Penaeus vannamei pond water quality comprehensive evaluation method based on deep learning

Abstract

The invention discloses a comprehensive evaluation method for water quality of a penaeus vannamei pond based on deep learning. The method comprises the following steps: firstly, acquiring multi-dimensional water quality monitoring data of a penaeus vannamei aquaculture pond; carrying out preprocessing operation on the obtained original data; then, stationary wavelet transform is executed on the processed water quality time sequence data, multi-scale characteristic decomposition and denoising are achieved, and low-frequency trend characteristics and high-frequency disturbance characteristics of water quality signals are extracted; and finally, inputting the features subjected to wavelet decomposition into a pre-trained deep learning model, extracting spatial features through a convolutional neural network, modeling a time dependency relationship through a long-term and short-term memory network, and obtaining five types of water quality grade evaluation results through a Softmax output layer. Based on the method, automatic, refined and real-time comprehensive evaluation of the water quality of the aquaculture pond can be achieved, the accuracy and stability of water quality evaluation are improved, and effective technical support is provided for intelligent monitoring and health management of the aquaculture environment of the penaeus vannamei boone.

Description

Technical Field

[0001] This invention belongs to the field of intelligent monitoring technology for aquaculture, and in particular relates to a comprehensive evaluation method for pond water quality of Litopenaeus vannamei based on deep learning. Background Technology

[0002] The Pacific white shrimp (Litopenaeus vannamei) is currently the most widely farmed economically important shrimp species globally. Its growth rate, immunity, and survival rate are all closely related to the water quality of the farming ponds. Fluctuations in water quality not only affect individual growth and feed conversion rates but also directly impact ecosystem stability and economic benefits. When key indicators such as dissolved oxygen (DO), pH, and ammonia nitrogen (NH₃-N) become abnormal, they can easily trigger stress responses, disease outbreaks, and even mass mortality.

[0003] Traditional water quality testing methods mainly rely on manual sampling and laboratory analysis, which are cumbersome, lack real-time performance, and are difficult to implement for continuous dynamic monitoring. While some statistical or shallow machine learning models can reflect water quality changes to a certain extent, their feature extraction capabilities are limited in multivariate, nonlinear, and time-series coupled data environments. Although single deep learning models such as CNN or LSTM have improved accuracy, they still suffer from problems such as insufficient multi-scale feature fusion, inadequate time-dependent modeling, and lack of frequency domain information.

[0004] Therefore, there is an urgent need for an intelligent water quality assessment method that can integrate multi-dimensional features of time, space and frequency to achieve accurate, real-time and robust assessment of water quality in Litopenaeus vannamei ponds. Summary of the Invention

[0005] In view of this, the present invention aims to solve the problems of poor real-time performance, insufficient feature extraction and low accuracy in existing aquaculture water quality assessment, and proposes a comprehensive evaluation method for pond water quality of Litopenaeus vannamei based on deep learning.

[0006] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0007] In a first aspect, the present invention provides a method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning, comprising the following steps:

[0008] Step 1: Obtain water quality monitoring data from Litopenaeus vannamei (whiteleg shrimp) farming ponds;

[0009] Step 2: Preprocess and serialize the acquired water quality monitoring data to obtain water quality time series data;

[0010] Step 3: Perform stationary wavelet transform on the water quality time series data to achieve multi-scale decomposition and extract multi-scale features, and at the same time, perform noise reduction and normalization on the decomposition coefficients.

[0011] Step 4: Input the data from Step 2 and Step 3 into the pre-trained SLCI model for detection and classification to obtain the water quality assessment results of the aquaculture pond.

[0012] Furthermore, in step 1, water quality data is collected in real time by deploying online sensors in the pond and uploaded at fixed sampling intervals. The water quality data includes water temperature, pH, dissolved oxygen, ammonia nitrogen, and nitrite.

[0013] Furthermore, step 2 includes:

[0014] Step 2.1: The original water quality data obtained in Step 1 is processed for missing values. A local neighborhood mean imputation strategy is adopted, and random perturbations are introduced into the imputed values.

[0015] Step 2.2: Perform dimensionless processing on each water quality characteristic and use Min-Max normalization to map the values ​​to the [0,1] interval;

[0016] Step 2.3: Construct time series samples based on the sliding window method, using a window length L, and combine L consecutive sampling points into a training sample;

[0017] Step 2.4: Use RandomOverSampler to oversample the minority class samples and add small random noise when copying samples.

