Cage culture environment monitoring system and method based on multi-source sensor

Abstract

The invention relates to the technical field of cage culture, and discloses a cage culture environment monitoring system and method based on a multi-source sensor, and the system comprises an environment data collection module, a data processing coordination module, an intelligent monitoring analysis module, an early warning decision generation module, a man-machine interaction display module, and a system optimization management module. A multi-source sensor is deployed to collect multi-modal environment data, a standardized processing flow is set for different environment parameters, the normalization and comparability of collection of various kinds of environment data are guaranteed, meanwhile, a frequency consistency judgment mechanism is adopted to carry out coordination processing on the multi-source data, and the accuracy of data processing is improved. The frequency difference problem in the data acquisition process is found and corrected in real time, the basic quality of multi-source data fusion is ensured, and the reliability of environment monitoring data is improved; cooperative analysis of multi-modal data such as water quality, weather and videos is realized by constructing multi-source fusion data and performing space-time alignment.

Description

Technical Field

[0001] This invention relates to the field of cage aquaculture technology, specifically to a cage aquaculture environment monitoring system and method based on multi-source sensors. Background Technology

[0002] Cage culture mainly relies on artificial feeding to accelerate the growth of farmed fish. However, some of the feed inevitably gets lost in the water, affecting the water's physicochemical properties. In cage culture, the loss of feed and fish excrement increases the solubility of nutrients and suspended solids, resulting in significantly lower water transparency compared to the control area. This reduced water transparency limits phytoplankton photosynthesis, leading to a decrease in oxygen production through photosynthesis.

[0003] Currently, in the field of cage aquaculture environmental monitoring, since the aquaculture environment involves multiple dimensions such as water quality, meteorology, and biological behavior, traditional monitoring methods mainly rely on single-type sensors or human experience for data collection and analysis, making it difficult to achieve collaborative perception and comprehensive evaluation of multi-source environmental data.

[0004] Therefore, a cage aquaculture environment monitoring system and method based on multi-source sensors is proposed to solve the above problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a cage aquaculture environment monitoring system and method based on multi-source sensors. This solves the problem mentioned in the background that traditional monitoring methods mainly rely on single-type sensors or human experience for data collection and analysis, making it difficult to achieve collaborative perception and comprehensive evaluation of multi-source environmental data.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a cage aquaculture environment monitoring system and method based on multi-source sensors, the method comprising the following steps: S1. Collect multimodal data of the cage aquaculture environment through multi-source sensors. The multimodal data includes water quality parameter data, meteorological environment data, and underwater video image data. S2. Perform data preprocessing and sampling frequency measurement on the multimodal data to generate water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data; S3. Based on water quality parameter frequency data, meteorological environment frequency data and underwater video image frequency data, perform multi-source sensor data sampling frequency consistency judgment processing, generate multi-source sensor data sampling frequency consistency judgment result, and if the frequencies are consistent, execute step S5. S4. When the frequencies are inconsistent, use a linear interpolation algorithm to adjust the sampling frequency of the data source with the lower frequency to generate adjusted multi-source sensor data. S5. Perform spatiotemporal alignment and feature fusion processing on the frequency-consistent and adjusted multi-source sensor data to construct multi-source fusion data of the cage aquaculture environment; S6. Based on the multi-source fusion data of the cage aquaculture environment and the preset aquaculture environment assessment standard data, perform anomaly monitoring algorithm matching processing to generate target anomaly monitoring algorithm type feature data; S7. Based on the target anomaly monitoring algorithm type characteristic data, perform real-time anomaly monitoring and early warning processing of the cage aquaculture environment, and generate anomaly early warning information for the cage aquaculture environment. S8. Visualize and push early warning information on abnormal cage aquaculture environments through the aquaculture supervision platform.

[0007] Preferably, the process of collecting multimodal data on the cage aquaculture environment in step S1 includes the following steps: S11. Collect water quality parameter data, including water temperature, pH value, dissolved oxygen, ammonia nitrogen content and salinity parameters, by deploying water quality sensors in the cage aquaculture area; S12. Collect meteorological environmental data, including air temperature, humidity, wind speed, wind direction and rainfall parameters, by means of meteorological sensors deployed in the cage aquaculture area; S13. Underwater video image data, including fish activity status, feeding behavior and water turbidity visual information, are collected by underwater camera equipment deployed in the cage aquaculture area.

