Fault detection method, system and device for intelligent water station

Abstract

The invention relates to the technical field of equipment fault detection, in particular to a fault detection method, system and device for an intelligent water station, the method provided by the invention is used in a device comprising a buoy body, a guide rail, a lifting module and a visual acquisition assembly, and whether an abnormal response exists or not is judged through combination of a position detection assembly and a first image sensor; and if the fault identification condition is satisfied, determining a target structure area of the buoy based on the response interruption point, controlling a second image sensor to collect a target image, and identifying a buoy fault type through structural feature extraction and a fault identification network. According to the invention, accurate cooperation of fault positioning and identification can be realized, and the detection efficiency and the identification accuracy are improved.

Description

Technical Field

[0001] The present application relates to the technical field of equipment fault detection, and in particular to a fault detection method, system and device for a smart water station. Background Art

[0002] In existing smart water station structures, to adapt to dynamic water level changes and ensure the continuous operation of water intake equipment, a buoy-type lifting structure is typically used to achieve real-time adjustment of the water intake depth. As the buoy rises and falls within the guide rails, its displacement status is often monitored by position detection components (such as encoders). However, over long-term operation or in complex water environments, the buoy may experience faults such as jamming, structural deformation, and guide rail misalignment. Traditional displacement detection methods cannot accurately identify the specific fault location and type.

[0003] Problems with existing technologies: Existing buoy monitoring systems are mostly based on fixed-angle, continuous monitoring methods. This approach has a single recognition logic and high computational complexity, making it difficult to achieve efficient and accurate fault identification. To address these issues, this application designs a fault detection method, system, and device for a smart water station. Summary of the Invention

[0004] The technical problem to be solved by the present application is to address the deficiencies of the existing technology and provide a fault detection method, system and device for a smart water station. The method proposed in the present application is used in a device including a buoy body, a guide rail, a lifting module and a visual acquisition component. The position detection component and the first image sensor are jointly used to determine whether there is an abnormal response; if the conditions for entering the fault identification are met, the target structural area of ​​the buoy is determined based on the response interruption point, the second image sensor is controlled to collect the target image, and the type of buoy fault is identified through structural feature extraction and fault identification network.

[0005] To achieve the above objectives, this application provides the following technical solutions:

[0006] A fault detection method for a smart water station is applied to a fault detection device. The fault detection device includes a lifting module, a visual acquisition component, a guide rail, and a buoy body. The lifting module is used to raise and lower the buoy body in the water in response to the water level of the water station to be detected. The visual acquisition component includes a first image sensor and a second image sensor. The first image sensor is configured in the guide rail at a position perpendicular to the buoy body, and the second image sensor is configured outside the buoy body. The method includes:

[0007] Controlling the lifting module to drive the buoy body to move up and down along the guide rail according to the acquired water level information;

[0008] Determining whether a condition for entering a fault identification phase is met based on the position detection component in the lifting module and the first image sensor;

[0009] If the condition is met, the buoy fault recognition logic is executed by the second image sensor to obtain the buoy fault type, wherein the buoy fault recognition logic locates the target area of ​​the buoy body based on the response interruption point, and performs structural feature extraction and fault type recognition processing on the target area.

[0010] The buoy fault identification logic includes:

[0011] Calculating a response interruption point based on feedback information from the lifting module during the current action cycle and detection information from the first image sensor;

[0012] determining a target area of ​​the buoy body according to the response interruption point, and obtaining a target image of the target area through the second image sensor;

[0013] Performing image enhancement, edge detection, and contour segmentation on the target image to obtain structural features;

[0014] The structural features are input into a preset fault identification network, and the structural features are processed by the fault identification network to output the buoy fault type, wherein the fault identification network is trained by using historical fault images of the buoy.

[0015] The step of calculating the response interruption point according to the feedback information of the lifting module in the current action cycle and the detection information of the first image sensor includes:

[0016] Obtaining the target displacement instruction and displacement feedback information of the lifting module in the current action cycle, and constructing the theoretical motion trajectory of the buoy body in the current action cycle;

[0017] Back-projecting the theoretical motion trajectory in image space according to the viewing angle parameter and calibration parameter of the first image sensor to obtain a theoretical change area of ​​the buoy top;

[0018] The theoretical change area is compared with the actual change area of ​​the buoy top in the detection information by the optical flow method, and the response trend change between consecutive image frames is detected to obtain the starting time of the interruption of the response of the buoy top image, and the control segment corresponding to the starting time is used as the response interruption point.

[0019] Determining a target area of ​​the buoy body according to the response interruption point includes:

[0020] Dividing the movement path of the buoy body along the guide rail into a plurality of preset structural sections, wherein each preset structural section corresponds one-to-one to a structural area of ​​the buoy body;

[0021] Determining a corresponding preset structural section according to the control section of the response interruption point;

[0022] The structural area of ​​the buoy body corresponding to the corresponding preset structural section is used as the target area.

[0023] The fault identification network includes a structure encoding layer, a state reasoning layer, and a fault classification output layer, wherein:

[0024] The structural coding layer is used to vectorize the structural features of the input, including edge continuity parameters, contour closure index and occlusion area ratio;

[0025] The state inference layer is provided with a graph structure attention mechanism for establishing spatial associations between structure encoding vectors and inferring the integrity state and force consistency characteristics of the buoy structure;

[0026] The fault classification output layer is used to output the fault type of the buoy through a Softmax classifier according to the integrity state and force consistency characteristics. The fault types include buoy stuck, guide rail stuck, bottom floating obstruction and buoy attitude tilt.

[0027] The lifting module includes a water level sensor, a position detection component and a lifting mechanism. The lifting module is controlled according to the acquired water level information to drive the buoy body to move up and down along the guide rail, including:

[0028] Acquiring water level information collected by the water level sensor;

[0029] Comparing the water level information with a preset target water extraction layer depth to determine the target position of the buoy body;

[0030] The lifting mechanism is controlled to move the buoy body to a height corresponding to the target position.

[0031] Determining whether a condition for entering a fault identification phase is met based on the position detection component in the lifting module and the first image sensor includes:

[0032] Acquiring first displacement data output by the position detection component;

[0033] Acquire a top image sequence captured by the first image sensor, wherein a capture period of the first image sensor is synchronously set with an action period of the lifting module;

[0034] performing edge detection on the top images in the top image sequence in sequence to obtain a contour feature sequence;

[0035] According to a preset spatial height mapping model, pixel size changes of the contour feature sequence in the image are converted into vertical displacement values ​​to obtain second displacement data, wherein the spatial height mapping model is constructed based on an image projection transformation relationship.

[0036] The conditions for entering the fault identification phase include at least one of the following:

[0037] An absolute value of a difference between the first displacement data and the second displacement data is greater than or equal to a preset error threshold;

[0038] The second displacement data does not change during the action cycle;

[0039] A changing direction of the second displacement data is opposite to a changing direction of the first displacement data.