[0018] Furthermore, step 3 includes:

[0019] Step 3.1: Perform stationary wavelet transform on each dimension of the water quality parameter sequence in the water quality time series data sample generated in Step 2;

[0020] Step 3.2: The stationary wavelet transform is decomposed using the Symlet4 wavelet basis;

[0021] Step 3.3: Extract and concatenate the low-frequency approximation coefficients and high-frequency detail coefficients obtained from each decomposition layer to form a multi-scale feature vector;

[0022] Step 3.4: Apply threshold denoising to the scale coefficients obtained from the decomposition to filter out noise components, and normalize the denoising coefficients.

[0023] Step 3.5: Align the processed multi-scale coefficient matrix with the original normalized time series data according to the feature dimension, and merge them to form the input feature matrix of the SLCI model.

[0024] Furthermore, step 4 includes:

[0025] Step 4.1: Build and train the SLCI deep learning model. Input the preprocessed water quality time series data and multi-scale features obtained in Step 2 and Step 3 into the model for training and testing to obtain the trained classification model.

[0026] Step 4.2: Use the trained SLCI model to evaluate the real-time or offline water quality data of the pond to be evaluated and output the water quality level.

[0027] Furthermore, the training of the SLCI model includes:

[0028] The multi-scale features decomposed by stationary wavelet transform are used as the input feature matrix and input into the one-dimensional convolutional neural network module. The one-dimensional convolutional neural network module extracts the local spatial features between water quality indicators through two convolutional layers. The first layer is used to capture the low-level change pattern, and the second layer further extracts the high-level semantic features. Batch normalization and Dropout are followed by the convolutional layers.

[0029] Spatial features extracted by a one-dimensional convolutional neural network are input into a long short-term memory network module to achieve dynamic modeling of time series. The long short-term memory network module consists of two stacked layers: the first layer learns short-term dependency information, and the second layer extracts long-term time series features.

[0030] The temporal features output by the Long Short-Term Memory (LSTM) network module are input into a fully connected layer for mapping, and the probability distribution of the five water quality levels is output through the Softmax activation function.

[0031] Furthermore, the model ultimately determines the corresponding water quality level based on the maximum probability value, thereby achieving a comprehensive evaluation of the water quality in Litopenaeus vannamei farming ponds.

[0032] Secondly, the present invention provides a deep learning-based device for comprehensive water quality evaluation in Litopenaeus vannamei ponds, comprising:

[0033] The data acquisition unit is used to acquire water quality monitoring data from Litopenaeus vannamei ponds.

[0034] The data processing unit is used to preprocess and serialize the acquired water quality monitoring data to obtain water quality time series data.

[0035] The wavelet transform unit is used to perform stationary wavelet transform on water quality time series data to achieve multi-scale decomposition and extract multi-scale features, while performing denoising and normalization on the decomposition coefficients.

[0036] The detection result unit is used to input the data from the data processing unit and the wavelet transform unit into the pre-trained SLCI model for detection and classification, and to obtain the water quality level assessment results of the aquaculture pond.

[0037] Thirdly, the present invention provides an electronic device, including a processor and a memory connected in communication with the processor and used to store executable instructions of the processor, wherein the processor is used to execute the above-mentioned method for comprehensive evaluation of pond water quality of Litopenaeus vannamei based on deep learning.

[0038] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the aforementioned method for comprehensive evaluation of pond water quality of Litopenaeus vannamei based on deep learning.

[0039] Compared with existing technologies, the deep learning-based comprehensive water quality evaluation method for Litopenaeus vannamei ponds described in this invention has the following advantages:

[0040] The SLCI model of this invention combines stationary wavelet transform (SWT), convolutional neural network (CNN), and long short-term memory network (LSTM) to achieve deep fusion of temporal, frequency, and spatial features, thereby significantly improving the accuracy and robustness of water quality classification and anomaly detection.

[0041] This invention addresses the problem of imbalance in water quality grade categories by employing RandomOverSampler to oversample minority class samples and adding small random noise when replicating samples to avoid overfitting and improve the model's generalization ability on the minority class.

[0042] This invention utilizes a CNN module to extract local spatial features between water quality indicators through a two-layer convolutional structure. The first layer is used to capture low-level change patterns, and the second layer further extracts high-level semantic features. Batch normalization and dropout are applied after the convolutional layers to improve the generalization performance of the model and prevent overfitting.

[0043] This invention inputs the spatial features extracted by CNN into the Long Short-Term Memory (LSTM) network module, which enhances the model's ability to capture the patterns of non-stationary water quality changes.

[0044] Experimental results show that the present invention exhibits good generalization ability. Attached Figure Description

[0045] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0046] Figure 1 The diagram shows the flowchart of the deep learning-based comprehensive water quality evaluation method for Litopenaeus vannamei ponds according to the present invention.

