Fish swimming posture wave node identification method and system and electronic device

Abstract

The application provides a fish swimming posture wave node recognition method and system and an electronic device, relates to the technical field of posture wave node recognition, and comprises the following steps: collecting data of a fish swimming process to obtain fish swimming data; performing posture recognition and wave node extraction according to the fish swimming data to obtain wave node recognition features; performing fish body posture analysis based on the wave node recognition features to obtain fish body posture results; performing fish body posture stability monitoring in combination with the fish body posture results, and outputting an abnormal early warning based on a monitoring result. Through the application, the technical problem that, in the prior art, the image quality is affected by factors such as water light, transparency and water flow, resulting in low recognition accuracy and poor real-time performance and further affecting the accuracy and efficiency of fish posture recognition can be solved, fish posture recognition technology based on multi-source data fusion and advanced algorithms is achieved, and the technical effects of improving recognition accuracy, enhancing real-time performance and adaptability are achieved.

Description

Technical Field

[0001] This application relates to the field of attitude node recognition technology, and in particular to methods, systems and electronic devices for fish swimming attitude node recognition. Background Technology

[0002] With the deepening of aquaculture and underwater biology research, the monitoring and analysis of fish swimming behavior has gradually become an important technical requirement. Fish movement patterns not only significantly influence their growth, development, and reproductive capacity, but are also closely related to environmental factors such as water flow and temperature. Therefore, how to accurately and in real-time monitor the swimming posture, movement patterns, and stability of fish has become a key technical issue in the field of aquatic science. Existing technologies mainly collect data through video monitoring and sensor acquisition, but these methods still have many shortcomings.

[0003] Currently, while video image-based fish posture recognition methods have made some progress in the field of visual recognition, the complexity of image processing, especially in complex underwater environments, means that image quality is often affected by factors such as water light, transparency, and current, leading to low recognition accuracy and poor real-time performance. Furthermore, traditional posture recognition methods largely rely on manually set calibration data and template matching, resulting in insufficient adaptability and flexibility to cope with different environmental conditions and fish species diversity. Most existing technologies lack comprehensive modeling of fish movement, particularly in node extraction and posture stability monitoring, failing to achieve accurate, real-time, and comprehensive recognition and feedback.

[0004] In summary, existing technologies suffer from low recognition accuracy and poor real-time performance due to the influence of factors such as water light, transparency, and water flow on image quality, which further affects the accuracy and efficiency of fish posture recognition. Summary of the Invention

[0005] The purpose of this application is to provide a method, system, and electronic device for fish swimming posture node recognition, in order to solve the technical problems in the prior art where the recognition accuracy is low and the real-time performance is poor due to the influence of factors such as water light, transparency, and water flow on image quality, which further affects the accuracy and efficiency of fish posture recognition.

[0006] In view of the above problems, this application provides a method, system and electronic device for identifying nodes of fish swimming posture.

[0007] In a first aspect, this application provides a method for identifying nodes in fish swimming postures, implemented through a fish swimming posture node identification system, comprising: collecting data on the fish swimming process to obtain fish swimming data; performing posture identification and node extraction based on the fish swimming data to obtain node identification features; performing fish posture analysis based on the node identification features to obtain fish posture results; and monitoring fish posture stability in conjunction with the fish posture results, and outputting anomaly warnings based on the monitoring results.

[0008] Secondly, this application also provides a fish swimming posture node recognition system for performing the fish swimming posture node recognition method as described in the first aspect, comprising: a data acquisition module for acquiring data on the fish swimming process to obtain fish swimming data; a node extraction module for performing posture recognition and node extraction based on the fish swimming data to obtain node recognition features; a posture analysis module for performing fish posture analysis based on the node recognition features to obtain fish posture results; and a stability monitoring module for monitoring fish posture stability in conjunction with the fish posture results and outputting anomaly warnings based on the monitoring results.