[0040] A fault detection system for a smart water station is applied to a fault detection device. The fault detection device includes a lifting module, a visual acquisition component, a guide rail, and a buoy body. The lifting module is used to raise and lower the buoy body in the water in response to the water level of the water station to be detected. The visual acquisition component includes a first image sensor and a second image sensor. The first image sensor is configured in the guide rail at a position perpendicular to the buoy body, and the second image sensor is configured outside the buoy body. The system includes:

[0041] A water level control module is used to obtain the water level information of the water station to be detected, and control the lifting module to move the buoy body to a height corresponding to the target position according to the water level information;

[0042] a displacement determination module, configured to determine whether a condition for entering a fault identification phase is met based on first displacement data output by a position detection component in the lifting module and a top image sequence acquired by the first image sensor;

[0043] an image acquisition module, configured to control the second image sensor to acquire a target image of a target area of ​​the buoy body after the condition for entering the fault identification phase is met;

[0044] A feature extraction module is used to perform image enhancement, edge detection and contour segmentation on the target image to obtain the structural features of the buoy;

[0045] The fault identification module is used to input the structural features into a preset fault identification network and output a corresponding buoy fault type based on the fault identification network.

[0046] A fault detection device for a smart water station, the fault detection device is used to detect the fault type of a buoy body, wherein the fault detection device includes:

[0047] A guide rail is used to limit the lifting and lowering movement of the buoy body along its length direction;

[0048] A lifting module is used to drive the buoy body in the water to move up and down along the guide rail in response to the water level of the water station to be detected;

[0049] A visual acquisition component includes a first image sensor and a second image sensor, wherein the first image sensor is arranged in the guide rail at a position perpendicular to the buoy body, and the second image sensor is arranged outside the buoy body;

[0050] A controller is connected to the lifting module and the visual acquisition component, and the controller is configured as follows:

[0051] Obtain water level information and control the rise and fall of buoys;

[0052] Based on the displacement feedback of the lifting module and the top image sequence of the first image sensor, determining whether the conditions for entering the fault identification stage are met;

[0053] When the conditions for entering the fault identification phase are met, the buoy response interruption point is calculated, the target area of ​​the buoy body is determined, and the second image sensor is controlled to capture the target image;

[0054] The structural features of the target image are extracted and input into the preset fault recognition network to output the fault type of the buoy.

[0055] Compared with the prior art, the present invention has the following advantages:

[0056] This application realizes dual-channel monitoring of the operating status of the buoy by combining the displacement feedback information of the lifting module with the top image sequence obtained by the first image sensor, and introduces a calculation mechanism for the response interruption point, which can accurately locate the time period and corresponding structural area where the buoy fails during operation. On this basis, the second image sensor is controlled to accurately capture the target area, and the specific fault type of the buoy is effectively identified by combining structural feature extraction and the trained fault recognition network. Compared with existing solutions that rely on a single sensor or passive monitoring, this method has significant improvements in the accuracy of fault location, recognition accuracy and overall processing efficiency, and can be widely used in smart water station scenarios with complex structures and changing operating environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0058] Figure 1 This is a schematic diagram of an exemplary application scenario of an embodiment of the present application;

[0059] Figure 2 This is a structural diagram of a fault detection device for a smart water station according to an embodiment of the present application;

[0060] Figure 3 This is a flow chart of a fault detection method for a smart water station according to an embodiment of the present application;

[0061] Figure 4 This is a schematic diagram of the displacement data calculation principle of the embodiment of the present application;

[0062] Figure 5 This is a schematic diagram of the logic flow of buoy fault identification according to an embodiment of the present application;

[0063] Figure 6 This is a schematic diagram of the calculation principle of the response interruption point in an embodiment of the present application.

[0064] Figure numerals: 101, float body; 102, guide rail; 103, visual acquisition component; 1031, first image sensor; 1032, second image sensor; 104, lifting module; 105, controller. DETAILED DESCRIPTION

[0065] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0066] References to "embodiments" herein mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It will be understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0067] This application is applicable to smart water station systems with high structural integration, electronically controlled water intake, and operating environments with complex underwater resistance or clogging risks. Application scenarios include but are not limited to:

[0068] Small and medium-sized distributed water stations that use electric lifting structures to control the depth of buoys;

[0069] Artificial wetland water stations or water source primary stations located in waters with dense aquatic plants and impurities;

[0070] An unmanned water station with video acquisition capabilities and the desire to achieve automatic fault identification and diagnosis.

[0071] It should be pointed out that the fault detection method proposed in this application mainly targets two core technical problems existing in the automatic control process of the existing water station water intake buoy, namely, motion feedback distortion and unclear fault location. A detection method is proposed that combines displacement data and image data, and introduces a response interruption judgment and structural target recognition mechanism.

[0072] See also Figure 1 , which is a schematic diagram of an exemplary application scenario provided in an embodiment of the present application.

[0073] like Figure 1 As shown, the water station usually obtains the current water level information through a liquid level sensor and inputs the information into the controller. The controller generates a target displacement instruction according to the water level change to control the lifting structure to drive the buoy in the water body to move up and down along the guide rail to adjust the water intake depth.

[0074] Figure 1 It is further shown that in order to realize feedback on the position of the float, an encoder is provided in the system. The encoder is connected to the lifting structure to detect the number of rotations or displacement of the lifting motor and feed back the displacement data to the controller as a basis for determining the actual displacement of the float.

[0075] Those skilled in the art will understand that in the control method of the prior art, the controller equates the displacement feedback by the encoder with the actual movement of the buoy body. However, in actual application, when there is blockage in the guide rail, entanglement in the underwater structure, or obstruction of the buoy mechanism, the buoy may not move according to the target instruction, and the encoder will still generate feedback of the displacement due to the action of the motor, resulting in distortion of the control system's judgment on the operating status of the buoy, thereby affecting the water intake accuracy and safe operation of the entire water station.

[0076] See also Figure 2 , which is a schematic structural diagram of a fault detection device for a smart water station provided in an embodiment of the present application, wherein the fault detection device is used to detect the fault type of the buoy body 101, wherein the fault detection device includes:

[0077] The guide rail 102 is used to limit the lifting and lowering movement of the buoy body 101 along its length direction, wherein the guide rail 102 can be made of a stainless steel frame, a carbon fiber slide, or other structural materials with water corrosion resistance and guiding functions;

[0078] The lifting module 104 is used to drive the buoy body 101 in the water to move up and down along the guide rail 102 in response to the water level of the water station to be detected;

[0079] The lifting module 104 includes a water level sensor, a position detection component and a lifting mechanism;

[0080] It is understood that the lifting mechanism may include an electrically controlled lifting structure such as an electric screw, a gear chain mechanism, or a synchronous belt mechanism. The position detection component is integrated with an encoder or magnetic scale for detecting displacement during the lifting process. The water level sensor may be an ultrasonic water level sensor, a pressure water level sensor, a capacitive liquid level gauge, or a radar water level sensor, etc., for real-time monitoring of water height information and feedback to the controller 105 to generate target control instructions for the buoy's lifting. The selection of the water level sensor can be flexibly adjusted according to the water quality conditions, installation space, and accuracy requirements of the water station's environment, and is not limited to this in this application.