[0047] Figure 2The diagram shown is a schematic of the SLCI model of the present invention;

[0048] Figure 3 The diagram shown is a schematic representation of the comprehensive evaluation results of pond water quality according to the present invention. Detailed Implementation

[0049] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0050] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0051] Example 1

[0052] See Figure 1 This invention provides a deep learning-based method for comprehensive evaluation of pond water quality for Litopenaeus vannamei, comprising the following steps:

[0053] Step 1: Obtain water quality monitoring data (including but not limited to water temperature, pH, dissolved oxygen, ammonia nitrogen, nitrite and other multi-dimensional water quality parameters) from the Litopenaeus vannamei farming ponds.

[0054] Step 2: Perform data preprocessing and serialization on the raw pond water quality data obtained in Step 1.

[0055] Specifically, step 2 includes:

[0056] S2.1: The original water quality data obtained in step 1 is processed for missing values. A local neighborhood mean imputation strategy is adopted (i.e., the mean of the effective values ​​at several times above and below the missing value is taken) and random perturbation ∈ ~U(-0.1σ,0.1σ) is introduced into the imputation value to enhance robustness.

[0057] S2.2: Dimensional elimination processing is performed on each water quality characteristic, and Min-Max normalization is used to map the values ​​to the [0,1] interval;

[0058] Specifically, the normalization formula is as follows:

[0059]

[0060] Where x is the original eigenvalue, x min and x max These are the minimum and maximum values ​​of the feature column, X. norm These are the normalized eigenvalues.

[0061] S2.3: Construct time-series samples based on the sliding window method, and use a window length L (preferably L=10) to combine L consecutive sampling points into a training sample to preserve time dependence;

[0062] Specifically, let the original data matrix be X∈R T×F If the window length is L = 10, then the data reshaped by the sliding window is:

[0063] X resgape ∈R (T-w+1)×w×F

[0064] Where T is the length of the original sequence, w is the length of the sliding window, F is the feature dimension at each time step, and X... reshape This represents the three-dimensional feature tensor consisting of (T-w+1) samples obtained after reshaping through a sliding window.

[0065] Each sample segment x i =[X i ,X i+1 ,...,X i+w-1 ]∈R w×F This method can reflect the dynamic changes over a continuous time period. It preserves the time dependency and provides richer input information for deep learning models.

[0066] S2.4: To address the imbalance in water quality class categories, RandomOverSampler is used to oversample minority class samples, and small random noise is added when replicating samples to avoid overfitting and improve the model's generalization ability on the minority class.

[0067] Specifically, a small amount of random noise ε from a uniform distribution is added to each eigenvalue, which can be expressed mathematically as:

[0068] x′ ij =x ij +ε ij ,ε ij ~U(-δ,δ)

[0069] Where, x ij ε represents the j-th feature value of the i-th sample in the original sample. ij The disturbance term follows a uniform distribution U(-δ,δ), where δ is the upper limit of the noise amplitude.

[0070] Step 3: Perform Stationary Wavelet Transform (SWT) on the collected water quality time-series data to achieve multi-scale decomposition and extract multi-scale features. Simultaneously, denoise and normalize the decomposition coefficients. Specifically, this includes:

[0071] S3.1: Perform stationary wavelet transform (SWT) on each dimension of the water quality parameter sequence in the time series sample generated in step 2;

[0072] S3.2: The SWT is decomposed using the Symlet4 (sym4) wavelet basis, and the preferred number of decomposition layers is 2.

[0073] Specifically, the formula is as follows:

[0074]

[0075] Where, φ j,k (n) is the scaling function of the wavelet basis function at scale j and position k.

[0076]

[0077] Where, ψ j,k (n) is the wavelet basis function at scale j and position k.

[0078] S3.3: Extract and concatenate the low-frequency approximation coefficients (A, representing low-frequency trend features) and high-frequency detail coefficients (D, representing short-term perturbation features) obtained from each decomposition layer to form a multi-scale feature vector;

[0079] S3.4: Threshold denoising is applied to the scale coefficients obtained from the decomposition to filter out noise components, and the denoised coefficients are normalized to facilitate subsequent model training.

[0080] S3.5: Align the processed multi-scale coefficient matrix with the original normalized time series data according to the feature dimension and merge them to form the input feature matrix of the SLCI model.

