Method and apparatus for determining sedimentation status in cable wells in low-light environments

Abstract

This disclosure presents a method and apparatus for determining the sedimentation state in cable wells under low-light conditions. One specific implementation of the method includes: acquiring weather information and a set of sensor signals; generating a trigger configuration information set based on the weather information and the sensor signal set; controlling a target camera to acquire a target image based on the trigger configuration information set; enhancing the target image using a pre-trained low-light image enhancement model to obtain an enhanced image; and determining the sedimentation state information corresponding to the cable well based on the enhanced image. This implementation achieves accurate determination of the sedimentation state in cable wells, thereby ensuring the service life of cables and reducing the probability of power equipment failure and damage caused by sedimentation in power wells.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the fields of computer technology and power technology, and specifically to a method and apparatus for determining the sedimentation state in cable wells under low-light conditions. Background Technology

[0002] Cable wells are structures designed to house cables and facilitate future maintenance and inspection. When the perimeter of a cable well is not properly sealed, rainwater and sediment can flow into it, causing sediment buildup. This sediment buildup can lead to "water treeing" of the cables inside, affecting their lifespan. Furthermore, as the cables age, they may experience leakage problems, ultimately causing electrical equipment malfunctions or even damage. Currently, the conventional method is to conduct regular manual inspections to check for sediment buildup in cable wells.

[0003] However, when using the above method, the following technical problems often arise:

[0004] Due to the large number of cable wells, it is difficult to effectively and comprehensively determine the siltation status of the cable wells by manual inspection, thus failing to guarantee the service life of the cables and reduce the probability of power equipment failure and damage caused by siltation in the cable wells.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not form prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] Some embodiments of this disclosure propose a method and apparatus for determining the sedimentation state in cable wells under low-light conditions, in order to solve the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide a method for determining the sedimentation state in a cable well under low-light conditions. The method includes: acquiring weather information and a sensor signal set, wherein the sensor signal set is acquired by a sensor assembly installed in the cable well, the sensor assembly including: a mounting frame and at least three fiber Bragg grating sensors, the at least three fiber Bragg grating sensors being parallel to each other and fixed to the mounting frame at preset intervals, the mounting frame including: a fixing groove, wherein the fixing groove faces the target side, the fixing groove being used to fix one of the at least three fiber Bragg grating sensors; generating a trigger based on the weather information and the sensor signal set. The configuration information group includes trigger configuration information such as a trigger frequency value and a trigger probability value. The trigger frequency value controls the image acquisition frequency of the target camera, and the trigger probability value controls the activation probability of the target camera when triggered. The target camera is located inside the cable well and faces the bottom of the cable well. According to the trigger configuration information group, the target camera is controlled to acquire target images. The target image is enhanced in low light using a pre-trained low-light image enhancement model to obtain an enhanced image. Based on the enhanced image, the sedimentation status information corresponding to the cable well is determined, including sedimentation type, sedimentation confidence level, and sedimentation risk level.

[0009] Secondly, some embodiments of this disclosure provide a device for determining the sedimentation status in cable wells under low-light conditions. The device includes: an acquisition unit configured to acquire weather information and a sensor signal group, wherein the sensor signal group is acquired by a sensor assembly disposed within the cable well. The sensor assembly includes: a mounting frame and at least three fiber Bragg grating sensors, the at least three fiber Bragg grating sensors being parallel to each other and fixed to the mounting frame at preset intervals. The mounting frame includes: a fixing groove, wherein the fixing groove faces the target side and is used to fix one of the fiber Bragg grating sensors; and a generation unit configured to generate a trigger configuration information group based on the weather information and the sensor signal group. The trigger configuration information includes: a trigger frequency value and a trigger probability value. The trigger frequency value is used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when triggered. The target camera is located inside the cable well and faces the bottom of the cable well. The control unit is configured to control the target camera to acquire target images according to the trigger configuration information. The low-light image enhancement unit is configured to enhance the target image in low light using a pre-trained low-light image enhancement model to obtain an enhanced image. The determination unit is configured to determine the sedimentation status information corresponding to the cable well based on the enhanced image. The sedimentation status information includes: sedimentation type, sedimentation confidence level, and sedimentation risk level.