[0009] Thirdly, this application also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the fish swimming posture node recognition method described in any of the first aspects above.

[0010] The technical solution provided in this application has at least the following technical effects or advantages: by collecting data on the swimming process of fish, fish swimming data is obtained; based on the fish swimming data, attitude recognition and node extraction are performed to obtain node recognition features; based on the node recognition features, fish posture analysis is performed to obtain fish posture results; combined with the fish posture results, fish posture stability is monitored, and anomaly warnings are output based on the monitoring results. In other words, by implementing fish posture recognition technology based on multi-source data fusion and advanced algorithms, the technical effects of improving recognition accuracy, enhancing real-time performance and adaptability, and enabling stable operation under different environmental conditions are achieved.

[0011] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0013] Figure 1 This is a flowchart illustrating the fish swimming posture node recognition method of this application;

[0014] Figure 2 This is a schematic diagram of the fish swimming posture node recognition system of this application;

[0015] Figure 3 This is a schematic diagram of the structure of an exemplary electronic device of this application.

[0016] Figure reference numerals: Data acquisition module 11, node extraction module 12, attitude analysis module 13, stability monitoring module 14, bus 300, receiver 301, processor 302, transmitter 303, memory 304, bus interface 305. Detailed Implementation

[0017] This application provides a method, system, and electronic device for fish swimming posture node recognition, solving the technical problems in existing technologies where low recognition accuracy and poor real-time performance are caused by factors such as water light, transparency, and water flow affecting image quality, further impacting the accuracy and efficiency of fish posture recognition. It achieves fish posture recognition technology based on multi-source data fusion and advanced algorithms, resulting in improved recognition accuracy, enhanced real-time performance and adaptability, and stable operation under different environmental conditions.

[0018] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0019] Example 1, please refer to the appendix. Figure 1 This application provides a method for identifying nodes in fish swimming postures, which is applied to a fish swimming posture node identification system, and specifically includes the following steps:

[0020] Step 1: Collect data on the swimming process of fish to obtain fish swimming data.

[0021] Specifically, this involves acquiring relevant information about fish swimming using various devices and technologies. This data includes the fish's movement trajectory, speed, and posture changes. For example, during the data collection process, high-speed cameras may be used to record images of the fish's movement, or underwater sensors may be used to measure the fish's swimming speed and water flow information. The purpose of acquiring fish swimming data is to gain a deeper understanding of the fish's swimming patterns and behavioral characteristics. For instance, when fish swim, their speed may vary between three and five meters per second; capturing these changes helps analyze whether the fish are swimming fast or slow. Continuously collecting this data provides an important foundation for subsequent motion analysis and posture assessment.

[0022] Step 2: Perform attitude recognition and node extraction based on the fish swimming data to obtain node recognition features.

[0023] Specifically, the aforementioned data, including the fish's trajectory, speed, and acceleration, are used to analyze the fish's posture in the water. Posture recognition refers to analyzing the fish's movement patterns to determine its current swimming state, such as whether it is swimming in a serpentine pattern or at a constant speed. Through these analyses, nodes related to the fish's body undulations, or nodes, can be extracted. Nodes are the points where the amplitude of the undulation reaches its maximum or minimum during the fish's swimming process; these nodes are closely related to the degree of curvature and movement pattern of the fish. For example, in serpentine swimming, the fish's nodes will appear as continuous bending points, which can be extracted through data analysis. Ultimately, through the results of posture recognition and node extraction, the node identification characteristics of fish swimming can be obtained. These characteristics include the frequency, amplitude, and position of the nodes, providing basic data for further research on fish swimming efficiency and behavior.

[0024] Step 3: Perform fish posture analysis based on the node recognition features to obtain the fish posture results.