[0081] Step 4: Input the preprocessed data and multi-scale features obtained in Steps 2 and 3 into the pre-trained SLCI (Stationary Wavelet–LSTM–CNNIntegrator) model for detection and classification to obtain the water quality assessment results for the aquaculture pond. Specifically, this includes:

[0082] S4.1: Build and train the SLCI (Stationary Wavelet–LSTM–CNNIntegrator) deep learning model. Input the preprocessed time-series data and multi-scale features obtained in steps 2 and 3 into the model for training and testing to obtain the trained classification model; the SLCI model diagram is shown below. Figure 2 As shown;

[0083] S4.2: Use the trained SLCI model to evaluate the real-time or offline water quality data of the pond to be evaluated and output the water quality level.

[0084] Specifically, the training process of the SLCI model includes:

[0085] S4.2.1: The multi-scale features decomposed by Stationary Wavelet Transform (SWT) are used as the input feature matrix and fed into a one-dimensional convolutional neural network (CNN) module. The CNN module extracts local spatial features between water quality indicators through a two-layer convolutional structure. The first layer captures low-level variation patterns, and the second layer further extracts high-level semantic features. Batch normalization and dropout are applied after the convolutional layers to improve the model's generalization performance and prevent overfitting.

[0086] S4.2.2: The spatial features extracted by the CNN are input into the Long Short-Term Memory (LSTM) network module to achieve dynamic modeling of time series. The LSTM module consists of two stacked layers. The first layer mainly learns short-term dependency information, and the second layer extracts long-term time series features, thereby enhancing the model's ability to capture non-stationary water quality change patterns.

[0087] S4.2.3: The temporal features output by the LSTM are input into a fully connected layer for mapping, and the probability distribution of five water quality levels is output through the Softmax activation function. The model finally determines the corresponding water quality level based on the maximum probability value, thus achieving a comprehensive evaluation of the water quality in Litopenaeus vannamei (whiteleg shrimp) farming ponds.

[0088] The detection results of the deep learning-based comprehensive water quality evaluation method for Litopenaeus vannamei ponds proposed in this invention are shown in [reference needed]. Figure 3 .

[0089] This invention was tested and verified on 12,226 automatically monitored samples collected from a Litopenaeus vannamei farming base in Wuqing District, Tianjin. The results showed that the SLCI model had a classification accuracy of 0.935, a precision of 0.962, a recall of 0.939, and an F1-score of 0.948, demonstrating good generalization ability.

[0090] Example 2

[0091] A deep learning-based comprehensive water quality assessment device for Litopenaeus vannamei ponds includes:

[0092] The data acquisition unit is used to acquire water quality monitoring data from Litopenaeus vannamei ponds.

[0093] The data processing unit is used to preprocess and serialize the acquired water quality monitoring data to obtain water quality time series data.

[0094] The wavelet transform unit is used to perform stationary wavelet transform on water quality time series data to achieve multi-scale decomposition and extract multi-scale features, while performing denoising and normalization on the decomposition coefficients.

[0095] The detection result unit is used to input the data from the data processing unit and the wavelet transform unit into the pre-trained SLCI model for detection and classification, and to obtain the water quality level assessment results of the aquaculture pond.

[0096] Example 3

[0097] An electronic device includes a processor and a memory communicatively connected to the processor and used to store processor-executable instructions, the processor being used to execute the aforementioned deep learning-based comprehensive water quality evaluation method for Litopenaeus vannamei ponds.

[0098] Example 4

[0099] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the aforementioned deep learning-based comprehensive water quality evaluation method for Litopenaeus vannamei ponds.

[0100] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in this disclosure all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0101] It should be noted that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should only be conducted after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy and procedures.

[0102] This disclosure is intended to provide implementation schemes for users to selectively prevent the use or access to their personal information data. Specifically, this disclosure is intended to provide hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, risks can be minimized by restricting data collection and deleting data. Furthermore, where applicable, such personal information is de-identified to protect user privacy.

[0103] The acquisition, transmission, storage, use, and processing of data in this disclosed technical solution all comply with the relevant provisions of national laws and regulations.

[0104] It should be noted that in the embodiments disclosed herein, certain software, components, models, and other existing solutions in the industry may be mentioned. These should be considered as exemplary and are intended only to illustrate the feasibility of implementing the technical solution of this application. However, they do not mean that the applicant has used or necessarily used such solutions.

[0105] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0106] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0107] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of this disclosure pertain.

[0108] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). In addition, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning paper or other media, followed by editing, interpreting or otherwise processing as necessary, and then stored in computer memory.