[0010] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0011] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0012] The above-described embodiments of this disclosure have the following beneficial effects: The method for determining the sedimentation state in cable wells under low-light environments, based on some embodiments of this disclosure, achieves accurate determination of the sedimentation state in cable wells, thereby ensuring the service life of cables and reducing the probability of power equipment failure and damage caused by sedimentation in power wells. Specifically, firstly, weather information and a sensor signal group are acquired. The sensor signal group is collected by a sensor assembly installed in the cable well. The sensor assembly includes a mounting frame and at least three fiber Bragg grating sensors. The at least three fiber Bragg grating sensors are parallel to each other and fixed to the mounting frame at a preset interval. The mounting frame includes a fixing groove facing the target side, and the fixing groove is used to fix one of the fiber Bragg grating sensors. Secondly, based on the weather information and the sensor signal group, a trigger configuration information group is generated. The trigger configuration information includes a trigger frequency value and a trigger probability value. The trigger frequency value is used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when triggered. The target camera is installed in the cable well and faces the bottom of the cable well. In practice, siltation is usually caused by the low-lying location of cable wells, especially during adverse weather conditions such as rain and snow, where rainwater and snowmelt carry silt and sand into the wells. Therefore, this disclosure generates trigger configuration information for the target camera by combining weather information and sensor signals. This avoids the problems of large data volume and high power consumption associated with conventional real-time monitoring via cameras. Next, based on the aforementioned trigger configuration information, the target camera is controlled to acquire target images. This disclosure effectively controls the activation and image acquisition of the target camera through trigger frequency and probability values ​​to ensure effective coverage. Furthermore, the target image is enhanced in low light using a pre-trained low-light image enhancement model, resulting in an enhanced image. In practice, cable wells are typical low-light environments, making it difficult to monitor siltation under low illumination. Therefore, this disclosure effectively enhances the image by designing a low-light image enhancement model. Finally, based on the enhanced image, the siltation status information corresponding to the cable well is determined, including: siltation type, siltation confidence level, and siltation risk level. In summary, this method enables accurate determination of the sedimentation status inside cable wells, thereby ensuring the service life of cables and reducing the probability of power equipment failure and damage caused by sedimentation in power wells. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a flowchart of some embodiments of the method for determining the sedimentation state in cable wells in low-light environments according to the present disclosure;

[0015] Figure 2 This is a schematic diagram of the sensor assembly.

[0016] Figure 3 This is a schematic diagram of the model structure for a low-light image enhancement model;

[0017] Figure 4 This is a schematic diagram of some embodiments of the device for determining the sedimentation state in cable wells in low-light environments according to the present disclosure;

[0018] Figure 5 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0019] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0020] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0021] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0022] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0023] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0024] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] refer to Figure 1The flowchart 100 illustrates some embodiments of a method for determining the sedimentation state in cable wells under low-light conditions according to the present disclosure. This method for determining the sedimentation state in cable wells under low-light conditions includes the following steps:

[0026] Step 101: Obtain weather information and sensor signal groups.

[0027] In some embodiments, the execution entity (e.g., a computing device) of the method for determining the sedimentation state in cable wells under low-light conditions can acquire weather information and a set of sensor signals via a wired or wireless connection. The sensor signal set is acquired by a sensor assembly installed inside the cable well. The sensor assembly includes a mounting bracket and at least three fiber Bragg grating sensors. The at least three fiber Bragg grating sensors are parallel to each other and fixed to the mounting bracket at predetermined intervals. The mounting bracket includes a fixing groove, which faces the target side. In practice, the mounting bracket is horizontally fixed to the inner wall of the cable well. The target side is the side facing the bottom of the cable well. The fixing groove is used to fix one of the at least three fiber Bragg grating sensors. The weather information characterizes the weather state of the area where the cable well is located. The sensor signals characterize the optical signals corresponding to the fiber Bragg grating sensors.

[0028] It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra wideband) connections, and other currently known or future wireless connection methods.

[0029] As an example, see Figure 2 The schematic diagram of the sensor assembly shown illustrates that the mounting bracket 2 can be made of lightweight galvanized steel to ensure good support strength and corrosion resistance. Figure 2 For example, four fiber Bragg grating sensors 1 are shown, one of which is disposed within a fixing groove facing the bottom of the cable well. The remaining three fiber Bragg grating sensors 1 are disposed in the middle of a mounting bracket. In practice, at least three fiber Bragg grating sensors 1 use the same sensor type and specifications. In particular, the side of the sensor assembly opposite the fixing groove 3 is the cable well mounting surface, which can be horizontally fixed to the inner wall of the cable well using expansion screws.