[0081] The visual acquisition component 103 includes a first image sensor 1031 and a second image sensor 1032. The first image sensor 1031 is arranged at a position perpendicular to the float body 101 in the guide rail 102, and the second image sensor 1032 is arranged on the outside of the float body 101. The above image sensors can be CMOS or CCD devices with industrial-grade resolution and frame rate requirements.

[0082] The controller 105 is connected to the lifting module 104 and the visual acquisition component 103. The controller 105 is configured as follows:

[0083] Obtain water level information and control the rise and fall of buoys;

[0084] Based on the displacement feedback of the lifting module 104 and the top image sequence of the first image sensor 1031, it is determined whether the conditions for entering the fault identification stage are met;

[0085] When the conditions for entering the fault identification phase are met, the response interruption point of the buoy body 101 is calculated, the target area of ​​the buoy body 101 is determined, and the second image sensor 1032 is controlled to capture the target image;

[0086] Extracting structural features from the target image and inputting them into a preset fault recognition network to output the fault type of the buoy body 101;

[0087] The controller 105 can adopt a general embedded processing unit (such as STM32, ARM Cortex series control chip), an industrial PC, an FPGA platform or an edge computing module with AI computing capabilities, which is not limited in this application.

[0088] Next, in conjunction with the accompanying drawings, a fault detection method for a smart water station provided by an embodiment of the present application is introduced. Figure 3The method shown is applied to a fault detection device, which includes a lifting module 104, a visual acquisition component 103, a guide rail 102, and a buoy body 101. The lifting module 104 is used to raise and lower the buoy body 101 in the water in response to the water level of the water station to be detected. The visual acquisition component 103 includes a first image sensor 1031 and a second image sensor 1032. The first image sensor 1031 is configured in the guide rail 102 at a position perpendicular to the buoy body 101, and the second image sensor 1032 is configured outside the buoy body 101. The method includes:

[0089] S1: Controlling the lifting module 104 to drive the buoy body 101 to move up and down along the guide rail 102 according to the acquired water level information;

[0090] In this embodiment, the lifting module 104 includes a water level sensor, a position detection component, and a lifting mechanism. The water level sensor is used to collect real-time liquid level data of the target water body, the position detection component is used to provide feedback on the actual displacement of the buoy, and the lifting mechanism can be based on an electric screw or a gear chain structure to achieve buoy lifting and lowering control.

[0091] S2: Determine whether the conditions for entering the fault identification phase are met based on the position detection component in the lifting module 104 and the first image sensor 1031;

[0092] In this embodiment, displacement feedback values ​​during the lifting process are collected along with a sequence of top images obtained by the first image sensor 1031. A preset spatial height mapping model is used to convert the top profile changes of the buoy in the images into actual displacement values, which are then compared and analyzed with the displacement sensor feedback. If the difference between the two types of displacement data exceeds a threshold, or if the image response trend is interrupted, a possible buoy failure is determined, meeting the conditions for entering the fault identification phase.

[0093] S3: If the conditions are met, the buoy fault identification logic is executed by the second image sensor 1032 to obtain the buoy fault type;

[0094] In this embodiment, when the fault identification phase is triggered, the controller 105 reverse-infers the target area of ​​the buoy structure based on the image response interruption point and controls the second image sensor 1032 to capture images of this target area. The captured images are then used through algorithms such as image enhancement, edge detection, and contour segmentation to extract structural feature information. This information is then fed into a fault identification network trained on historical buoy fault images, ultimately outputting a fault type label. Fault types may include, but are not limited to, buoy jamming, guide rail 102 obstruction, tilted attitude, or bottom flotation obstruction.

[0095] It's easy to understand that this application assumes that the lifting module 104 can stably drive the buoy along the guide rail 102 after receiving control commands, and that the encoder or magnetic scale position detection component can output the theoretical motion trajectory of the buoy in real time. Simultaneously, the first image sensor 1031 can stably capture continuous images of the buoy's top profile in a synchronized time sequence. Based on these assumptions, the system can achieve consistent displacement feedback and image response under normal operating conditions.

[0096] In traditional water station buoy control systems, the position of the buoy is often determined by displacement sensors (such as cable encoders or electric displacement feedback devices). However, when the guide rail 102 becomes stuck, the buoy deflects, or there is local resistance underwater, such as weed entanglement or rail corrosion, the system may still misjudge due to the motor performing no-load idling, causing the system to continue to feedback the erroneous state of normal buoy movement. In such scenarios, existing technologies often lack redundant channels to determine whether it is a false movement, and even lack the recognition logic to determine whether a local response interruption has occurred in a specific structural area of ​​the buoy. As a result, fault diagnosis remains at the level of the occurrence of the abnormality rather than the precise location of the source of the abnormality. In addition, traditional visual fusion systems often use a full-structure image acquisition solution, obtaining a complete image before performing full-image analysis. This has high imaging redundancy, low recognition efficiency, and difficulty in ensuring image recognition stability in various complex underwater environments.

[0097] In this embodiment, a first image sensor 1031 is introduced to capture a continuous image sequence of the buoy's top. This image sequence is then mapped to the displacement feedback from the lifting module 104. Using optical flow, the image response discontinuity points of the buoy's top are determined. This not only monitors the timing of structural response interruptions in real time but also provides precise anchor points in the temporal dimension for subsequent identification. Based on these response discontinuity points and in conjunction with a segment mapping model of the guide rail 102 structure, a target area can be identified on the buoy structure through reverse engineering. The second image sensor 1032 is then controlled to image only this target area.

[0098] Next, the method of the present application is further expanded on the part in which the lifting module 104 drives the buoy body 101 to move up and down along the guide rail 102 .

[0099] In traditional water intake systems, floats often rely on the water's own fluctuations to rise and fall, working in conjunction with a follow-up hose or connecting rod to drain water samples. However, this passive floating mode no longer meets the needs of complex water station operations, such as precise water intake, water layer selection, and water quality differential analysis. For example, in scenarios with multi-layer water intake, frequent fluctuations in reservoir water levels, or manual adjustments to the water intake depth, relying solely on the float's natural floating height makes it difficult to control the actual water intake layer depth. This can easily lead to unclear water intake layers, insufficient sample representation, or even a complete loss of water intake capability at low water levels.

