River water quality detection method and system based on artificial intelligence

Abstract

The embodiment of the present application provides a river water quality detection method and system based on artificial intelligence, which significantly improves the accuracy and efficiency of water quality detection by obtaining on-site collected data of the river water quality detection system as regional water quality monitoring samples, and using a shared learning network to predict water quality status labels for these regional water quality monitoring samples. Specifically, by selecting the deep learning network with the best training effect as a shared learning network from multiple training entities participating in the linkage network parameter learning, the reliability and high precision of the water quality status label prediction data are ensured. Furthermore, based on these prediction data, the parameters of the target water quality assessment model are learned, and the optimized target application is determined, so that more accurate water quality status label prediction data can be generated. This method not only improves the intelligence level of water quality detection, but also effectively reduces the human errors and time costs that may exist in traditional water quality detection methods.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to a river water quality detection method and system based on artificial intelligence. Background Art

[0002] With the rapid development of industrialization and urbanization, river water pollution is becoming increasingly serious, posing a significant threat to the ecological environment and human health. To effectively monitor and assess river water quality, traditional water quality testing methods rely primarily on manual sampling and laboratory analysis. However, this method is not only time-consuming and labor-intensive, but also significantly affected by human factors, making it difficult to ensure the accuracy and timeliness of test results. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a river water quality detection method and system based on artificial intelligence.

[0004] In conjunction with the first aspect of the present application, a river water quality detection method based on artificial intelligence is provided, which is applied to a river water quality detection system based on artificial intelligence, and the method comprises:

[0005] Acquire a regional water quality monitoring sample, where the regional water quality monitoring sample is field-collected data from a river water quality detection system, and the river water quality detection system is any training entity that performs linkage network parameter learning;

[0006] Predicting water quality status labels for the regional water quality monitoring samples based on a shared learning network to generate water quality status label prediction data, wherein the shared learning network is obtained by the river water quality detection system from other training entities participating in the linkage network parameter learning, and the shared learning network is the deep learning network with the best training effect among the other training entities;

[0007] Performing parameter learning on a target water quality assessment model based on the water quality status label prediction data, and determining a target application program corresponding to the target water quality assessment model for which parameter learning has been completed;

[0008] The regional water quality monitoring samples are loaded into the target application corresponding to the target water quality assessment model that has completed parameter learning to generate optimized water quality status label prediction data.

[0009] In a possible implementation of the first aspect, performing parameter learning on a target water quality assessment model based on the water quality status label prediction data and determining a target application corresponding to the target water quality assessment model for which parameter learning is completed specifically includes:

[0010] Based on the water quality status label prediction data, parameter learning is performed on a target water quality assessment model to generate a first water quality detection network in which parameter learning is completed, and the target water quality assessment model performs deep learning based on the water quality status label prediction data;

[0011] Based on the first on-site verification collected data, the first water quality detection network that has completed parameter learning is tested to generate first water quality status detection data, wherein the first on-site verification collected data is water quality parameter collection data in the on-site water quality parameter collection data sequence of the river water quality detection system;

[0012] If the first water quality status detection data meets the target requirements, the first water quality detection network that has completed parameter learning is output as a target application program that has completed parameter learning corresponding to the target water quality assessment model.

[0013] In a possible implementation of the first aspect, the shared learning network is obtained by the river water quality detection system from other training entities participating in linkage network parameter learning, specifically including:

[0014] Get the most recent iteration network of each other training entity;

[0015] Based on the second on-site verification collected data, each of the most recent iteration networks is tested to generate second water quality status detection data, wherein the second on-site verification collected data is part of the on-site collected data of the river water quality detection system;

[0016] According to the second water quality status detection data, the deep learning network with the best training effect in each recent iterative network is output as the shared learning network.

[0017] In a possible implementation of the first aspect, performing parameter learning on a target water quality assessment model based on the water quality status label prediction data to generate a first water quality detection network in which parameter learning is completed specifically includes:

[0018] Acquire the remaining on-site collected data; the remaining on-site collected data is part or all of the on-site collected data except the second on-site verification collected data;

[0019] The water quality status label prediction data and the remaining on-site collected data are used as target sample learning data to perform parameter learning on the target water quality assessment model.