[0030] Optionally, the sensor assembly may further include a receiver, a microprocessor, and a power supply unit. The receiver, microprocessor, and power supply unit can be embedded within a mounting frame to perform signal processing and image processing on sensor signal groups acquired by at least three fiber Bragg grating sensors. Specifically, considering power consumption, a wired receiver is used. The power supply unit supplies power to the receiver, microprocessor, and fiber Bragg grating sensors via a wired connection, with the power supply unit connected to a power cable.

[0031] It should be noted that the aforementioned computing device can be either hardware or software. When the computing device is hardware, it can be implemented as a single terminal device. When the computing device is software, it can be installed within the hardware devices listed above. It can be implemented as a single software program or a software module. No specific limitations are made here. In particular, the computing device can be a microprocessor included in a sensor assembly.

[0032] Optionally, when the weather changes in the area where the cable wells are located, the weather information is actively distributed and synchronized by a remote server. The remote server can be a server used to synchronize weather information with the sensor components installed in each cable well. To reduce the power consumption of the sensor components, a passive update method is adopted for the weather information. The remote server monitors weather changes in the area where the cable wells are located in real time. When the remote server detects a change in the weather in the area where the cable wells are located, it synchronizes the weather information with the sensor components in the cable wells via a wired connection.

[0033] Optionally, the weather information includes: weather type, predicted precipitation, predicted precipitation probability, predicted temperature, and predicted duration. The weather type represents the type of weather in the area where the cable well is located. In practice, the weather type may include, but is not limited to, precipitation type and snowfall type. The precipitation type or snowfall type can be further refined based on the corresponding precipitation amount or snowfall amount. For example, the precipitation type can be further refined into: light rain, moderate rain, heavy rain, and torrential rain. The snowfall type can be further refined into: light snow, moderate snow, heavy snow, and blizzard. The predicted precipitation represents the predicted precipitation amount in the area where the cable well is located within a preset future duration. The predicted precipitation probability represents the predicted probability of precipitation in the area where the cable well is located within a preset future duration. The predicted temperature represents the predicted temperature in the area where the cable well is located within a preset future duration. The predicted duration represents the predicted duration of the area where the cable well is located within the predicted future duration. In particular, to improve the weather state representation capability of the weather information, the predicted precipitation, predicted precipitation probability, and predicted temperature are all time-varying value sequences. For example, taking precipitation prediction as an example, if the preset duration is 24 hours, then the predicted precipitation can be a sequence of precipitation values ​​within 24 hours, with a time granularity of 1 hour. Similarly, taking precipitation probability prediction as an example, if the preset duration is 24 hours, then the predicted precipitation probability can be a sequence of precipitation probability values ​​within 24 hours, with a time granularity of 1 hour. And again, taking temperature prediction as an example, if the preset duration is 24 hours, then the predicted temperature can be a sequence of temperature values ​​within 24 hours, with a time granularity of 1 hour.

[0034] Step 102: Generate a trigger configuration information group based on weather information and sensor signal group.

[0035] In some embodiments, the aforementioned executing entity can generate a trigger configuration information group based on weather information and sensor signal groups. The trigger configuration information includes a trigger frequency value and a trigger probability value. The trigger frequency value controls the image acquisition frequency of the target camera. The trigger probability value controls the activation probability of the target camera when triggered. The target camera is located inside the cable well and faces the bottom of the cable well. In practice, the trigger configuration information group may include two trigger configuration information sets, namely trigger configuration information A and trigger configuration information B. Trigger configuration information A corresponds to the weather information. Trigger configuration information B corresponds to the sensor signal groups. In practice, the weather information and sensor signal groups can be mapped to corresponding trigger configuration information according to configured mapping rules to obtain the aforementioned trigger configuration information group.