[0100] This application uses an electrically controlled lifting module 104 to rigidly drive and control the float body 101, limits the float's movement direction through a guide rail 102, and uses a height feedback device for displacement sensing to achieve precise control of the float's lifting height. Compared with the traditional floating method that relies on water level, this application has the following advantages: First, it can actively respond to the target water layer setting, allowing the float to still align with the target water intake position under unnatural water level conditions; second, because the structure uses a linear guide rail 102 or a slide rail to limit the float's movement direction, it can effectively avoid undesirable movements such as posture deflection, drift, and collision, thereby improving structural stability and repeatability.

[0101] In an example, the specific steps of S1 are as follows:

[0102] S1.1: Obtaining water level information collected by the water level sensor;

[0103] Specifically, the lifting module 104 is provided with at least one set of water level sensors for real-time monitoring of the current water level of the water station to be inspected.

[0104] Furthermore, in order to improve the accuracy of water level monitoring, the sensor output value can be filtered to remove the interference caused by water surface fluctuations, thereby obtaining more stable water level reference information.

[0105] S1.2: Compare the water level information with the preset target water intake layer depth to determine the target position of the buoy body 101;

[0106] Specifically, after obtaining the current water level, the controller 105 calculates the difference between the current water surface height and the target water layer depth according to the set target water layer logic. The target water layer depth can be set to a fixed depth or a depth calibration value dynamically adjusted according to water quality parameters, which is not limited in this application.

[0107] Furthermore, after calculating the target position, controller 105 uses the target position as the desired position of the buoy, generating a specific displacement command value. Specifically, to avoid frequent minor adjustments, a position control deadband threshold can be set in this application. Only when the deviation between the current buoy position and the target position exceeds the set position control deadband threshold will the lifting mechanism be activated for adjustment, thereby improving system stability.

[0108] How to form a specific displacement instruction value according to the desired position of the float belongs to the existing technology and will not be described in detail in this application.

[0109] S1.3: Control the lifting mechanism to move the buoy body 101 to a height corresponding to the target position;

[0110] Specifically, controller 105 generates PWM control signals or stepper motor drive pulses based on the displacement command value calculated in S1.2, and controls the motor's motion. To ensure the float is actually in place, position detection devices such as encoders or magnetic scales can provide position feedback, performing a closed-loop comparison between the actual lift value and the target value. During the lift process, if characteristic signals such as abnormal speed or sudden load changes are detected, the lift process can be interrupted, triggering a fault warning or protection mechanism.

[0111] Next, the part of the method of the present application regarding the determination of the buoy position will be further expanded.

[0112] In existing technical solutions for transmitting objects through motors, the theoretical feedback value of mechanical transmission is generally used as the main basis for determining whether the position of the object has changed. However, in the application scenario of the present application, especially when foreign objects are adsorbed on the bottom of the float, the underwater guide rail 102 is blocked, and there is interference from turbulent water flow, although the float has become stuck or even completely stationary, the drive motor is still moving, and the feedback data obtained by the position detection component still shows a normal movement trend, misleading the system to believe that the float has been running according to the instructions, resulting in the system failing to trigger the abnormal detection logic, and ultimately causing the water intake accuracy to deviate or the water pump to run idle.

[0113] This application introduces a first image sensor 1031 with a vertical viewing angle configuration in the top guide rail 102 of the pontoon, and designs an image displacement estimation auxiliary judgment path to determine whether the pontoon has actually undergone spatial displacement, which serves as one of the core bases for entering the fault identification stage.

[0114] It is understood that the imaging direction is consistent with the direction of the buoy body 101's elevation, enabling the capture of a sequence of images of the buoy's top area during the elevation process from a bird's-eye view. The vertical arrangement allows the buoy's vertical motion trajectory to be accurately reflected in the scaling changes of the circular outline in the image, thereby stably converting pixel changes into the buoy's actual vertical displacement through a preset spatial height mapping model. The adoption of a bird's-eye view, rather than a traditional lateral perspective, is based on the buoy structure's inherent high degree of axial symmetry and the difficulty of imaging its lateral profile, thereby improving the stability of the image response and measurement accuracy.

[0115] Taking displacement calculation as an example, you can refer to Figure 4 To understand, Figure 4 This is a schematic diagram of the displacement data calculation principle of an embodiment of the present application, which is used to exemplarily illustrate the calculation process of the first displacement data and the second displacement data in the buoy displacement detection and their difference characteristics.

[0116] Figure 4The position detection component is shown as encoding feedback on the displacement of the lifting module 104, outputting corresponding first displacement data. This data represents the theoretical stroke calculated within the control system based on the displacement command and encoder feedback. It offers high responsiveness and real-time performance. However, because it relies solely on feedback from the drive end, it is susceptible to factors such as structural stagnation and slippage, potentially deviating from the actual buoy state.

[0117] Figure 4 The figure further illustrates a sequence of images of the buoy's top captured by the first image sensor 1031 at a fixed acquisition cycle, and an edge detection algorithm is used to extract a sequence of top contour features. The figure shows three different frames that demonstrate the scaling effect of the buoy's contour at three different moments. The number of frames can be customized. The contour shows a clear trend of shrinking as the buoy descends. Using a spatial height mapping model, the pixel size of the buoy's top contour in the image frames is mapped to actual vertical height changes, thereby generating second displacement data.

[0118] It's important to note that because the image feedback path more directly reflects the actual state of the buoy body 101, the second displacement data is more accurate than the first displacement data in practical applications, making it particularly valuable for identifying structural anomalies. In this application, the second displacement data is used by default as the evaluation standard for actual motion status, while the first displacement data is used as a reference for the control link. These two data serve as a comparison benchmark for determining whether a fault response interruption has occurred.

[0119] It is worth understanding that since encoder feedback is usually continuous and high-frequency, while image acquisition is limited by frame rate and processing cycle, there are natural differences between the two in terms of data length, sampling period and response granularity. This is also one of the reasons why the two displacement data are stored separately in different data cache structures in the diagram.

[0120] In an example, the specific steps of S2 are as follows:

[0121] S2.1: Obtain first displacement data output by the position detection component;

[0122] Specifically, the position detection component utilizes an encoder sensor, mounted on the drive shaft or associated driven structure of the lifting mechanism. This sensor captures displacement information as the buoy moves along the guide rail 102 by reading changes in rotational angle or linear displacement. The primary displacement data reflects the theoretical displacement path of the buoy under ideal, unobstructed motion conditions, providing direct, quantitative feedback between control signals and physical execution.

[0123] In this embodiment, the encoder can be an incremental or absolute rotary encoder, which uses high-resolution signal pulse feedback to achieve discrete acquisition of displacement. Alternatively, a linear displacement sensor with a magnetic scale can be used to adapt to different types of guide rail 102 structures. The encoder output signal is converted into a numerical form by the acquisition module and sent to the controller 105 for cross-comparison with subsequent image data.