[0020] In a possible implementation of the first aspect, obtaining the most recent iterative network of each other training entity specifically includes:

[0021] Obtaining a network index sequence; the network index sequence is used to record the neural network information run by each training entity participating in the linkage network parameter learning, the neural network information including the network iteration identifier and the network index information;

[0022] Based on the network iteration identifier, determining the most recent iteration network of each other training entity;

[0023] Based on the network index information of the latest iteration network of each other training entity, the latest iteration network of each other training entity is obtained.

[0024] In a possible implementation of the first aspect, the target water quality assessment model includes a neural network to be optimized in the river water quality detection system; the target application includes an optimized neural network corresponding to the neural network to be optimized; and the method further includes:

[0025] The neural network to be optimized is tested based on the first on-site verification and collection data to generate third water quality status detection data.

[0026] In a possible implementation of the first aspect, the method further includes:

[0027] Acquire first on-site collected data, perform parameter learning on the target water quality assessment model, and generate a second water quality detection network in which parameter learning has been completed; the first on-site collected data is part or all of the water quality parameter collection data in the on-site water quality parameter collection data sequence of the river water quality detection system;

[0028] Based on the first on-site verification and collection data, the second water quality detection network that has completed parameter learning is tested to generate fourth water quality status detection data.

[0029] In a possible implementation of the first aspect, the method further includes:

[0030] The iteration identifier and network index information of the target application are saved in a network index sequence; the network index sequence is used to store the neural network information run by each training entity for linked network parameter learning; the neural network information includes the network iteration identifier and the network index information.

[0031] In combination with the second aspect of the present application, an artificial intelligence-based river water quality detection system is provided, wherein the artificial intelligence-based river water quality detection system includes a machine-readable storage medium and a processor, wherein the machine-readable storage medium stores machine-executable instructions, and when the processor executes the machine-executable instructions, the artificial intelligence-based river water quality detection system implements the aforementioned artificial intelligence-based river water quality detection method.

[0032] In combination with the third aspect of the present application, a computer-readable storage medium is provided, in which computer-executable instructions are stored. When the computer-executable instructions are executed, the aforementioned artificial intelligence-based river water quality detection method is implemented.

[0033] In combination with any of the above aspects, the embodiment of the present application obtains the on-site collected data of the river water quality detection system as regional water quality monitoring samples, and uses a shared learning network to predict the water quality status labels of these regional water quality monitoring samples, thereby significantly improving the accuracy and efficiency of water quality detection. Specifically, by selecting the deep learning network with the best training effect as a shared learning network from multiple training entities participating in the linkage network parameter learning, the reliability and high precision of the water quality status label prediction data are ensured. Furthermore, based on these prediction data, the parameters of the target water quality assessment model are learned, and the optimized target application is determined, so that more accurate water quality status label prediction data can be generated. This method not only improves the intelligence level of water quality detection, but also effectively reduces the human errors and time costs that may exist in traditional water quality detection methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained by combining these drawings without paying any creative work.

[0035] Figure 1 A flow chart of the artificial intelligence-based river water quality detection method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0037] The terms "first," "second," and so on, in the specification and claims of the present invention and the accompanying drawings are used to distinguish between different items, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or end comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed therein, or may optionally include other steps or elements inherent to such process, method, product, or end.

[0038] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0039] Figure 1 The following is a flow chart of the artificial intelligence-based river water quality detection method provided in an embodiment of the present application. It should be understood that in other embodiments, the order of some steps in the artificial intelligence-based river water quality detection method of this embodiment can be shared based on actual needs, or some steps can be omitted or maintained. The details of the artificial intelligence-based river water quality detection method include:

[0040] Step S110, obtaining regional water quality monitoring samples, wherein the regional water quality monitoring samples are on-site collected data of a river water quality detection system, and the river water quality detection system is any training entity that performs linkage network parameter learning.