[0036] In some optional implementations of certain embodiments, the execution entity generates a trigger configuration information group based on the weather information and the sensor signal group, including:

[0037] The first step is to encode the weather type, the predicted precipitation, the predicted precipitation probability, the predicted temperature, and the predicted duration of the weather type, precipitation, precipitation probability, temperature, and duration of the weather type, precipitation, precipitation probability, temperature, and duration of the weather type, precipitation probability ... precipitation duration of the weather type, precipitation probability, precipitation probability, and precipitation probability of the weather type, precipitation probability, precipitation probability, and precipitation probability of the weather type, precipitation probability, precipitation probability, and precipitation probability of the weather type, precipitation probability, precipitation probability, and precipitation probability of the weather type, precipitation probability, precipitation probability, precipitation probability, and precipitation probability of the weather type, precipitation probability, precipitation probability, precipitation probability, precipitation probability, precipitation probability, precipitation probability, precipitation probability, precipitation probability, precipitation probability, precipitation probability

[0038] In practice, for weather types, since the number of weather types is limited, one-heat encoding can be used for feature encoding to obtain the aforementioned weather type features. For predicted precipitation, predicted precipitation probability, and predicted temperature, since they are all value sequences with different dimensions, the Z-Score method is used for feature encoding to obtain precipitation features. Precipitation probability features and temperature features are also encoded using the Z-Score method to ensure consistency in dimensions with the other value features.

[0039] The second step involves generating trigger configuration information corresponding to the weather information in the trigger configuration information group based on the pre-trained first trigger frequency classification model, the aforementioned weather type features, the aforementioned precipitation features, the aforementioned precipitation probability features, the aforementioned temperature features, and the aforementioned duration features.

[0040] The aforementioned first trigger frequency classification model includes: one input layer, five hidden layers, and one output layer. The input layer includes 5×N neurons. The hidden layers include: a first hidden layer, a second hidden layer, a third hidden layer, a fourth hidden layer, and a fifth hidden layer. The first and fifth hidden layers each include 5×N neurons, the second and fourth hidden layers each include 7×N neurons, the third hidden layer includes 9×N neurons, and the output layer includes 5×N neurons. The five neurons in the output layer correspond to five preset trigger frequency values ​​at different time scales.

[0041] Specifically, considering that the first trigger frequency classification model accepts numerical input, and that the weather type feature is a feature vector after unique thermal encoding, while the precipitation feature, the aforementioned precipitation probability feature, and the aforementioned temperature feature are value sequences, N is taken as the length of the feature vector formed by connecting the aforementioned weather type feature, the aforementioned precipitation feature, the aforementioned precipitation probability feature, the aforementioned temperature feature, and the aforementioned duration feature, thus corresponding to a numerical input for each neuron. Taking the input layer as an example, 5×N neurons represent the input layer, which includes 5×N parallelly configured neurons; the hidden layer and output layer are similarly configured. Furthermore, to avoid overfitting, the first trigger frequency classification model employs L2 regularization. The output layer includes 5 neurons corresponding to time scales of seconds, minutes, quarters, hours, and days, respectively. For example, the 5 neurons in the output layer can be: neuron O1, neuron O2, neuron O3, neuron O4, and neuron O5. Among them, the preset trigger frequency value corresponding to neuron O1 is 1 second. The preset trigger frequency value corresponding to neuron O2 is 1 minute. The preset trigger frequency value for neuron O3 is 15 minutes. The preset trigger frequency value for neuron O4 is 1 hour. The preset trigger frequency value for neuron O5 is 1 day. The above-mentioned execution entity can use the preset trigger frequency value corresponding to the time scale with the highest confidence as the trigger frequency, and the (highest) confidence value corresponding to the trigger frequency as the trigger probability value, thereby obtaining the trigger configuration information corresponding to the weather information.

[0042] The third step is to determine the reference strain change for each sensor signal in the above sensor signal group other than the target sensor signal, based on the wavelength offset corresponding to the above sensor signal, the sensitivity coefficient group corresponding to the above sensor signal, the wavelength offset corresponding to the above target sensor signal, and the sensitivity coefficient group corresponding to the above target sensor signal.

[0043] The target sensor signal mentioned above is the sensor signal collected by the fiber Bragg grating sensor fixed in the aforementioned mounting groove. The sensitivity coefficient set includes: temperature sensitivity coefficient and strain sensitivity coefficient. All at least three fiber Bragg grating sensors mentioned above have undergone sensitivity coefficient calibration before use.

