirements, environmental tolerance and industry standards of aquaculture organisms. If the predicted aquatic environment value exceeds the preset threshold, it means that the current or upcoming environmental conditions may have an adverse effect on aquaculture organisms. At this time, the central processing system triggers an alarm mechanism. The central processing system sends a message to the connected mobile network or other communication methods. The mobile device of the aquaculture personnel sends an alarm message; by inputting the predicted value of the aquatic environment into the environmental index model in real time, the system can continuously monitor the status of the aquaculture environment and issue a timely warning when unfavorable conditions are predicted; since the alarm information is sent directly to the mobile device of the aquaculture personnel, this ensures that the aquaculture personnel can be quickly notified of environmental abnormalities, so that they can take necessary measures as soon as possible to reduce or avoid potential losses; through timely environmental warnings and rapid responses, aquaculture personnel can better manage the aquaculture environment, improve the survival rate and growth efficiency of aquaculture organisms, and reduce aquaculture risks caused by environmental mutations; the entire process is highly automated, which reduces the frequency and intensity of manual inspections, improves the level of intelligence in aquaculture, and helps promote the modernization of the aquaculture industry. The model is used to calculate the Comprehensive Environmental Index (CEI) to evaluate the suitability of the aquaculture living environment. The first part is the standardization of the jth environmental parameter. This step converts the parameter value to the interval [0,1] to facilitate the comparison and weighting between different parameters. j represents the weight of the jth parameter, reflecting the importance of the parameter to the suitability of the aquaculture environment. The larger the weight, the greater the impact of the parameter on the comprehensive index.j is the exponential coefficient of the jth parameter, which is used to adjust the nonlinear effect of the parameter on the comprehensive index. j The value of can magnify or reduce the impact of parameter changes on the comprehensive index. By adjusting the weights and index coefficients, it can flexibly adapt to different aquaculture environments and fish needs. The formula takes into account multiple environmental parameters and provides a comprehensive assessment framework that can more accurately reflect the overall suitability of the aquaculture environment. The formula allows multiple environmental parameters to be integrated into a single index, which helps to provide a comprehensive and concise environmental status assessment indicator, making it easier for decision makers to quickly understand the overall status of the current environment.

[0150] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for monitoring an aquaculture environment, characterized in that: The method comprises: The aquaculture environment is divided into a number of spatial areas of equal volume and interconnected, and sensor groups are respectively deployed in the interconnected spatial areas to obtain aquaculture environment parameters; The aquaculture environment parameters are transmitted to a central processing system via LoRa, and a built-in prediction model of the central processing system processes the aquaculture environment parameters; After processing, the predicted value of the aquaculture environment is obtained, and the comprehensive aquaculture environment index model is used to calculate whether the predicted value of the aquaculture environment exceeds a preset threshold. If it exceeds the preset threshold, the central processing system sends an alarm message to the equipment of the aquaculture personnel.

2. The aquaculture environment monitoring method according to claim 1, characterized in that: The aquaculture environment is divided into several interconnected spatial areas of equal volume, and sensor groups are deployed in the interconnected spatial areas to obtain aquaculture environment parameters, including: The aquaculture environment is divided into a number of spatial areas of equal volume and interconnected with each other, and sensor groups are respectively deployed in the interconnected spatial areas, wherein the sensor groups include a dissolved oxygen sensor, a temperature sensor, a pH sensor, an ammonia nitrogen sensor, a total phosphorus sensor, a total nitrogen sensor and a turbidity sensor; Collecting environmental factor data measured by the sensor group, and preprocessing the environmental factor data, wherein the preprocessing includes filtering, denoising and smoothing; After preprocessing, the weight of each sensor is dynamically adjusted based on the sensor performance evaluation results and environmental difference analysis; The sensor data from each area are weightedly fused using the adjusted weights to obtain various aquaculture environmental parameters.

3. The aquaculture environment monitoring method according to claim 2, characterized in that: Based on the sensor performance evaluation results and environmental difference analysis, the weight of each sensor is dynamically adjusted, including: The sensor is placed in an environment with constant environmental parameters for testing, and the stability of the sensor is obtained by calculating the standard deviation of the sensor readings over a period of time; The mean absolute error is used to calculate the accuracy evaluation value of the sensor, and the sensor performance evaluation result is obtained according to the stability and accuracy evaluation value of the sensor; For each of the environmental factor data, the difference between the data and the standard value is calculated, and the differences are standardized to the same scale, and the environmental difference coefficient is calculated by combining multiple environmental factors; The weight of the sensor is obtained according to the sensor performance evaluation results and the environmental difference coefficient.

4. The aquaculture environment monitoring method according to claim 1, characterized in that: The aquaculture environment parameters are transmitted to a central processing system via LoRa, and the prediction model built into the central processing system processes the aquaculture environment parameters, including: The central processing system acquires historical data of the sensor group, and performs preliminary processing on the historical data of the sensor group, wherein the preliminary processing includes data cleaning, data transformation and feature extraction, wherein the feature extraction includes extracting statistical features and time series features from the original data, and after the preliminary processing, the data is divided into a training set and a test set; Building an aquaculture environment prediction model through a long short-term memory network (LSTM), and training the aquaculture environment prediction model using a training set in the preliminarily processed data; Using the test set in the initially processed data to evaluate and verify the trained aquaculture environment prediction model; The aquaculture environment parameters are input into the aquaculture environment prediction model to obtain an aquaculture environment prediction value.