[0041] In this embodiment, in a large-scale water resource management system, the server is responsible for data analysis and model training for the entire river water quality monitoring system. Assume that there are multiple river water quality monitoring systems distributed in different regions. These monitoring systems are capable of collecting water quality data on-site and serve as training entities for performing linked network parameter learning. For example, on a large river that flows through multiple cities, river water quality monitoring systems are set up in cities A, B, and C, from upstream to downstream. Each monitoring system contains a series of sensors that can collect multiple water quality parameters such as water temperature, pH value, dissolved oxygen content, chemical oxygen demand (COD), biochemical oxygen demand (BOD), and so on.

[0042] The server communicates with these river water quality monitoring systems. For example, in City A, sensors collect comprehensive data on the river water quality in the area at regular intervals (e.g., every hour). This data includes information such as a water temperature of 20 degrees Celsius, a pH of 7.2, dissolved oxygen levels of 8 mg / L, COD of 20 mg / L, and BOD of 5 mg / L at a specific location in the river at a specific moment. This data, taken as a whole, constitutes a regional water quality monitoring sample.

[0043] The server obtains these regional water quality monitoring samples from the river water quality monitoring system in City A via a network communication protocol (such as TCP / IP). Similarly, the server can also obtain regional water quality monitoring samples from the river water quality monitoring systems in City B and City C. These samples obtained from different river water quality monitoring systems serve as the basis for subsequent analysis and processing, and are used for water quality status analysis, model training, and other operations throughout the linkage network parameter learning process.

[0044] Step S120, predicting the water quality status labels of the regional water quality monitoring samples based on the shared learning network to generate water quality status label prediction data. The shared learning network is obtained by the river water quality detection system from other training entities participating in the linkage network parameter learning. The shared learning network is the deep learning network with the best training effect among other training entities.

[0045] In this water resource management scenario, multiple river water quality monitoring systems are linked to each other for network parameter learning. Assume that, in addition to the river water quality monitoring systems in cities A, B, and C mentioned above, the river water quality monitoring systems in cities D and E also participate in this linked learning process.

[0046] Each river water quality monitoring system continuously trains its own deep learning network. For example, the river water quality monitoring system in City D uses a large amount of locally collected historical water quality data to train a deep learning network. Through continuous iterative optimization, this network can now predict water quality to a certain degree. Similarly, the river water quality monitoring system in City E also has its own deep learning network.

[0047] The server needs to obtain a shared learning network for City A's river water quality monitoring system. First, the server obtains the most recently iterated network of each training entity (here, the river water quality monitoring systems in Cities D and E). For City D's river water quality monitoring system, the server queries its network index sequence (this sequence records the neural network information running in City D's river water quality monitoring system, including network iteration identifiers and network index information) and finds its most recently iterated network based on the network iteration identifier. Then, based on the network index information, the server obtains the specific structure and parameters of this most recently iterated network. The same operation is performed for City E's river water quality monitoring system.

[0048] Next, the server needs to test the recently acquired network iterations of the river water quality monitoring systems in cities D and E based on the second on-site verification data to generate the second water quality status detection data. Assume that the second on-site verification data is a subset of the on-site data collected by the river water quality monitoring system in city A, for example, water quality data from different time periods on a single day. This data is then input into the recently acquired network iterations of the river water quality monitoring systems in cities D and E, generating the corresponding water quality status detection results. These results serve as the second water quality status detection data.

[0049] Then, based on this second water quality data, the server selects the deep learning network with the best training results among the most recently iterated networks as the shared learning network. Suppose, after comparison, the most recently iterated network of City E's river water quality monitoring system demonstrates higher accuracy and lower false positive rates when performing water quality monitoring on the second on-site verification data collected in City A. In this case, the most recently iterated network of City E's river water quality monitoring system is designated as the shared learning network.

[0050] Finally, the server uses this determined shared learning network to predict water quality status labels for the regional water quality monitoring samples from City A's river water quality monitoring system (obtained in step S110). For example, parameters such as water temperature, pH, and dissolved oxygen content from the regional water quality monitoring samples are input into the shared learning network. The network then undergoes a series of computations (such as convolutional layers, pooling layers, and fully connected layers) to ultimately output predicted water quality status labels. This predicted data may include information such as whether the water is polluted, and whether the pollution level is mild, moderate, or severe.