[0025] Specifically, the system utilizes identified node features (such as node position, frequency, and amplitude) to analyze the overall posture of fish during swimming. Fish posture analysis refers to the assessment and interpretation of the fish's movement in the water. For example, a high node frequency, large amplitude, and strong curvature may indicate that the fish is performing a relatively fast and significantly curved serpentine swimming motion. Conversely, a low node frequency and small amplitude may suggest that the fish is moving more smoothly and at a constant speed. By analyzing these changes in node features, the system can accurately infer the fish's posture, thus obtaining posture results. These results can help analyze characteristics such as fish swimming efficiency and stability, further providing data support for research.

[0026] Step 4: Monitor the stability of the fish's posture based on the fish posture results, and output anomaly warnings based on the monitoring results.

[0027] Specifically, by analyzing the fish's posture, the system can monitor the stability of the fish's movement in the water in real time. Fish posture stability refers to whether the fish maintains a constant state of motion during swimming; for example, frequent, abrupt changes or irregular fluctuations may indicate posture instability. By analyzing characteristics such as the fish's wave node frequency, amplitude, and curvature, the system can determine whether the fish's movement is smooth. If the fish's posture is within a set stable range, the monitoring system considers it stable; if it exceeds the normal fluctuation range, the system will warn of potential anomalies.

[0028] When monitoring and analysis detect abnormal fluctuations in the fish's posture, such as frequency, amplitude, or curvature exceeding normal ranges, the system will automatically generate an early warning signal to alert relevant personnel. For example, if the frequency of the fish's nodes increases abnormally or the amplitude increases significantly, it indicates that the fish's swimming state may be abnormal, such as rapid swimming or struggling, triggering an alarm for further observation or adjustment.

[0029] The described fish swimming posture node recognition method is applied to a fish swimming posture node recognition system. It can realize fish posture recognition technology based on multi-source data fusion and advanced algorithms, thereby improving recognition accuracy, enhancing real-time performance and adaptability, and enabling stable operation under different environmental conditions.

[0030] Furthermore, this application also includes: acquiring fish swimming videos using a high-speed camera and extracting fish swimming images; capturing fish swimming sensor data using an underwater sensor; detecting water flow change data of fish coordinates based on an optical current velocity sensor; and obtaining fish swimming data by combining the fish swimming images, the fish swimming sensor data, and the water flow change data.

[0031] Specifically, high-speed cameras are used to capture video of fish swimming and extract images of the fish swimming. A high-speed camera is a device that can capture a large amount of footage in a short time and store it quickly; it is typically used to capture fast-moving objects. In this process, the video of fish swimming captured by the camera is converted into still images. These images provide the basic data for observing the fish's posture, movement trajectory, and other characteristics. By analyzing these images, we can further understand the fish's movement patterns and swimming states in the water, such as whether there are undulating movements or serpentine swimming patterns.

[0032] Next, underwater sensors are used to capture data on fish swimming. Underwater sensors are specialized devices used in water to measure environmental parameters such as water flow velocity, temperature, and pressure. Here, the sensors are primarily used to acquire dynamic data about the fish's swimming process, including the fish's movement speed and angle changes. This sensor data provides reliable real-time data for subsequent attitude analysis and behavior classification. For example, the sensors might measure the fish's swimming speed to be between five and eight meters per second, helping to identify whether the fish is swimming fast or slow.

[0033] Then, optical current sensors are used to detect changes in water flow at the fish's coordinates. An optical current sensor is a device that calculates flow velocity by analyzing the movement of particles in the water, accurately capturing the flow state of a body of water. In this case, the optical current sensor helps detect the changes in water flow experienced by the fish as it swims, providing more detailed dynamic information. This data on water flow changes reveals the interaction between the fish and the surrounding water flow, thus affecting the efficiency and manner of the fish's swimming. For example, if a current velocity of two meters per second is detected, the fish's swimming trajectory may be significantly affected, causing changes in its swimming posture.