[0008] Preferably, the data preprocessing and sampling frequency measurement in S2 include the following steps: S21. Transmit the collected water quality parameter data, meteorological environment data, and underwater video image data to the aquaculture supervision platform respectively; S22. Use time series analysis to measure the sampling frequency of various types of data to generate water quality parameter frequency data, meteorological environment frequency data and underwater video image frequency data; S23. Perform frame extraction and image enhancement processing on the underwater video image data to generate standardized underwater image data.

[0009] Preferably, the process for determining the consistency of multi-source sensor data sampling frequencies in step S3 includes the following steps: S31. Acquire water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data; S32. Perform pairwise comparisons of the three types of frequency data to generate multiple sets of frequency comparison results; S33. Based on a comprehensive judgment of multiple frequency comparison results, generate a consistency judgment result for the sampling frequency of multi-source sensor data: When the sampling frequency of all data sources is consistent, the output frequency consistency judgment result is consistent; When any two data sources have inconsistent sampling frequencies, the output frequency consistency judgment result is inconsistent.

[0010] Preferably, the sampling frequency adjustment in S4 includes the following steps: S41. When the frequency consistency judgment result is inconsistent, identify the data source with the lowest sampling frequency; S42. Use a linear interpolation algorithm to boost the frequency of the data source with the lowest sampling frequency. S43. Generate adjusted multi-source sensor data, including adjusted water quality parameter data, meteorological environment data, and underwater video image data.

[0011] Preferably, the spatiotemporal alignment and feature fusion processing in S5 includes the following steps: S51. Establish a unified data coordinate system with timestamps and spatial coordinates as references; S52. Align the multi-source sensor data with consistent frequency and adjustment according to timestamps and spatial coordinates; S53. Extract feature parameters from various types of data, including water quality feature parameters, meteorological feature parameters, and image visual feature parameters; S54. The feature-level fusion method is used to fuse multi-source feature parameters into unified multi-source fusion data of cage aquaculture environment.

[0012] Preferably, the anomaly detection algorithm matching process in S6 includes the following steps: S61. Establish a database of algorithms for monitoring aquaculture environment anomalies, including algorithms for monitoring water quality anomalies, meteorological anomalies, biological behavior anomalies, and comprehensive assessment and monitoring algorithms. S62. Construct a standard environmental data model corresponding to each anomaly detection algorithm; S63. Perform similarity matching between multi-source fusion data of cage aquaculture environment and various standard environmental data patterns; S64. Select the anomaly detection algorithm corresponding to the standard environmental data pattern with the highest similarity as the target anomaly detection algorithm. S65. Generate target anomaly monitoring algorithm type feature data, including algorithm identifier and parameter configuration information.

[0013] Preferably, the real-time anomaly monitoring and early warning processing in S7 includes the following steps: S71. Call the corresponding anomaly monitoring program based on the target anomaly monitoring algorithm type characteristic data; S72. Extract feature parameters from multi-source fusion data of cage aquaculture environment; S73. Use dynamic threshold detection method to judge the anomalies of environmental parameters; S74. When an environmental anomaly is detected, generate an early warning message for an abnormal cage aquaculture environment, including the anomaly type, anomaly level, time of occurrence, and location. S75. Activate the corresponding early warning and handling process according to the anomaly level.

[0014] Preferably, the visualization and early warning notification push in S8 includes the following steps: S81. Visualize early warning information about abnormal cage aquaculture environment on the map interface of the aquaculture supervision platform; S82. Send early warning notifications to aquaculture managers through various communication methods, including SMS, mobile application push and email; S83. Generate abnormal handling suggestions, including oxygenation suggestions, feeding adjustment suggestions, and emergency handling suggestions; S84. Record the entire process of early warning handling and form a monitoring and early warning log for cage aquaculture environment.