5. The aquaculture environment monitoring method according to claim 1, characterized in that: After processing, the predicted value of the aquaculture environment is obtained, and the predicted value of the aquaculture environment is calculated by the comprehensive aquaculture environment index model to determine whether it exceeds a preset threshold. If it exceeds the preset threshold, the central processing system sends an alarm message to the equipment of the aquaculture personnel, including: Inputting the aquatic environment prediction value into the aquaculture environment index model, and the aquaculture environment model calculates whether the aquatic environment prediction value exceeds a preset threshold; If the predicted value of the aquaculture environment exceeds a preset threshold, the central processing system sends an alarm message to the mobile device of the aquaculture personnel.

6. An aquaculture environment monitoring system, characterized in that: The system comprises: The aquaculture environment monitoring module is used to divide the aquaculture environment into a number of spatial areas of equal volume and interconnected, and to deploy sensor groups in the interconnected spatial areas to obtain aquaculture environment parameters; A module for predicting aquaculture environmental parameters, used for transmitting the aquaculture environmental parameters to a central processing system via LoRa, and a prediction model built into the central processing system processes the aquaculture environmental parameters; The alarm module is used to obtain the predicted value of the aquaculture environment after processing, and calculate whether the predicted value of the aquaculture environment exceeds a preset threshold through a comprehensive aquaculture environment index model. If it exceeds the preset threshold, the central processing system sends an alarm message to the equipment of the aquaculture personnel.

7. An aquaculture environment monitoring system according to claim 6, characterized in that: The aquaculture environment monitoring module includes: A sensor group deployment module is used to divide the aquaculture environment into a number of spatial areas of equal volume and interconnected, and respectively deploy sensor groups in the interconnected spatial areas, wherein the sensor groups include a dissolved oxygen sensor, a temperature sensor, a pH sensor, an ammonia nitrogen sensor, a total phosphorus sensor, a total nitrogen sensor, and a turbidity sensor; A preprocessing module, used for collecting the environmental factor data measured by the sensor group, and preprocessing the environmental factor data, wherein the preprocessing includes filtering, denoising and smoothing; Dynamically adjust sensor weight module, which is used to dynamically adjust the weight of each sensor according to sensor performance evaluation results and environmental difference analysis after preprocessing; The module for obtaining aquaculture environmental parameters is used to perform weighted fusion on the sensor data of each area using the adjusted weights to obtain various aquaculture environmental parameters.

8. An aquaculture environment monitoring system according to claim 7, characterized in that: The module for dynamically adjusting sensor weights includes: The sensor stability acquisition module is used to test the sensor in an environment with constant environmental parameters and obtain the sensor stability by calculating the standard deviation of the sensor readings over a period of time; A module for obtaining sensor performance evaluation results is used to calculate the accuracy evaluation value of the sensor using the mean absolute error, and obtain the sensor performance evaluation result according to the stability and accuracy evaluation value of the sensor; An environment difference coefficient acquisition module is used to calculate the difference between each environmental factor data and the standard value, standardize these differences to the same scale, and combine multiple environmental factors to calculate the environment difference coefficient; The sensor weight obtaining module is used to obtain the sensor weight according to the sensor performance evaluation result and the environmental difference coefficient.

9. An aquaculture environment monitoring system according to claim 6, characterized in that: The module for predicting aquaculture environmental parameters comprises: A preliminary processing module, used for the central processing system to obtain historical data of the sensor group, and to perform preliminary processing on the historical data of the sensor group, wherein the preliminary processing includes data cleaning, data transformation and feature extraction, wherein the feature extraction includes extracting statistical features and time series features from the original data, and after the preliminary processing, the data is divided into a training set and a test set; A training model module is used to construct an aquaculture environment prediction model through a long short-term memory network (LSTM), and train the aquaculture environment prediction model using a training set in the preliminarily processed data; A testing module, used to evaluate and verify the trained aquaculture environment prediction model using a test set in the preliminarily processed data; The module for obtaining the predicted value of the aquaculture environment is used to input the aquaculture environment parameters into the aquaculture environment prediction model to obtain the predicted value of the aquaculture environment.

10. An aquaculture environment monitoring system according to claim 6, characterized in that: The alarm module comprises: A module for calculating whether the value exceeds a preset threshold value is used to input the predicted value of the aquaculture environment into the aquaculture environment index model, and the aquaculture environment model calculates whether the predicted value of the aquaculture environment exceeds the preset threshold value; The module for sending alarm information is used for the central processing system to send alarm information to the mobile device of the aquaculture personnel if the predicted value of the aquaculture environment exceeds a preset threshold.

Images