[0124] S2.2: Acquire a top image sequence captured by the first image sensor 1031 , wherein the acquisition cycle of the first image sensor 1031 is synchronized with the action cycle of the lifting module 104 ;

[0125] Specifically, the first image sensor 1031 is rigidly mounted on the vertical observation position of the guide rail 102 structure, with its optical axis pointing vertically toward the top of the pontoon, so as to capture a plan view of the top of the pontoon.

[0126] Furthermore, in order to ensure that the image sequence can completely cover the changing trajectory of the float during the entire lifting cycle, the acquisition cycle of the first image sensor 1031 is set to be synchronized with the action cycle of the lifting module 104, that is, every time the lifting module 104 completes a unit control cycle, the image sensor triggers a shooting action until the displacement detection component feedback completes the lifting action and takes the last top image, thereby forming a top image sequence.

[0127] S2.3: performing edge detection on the top images in the top image sequence in sequence to obtain a contour feature sequence;

[0128] Specifically, to extract quantifiable motion features from consecutive image frames, this embodiment uses an edge detection algorithm to process the top image sequence. The goal of edge detection is to accurately identify the boundary position of the buoy's top contour and extract its two-dimensional projection dimensions within the image frame, providing a reliable basis for subsequent pixel-to-real-space mapping. In practical applications, preferred edge detection algorithms include the Canny edge operator or the Sobel operator to adapt to image contrast and noise environments under different shooting conditions.

[0129] Those skilled in the art will understand that how to extract contour features through edge detection is a prior art, such as extracting the pixel size value of its circumscribed rectangle or circumscribed circle, and this application will not elaborate on this.

[0130] S2.4: Converting pixel size changes of the contour feature sequence in the image into vertical displacement values ​​according to a predetermined spatial height mapping model, to obtain second displacement data, wherein the spatial height mapping model is constructed based on an image projection transformation relationship;

[0131] Specifically, to convert the sequence of contour features in the image into meaningful displacements, this embodiment introduces a spatial height mapping model. This model, based on the projective geometry of a fixed viewing angle, maps the pixel size of the contour in the image to the physical distance of the actual buoy top from the image sensor. In this embodiment, physical calibration is used to sample images of the buoy at different heights, and a calibration curve or interpolation model is established between pixel size and height.

[0132] In this embodiment, the outline size of each frame of image is reversely calculated using a spatial height mapping model to obtain a corresponding vertical displacement value, and finally construct second displacement data synchronized with the image frame.

[0133] Furthermore, the spatial height mapping model may also have scalability and recalibration capabilities, and may be configured based on first image sensors 1031 of different sizes or installation heights.

[0134] In one example, the conditions for determining whether to enter the fault identification phase based on displacement data include the following two aspects:

[0135] In a first aspect, the judgment is made based on the absolute value of the difference between the first displacement data and the second displacement data.

[0136] In one case, if the absolute value of the difference between the first displacement data and the second displacement data is greater than or equal to a preset error threshold, it can be determined that the fault identification stage has begun, where the error threshold can be obtained by performing a retrospective statistical analysis of the buoy lifting process in the historical operating condition data.

[0137] Preferably, the maximum deviation value between the encoder feedback displacement and the image measurement displacement is extracted under multiple fault-free states and used as the initial setting of the empirical threshold, which can then be dynamically fine-tuned based on on-site feedback.

[0138] Understandably, when the buoy body 101 experiences unexpected disturbances during the lifting process, such as structural jamming, posture deviation, or obstruction by surface debris, its motion behavior in image space often fails to form a completely consistent response curve with the encoder feedback, resulting in significant discrepancies in the displacement estimates. In these cases, the true response reflected by the image will lag or even be lost, while the encoder signal continues to output the ideal displacement, resulting in an abnormal deviation. By setting an error threshold to trigger a judgment on the deviation, it is possible to promptly detect problems such as mechanical or structural failures in the system.

[0139] In another case, if the absolute value of the difference between the first and second displacement data is less than the preset error threshold, there's no need to proceed to the fault identification phase. In this case, it can be assumed that the buoy's movement within the lifting path remains stable, the visual response is well synchronized with the electronic control execution path, and there are no signs of anomalies such as drift, sticking, or voiding. Therefore, further structural image analysis is unnecessary.

[0140] Understandably, when the buoy moves smoothly within the lifting guide rails 102, the mechanical structure is functioning properly, the sensors are operating stably, and the lighting and image quality meet expectations, the system will exhibit a high degree of overlap between the encoder feedback path and the image estimation path. In this case, even minor errors can be tolerated by properly setting the error threshold. This prevents the system from frequently entering the image recognition phase due to minor errors, conserving computing resources and improving overall recognition efficiency and stability.

[0141] In a second aspect, a judgment is made based on changes in the second displacement data within the action cycle.

[0142] In one case, when the second displacement data does not change within the action cycle or the rate of change is lower than the set change threshold, it can be determined that the fault identification stage has begun. The change threshold can be obtained by statistically fitting the buoy lifting and lowering processes under a large number of normal working conditions, and is optimized according to the motion inertia of the buoy body 101 and the image sampling frequency.

[0143] It's understandable that when the buoy body 101 executes a lift command, if the top profile features reflected in the top image sequence remain stationary or experience only minimal vibration, this means the buoy hasn't completed its intended movement. Because the first image sensor 1031 is always located atop the guide rail 102, its continuous monitoring of the top image profile can sensitively detect stagnation, thereby using the lack of rate of change as an effective trigger for transitioning to the next stage of judgment.

[0144] Alternatively, if the rate of change of the second displacement data within the motion cycle is greater than or equal to the set change threshold, there's no need to enter the fault identification phase. In this case, the buoy's response in image space can be considered consistent with the lift command, and its motion trend shows no obvious structural anomalies or signs of motion failure.

[0145] It can be understood that when the top profile in the top image sequence exhibits continuous and stable scale changes, it indicates that the float is indeed moving in response to the lifting mechanism, the system structure is not disturbed, and the lifting control closed loop is considered normal. In this case, even if there are slight errors in the encoder or other components, the dynamic response of the displacement change rate can still eliminate interference signals and avoid misjudgment.

[0146] Furthermore, the first aspect and the second aspect may be combined to determine whether to enter the fault identification stage.

[0147] When determining whether to enter the fault identification phase through the combination of the first and second aspects, if the results of the judgment based on the first and second aspects are inconsistent, the judgment based on the second aspect prevails, i.e., the judgment based on the second aspect takes precedence over the judgment based on the first aspect. This is because the second aspect directly reflects the actual response behavior of the buoy top based on the image change trend, and its stability and anti-interference capabilities are relatively strong. In particular, when the encoder has zero drift, sampling lag, local failure, or errors caused by external vibration, the first displacement data may be biased, but the vertical change trend of the image profile can still accurately reflect the dynamic state of the buoy. The image response is essentially the projection of the buoy's physical movement within the visual domain. Its change trend forms a continuous trajectory and is constrained by the fixed position and viewing angle of the first image sensor 1031, making it less susceptible to external occasional disturbances.