[0015] Preferably, the system includes: The environmental data acquisition module uses a multi-source sensor unit to collect multimodal environmental data, performs data standardization processing through a sensor calibration unit, and outputs the sampling frequency information of each data source through a frequency measurement unit. The data processing coordination module receives the sampling frequency information, performs data coordination analysis through the frequency consistency judgment unit, performs frequency adjustment operation using the data interpolation processing unit, and outputs standardized multi-source data through the spatiotemporal alignment unit. The intelligent monitoring and analysis module receives the standardized multi-source data, obtains environmental feature parameters through the feature extraction unit, generates environmental fusion data using the multi-source data fusion unit, and outputs the optimal monitoring scheme through the anomaly monitoring algorithm matching unit. The early warning decision generation module receives the environmental fusion data, performs anomaly detection through the dynamic threshold analysis unit, determines the early warning level through the risk assessment unit, and outputs early warning decision information through the early warning information generation unit. The human-computer interaction display module receives the early warning decision information, generates an environmental monitoring view through the visualization display unit, pushes early warning information through the early warning notification unit, and receives user commands through the interactive feedback unit. The system optimization and management module receives all data during the monitoring process, analyzes the system's operating status through the performance evaluation unit, adjusts the system configuration using the parameter optimization unit, and outputs system optimization solutions through the maintenance management unit. Beneficial effects

[0016] Compared with existing technologies, this invention provides a cage aquaculture environment monitoring system and method based on multi-source sensors, which has the following beneficial effects: 1. In this invention, multi-source sensors are deployed to collect multi-modal environmental data, and standardized processing procedures are set for different environmental parameters to ensure the standardization and comparability of various environmental data collection. At the same time, a frequency consistency judgment mechanism is used to coordinate the processing of multi-source data, detect and correct frequency differences in the data collection process in real time, ensure the basic quality of multi-source data fusion, and improve the reliability of environmental monitoring data.

[0017] 2. In this invention, by constructing multi-source fusion data and performing spatiotemporal alignment, collaborative analysis of multimodal data such as water quality, meteorology, and video is achieved, enabling the system to perceive the state of the aquaculture environment. When data anomalies occur, the optimal monitoring scheme is quickly identified through feature extraction and algorithm matching, ensuring the timeliness and accuracy of environmental anomaly detection and improving the risk warning capability of the aquaculture environment.

[0018] 3. In this invention, early warning information is pushed through visual display and multiple communication methods, providing intuitive environmental monitoring results and early warning prompts to aquaculture managers in real time. This enables users to grasp the status of the aquaculture environment in a timely manner and take corresponding measures, thereby improving the efficiency and response speed of aquaculture management and enhancing the system's practicality and user experience. Attached Figure Description

[0019] Figure 1 This is a flowchart of the cage aquaculture environment monitoring method based on multi-source sensors according to the present invention; Figure 2 This is a schematic diagram of the cage aquaculture environment monitoring system based on multi-source sensors according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Specific embodiment: A cage aquaculture environment monitoring system and method based on multi-source sensors, the method including the following steps: S1. Collect multimodal data of the cage aquaculture environment through multi-source sensors. The multimodal data includes water quality parameter data, meteorological environment data, and underwater video image data. S2. Perform data preprocessing and sampling frequency measurement on the multimodal data to generate water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data; S3. Based on water quality parameter frequency data, meteorological environment frequency data and underwater video image frequency data, perform multi-source sensor data sampling frequency consistency judgment processing, generate multi-source sensor data sampling frequency consistency judgment result, and if the frequencies are consistent, execute step S5. S4. When the frequencies are inconsistent, use a linear interpolation algorithm to adjust the sampling frequency of the data source with the lower frequency to generate adjusted multi-source sensor data. S5. Perform spatiotemporal alignment and feature fusion processing on the frequency-consistent and adjusted multi-source sensor data to construct multi-source fusion data of the cage aquaculture environment; S6. Based on the multi-source fusion data of the cage aquaculture environment and the preset aquaculture environment assessment standard data, perform anomaly monitoring algorithm matching processing to generate target anomaly monitoring algorithm type feature data; S7. Based on the target anomaly monitoring algorithm type characteristic data, perform real-time anomaly monitoring and early warning processing of the cage aquaculture environment, and generate anomaly early warning information for the cage aquaculture environment. S8. Visualize and push early warning information on abnormal cage aquaculture environments through the aquaculture supervision platform.