[0051] Step S130 , performing parameter learning on a target water quality assessment model based on the water quality status tag prediction data, and determining a target application program corresponding to the target water quality assessment model that has completed parameter learning.

[0052] In this scenario, assume that the target water quality assessment model in City A's river water quality monitoring system is a neural network-based model. The model's initial structure consists of an input layer, several hidden layers, and an output layer. The input layer receives various water quality parameters, the hidden layers extract and transform data features, and the output layer outputs the water quality assessment results.

[0053] The server first learns the parameters of the target water quality assessment model based on the water quality status label prediction data. For example, the server obtains the remaining field data. This refers to some or all of the field data collected from the river water quality monitoring system in City A, excluding the second field verification data previously used to determine the shared learning network. If the second field verification data was collected on a specific day, the remaining field data could be data from another date.

[0054] The water quality status label prediction data and the remaining field-collected data are used as target sample learning data to perform parameter learning on the target water quality assessment model. During the learning process, the model uses the backpropagation algorithm to adjust the weights and bias parameters in the neural network based on the input target sample learning data. For example, if the water quality status label prediction data indicates that the water quality in a certain area is slightly polluted, and the dissolved oxygen content in the target sample learning data is low and the COD value is high, the model will adjust the weight parameters related to dissolved oxygen and COD in the neural network based on this information, so that the model can more accurately reflect this relationship in subsequent predictions.

[0055] The server then tests the first water quality detection network, which has completed parameter learning, based on the first field verification data (here, the first water quality detection network is generated by the target water quality assessment model through deep learning based on the water quality status label prediction data), generating the first water quality status detection data. The first field verification data is from the sequence of field water quality parameter data collected by the river water quality monitoring system in City A. Assume that the first field verification data is the most recently collected data set, including parameters such as water temperature, pH value, and dissolved oxygen content. This data is input into the first water quality detection network, which has completed parameter learning. The network then calculates and outputs a detection result regarding the water quality status, which is the first water quality status detection data.

[0056] If the first water quality test data meets the target requirements—for example, if the predicted water quality matches the actual water quality test results (obtained through more precise laboratory analysis or other high-precision testing methods) with an accuracy rate exceeding a set threshold (e.g., 90%)—the server will output the parameter-learned first water quality test network as the target application corresponding to the target water quality assessment model. This target application can then be used to more accurately assess and predict the water quality of City A's rivers.

[0057] Step S140 , loading the regional water quality monitoring sample into the target application program corresponding to the target water quality assessment model that has completed parameter learning, to generate optimized water quality status label prediction data.

[0058] In this step, the server has determined the target application for which parameter learning has been completed for the target water quality assessment model. The server then retrieves the regional water quality monitoring samples from City A's river water quality monitoring system (similar to the regional water quality monitoring samples in step S110, but likely newly collected data).

[0059] Load this regional water quality monitoring sample into the target application. For example, the sample's water temperature is 22 degrees Celsius, pH is 7.3, dissolved oxygen is 7 mg / L, and COD is 18 mg / L. These parameters are then fed into the target application (i.e., the neural network structure corresponding to the target water quality assessment model that has completed parameter learning).

[0060] The neural network structure in the target application begins processing this input data. The input layer receives these water quality parameters, which are then passed through hidden layers for feature extraction and transformation. Within these hidden layers, neurons perform calculations on the input data based on previously learned weights and biases. For example, the neuron associated with dissolved oxygen content calculates the value of 7 mg / L based on previously adjusted weights. After processing through multiple hidden layers, the data ultimately reaches the output layer.

[0061] The output layer outputs optimized water quality label predictions based on previous training results. This optimized prediction may be more accurate than the previous one (the water quality label predictions in step S120). For example, the previous prediction may have simply categorized water quality as polluted or unpolluted. However, after processing by the target application, more detailed water quality label predictions can be generated, such as more accurately predicting that the water is slightly polluted, with the primary pollutant being a slightly low dissolved oxygen content. This optimized data is crucial for accurately assessing river water quality and implementing targeted remediation measures.