[0148] In addition, in addition to the above two aspects, in the embodiment of the present application, whether to enter the fault identification stage can also be determined by other methods or in combination with other methods.

[0149] For example, the judgment is made according to the change direction of the second displacement data and the change direction of the first displacement data.

[0150] Optionally, when the direction of change of the second displacement data is opposite to that of the first displacement data, it can be determined that the actual response of the buoy during the current motion cycle is abnormal, meeting the conditions for entering the fault identification stage. For example, when the lifting instruction is to rise, the first displacement data recorded by the encoder shows an upward trend, but the second displacement data calculated by the image sequence shows a downward trend. This indicates that the actual movement direction of the buoy has deviated from the theoretical instruction. This may be caused by a jamming of the guide rail 102, slippage of the pulley, or abnormal force on the buoy, requiring further fault identification.

[0151] Conversely, when the two directions remain consistent and fluctuate within a reasonable threshold, even with a small error, the float response can be considered normal, eliminating the need to enter the fault identification phase. Directional consistency, as a complementary approach to trend matching, effectively avoids boundary conditions that are difficult to identify due to small displacements, such as micro-movements at the float's initial start or near its final position.

[0152] Next, the part of the method of the present application regarding buoy fault identification is further expanded.

[0153] Those skilled in the art understand that during the long-term operation of the buoy lifting structure, due to the interaction of factors such as the water environment, guide rail 102 wear, entanglement with aquatic plants, and structural deformation, common faults include: travel interruption due to partial jamming of the guide rail 102, attitude deviation due to buoy shell deformation, suspension blockage caused by bottom constraints, and non-uniform motion caused by loose connectors. These faults not only occur in different areas of the buoy structure, but also correspond to different stages of the operation during the lifting process.

[0154] Taking a typical scenario as an example, if the bottom of the buoy is restricted due to entanglement of aquatic plants or accumulation of sediment, the abnormality usually manifests as abnormal structural response at the end of the descent; and if there is metal fatigue or foreign matter stuck in the middle section of the buoy guide rail 102, obvious reverse resistance can be triggered as soon as the movement starts, which manifests as a serious discrepancy between the upper displacement feedback and the image response.

[0155] Therefore, the present application believes that the time point of occurrence of buoy structure abnormalities and their corresponding position heights often have strong spatial directivity. In other words, the time point of occurrence of different faults implicitly corresponds to the spatial distribution position of the structural abnormality parts.

[0156] For reference Figure 5 To understand, Figure 5 This is a logical flow chart of buoy fault identification according to an embodiment of the present application.

[0157] In an example, the specific steps for S3 are as follows:

[0158] S3.1: Calculate the response interruption point based on the feedback information of the lifting module 104 in the current action cycle and the detection information of the first image sensor 1031;

[0159] It's understandable that the response breakpoint isn't an absolute coordinate of the structure's position. Rather, it refers to the initial segment within a lift cycle where the image trend first deviates from the theoretical displacement trend. Because image information possesses a higher degree of real-world responsiveness, the response breakpoint identifies the location where the structure first loses effective linkage during the current movement, serving as a key focus area for subsequent visual inspection.

[0160] Furthermore, the response interruption point not only has a clear temporal attribute (corresponding to a frame number in the image sequence), but also inherently carries a high degree of spatial semantics. The corresponding theoretical displacement segment can be mapped to a specific area of ​​the buoy structure. This feature eliminates the need for the second image sensor 1032 to perform an indiscriminate full-field scan during the fault identification phase. Instead, it captures targeted images based on the structural area where the response interruption point is located.

[0161] Taking the response interruption point as an example, you can refer to Figure 6 To understand, Figure 6This is a schematic diagram of the calculation principle of the response interruption point in an embodiment of the present application.

[0162] Figure 6 The figure shows how the theoretical motion trajectory of the buoy is constructed based on the target displacement command and feedback displacement data of the lifting module 104 during the current motion cycle. This trajectory reflects the continuous displacement of the buoy within the guide rail 102 under ideal operating conditions. Subsequently, combining the calibration parameters and viewing angle configuration of the first image sensor 1031, this theoretical motion trajectory is mapped into image space through image backprojection, resulting in the theoretical change region of the buoy top in the image sequence.

[0163] It is understandable that Figure 6 Although only one theoretical change region of the buoy top is marked in the figure, in actual operation, the theoretical change region is a set of regions generated by projecting frame by frame on the entire displacement time axis based on the continuous target displacement instructions of the buoy. Each theoretical change region corresponds to the top contour of the buoy that should appear in the image space at a certain theoretical height, reflecting how the image contour of the buoy should change over time under different displacement states. Therefore, Figure 6 The single theoretical change region shown in the figure is only a schematic representation of a specific frame in the image sequence, used to intuitively illustrate the principle. Overall, the image backprojection process is a dynamic modeling result based on temporal continuity, and its output is a multi-frame sequence of theoretical image contour evolution.

[0164] Figure 6 The figure further illustrates the extraction of the actual changing region of the buoy's top from the sequence of captured top images using image processing techniques such as edge detection and contour analysis. This region reflects the actual image response behavior of the buoy's top during the raising and lowering process, and may include missing or interrupted responses due to abnormal factors such as structural jamming and interference with the guide rail 102.

[0165] Furthermore, an optical flow method is used to compare the response trends between the theoretical and actual change regions frame by frame, detecting the cumulative change trend of the response offset. Starting from a certain image frame, if the actual response image no longer changes as expected with the theoretical displacement, that is, if the response trend is interrupted or significantly delayed, the control instruction segment corresponding to that frame can be identified, which is the response interruption point.

[0166] In one example, the calculation steps for responding to an interruption point are as follows:

[0167] S3.1.1: Obtain the target displacement instruction and displacement feedback information of the lifting module 104 in the current motion cycle, and construct the theoretical motion trajectory of the buoy body 101 in the current motion cycle;

[0168] It is understood that the target displacement command originates from controller 105, while the displacement feedback information originates from the displacement detection component. The specific construction of the theoretical motion trajectory is a matter of existing technology. For example, a timestamp-based target-feedback motion comparison table is constructed based on the temporal alignment of the displacement command and the feedback displacement. Furthermore, by compensating for the response time characteristics of the lifting module 104 (such as acceleration delay and braking time), the expected movement behavior of the buoy within the current cycle can be precisely modeled, ultimately generating a theoretical motion trajectory with continuous temporal sequence and segmented traceability. This application does not elaborate on this in detail.