[0022] The steps involved in collecting multimodal data on the cage aquaculture environment in S1 are as follows: S11. Collect water quality parameter data, including water temperature, pH value, dissolved oxygen, ammonia nitrogen content and salinity parameters, by deploying water quality sensors in the cage aquaculture area; S12. Collect meteorological environmental data, including air temperature, humidity, wind speed, wind direction and rainfall parameters, by means of meteorological sensors deployed in the cage aquaculture area; S13. Underwater video image data, including fish activity status, feeding behavior and water turbidity visual information, are collected by underwater camera equipment deployed in the cage aquaculture area.

[0023] Data preprocessing and sampling frequency measurement in S2 include the following steps: S21. Transmit the collected water quality parameter data, meteorological environment data, and underwater video image data to the aquaculture supervision platform respectively; S22. Use time series analysis to measure the sampling frequency of various types of data to generate water quality parameter frequency data, meteorological environment frequency data and underwater video image frequency data; In the specific implementation of time series analysis methods, a discrete time series model is first established to mathematically describe the sampling process. Let the collected sensor data sequence be... ,in This represents the data value of the nth sampling point, where n is the total number of sampling points; the sampling interval. It is obtained by calculating the average time difference between adjacent sampling points. The calculation formula is as follows: ; in The sampling interval is... This represents the timestamp of the i-th sampling point; the sampling frequency f is the reciprocal of the sampling interval, i.e.: ; Using the above methods, water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data can be calculated, providing a data basis for subsequent frequency consistency judgment. S23. Perform frame extraction and image enhancement processing on the underwater video image data to generate standardized underwater image data; Frame extraction and image enhancement processing are first implemented using an equal-interval sampling method to extract keyframes from the video stream, with an extraction interval of [missing information]. Then the timestamp of the kth extracted frame is: ; in The starting time is used; the extracted frame images are enhanced using histogram equalization, and the enhancement function is: ; in To input grayscale level, To output grayscale levels, where L is the number of grayscale levels. Let N be the number of pixels at gray level j, N be the total number of pixels, and j be the gray level index. This is a grayscale transformation function; after processing, it generates standardized underwater image data, improving image quality and support for feature recognition.

[0024] The process for determining the consistency of multi-source sensor data sampling frequency in S3 includes the following steps: S31. Acquire water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data; S32. Perform pairwise comparisons of the three types of frequency data to generate multiple sets of frequency comparison results; S33. Based on a comprehensive judgment of multiple frequency comparison results, generate a consistency judgment result for the sampling frequency of multi-source sensor data: When the sampling frequency of all data sources is consistent, the output frequency consistency judgment result is consistent; When any two data sources have inconsistent sampling frequencies, the output frequency consistency judgment result is inconsistent.

[0025] The sampling frequency adjustment in S4 includes the following steps: S41. When the frequency consistency judgment result is inconsistent, identify the data source with the lowest sampling frequency; S42. Use a linear interpolation algorithm to boost the frequency of the data source with the lowest sampling frequency. Let the original low-frequency data sequence be For any target time point t, find the two nearest original data points before and after it. and The corresponding time point is and The interpolation formula is: ; Where y(t) is the interpolation at time t. For the measurement value of the i-th original data point, For timestamps; The above method increases the sampling frequency of low-frequency data to match that of the highest sampling frequency data source, generating adjusted multi-source sensor data and ensuring the accuracy of subsequent data fusion.

[0026] S43. Generate adjusted multi-source sensor data, including adjusted water quality parameter data, meteorological environment data, and underwater video image data.

[0027] The spatiotemporal alignment and feature fusion processing in S5 includes the following steps: S51. Establish a unified data coordinate system with timestamps and spatial coordinates as references; S52. Align the multi-source sensor data with consistent frequency and adjustment according to timestamps and spatial coordinates; In the spatiotemporal alignment process, the spatiotemporal coordinates of each data point are set as (t, x, y, z), where t is the timestamp and (x, y, z) are the three-dimensional spatial coordinates. The alignment process uses the nearest neighbor interpolation method to map data from different sensors onto a unified spatiotemporal grid. For each grid point... Its data value is obtained by weighting the average of all data points within the range, and the weighting function is: ; in This represents the weight coefficient of the i-th data point. Represents the time coordinates of the grid points. The spatiotemporal scale balance coefficient. To represent the timestamp of the i-th data point, , , Let i be the spatial three-dimensional coordinates of the i-th data point. , , The spatial three-dimensional coordinates of the grid points are used; the above processing achieves the alignment of multi-source sensor data in the spatiotemporal dimension.