[0062] Based on the above steps, the embodiment of the present application obtains the on-site collected data of the river water quality detection system as regional water quality monitoring samples, and uses a shared learning network to predict the water quality status labels of these regional water quality monitoring samples, thereby significantly improving the accuracy and efficiency of water quality detection. Specifically, by selecting the deep learning network with the best training effect as a shared learning network from multiple training entities participating in the linkage network parameter learning, the reliability and high precision of the water quality status label prediction data are ensured. Furthermore, based on these prediction data, the parameters of the target water quality assessment model are learned, and the optimized target application is determined, so that more accurate water quality status label prediction data can be generated. This method not only improves the intelligence level of water quality detection, but also effectively reduces the human errors and time costs that may exist in traditional water quality detection methods.

[0063] In a possible implementation, performing parameter learning on a target water quality assessment model based on the water quality status tag prediction data and determining a target application corresponding to the target water quality assessment model for which parameter learning is completed specifically includes:

[0064] Based on the water quality status label prediction data, parameter learning is performed on the target water quality assessment model to generate a first water quality detection network that has completed parameter learning. The target water quality assessment model performs deep learning based on the water quality status label prediction data.

[0065] Based on the first on-site verification collection data, the first water quality detection network that has completed parameter learning is tested to generate first water quality status detection data. The first on-site verification collection data is the water quality parameter collection data in the on-site water quality parameter collection data sequence of the river water quality detection system.

[0066] If the first water quality status detection data meets the target requirements, the first water quality detection network that has completed parameter learning is output as a target application program that has completed parameter learning corresponding to the target water quality assessment model.

[0067] In a possible implementation, the shared learning network is obtained by the river water quality detection system from other training entities participating in the linkage network parameter learning, specifically including:

[0068] Get the most recent iteration network for each other training entity.

[0069] Based on the second on-site verification collected data, each latest iteration network is detected to generate second water quality status detection data, where the second on-site verification collected data is part of the on-site collected data of the river water quality detection system.

[0070] According to the second water quality status detection data, the deep learning network with the best training effect in each recent iterative network is output as the shared learning network.

[0071] In a possible implementation, performing parameter learning on a target water quality assessment model based on the water quality status label prediction data to generate a first water quality detection network in which parameter learning is completed specifically includes:

[0072] Obtaining the remaining on-site collected data, wherein the remaining on-site collected data is part of or all of the on-site collected data except the second on-site verification collected data.

[0073] The water quality status label prediction data and the remaining on-site collected data are used as target sample learning data to perform parameter learning on the target water quality assessment model.

[0074] In a possible implementation, obtaining the most recent iterative network of each other training entity specifically includes:

[0075] Obtain a network index sequence. The network index sequence is used to record the neural network information run by each training entity participating in the linkage network parameter learning, and the neural network information includes a network iteration identifier and network index information.

[0076] Based on the network iteration identifier, the most recent iteration network of each other training entity is determined.

[0077] Based on the network index information of the latest iteration network of each other training entity, the latest iteration network of each other training entity is obtained.

[0078] In one possible implementation, the target water quality assessment model includes the neural network to be optimized in the river water quality detection system. The target application includes an optimized neural network corresponding to the neural network to be optimized. The method further includes:

[0079] The neural network to be optimized is tested based on the first on-site verification and collection data to generate third water quality status detection data.

[0080] In one possible implementation, the method further includes:

[0081] Acquire first on-site collected data, perform parameter learning on the target water quality assessment model, and generate a second water quality detection network that has completed parameter learning. The first on-site collected data is part or all of the water quality parameter collection data in the on-site water quality parameter collection data sequence of the river water quality detection system.

[0082] Based on the first on-site verification and collection data, the second water quality detection network that has completed parameter learning is tested to generate fourth water quality status detection data.