[0169] S3.1.2: Back-project the theoretical motion trajectory into image space based on the viewing angle parameters and calibration parameters of the first image sensor 1031 to obtain a theoretical change area on the top of the buoy;

[0170] In this embodiment, the viewing angle parameters include but are not limited to focal length, principal point position and viewing angle, and the calibration parameters include but are not limited to distortion correction matrix, intrinsic parameter matrix and posture information.

[0171] Those skilled in the art will appreciate that, given known viewing angle and calibration parameters, a projection transformation matrix can be constructed from three-dimensional space to a two-dimensional image plane. Based on this, each discrete position in the theoretical motion trajectory is used as input, and a back-projection transformation is performed to calculate the edge contour range in the image frame corresponding to each theoretical position of the buoy top.

[0172] Furthermore, in order to adapt to the actual situation where the top structure of the buoy has a certain curvature or is affected by mirror reflection, a set of compensation parameters can be introduced in the projection operation, such as the geometric boundary expansion factor obtained based on the top structure modeling, which is used to dynamically adjust the size range of the theoretical change area, improve the fault tolerance of the matching between the theoretical area and the actual image, and finally output the theoretical change area in an image sequence.

[0173] S3.1.3: Compare the theoretical change region with the actual change region of the buoy top in the detection information using an optical flow method, detect the response trend change between consecutive image frames, and determine the starting time of the buoy top image response interruption. The control segment corresponding to the starting time is used as the response interruption point.

[0174] S3.2: Determine a target area of ​​the buoy body 101 according to the response interruption point, and obtain a target image of the target area through the second image sensor 1032;

[0175] Determining the target area of ​​the buoy body 101 according to the response interruption point includes:

[0176] Dividing the movement path of the buoy body 101 along the guide rail 102 into a plurality of preset structural sections, wherein each preset structural section corresponds one-to-one to a structural area of ​​the buoy body 101;

[0177] It is understood that the overall structure of the buoy body 101 is composed of multiple functional modules or structural units connected in series longitudinally, including but not limited to: a top guide cap, a main buoyancy chamber section, a counterweight connection section, and a bottom guide rail limit structure. Based on the spatial position and distribution characteristics of these physical structures, the guide rail 102 can be divided into several preset structural sections along its length. The height range of each section is completely consistent with the size of its corresponding structural unit or has a precision tolerance setting, and the section division boundary can be accurately determined by calibrating the buoy structure design drawings. In actual deployment, the number and position of the preset structural sections do not need to be generated in real time, but are stored in the controller 105 in the form of a configuration file during the system initialization phase for subsequent section identification and acquisition logic mapping.

[0178] Furthermore, once the response interruption point is calculated, the control segment to which it belongs can be located. This control segment refers to the displacement interval of the buoy in the theoretical motion trajectory corresponding to the response interruption point and has a unique identifier, such as a segment number or height tag. The controller 105 can quickly locate the target structural area corresponding to the response interruption point by matching the numbering or identification mapping between the control segment and the preset structural segments. This target structural area is the target of image acquisition and represents the location on the buoy body 101 where structural anomalies are most likely to occur.

[0179] Determining a corresponding preset structural section according to the control section of the response interruption point;

[0180] The structural area of ​​the buoy body 101 corresponding to the corresponding preset structural section is used as the target area.

[0181] It is important to note that to ensure efficient acquisition of target images, the second image sensor 1032 is positioned outside the buoy, providing an imaging angle along the buoy's axis and automatically capturing and caching image frames as the target area passes through its field of view. In an optional embodiment, the second image sensor 1032 can be coupled with a limit trigger or time synchronizer to automatically trigger the shutter action when the target area passes, ensuring image acquisition accuracy and consistent timing of response point identification.

[0182] S3.3: performing image enhancement, edge detection, and contour segmentation on the target image to obtain structural features;

[0183] Those skilled in the art will understand that after obtaining an image of the target area, how to identify and segment it belongs to the existing technology and will not be described in detail in this application.

[0184] It is important to note that the structural features of this application refer to a collection of image information extracted through multi-layer processing of the target area image of the buoy body 101, which characterizes the geometric structure, contour morphology, and local deformation state of that area. This structural feature not only reflects the visual information of the buoy surface or boundary, but also includes the distortion patterns, crack trends, seizure marks, or other microscopic defect outlines that may occur in the buoy structure under abnormal stress or deformation conditions, supporting subsequent fault type identification operations.

[0185] S3.4: Inputting the structural features into a preset fault identification network, processing the structural features through the fault identification network, and outputting a buoy fault type, wherein the fault identification network is trained using historical buoy fault images;

[0186] In one example, the fault identification network includes a structure encoding layer, a state reasoning layer, and a fault classification output layer, wherein:

[0187] The structural coding layer is used to vectorize the structural features of the input, including edge continuity parameters, contour closure index and occlusion area ratio;

[0188] The state inference layer is provided with a graph structure attention mechanism for establishing spatial associations between structure encoding vectors and inferring the integrity state and force consistency characteristics of the buoy structure;

[0189] The fault classification output layer is used to output the fault type of the buoy through a Softmax classifier according to the integrity state and force consistency characteristics. The fault types include buoy stuck, guide rail 102 stuck, bottom floating obstruction and buoy attitude tilt.

[0190] In one example, the present application provides a fault detection system for a smart water station, which is applied to a fault detection device. The fault detection device includes a lifting module 104, a visual acquisition component 103, a guide rail 102, and a buoy body 101. The lifting module 104 is used to raise and lower the buoy body 101 in the water in response to the water level of the water station to be detected. The visual acquisition component 103 includes a first image sensor 1031 and a second image sensor 1032. The first image sensor 1031 is configured in the guide rail 102 at a position perpendicular to the buoy body 101, and the second image sensor 1032 is configured on the outside of the buoy body 101. The system includes:

[0191] A water level control module is used to obtain the water level information of the water station to be detected, and control the lifting module 104 to move the buoy body 101 to a height corresponding to the target position according to the water level information;

[0192] a displacement determination module, configured to determine whether a condition for entering a fault identification phase is met based on the first displacement data output by the position detection component in the lifting module 104 and the top image sequence captured by the first image sensor 1031;

[0193] An image acquisition module, configured to control the second image sensor 1032 to acquire a target image of the buoy body 101 after the condition for entering the fault identification phase is met;

[0194] A feature extraction module is used to perform image enhancement, edge detection and contour segmentation on the target image to obtain the structural features of the buoy;

[0195] The fault identification module is used to input the structural features into a preset fault identification network and output a corresponding buoy fault type based on the fault identification network.