[0034] By integrating images, sensor data, and water flow change data obtained from different sensors and devices, a comprehensive analysis of fish swimming behavior and its interaction with the environment can be achieved. For example, by combining the movement trajectory of a fish in an image with changes in water flow, it is possible to determine whether the fish's posture is affected by the water flow, or whether its swimming style (such as serpentine swimming) is as expected. This fusion approach makes the monitoring of fish behavior more accurate and comprehensive.

[0035] Furthermore, this application also includes: constructing a geometric model based on the fish swimming sensor data to obtain a fish body geometric model; performing an interaction analysis between the fish body and the water flow based on the fish body geometric model and the water flow change data to obtain interaction data; obtaining fish body motion data through the interaction data, and establishing a wave propagation model based on the fish body motion data; calculating the wave path in conjunction with the wave propagation model, and obtaining fish body wave nodes through wave path analysis; verifying the fish body wave nodes based on the fish swimming image, and if the verification is successful, obtaining the wave node recognition features.

[0036] Specifically, a geometric model is constructed based on the fish swimming sensor data to obtain a fish body geometric model. The fish swimming sensor data includes information such as the fish's position, velocity, and orientation in the water. This data helps to build a mathematical model describing the fish's geometric morphology. The geometric model typically represents the fish's shape in three-dimensional space using mathematical formulas or computer graphics algorithms. For example, the system can use this data to construct a fish model in a three-dimensional coordinate system, reflecting the fish's size, shape, and dynamic characteristics. This geometric model provides a clear understanding of the fish's posture and movement in the water, laying the foundation for subsequent analysis.

[0037] Next, based on the fish geometry model and the water flow variation data, an interaction analysis of the fish and the water flow is performed to obtain interaction data. The water flow variation data provides the flow state of the water body, including water flow velocity and direction, while the fish geometry model describes the shape and size of the fish. By combining this information, analyzing the interaction between the fish and the water flow can reveal how the fish is affected by the water flow and how it adjusts its swimming posture. For example, if the water flow velocity reaches three meters per second, the fish's swimming trajectory may be significantly affected by the water flow, causing the fish to adjust its swimming pattern to adapt to the environment. Through this analysis, the obtained interaction data helps to further understand the fish's movement performance under different water flow conditions.

[0038] Then, fish motion data is obtained through the interaction data, and a wave propagation model is established based on this data. The interaction data provides the dynamic relationship between the fish and the water flow, reflecting details of the fish's motion, such as acceleration and changes in direction. Using this motion data, a wave propagation model can be built to describe the wave motion of the fish in the water. The wave propagation model simulates the propagation process of water waves generated by the fish as it swims, predicting how waves propagate in the water and affect the fish's posture. This model helps us understand how fish use wave propagation to generate propulsion and how they adjust their posture and speed through wave motion while swimming.

[0039] Next, the wave propagation model is used to calculate the wave path, and wave path analysis is used to obtain the wave nodes of the fish body. The wave path refers to the trajectory of wave propagation generated by the fish body during swimming in water, which is usually related to the fish's swimming pattern and water flow conditions. By calculating the wave path, we can further analyze how waves affect the fish's movement, while wave nodes refer to the parts of the fish body that exhibit nodal behavior during wave propagation, usually the locations where the fish's curvature is at its maximum or minimum. Through these nodes, we can identify the wave patterns of the fish body under different swimming states and provide a basis for further motion optimization.

[0040] Finally, the fish body wave nodes are verified based on the fish swimming images. If the verification is successful, the node recognition features are obtained. Fish swimming images are image data acquired through devices such as high-speed cameras, providing information on the fish's posture and movement in the water. By comparing these images with the previously calculated wave nodes, the system can determine whether the wave nodes match the actual situation. If the verification is successful, the node recognition features can be obtained, including data such as the frequency and amplitude of the fish body waves. Verification ensures the accuracy of the established wave propagation model and allows for further optimization of the analysis of fish movement.