[0083] In one possible implementation, the method further includes:

[0084] The iteration identifier and network index information of the target application are saved in a network index sequence. The network index sequence is used to store the neural network information run by each training entity for linked network parameter learning. The neural network information includes the network iteration identifier and network index information.

[0085] In the above embodiments, the artificial intelligence-based river water quality detection system for executing the above method embodiments has at least one processor, a control module (chip set) coupled to at least one of the (at least one) processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one load / output device coupled to the control module, and a network interface coupled to the control module.

[0086] The processor may include at least one single-core or multi-core processor, and the processor may include any combination of general-purpose processors or dedicated processors (such as graphics processors, application processors, baseband processors, etc.). For some alternative implementations, the artificial intelligence-based river water quality detection system can be used as the electronic device such as the gateway described in the embodiments of this application.

[0087] For some alternative embodiments, the artificial intelligence-based river water quality detection system may include at least one computer-readable medium (e.g., a memory or NVM / storage device) having instructions and at least one processor integrated with the at least one computer-readable medium and configured to execute the instructions to implement a module to perform the actions described in the present disclosure.

[0088] For one embodiment, the control module may include any suitable interface controller to provide any suitable interface to at least one of the processor(s) and / or any suitable device or component in communication with the control module.

[0089] The control module may include a memory controller module to provide an interface to the memory. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0090] The memory may be used, for example, to load and store data and / or instructions for an artificial intelligence-based river water quality detection system. For one embodiment, the memory may include any suitable volatile memory, for example, a suitable DRAM.

[0091] For one embodiment, the control module may include at least one load / output controller to provide an interface to the NVM / storage device and the (at least one) load / output device.

[0092] For example, NVM / storage devices may be used to store data and / or instructions. The NVM / storage devices may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one compact disc (CD) drive, and / or at least one digital versatile disc (DVD) drive).

[0093] The NVM / storage device may include storage resources that are physically part of the device on which the artificial intelligence-based river water quality detection system is installed, or may be accessible to the device without being part of the device. For example, the NVM / storage device may be accessed via (at least one) load / output device over a network.

[0094] The (at least one) load-in / output device may provide an interface for the artificial intelligence-based river water quality detection system to communicate with any other appropriate device, and the load-in / output device may include a communication component, a phonetic component, a sensor component, etc. The network interface may provide an interface for the artificial intelligence-based river water quality detection system to communicate based on at least one network, and the artificial intelligence-based river water quality detection system may wirelessly communicate with at least one component of a wireless network based on any of at least one wireless network priors and / or protocols, such as accessing a wireless network based on a communication prior.

[0095] For one embodiment, at least one of the (at least one) processors may be loaded together with the logic of at least one controller of a control module (e.g., a memory controller module). For one embodiment, at least one of the (at least one) processors may be loaded together with the logic of at least one controller of a control module to form a system-level load. For one embodiment, at least one of the (at least one) processors may be fused on the same die with the logic of at least one controller of a control module. For one embodiment, at least one of the (at least one) processors may be fused on the same die with the logic of at least one controller of a control module to form a system-on-chip (SoC).

[0096] The above is a detailed introduction to the embodiments of the present application. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for those skilled in the art, based on the ideas of the present application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0097] An embodiment of the present invention discloses a computer-readable storage medium that stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the artificial intelligence-based river water quality detection method described in the aforementioned embodiment.

[0098] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps of the artificial intelligence-based river water quality detection method described in the aforementioned embodiment.

[0099] The device embodiments described above are merely illustrative, wherein the modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple modules. Some or all of the modules may be selected based on actual needs to achieve the purpose of the present embodiment. Those skilled in the art can understand and implement the present invention without inventive effort.

[0100] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory, or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of storing or storing data.