[0028] S53. Extract feature parameters from various types of data, including water quality feature parameters, meteorological feature parameters, and image visual feature parameters; S54. Employ a feature-level fusion method to fuse multi-source feature parameters into unified multi-source fusion data of cage aquaculture environment; The feature-level fusion method employs a weighted average-based feature fusion algorithm, where the feature vectors extracted from M data sources are... The weight of each feature vector is This reflects the importance of the data source, and the fused feature vector is: ; in The feature vector is the fused vector, and M is the total number of data sources participating in the fusion. This is the feature vector extracted from the m-th data source; Weight Dynamically adjust based on data quality and feature reliability to meet [the requirements]. ; Multi-source fusion data of cage aquaculture environment is generated by feature-level fusion method, providing high-quality data input for subsequent anomaly monitoring.

[0029] The anomaly detection algorithm matching process in S6 includes the following steps: S61. Establish a database of algorithms for monitoring aquaculture environment anomalies, including algorithms for monitoring water quality anomalies, meteorological anomalies, biological behavior anomalies, and comprehensive assessment and monitoring algorithms. The aquaculture environment anomaly monitoring algorithm library includes four types of monitoring algorithms: water quality anomaly monitoring algorithm, which uses a statistical process control-based approach to set upper and lower control limits for water quality parameters; meteorological anomaly monitoring algorithm, which uses an extreme weather identification model and a sliding window to detect abnormal fluctuations in meteorological parameters; biological behavior anomaly monitoring algorithm, which uses computer vision technology to identify abnormal patterns in fish activity; and comprehensive evaluation monitoring algorithm, which uses a multi-indicator fusion method to comprehensively consider the degree of anomaly of various parameters. The algorithm library is built in a modular manner, with each algorithm module providing a unified interface specification and supporting dynamic loading and configuration updates.

[0030] S62. Construct a standard environmental data model corresponding to each anomaly detection algorithm; S63. Perform similarity matching between multi-source fusion data of cage aquaculture environment and various standard environmental data patterns; Similarity matching employs a cosine similarity-based matching algorithm. Let the feature vector of the current environment's multi-source fusion data be... The feature vector of the standard environmental data pattern is The similarity calculation formula is: ; in Similarity is used; by calculating the similarity between the current data and each standard pattern, the anomaly detection algorithm corresponding to the standard environmental data pattern with the highest similarity is selected as the target anomaly detection algorithm, ensuring the accuracy and adaptability of the algorithm selection.

[0031] S64. Select the anomaly detection algorithm corresponding to the standard environmental data pattern with the highest similarity as the target anomaly detection algorithm. S65. Generate target anomaly monitoring algorithm type feature data, including algorithm identifier and parameter configuration information.

[0032] The real-time anomaly monitoring and early warning processing in S7 includes the following steps: S71. Call the corresponding anomaly monitoring program based on the target anomaly monitoring algorithm type characteristic data; The anomaly monitoring program is implemented using dynamic link library technology. Based on the algorithm identifier in the target anomaly monitoring algorithm type feature data, the corresponding dynamic link library file is loaded from the algorithm library to initialize the algorithm instance. Multi-source fusion data of the cage aquaculture environment is transmitted through a unified application programming interface to call the anomaly monitoring function and obtain the monitoring results. The entire process adopts an anomaly handling mechanism to ensure the stability and reliability of the call.

[0033] S72. Extract feature parameters from multi-source fusion data of cage aquaculture environment; S73. Use dynamic threshold detection method to judge the anomalies of environmental parameters; The dynamic threshold detection method employs an adaptive threshold adjustment mechanism. The threshold is dynamically calculated based on the historical statistical characteristics of environmental data, with the historical average being [value missing]. The standard deviation is The dynamically calculated anomaly detection threshold is: ; in The anomaly detection threshold is dynamically calculated, and q is the sensitivity coefficient, which is dynamically adjusted according to environmental stability and early warning requirements. The above method is used to detect environmental anomalies and reduce false alarm and false alarm rates. S74. When an environmental anomaly is detected, generate an early warning message for an abnormal cage aquaculture environment, including the anomaly type, anomaly level, time of occurrence, and location. S75. Activate the corresponding early warning and handling procedures according to the anomaly level; The early warning process adopts a tiered early warning mechanism, which initiates corresponding procedures based on the level of abnormality: Level 1 early warning triggers data review and confirmation procedures; Level 2 early warning activates automatic control equipment to adjust environmental parameters; Level 3 early warning notifies management personnel to intervene; each early warning level corresponds to different response time limits and processing procedures to ensure that early warning information is handled in a timely and appropriate manner.