[0041] Furthermore, this application also includes: dividing the fish swimming image into parts by image segmentation, contour recognition, and key point detection to obtain fish body part images; calculating fish node positions based on the fish body geometric features and fish movement patterns to obtain fish node positions; assigning fish node positions to the fish body part images to obtain fish node position images; identifying fish body undulation nodes through the fish node position images, and obtaining a verification pass result if the recognition accuracy meets the recognition threshold.

[0042] Specifically, image processing techniques are used to analyze images of swimming fish. Image segmentation divides the image into regions to analyze different parts of the fish body separately; contour recognition helps extract the outer edges of the fish's shape, thus determining its contour; keypoint detection identifies important points in specific parts of the fish (such as the head, tail, or abdomen) as the basis for analysis and calculation. Through these techniques, images of different parts of the fish can be clearly extracted. For example, segmentation techniques can be used to obtain images of the fish's tail, laying the foundation for subsequent pose recognition and motion analysis.

[0043] The system calculates the positions of fish nodes based on their geometric features and movement patterns. Fish geometric features refer to the shape, size, and relative positions of the fish's parts, while movement patterns describe the fish's swimming behavior, such as uniform swimming or serpentine swimming. By combining these geometric features with movement patterns, the system can calculate the positions of each node during swimming. A node is the point of maximum or minimum undulation during swimming, such as the extreme point of bending during serpentine swimming. Through calculation, the positions and distribution of nodes during swimming can be determined, providing positional data for further analysis.

[0044] The fish's node positions are assigned to the images of the fish's body parts to obtain images of the fish's node positions. Assignment refers to applying the previously calculated fish node positions to images of various parts of the fish's body. By mapping each part to its corresponding node position, an image containing node information can be generated. This method allows for a more intuitive view of the relationship between each part and its node position. For example, if the fish's belly has the greatest curvature during serpentine swimming, then the image of that part will display the node position, clearly showing the undulation of each part.

[0045] The fish's body undulation nodes are identified using the fish's node position image. If the identification accuracy meets the identification threshold, the verification result is obtained. The core of this step lies in identifying undulation nodes using the previously generated fish node position image. Evasion nodes are the points where the fish's curvature reaches its maximum or minimum during swimming; identifying these nodes helps in understanding the fish's swimming posture and patterns. The identification accuracy is the standard for measuring the effectiveness of this process. If the accuracy reaches the preset identification threshold, it means that the system has successfully identified the fish's body undulation nodes, thus verifying the effectiveness of the entire identification process.

[0046] Furthermore, this application also includes: calculating frequency, amplitude, and curvature based on the node identification features to obtain node frequency, node amplitude, and node curvature; inputting the node frequency, node amplitude, and node curvature into the fish posture judgment model to output the fish posture result.

[0047] Specifically, based on the node identification features, frequency, amplitude, and curvature are calculated to obtain node frequency, node amplitude, and node curvature. This step quantifies the identified node features. Frequency refers to the number of times the fish completes a wave cycle per unit time, reflecting the fish's swimming rhythm. For example, if the fish completes five waves per second, its frequency is five times per second. Amplitude refers to the amplitude of the fish's wave during swimming, usually indicating the degree of curvature during the wave. The larger the amplitude, the greater the curvature of the fish, and the stronger the swimming force may be. For example, a node with an amplitude of three centimeters may indicate a large degree of curvature, while a node with an amplitude of one centimeter represents a smaller curvature. Curvature refers to the degree of curvature of a certain part of the fish's body; specifically, it refers to the angle of curvature. The larger the angle, the more pronounced the curvature. Through these calculations, the wave characteristics of the fish can be quantified in detail, thus providing accurate data for subsequent analysis.