[0101] Finally, it should be noted that what is disclosed above is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A river water quality detection method based on artificial intelligence, characterized in that: The method comprises: Acquire a regional water quality monitoring sample, where the regional water quality monitoring sample is field-collected data from a river water quality detection system, and the river water quality detection system is any training entity that performs linkage network parameter learning; Predicting water quality status labels for the regional water quality monitoring samples based on a shared learning network to generate water quality status label prediction data, wherein the shared learning network is obtained by the river water quality detection system from other training entities participating in the linkage network parameter learning, and the shared learning network is the deep learning network with the best training effect among the other training entities; Performing parameter learning on a target water quality assessment model based on the water quality status label prediction data, and determining a target application program corresponding to the target water quality assessment model for which parameter learning has been completed; Loading the regional water quality monitoring samples into the target application program corresponding to the target water quality assessment model after completing parameter learning to generate optimized water quality status label prediction data; The step of performing parameter learning on a target water quality assessment model based on the water quality status tag prediction data and determining a target application corresponding to the target water quality assessment model for which parameter learning is completed specifically includes: Based on the water quality status label prediction data, parameter learning is performed on a target water quality assessment model to generate a first water quality detection network in which parameter learning is completed, and the target water quality assessment model performs deep learning based on the water quality status label prediction data; The first water quality detection network that has completed parameter learning is tested based on the first on-site verification collected data to generate first water quality status detection data, wherein the first on-site verification collected data is water quality parameter collection data in the on-site water quality parameter collection data sequence of the river water quality detection system; If the first water quality status detection data meets the target requirements, outputting the first water quality detection network that has completed parameter learning as a target application program that has completed parameter learning corresponding to the target water quality assessment model; The shared learning network is obtained by the river water quality detection system from other training entities participating in the linkage network parameter learning, specifically including: Get the most recent iteration network of each other training entity; Based on the second on-site verification collected data, each of the most recent iteration networks is tested to generate second water quality status detection data, wherein the second on-site verification collected data is part of the on-site collected data of the river water quality detection system; Outputting the deep learning network with the best training effect among the most recently iterated networks as the shared learning network based on the second water quality status detection data; The step of performing parameter learning on a target water quality assessment model based on the water quality status label prediction data to generate a first water quality detection network in which parameter learning is completed specifically includes: Acquire the remaining on-site collected data; the remaining on-site collected data is part or all of the on-site collected data except the second on-site verification collected data; Using the water quality status label prediction data and the remaining field collected data as target sample learning data, performing parameter learning on the target water quality assessment model; The obtaining of the most recent iterative network of each other training entity specifically includes: Obtaining a network index sequence; the network index sequence is used to record the neural network information run by each training entity participating in the linkage network parameter learning, the neural network information including the network iteration identifier and the network index information; Based on the network iteration identifier, determining the most recent iteration network of each other training entity; Based on the network index information of the latest iteration network of each other training entity, the latest iteration network of each other training entity is obtained.

2. The method for detecting river water quality based on artificial intelligence according to claim 1, characterized in that: The target water quality assessment model includes the neural network to be optimized in the river water quality detection system; the target application includes an optimized neural network corresponding to the neural network to be optimized; and the method further includes: The neural network to be optimized is tested based on the first on-site verification and collection data to generate third water quality status detection data.

3. The river water quality detection method based on artificial intelligence according to claim 1 is characterized in that: The method further comprises: Acquire first on-site collected data, perform parameter learning on the target water quality assessment model, and generate a second water quality detection network in which parameter learning has been completed; the first on-site collected data is part or all of the water quality parameter collection data in the on-site water quality parameter collection data sequence of the river water quality detection system; Based on the first on-site verification and collection data, the second water quality detection network that has completed parameter learning is tested to generate fourth water quality status detection data.

4. The method for detecting river water quality based on artificial intelligence according to claim 1, characterized in that: The method further comprises: The iteration identifier and network index information of the target application are saved in a network index sequence; the network index sequence is used to store the neural network information run by each training entity for linked network parameter learning; the neural network information includes the network iteration identifier and the network index information.

5. A computer-readable storage medium, characterized in that The computer-readable storage medium stores machine-executable instructions, which, when executed by a computer, implement the artificial intelligence-based river water quality detection method described in any one of claims 1 to 4.

6. A river water quality detection system based on artificial intelligence, characterized in that: It includes a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the river water quality detection method based on artificial intelligence as described in any one of claims 1 to 4 is implemented.

Images