[0044] In practice, fiber Bragg grating sensors are affected by both temperature and strain, thus requiring the decomposition of sensitivity coefficients corresponding to temperature and strain. Specifically, wavelength shift, temperature change, strain change, temperature sensitivity coefficient, and strain sensitivity coefficient satisfy the following relationship:

[0045] Wavelength offset = Temperature sensitivity coefficient × Temperature change + Strain sensitivity coefficient × Strain change

[0046] As an example, to minimize errors, this disclosure employs multiple grating optical sensors. Taking three fiber Bragg grating sensors as an example: fiber Bragg grating sensor A, fiber Bragg grating sensor B, and fiber Bragg grating sensor C. For instance, fiber Bragg grating sensor A is used as the grating optical sensor corresponding to the target sensor signal. Since fiber Bragg grating sensor A is placed within a fixed groove for strain constraint, it can be understood as being only affected by stability. Therefore, by combining the wavelength shift, temperature sensitivity coefficient, and strain sensitivity coefficient of the sensor signal corresponding to fiber Bragg grating sensor B, and the wavelength shift, temperature sensitivity coefficient, and strain sensitivity coefficient of the sensor signal corresponding to fiber Bragg grating sensor A, a strain change is obtained as a reference strain change B. Similarly, by combining the wavelength shift, temperature sensitivity coefficient, and strain sensitivity coefficient of the sensor signal corresponding to fiber Bragg grating sensor C, and the wavelength shift, temperature sensitivity coefficient, and strain sensitivity coefficient of the sensor signal corresponding to fiber Bragg grating sensor A, a strain change is obtained as a reference strain change C.

[0047] As another example, since fiber Bragg grating sensors can monitor changes in temperature and stress, the temperature change can be further calculated based on the scheme disclosed herein. In practice, when the temperature change is large, or when the real-time temperature value obtained by superimposing the temperature change on the base temperature value is greater than the preset value, a corresponding temperature anomaly alert can be triggered within the cable well.

[0048] As another example, fiber Bragg grating sensors, as a type of fiber optic sensor, can also detect vibrations transmitted into cable wells. When an abnormal vibration occurs, a corresponding vibration anomaly alert can be triggered.

[0049] Therefore, by combining fiber optic grating sensors, effective monitoring of abnormal temperatures, vibrations, and stresses can be achieved.

[0050] The fourth step is to determine the average value of the changes in each reference strain in the obtained set of reference strain changes, and use it as the target strain change.

[0051] As an example, the set of reference strain changes can include: reference strain change B and reference strain change C. Therefore, the target strain change = (reference strain change B + reference strain change C) / 2. This method can effectively eliminate the subtle sensor errors of each fiber Bragg grating sensor.

[0052] Fifth step: Based on the target strain change and the preset second trigger frequency classification model, generate the trigger configuration information corresponding to the sensor signal group in the trigger configuration information group.

[0053] The second trigger frequency classification model mentioned above is a pre-constructed decision tree model. In practice, the second trigger frequency classification model also includes five preset trigger frequency values ​​at different time scales. These five different time scales are the second, minute, quarter, hour, and day time scales. The executing entity can use the preset trigger frequency value corresponding to the time scale with the highest confidence as the trigger frequency, and the (highest) confidence value corresponding to the trigger frequency as the trigger probability value, thereby obtaining the trigger configuration information corresponding to the sensor signal group.

[0054] Step 103: Control the target camera to acquire target images according to the trigger configuration information group.

[0055] In some embodiments, the aforementioned execution entity can control the target camera to acquire a target image based on the trigger configuration information group. The target image is an image containing the bottom of the cable well. In practice, the aforementioned execution entity can use the trigger frequency value included in the trigger configuration information as the image acquisition frequency of the target camera, and the trigger probability included in the trigger configuration information as the probability of whether the target camera performs image acquisition after the trigger is activated, thereby controlling the target camera to acquire the target image.

[0056] In some optional implementations of some embodiments, the aforementioned execution entity controls the target camera to acquire target images according to the trigger configuration information group, including:

[0057] First, in response to the fact that the trigger frequency values ​​included in the various trigger configuration information in the above trigger configuration information group are different, the following first processing step is performed:

[0058] The first sub-step involves generating a timing clock corresponding to each trigger configuration information in the trigger configuration information group, based on the trigger frequency value included in the trigger configuration information.

[0059] The timing clock is used to trigger the target camera at a fixed frequency. Specifically, the timing clock is implemented using a timer to trigger the target camera to turn on at a fixed frequency.

[0060] The second sub-step is to determine whether to activate the target camera based on the trigger probability value included in the trigger configuration information corresponding to the target timer, in response to the existence of a target timer in the obtained timer group.