[0048] The node frequency, node amplitude, and node curvature are input into the fish posture judgment model, which outputs the fish posture result. The fish posture judgment model is a mathematical or computer model that uses previously calculated data such as frequency, amplitude, and curvature to analyze the overall posture of the fish. The posture judgment model may include machine learning algorithms or rule-based inference systems. Based on the input node data, the model can predict the fish's swimming pattern and determine whether it is in a serpentine swimming, uniform swimming, or other posture. For example, if the node frequency is high, the amplitude is large, and the curvature is within a certain range, the model may determine that the fish is in a rapid serpentine swimming state. In this way, the system can output the fish's motion posture in real time, helping researchers understand the swimming characteristics and behavioral patterns of fish.

[0049] Furthermore, this application also includes: constructing input data based on fish body node information and fish swimming patterns; dividing the input data into training data and validation data; training a pre-constructed fish posture judgment model architecture using the training data; and validating the fish posture judgment model architecture using the validation data; if the output accuracy of the fish posture judgment model architecture meets the output accuracy threshold, the construction of the fish posture judgment model is completed.

[0050] Specifically, the model uses node information (such as node position, frequency, amplitude, and curvature) and swimming patterns (such as serpentine swimming and uniform swimming) identified during the fish's swimming process as input data. Node information reflects the pattern of the fish's movement in the water, while the swimming pattern describes the overall motion of the fish. This information provides key features for further attitude determination, and the construction of the input data is the foundation for model training.

[0051] The input data is divided into two parts: training data and validation data. Training data is used to train the model architecture; the model learns how to recognize fish poses by analyzing features in the training data. Validation data is used to evaluate the model's performance during training, ensuring that the model can make effective predictions on unknown data. Training data helps the model adjust its internal parameters, while validation data is used to test whether the model can accurately determine the fish's pose.

[0052] After training and validation, the model's output accuracy reached the preset threshold. When the model can make correct predictions on the test data and the accuracy is high enough, the model can be considered to have been successfully built and can stably determine the fish's posture.

[0053] Furthermore, this application also includes: obtaining a preset fish body node threshold for the fish body posture result; determining whether the preset fish body node threshold is satisfied by the node frequency, node amplitude, and node curvature; if satisfied, obtaining a monitoring stability result.

[0054] Specifically, when analyzing fish posture, a standard value or threshold is first set. This threshold represents the expected range of certain characteristics of the fish's nodes (such as frequency, amplitude, and curvature) under specific conditions. The threshold is a reference value used to measure whether the fish's undulation characteristics are normal. For example, assuming the node frequency threshold is five times per second, the amplitude threshold is three centimeters, and the curvature threshold is thirty degrees, then during the analysis, these thresholds serve as standards for judging whether the fish's posture is stable or normal.

[0055] The system compares the identified node characteristics (such as frequency, amplitude, and curvature) with preset thresholds. The determination process involves comparing the actual measured values ​​with the preset values ​​to decide whether the threshold requirements are met. For example, if the actual measured frequency is four times per second, the amplitude is four centimeters, and the curvature is forty-five degrees, the system will compare these values ​​with the preset thresholds to determine if they meet the predetermined standards.

[0056] If the node characteristics of the fish's body meet the preset thresholds, the system will determine that the fish's movement is stable, thus obtaining a stable monitoring result. This result reflects whether the fish's swimming is normal and stable. For example, when the frequency, amplitude, and curvature are all within the set threshold range, the system will consider the fish's posture to be stable and obtain a "stable" result; conversely, if these characteristics exceed the set range, it may mean that the fish's posture is unstable, and the result will be "unstable".

[0057] In summary, the fish swimming posture node recognition method provided in this application has the following technical effects: It acquires fish swimming data by collecting data on the fish swimming process; it performs posture recognition and node extraction based on the fish swimming data to obtain node recognition features; it analyzes the fish posture based on the node recognition features to obtain fish posture results; and it monitors the stability of the fish posture in conjunction with the fish posture results, outputting anomaly warnings based on the monitoring results. In other words, by implementing fish posture recognition technology based on multi-source data fusion and advanced algorithms, it achieves the technical effects of improving recognition accuracy, enhancing real-time performance and adaptability, and enabling stable operation under different environmental conditions.