[0034] Visualization and alert notification push in S8 include the following steps: S81. Visualize early warning information about abnormal cage aquaculture environment on the map interface of the aquaculture supervision platform; S82. Send early warning notifications to aquaculture managers through various communication methods, including SMS, mobile application push and email; S83. Generate abnormal handling suggestions, including oxygenation suggestions, feeding adjustment suggestions, and emergency handling suggestions; S84. Record the entire process of early warning handling and form a monitoring and early warning log for cage aquaculture environment.

[0035] The system includes: The environmental data acquisition module uses a multi-source sensor unit to collect multimodal environmental data, performs data standardization processing through a sensor calibration unit, and outputs the sampling frequency information of each data source through a frequency measurement unit. The data processing coordination module receives the sampling frequency information, performs data coordination analysis through the frequency consistency judgment unit, performs frequency adjustment operation using the data interpolation processing unit, and outputs standardized multi-source data through the spatiotemporal alignment unit. The intelligent monitoring and analysis module receives the standardized multi-source data, obtains environmental feature parameters through the feature extraction unit, generates environmental fusion data using the multi-source data fusion unit, and outputs the optimal monitoring scheme through the anomaly monitoring algorithm matching unit. The early warning decision generation module receives the environmental fusion data, performs anomaly detection through the dynamic threshold analysis unit, determines the early warning level through the risk assessment unit, and outputs early warning decision information through the early warning information generation unit. The human-computer interaction display module receives the early warning decision information, generates an environmental monitoring view through the visualization display unit, pushes early warning information through the early warning notification unit, and receives user commands through the interactive feedback unit. The system optimization and management module receives all data during the monitoring process, analyzes the system's operating status through the performance evaluation unit, adjusts the system configuration using the parameter optimization unit, and outputs system optimization solutions through the maintenance management unit.

[0036] The operation steps of this system and method are as follows: First, multimodal data, including water quality parameters, meteorological data, and underwater video image data, are collected using multi-source sensors deployed in the cage aquaculture area. Water quality parameters include water temperature, pH, dissolved oxygen, ammonia nitrogen content, and salinity; meteorological data includes air temperature, humidity, wind speed, wind direction, and rainfall; underwater video image data involves visual information on fish activity, feeding behavior, and water turbidity. The multimodal data undergoes preprocessing and sampling frequency measurement. The collected data is transmitted to the aquaculture monitoring platform, where time-series analysis is used to generate frequency data for water quality parameters, meteorological data, and underwater video images. Simultaneously, frame extraction and image enhancement processing are performed on the underwater video image data to generate standardized underwater image data.

[0037] Next, based on the aforementioned frequency data, a consistency judgment process for the sampling frequency of multi-source sensor data is performed. By comparing the three types of frequency data pairwise, multiple sets of frequency comparison results are generated, and a comprehensive judgment is made to output a consistency result. If the frequencies are consistent, subsequent steps are directly executed; if they are inconsistent, a linear interpolation algorithm is used to adjust the sampling frequency of the data source with the lower frequency, generating adjusted multi-source sensor data. The frequency-consistent and adjusted multi-source sensor data are then subjected to spatiotemporal alignment and feature fusion processing. A unified data coordinate system is established, data is aligned according to timestamps and spatial coordinates, water quality characteristic parameters, meteorological characteristic parameters, and image visual characteristic parameters are extracted, and a feature-level fusion method is used to construct multi-source fused data of the cage aquaculture environment.