[0109] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0110] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0111] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0112] The storage medium mentioned above may be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A comprehensive water quality evaluation method for Litopenaeus vannamei ponds based on deep learning, characterized in that: Includes the following steps: Step 1: Obtain water quality monitoring data from Litopenaeus vannamei (whiteleg shrimp) farming ponds; Step 2: Preprocess and serialize the acquired water quality monitoring data to obtain water quality time series data; Step 3: Perform stationary wavelet transform on the water quality time series data to achieve multi-scale decomposition and extract multi-scale features, and at the same time, perform noise reduction and normalization on the decomposition coefficients. Step 4: Input the data from Step 2 and Step 3 into the pre-trained SLCI model for detection and classification to obtain the water quality assessment results of the aquaculture pond.

2. The method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning according to claim 1, characterized in that: In step 1, water quality data is collected in real time by deploying online sensors in the pond and uploaded at fixed sampling intervals. The water quality data includes water temperature, pH, dissolved oxygen, ammonia nitrogen, and nitrite.

3. The method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning as described in claim 1, characterized in that: Step 2 includes: Step 2.1: The original water quality data obtained in Step 1 is processed for missing values. A local neighborhood mean imputation strategy is adopted, and random perturbations are introduced into the imputed values. Step 2.2: Perform dimensionless processing on each water quality characteristic and use Min-Max normalization to map the values ​​to the [0,1] interval; Step 2.3: Construct time series samples based on the sliding window method, using a window length L, and combine L consecutive sampling points into a training sample; Step 2.4: Use RandomOverSampler to oversample the minority class samples and add small random noise when copying samples.

4. The method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning as described in claim 1, characterized in that: Step 3 includes: Step 3.1: Perform stationary wavelet transform on each dimension of the water quality parameter sequence in the water quality time series data sample generated in Step 2; Step 3.2: The stationary wavelet transform is decomposed using the Symlet4 wavelet basis; Step 3.3: Extract and concatenate the low-frequency approximation coefficients and high-frequency detail coefficients obtained from each decomposition layer to form a multi-scale feature vector; Step 3.4: Apply threshold denoising to the scale coefficients obtained from the decomposition to filter out noise components, and normalize the denoising coefficients. Step 3.5: Align the processed multi-scale coefficient matrix with the original normalized time series data according to the feature dimension, and merge them to form the input feature matrix of the SLCI model.

5. The method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning according to claim 1, characterized in that: Step 4 includes: Step 4.1: Build and train the SLCI deep learning model. Input the preprocessed water quality time series data and multi-scale features obtained in Step 2 and Step 3 into the model for training and testing to obtain the trained classification model. Step 4.2: Use the trained SLCI model to evaluate the real-time or offline water quality data of the pond to be evaluated and output the water quality level.

6. The method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning according to claim 1, characterized in that: Training the SLCI model includes: The multi-scale features decomposed by stationary wavelet transform are used as the input feature matrix and input into the one-dimensional convolutional neural network module. The one-dimensional convolutional neural network module extracts the local spatial features between water quality indicators through two convolutional layers. The first layer is used to capture the low-level change pattern, and the second layer further extracts the high-level semantic features. Batch normalization and Dropout are followed by the convolutional layers. Spatial features extracted by a one-dimensional convolutional neural network are input into a long short-term memory network module to achieve dynamic modeling of time series. The long short-term memory network module consists of two stacked layers: the first layer learns short-term dependency information, and the second layer extracts long-term time series features. The temporal features output by the Long Short-Term Memory (LSTM) network module are input into a fully connected layer for mapping, and the probability distribution of the five water quality levels is output through the Softmax activation function.

7. The method for comprehensive evaluation of pond water quality for Litopenaeus vannamei based on deep learning according to claim 6, characterized in that: The model ultimately determines the corresponding water quality level based on the maximum probability value, thus achieving a comprehensive evaluation of the water quality in Litopenaeus vannamei ponds.

8. A deep learning-based comprehensive water quality evaluation device for Litopenaeus vannamei ponds, characterized in that: include: The data acquisition unit is used to acquire water quality monitoring data from Litopenaeus vannamei ponds. The data processing unit is used to preprocess and serialize the acquired water quality monitoring data to obtain water quality time series data. The wavelet transform unit is used to perform stationary wavelet transform on water quality time series data to achieve multi-scale decomposition and extract multi-scale features, while performing denoising and normalization on the decomposition coefficients. The detection result unit is used to input the data from the data processing unit and the wavelet transform unit into the pre-trained SLCI model for detection and classification, and to obtain the water quality level assessment results of the aquaculture pond.

9. An electronic device comprising a processor and a memory communicatively connected to the processor and used for storing processor-executable instructions, characterized in that: The processor is used to execute the method described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the method described in any one of claims 1-7.

Images