[0058] Example 2: Based on the same inventive concept as the fish swimming posture node recognition method in the foregoing examples, this application also provides a fish swimming posture node recognition system. Please refer to the appendix. Figure 2 The system includes: a data acquisition module 11, which is used to acquire data on the swimming process of fish and obtain fish swimming data; a node extraction module 12, which is used to perform attitude recognition and node extraction based on the fish swimming data and obtain node recognition features; an attitude analysis module 13, which is used to perform fish attitude analysis based on the node recognition features and obtain fish attitude results; and a stability monitoring module 14, which is used to monitor the stability of fish attitude in conjunction with the fish attitude results and output abnormal warnings based on the monitoring results.

[0059] Furthermore, the fish swimming posture node recognition system is also used for: acquiring fish swimming videos through a high-speed camera and extracting fish swimming images; capturing fish swimming sensor data based on underwater sensors; detecting water flow change data of fish coordinates based on an optical current velocity sensor; and obtaining fish swimming data by combining the fish swimming images, the fish swimming sensor data, and the water flow change data.

[0060] Furthermore, the fish swimming posture node recognition system is also used for: constructing a geometric model based on the fish swimming sensor data to obtain a fish body geometric model; performing interaction analysis between the fish body and the water flow based on the fish body geometric model and the water flow change data to obtain interaction data; obtaining fish body motion data through the interaction data, and establishing a wave propagation model based on the fish body motion data; calculating the wave path in combination with the wave propagation model, and obtaining fish body wave nodes through wave path analysis; verifying the fish body wave nodes based on the fish swimming image, and if the verification is successful, obtaining the node recognition feature.

[0061] Furthermore, the fish swimming posture node recognition system is also used for: dividing the fish swimming image into parts through image segmentation, contour recognition, and key point detection to obtain fish body part images; calculating fish node positions based on the fish body geometric features and fish movement patterns to obtain fish node positions; assigning fish node positions to the fish body part images to obtain fish node position images; and recognizing the fish body undulation nodes through the fish node position images. If the recognition accuracy meets the recognition threshold, a verification pass result is obtained.

[0062] Furthermore, the fish swimming posture node recognition system is also used to: calculate frequency, amplitude and curvature based on the node recognition features to obtain node frequency, node amplitude and node curvature; input the node frequency, node amplitude and node curvature into the fish posture judgment model, and output the fish posture result.

[0063] Furthermore, the fish swimming posture node recognition system is also used for: constructing input data based on fish body node information and fish swimming patterns; dividing the input data into training data and validation data; training a pre-constructed fish posture judgment model architecture using the training data; and validating the fish posture judgment model architecture using the validation data; if the output accuracy of the fish posture judgment model architecture meets the output accuracy threshold, the construction of the fish posture judgment model is completed.

[0064] Furthermore, the fish swimming posture node recognition system is also used to: obtain a preset fish body node threshold for the fish posture result; determine whether the preset fish body node threshold is met by the node frequency, node amplitude and node curvature; if it is met, obtain the monitoring stability result.

[0065] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The fish swimming posture node recognition method and specific examples in the aforementioned embodiment one are also applicable to the fish swimming posture node recognition system in this embodiment. Through the foregoing detailed description of the fish swimming posture node recognition method, those skilled in the art can clearly understand the fish swimming posture node recognition system in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0066] Example 3: Based on the inventive concept of the fish swimming posture node recognition method in the foregoing embodiments, this application also provides an electronic device, including: at least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the fish swimming posture node recognition method described in any one of the above Examples 1.

[0067] Appendix Figure 3 This is a schematic diagram of the structure of an exemplary electronic device of this application. Figure 3 In this document, the bus architecture is represented by bus 300. Bus 300 may include any number of interconnected buses and bridges, and bus 300 connects various circuits including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 305 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, i.e., a transceiver, providing a unit for communicating with various other devices over a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 can be used to store data used by processor 302 during operation.