[0038] Based on this, anomaly monitoring algorithms are matched using multi-source fusion data of the cage aquaculture environment and pre-set aquaculture environment assessment standard data. By establishing an algorithm library and standard environmental data patterns, similarity matching is performed to select the optimal target anomaly monitoring algorithm, and type feature data containing algorithm identifiers and parameter configuration information is generated. Real-time anomaly monitoring and early warning processing of the cage aquaculture environment is then performed based on the target anomaly monitoring algorithm type feature data. The corresponding anomaly monitoring program is invoked, feature parameters are extracted, and a dynamic threshold detection method is used to determine anomalies, generating early warning information including anomaly type, level, time, and location, and initiating the corresponding early warning processing flow.

[0039] Finally, the aquaculture monitoring platform visualizes and pushes early warning information on abnormal cage aquaculture environments. Early warning information is displayed on a map interface, and notifications are sent via SMS, mobile application push notifications, and email. Anomaly handling suggestions are generated, and the entire early warning handling process is recorded to form a monitoring and early warning log. The corresponding monitoring system includes an environmental data acquisition module, a data processing and coordination module, an intelligent monitoring and analysis module, an early warning decision generation module, a human-computer interaction display module, and a system optimization and management module. These modules work together to implement the above methods and processes, ensuring the reliability and effectiveness of cage aquaculture environmental monitoring.

[0040] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0041] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring the cage aquaculture environment based on multi-source sensors, characterized in that: The method includes the following steps: S1. Collect multimodal data of the cage aquaculture environment through multi-source sensors. The multimodal data includes water quality parameter data, meteorological environment data, and underwater video image data. S2. Perform data preprocessing and sampling frequency measurement on the multimodal data to generate water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data; S3. Based on water quality parameter frequency data, meteorological environment frequency data and underwater video image frequency data, perform multi-source sensor data sampling frequency consistency judgment processing, generate multi-source sensor data sampling frequency consistency judgment result, and if the frequencies are consistent, execute step S5. S4. When the frequencies are inconsistent, use a linear interpolation algorithm to adjust the sampling frequency of the data source with the lower frequency to generate adjusted multi-source sensor data. S5. Perform spatiotemporal alignment and feature fusion processing on the frequency-consistent and adjusted multi-source sensor data to construct multi-source fusion data of the cage aquaculture environment; S6. Based on the multi-source fusion data of the cage aquaculture environment and the preset aquaculture environment assessment standard data, perform anomaly monitoring algorithm matching processing to generate target anomaly monitoring algorithm type feature data; S7. Based on the target anomaly monitoring algorithm type characteristic data, perform real-time anomaly monitoring and early warning processing of the cage aquaculture environment, and generate anomaly early warning information for the cage aquaculture environment. S8. Visualize and push early warning information on abnormal cage aquaculture environments through the aquaculture supervision platform.

2. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The process of collecting multimodal data on the cage aquaculture environment in S1 includes the following steps: S11. Collect water quality parameter data, including water temperature, pH value, dissolved oxygen, ammonia nitrogen content and salinity parameters, by deploying water quality sensors in the cage aquaculture area; S12. Collect meteorological environmental data, including air temperature, humidity, wind speed, wind direction and rainfall parameters, by means of meteorological sensors deployed in the cage aquaculture area; S13. Underwater video image data, including fish activity status, feeding behavior and water turbidity visual information, are collected by underwater camera equipment deployed in the cage aquaculture area.

3. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The data preprocessing and sampling frequency measurement in S2 include the following steps: S21. Transmit the collected water quality parameter data, meteorological environment data, and underwater video image data to the aquaculture supervision platform respectively; S22. Use time series analysis to measure the sampling frequency of various types of data to generate water quality parameter frequency data, meteorological environment frequency data and underwater video image frequency data; S23. Perform frame extraction and image enhancement processing on the underwater video image data to generate standardized underwater image data.

4. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The process for determining the consistency of multi-source sensor data sampling frequency in S3 includes the following steps: S31. Acquire water quality parameter frequency data, meteorological environment frequency data, and underwater video image frequency data; S32. Perform pairwise comparisons of the three types of frequency data to generate multiple sets of frequency comparison results; S33. Based on a comprehensive judgment of multiple frequency comparison results, generate a consistency judgment result for the sampling frequency of multi-source sensor data: When the sampling frequency of all data sources is consistent, the output frequency consistency judgment result is consistent; When any two data sources have inconsistent sampling frequencies, the output frequency consistency judgment result is inconsistent.

5. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The sampling frequency adjustment in S4 includes the following steps: S41. When the frequency consistency judgment result is inconsistent, identify the data source with the lowest sampling frequency; S42. Use a linear interpolation algorithm to boost the frequency of the data source with the lowest sampling frequency. S43. Generate adjusted multi-source sensor data, including adjusted water quality parameter data, meteorological environment data, and underwater video image data.

6. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The spatiotemporal alignment and feature fusion processing in S5 includes the following steps: S51. Establish a unified data coordinate system with timestamps and spatial coordinates as references; S52. Align the multi-source sensor data with consistent frequency and adjustment according to timestamps and spatial coordinates; S53. Extract feature parameters from various types of data, including water quality feature parameters, meteorological feature parameters, and image visual feature parameters; S54. The feature-level fusion method is used to fuse multi-source feature parameters into unified multi-source fusion data of cage aquaculture environment.

7. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The anomaly detection algorithm matching process in S6 includes the following steps: S61. Establish a database of algorithms for monitoring aquaculture environment anomalies, including algorithms for monitoring water quality anomalies, meteorological anomalies, biological behavior anomalies, and comprehensive assessment and monitoring algorithms. S62. Construct a standard environmental data model corresponding to each anomaly detection algorithm; S63. Perform similarity matching between multi-source fusion data of cage aquaculture environment and various standard environmental data patterns; S64. Select the anomaly detection algorithm corresponding to the standard environmental data pattern with the highest similarity as the target anomaly detection algorithm. S65. Generate target anomaly monitoring algorithm type feature data, including algorithm identifier and parameter configuration information.

8. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The real-time anomaly monitoring and early warning processing in S7 includes the following steps: S71. Call the corresponding anomaly monitoring program based on the target anomaly monitoring algorithm type characteristic data; S72. Extract feature parameters from multi-source fusion data of cage aquaculture environment; S73. Use dynamic threshold detection method to judge the anomalies of environmental parameters; S74. When an environmental anomaly is detected, generate an early warning message for an abnormal cage aquaculture environment, including the anomaly type, anomaly level, time of occurrence, and location. S75. Activate the corresponding early warning and handling process according to the anomaly level.

9. The method for monitoring the cage aquaculture environment based on multi-source sensors according to claim 1, characterized in that: The visualization and early warning notification push in S8 include the following steps: S81. Visualize early warning information about abnormal cage aquaculture environment on the map interface of the aquaculture supervision platform; S82. Send early warning notifications to aquaculture managers through various communication methods, including SMS, mobile application push and email; S83. Generate abnormal handling suggestions, including oxygenation suggestions, feeding adjustment suggestions, and emergency handling suggestions; S84. Record the entire process of early warning handling and form a monitoring and early warning log for cage aquaculture environment.

10. A cage aquaculture environment monitoring system based on multi-source sensors, used to implement the cage aquaculture environment monitoring method based on multi-source sensors as described in any one of claims 1-9, characterized in that: The system includes: The environmental data acquisition module uses a multi-source sensor unit to collect multimodal environmental data, performs data standardization processing through a sensor calibration unit, and outputs the sampling frequency information of each data source through a frequency measurement unit. The data processing coordination module receives the sampling frequency information, performs data coordination analysis through the frequency consistency judgment unit, performs frequency adjustment operation using the data interpolation processing unit, and outputs standardized multi-source data through the spatiotemporal alignment unit. The intelligent monitoring and analysis module receives the standardized multi-source data, obtains environmental feature parameters through the feature extraction unit, generates environmental fusion data using the multi-source data fusion unit, and outputs the optimal monitoring scheme through the anomaly monitoring algorithm matching unit. The early warning decision generation module receives the environmental fusion data, performs anomaly detection through the dynamic threshold analysis unit, determines the early warning level through the risk assessment unit, and outputs early warning decision information through the early warning information generation unit. The human-computer interaction display module receives the early warning decision information, generates an environmental monitoring view through the visualization display unit, pushes early warning information through the early warning notification unit, and receives user commands through the interactive feedback unit. The system optimization and management module receives all data during the monitoring process, analyzes the system's operating status through the performance evaluation unit, adjusts the system configuration using the parameter optimization unit, and outputs system optimization solutions through the maintenance management unit.

Images