[0196] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A fault detection method for a smart water station, applied to a fault detection device, wherein the fault detection device comprises a lifting module, a visual acquisition component, a guide rail, and a buoy body. The lifting module is used to raise and lower the buoy body in the water in response to the water level of the water station to be detected. The visual acquisition component comprises a first image sensor and a second image sensor. The first image sensor is arranged in the guide rail at a position perpendicular to the buoy body, and the second image sensor is arranged outside the buoy body. The method is characterized in that: The method comprises: Controlling the lifting module to drive the buoy body to move up and down along the guide rail according to the acquired water level information; Determining whether a condition for entering a fault identification phase is met based on the position detection component in the lifting module and the first image sensor; If the condition is met, the buoy fault recognition logic is executed by the second image sensor to obtain the buoy fault type, wherein the buoy fault recognition logic locates the target area of ​​the buoy body based on the response interruption point, and performs structural feature extraction and fault type recognition processing on the target area.

2. A fault detection method for a smart water station according to claim 1, characterized in that: The buoy fault identification logic includes: Calculating a response interruption point based on feedback information from the lifting module during the current action cycle and detection information from the first image sensor; determining a target area of ​​the buoy body according to the response interruption point, and obtaining a target image of the target area through the second image sensor; Performing image enhancement, edge detection, and contour segmentation on the target image to obtain structural features; The structural features are input into a preset fault identification network, and the structural features are processed by the fault identification network to output the buoy fault type, wherein the fault identification network is trained by using historical fault images of the buoy.

3. A fault detection method for a smart water station according to claim 2, characterized in that: The step of calculating the response interruption point according to the feedback information of the lifting module in the current action cycle and the detection information of the first image sensor includes: Obtaining the target displacement instruction and displacement feedback information of the lifting module in the current action cycle, and constructing the theoretical motion trajectory of the buoy body in the current action cycle; Back-projecting the theoretical motion trajectory in image space according to the viewing angle parameter and calibration parameter of the first image sensor to obtain a theoretical change area of ​​the buoy top; The theoretical change area is compared with the actual change area of ​​the buoy top in the detection information by the optical flow method, and the response trend change between consecutive image frames is detected to obtain the starting time of the interruption of the response of the buoy top image, and the control segment corresponding to the starting time is used as the response interruption point.

4. A fault detection method for a smart water station according to claim 2, characterized in that: Determining a target area of ​​the buoy body according to the response interruption point includes: Dividing the movement path of the buoy body along the guide rail into a plurality of preset structural sections, wherein each preset structural section corresponds one-to-one to a structural area of ​​the buoy body; Determining a corresponding preset structural section according to the control section of the response interruption point; The structural area of ​​the buoy body corresponding to the corresponding preset structural section is used as the target area.

5. A fault detection method for a smart water station according to claim 2, characterized in that: The fault identification network includes a structure encoding layer, a state reasoning layer, and a fault classification output layer, wherein: The structural coding layer is used to vectorize the structural features of the input, including edge continuity parameters, contour closure index and occlusion area ratio; The state inference layer is provided with a graph structure attention mechanism for establishing spatial associations between structure encoding vectors and inferring the integrity state and force consistency characteristics of the buoy structure; The fault classification output layer is used to output the fault type of the buoy through a Softmax classifier according to the integrity state and force consistency characteristics. The fault types include buoy stuck, guide rail stuck, bottom floating obstruction and buoy attitude tilt.

6. A fault detection method for a smart water station according to claim 5, characterized in that: The lifting module includes a water level sensor, a position detection component and a lifting mechanism. The lifting module is controlled according to the acquired water level information to drive the buoy body to move up and down along the guide rail, including: Acquiring water level information collected by the water level sensor; Comparing the water level information with a preset target water extraction layer depth to determine the target position of the buoy body; The lifting mechanism is controlled to move the buoy body to a height corresponding to the target position.

7. A fault detection method for a smart water station according to claim 1, characterized in that: Determining whether a condition for entering a fault identification phase is met based on the position detection component in the lifting module and the first image sensor includes: Acquiring first displacement data output by the position detection component; Acquire a top image sequence captured by the first image sensor, wherein a capture period of the first image sensor is synchronously set with an action period of the lifting module; performing edge detection on the top images in the top image sequence in sequence to obtain a contour feature sequence; According to a preset spatial height mapping model, pixel size changes of the contour feature sequence in the image are converted into vertical displacement values ​​to obtain second displacement data, wherein the spatial height mapping model is constructed based on an image projection transformation relationship.

8. A fault detection method for a smart water station according to claim 7, characterized in that: The conditions for entering the fault identification phase include at least one of the following: An absolute value of a difference between the first displacement data and the second displacement data is greater than or equal to a preset error threshold; The second displacement data does not change during the action cycle; A changing direction of the second displacement data is opposite to a changing direction of the first displacement data.

9. A fault detection system for a smart water station, used to implement a fault detection method for a smart water station as described in any one of claims 1 to 8, wherein the system is applied to a fault detection device, the fault detection device comprising a lifting module, a visual acquisition component, a guide rail and a buoy body, the lifting module being used to raise and lower the buoy body in the water in response to the water level of the water station to be detected, the visual acquisition component comprising a first image sensor and a second image sensor, the first image sensor being arranged at a position perpendicular to the buoy body in the guide rail, and the second image sensor being arranged outside the buoy body, characterized in that The system comprises: A water level control module is used to obtain the water level information of the water station to be detected, and control the lifting module to move the buoy body to a height corresponding to the target position according to the water level information; a displacement determination module, configured to determine whether a condition for entering a fault identification phase is met based on first displacement data output by a position detection component in the lifting module and a top image sequence acquired by the first image sensor; an image acquisition module, configured to control the second image sensor to acquire a target image of a target area of ​​the buoy body after the condition for entering the fault identification phase is met; A feature extraction module is used to perform image enhancement, edge detection and contour segmentation on the target image to obtain the structural features of the buoy; The fault identification module is used to input the structural features into a preset fault identification network and output a corresponding buoy fault type based on the fault identification network.

10. A fault detection device for a smart water station, used to implement the fault detection method for a smart water station according to any one of claims 1 to 8, characterized in that: The fault detection device is used to detect the fault type of the buoy body, wherein the fault detection device includes: A guide rail is used to limit the lifting and lowering movement of the buoy body along its length direction; A lifting module is used to drive the buoy body in the water to move up and down along the guide rail in response to the water level of the water station to be detected; A visual acquisition component includes a first image sensor and a second image sensor, wherein the first image sensor is arranged in the guide rail at a position perpendicular to the buoy body, and the second image sensor is arranged outside the buoy body; A controller is connected to the lifting module and the visual acquisition component, and the controller is configured as follows: Obtain water level information and control the rise and fall of buoys; Based on the displacement feedback of the lifting module and the top image sequence of the first image sensor, determining whether the conditions for entering the fault identification stage are met; When the conditions for entering the fault identification phase are met, the buoy response interruption point is calculated, the target area of ​​the buoy body is determined, and the second image sensor is controlled to capture the target image; The structural features of the target image are extracted and input into the preset fault recognition network to output the fault type of the buoy.

Images