[0068] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0069] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for identifying nodes in fish swimming posture, characterized in that, include: Data is collected on the swimming process of fish to obtain fish swimming data; Based on the fish swimming data, attitude recognition and node extraction are performed to obtain node recognition features; Fish posture analysis is performed based on the node recognition features to obtain fish posture results; Combine the fish posture results to monitor the stability of the fish posture, and output anomaly warnings based on the monitoring results; The process of collecting data on the swimming process of fish to obtain fish swimming data includes: High-speed cameras are used to capture videos of fish swimming and extract images of the fish swimming. Capture fish swimming sensor data based on underwater sensors; Water flow change data based on the detection of fish coordinates using an optical flow velocity sensor; The fish swimming data is obtained by combining the fish swimming images, the fish swimming sensor data, and the water flow change data. The step of performing attitude recognition and node extraction based on the fish swimming data to obtain node recognition features includes: A geometric model is constructed based on the fish swimming sensor data to obtain a fish body geometric model; Based on the fish body geometric model and the water flow change data, the interaction between the fish body and the water flow is analyzed to obtain interaction data; Fish movement data is obtained through the interaction data, and a wave propagation model is established based on the fish movement data. The wave propagation model is used to calculate the wave path, and the wave path analysis is used to obtain the wave nodes of the fish body. The fish body undulation nodes are verified based on the fish swimming image. If the verification is successful, the node recognition features are obtained.

2. The method for identifying nodes of fish swimming posture as described in claim 1, characterized in that, The verification of the fish body undulation nodes based on the fish swimming image includes: The fish swimming images are segmented into body parts by image segmentation, contour recognition and key point detection to obtain images of fish body parts. The fish node positions are calculated based on the fish's geometric features and movement patterns to obtain the fish node positions. The fish node positions are assigned based on the fish body part images to obtain fish node position images; The fish body wave nodes are identified by the fish node position image. If the identification accuracy meets the identification threshold, the verification result is obtained.

3. The method for identifying nodes of fish swimming posture as described in claim 1, characterized in that, The fish posture analysis based on the node recognition features to obtain the fish posture result includes: Based on the node identification features, frequency, amplitude, and curvature are calculated to obtain node frequency, node amplitude, and node curvature. The node frequency, node amplitude, and node curvature are input into the fish posture judgment model, and the fish posture result is output.

4. The method for identifying nodes of fish swimming posture as described in claim 3, characterized in that, The fish posture determination model includes: Input data is constructed based on fish body node information and fish swimming patterns; The input data is divided into training data and validation data. The pre-built fish posture judgment model architecture is trained using the training data, and the fish posture judgment model architecture is validated using the validation data. If the output accuracy of the fish posture judgment model architecture meets the output accuracy threshold, the construction of the fish posture judgment model is complete.

5. The method for identifying nodes of fish swimming posture as described in claim 3, characterized in that, The monitoring of fish posture stability based on the fish posture results includes: The preset fish body node threshold is used to obtain the fish body posture result; The preset fish body node threshold is determined by the node frequency, node amplitude, and node curvature. If the conditions are met, then obtain the stable monitoring results.

6. A fish swimming posture node recognition system, characterized in that, The steps for implementing the fish swimming posture node recognition method according to any one of claims 1 to 5 include: The data acquisition module is used to collect data on the swimming process of fish and obtain fish swimming data. A node extraction module is used to perform attitude recognition and node extraction based on the fish swimming data to obtain node recognition features. The attitude analysis module is used to perform fish attitude analysis based on the node recognition features and obtain fish attitude results. A stability monitoring module is used to monitor the stability of the fish's posture by combining the fish's posture results, and output anomaly warnings based on the monitoring results.

7. An electronic device, comprising: At least one processor; A memory that is communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the steps of the fish swimming posture node recognition method according to any one of claims 1 to 5.

Images