[0061] The aforementioned target timer is the timer that reaches the corresponding trigger moment. Specifically, the timer is released upon reaching the trigger moment. The executing entity can randomly determine whether the target camera is turned on based on the trigger probability value. The higher the trigger probability value, the higher the probability that the target camera is turned on.

[0062] In practice, for scenarios with minimal changes within cable wells, conventional fixed-frequency image acquisition methods result in a large number of invalid and repetitive images due to the small differences between them. Furthermore, real-time fixed-frequency acquisition leads to high power consumption, especially with a large number of cable wells, where it generates significant power consumption. Therefore, this disclosure, by combining weather information and sensor signal groups to determine the trigger frequency and probability values, effectively reduces the number of image samples and camera runtime, significantly lowering device power consumption. Simultaneously, combining weather changes and sensor signals effectively filters images acquired in invalid scenarios, avoiding meaningless analysis of invalid images and preventing the inefficient use of computing resources.

[0063] The third sub-step is to control the target camera to turn on and to control the target camera to acquire the target image in response to determining that the target camera is turned on.

[0064] In practice, considering the brief increase in power consumption when the camera is turned on, the target camera can be put into sleep mode when not turned on for short periods. When not turned on for extended periods, it can be turned off, further improving energy efficiency.

[0065] The second step, in response to the fact that the trigger frequency values ​​included in each of the trigger configuration information in the above trigger configuration information group are the same, is to perform the following second processing step:

[0066] The first sub-step is to generate the timing clock corresponding to the trigger configuration information group mentioned above.

[0067] In practice, when the trigger frequency values ​​included in each trigger configuration information are the same, a timing clock can be generated for the trigger configuration information group to avoid timing clock conflicts.

[0068] The second sub-step involves determining whether to activate the target camera based on the average trigger probability when the timer reaches the corresponding trigger moment.

[0069] The average trigger probability is the average of the trigger probabilities included in each trigger configuration information in the aforementioned trigger configuration information group.

[0070] The third sub-step is to control the target camera to turn on and to control the target camera to acquire the target image in response to determining that the target camera is turned on.

[0071] Step 104: Use a pre-trained low-light image enhancement model to perform low-light image enhancement on the target image to obtain the enhanced image.

[0072] In some embodiments, the aforementioned execution entity can perform low-light image enhancement on the target image using a pre-trained low-light image enhancement model to obtain an enhanced image. The images before and after enhancement refer to the target image before and after brightness enhancement. In practice, the low-light image enhancement model can employ the LLNet (Low Light Network) model.

[0073] Optionally, the aforementioned low-light image enhancement model includes: a shallow image feature extraction network and a deep image feature extraction network. The shallow image feature extraction network includes a cross-coding module A1, and the deep image feature extraction network includes an encoder network and a decoder network. The encoder network includes cross-coding modules B1, B2, B3, and B4, which are serially connected. The decoder network includes cross-coding modules C1, C2, C3, and C4. Cross-coding module C4, wherein cross-coding modules C1, C2, C3 and C4 are connected in series, an attention fusion module D1 is provided between cross-coding modules B1 and C1, an attention fusion module D2 is provided between cross-coding modules B2 and C2, an attention fusion module D3 is provided between cross-coding modules B3 and C3, an attention fusion module D4 is provided between cross-coding modules B4 and C4, and a cross-coding module E1 is connected after the attention fusion module D1.

[0074] In some optional implementations of some embodiments, the execution entity performs low-light image enhancement on the target image using a pre-trained low-light image enhancement model to obtain an enhanced image, including:

[0075] The first step is to extract shallow image features from the target image using the shallow image feature extraction network described above, in order to generate shallow image features.

[0076] The second step involves encoding the shallow image features using the encoder network described above to generate encoded image features.

[0077] The third step involves generating the enhanced image using the decoder network, the attention fusion module located between the encoder network and the decoder network, and the cross-coding module E1.

[0078] As an example, see Figure 3The diagram shows the model structure of the low-light image enhancement model. The input to cross-coding module A1 is the target image. The input to cross-coding module B1 is the output of cross-coding module A1. The input to cross-coding module B2 is the output of cross-coding module B1. The input to cross-coding module B3 is the output of cross-coding module B2. The input to cross-coding module B4 is the output of cross-coding module B3. The input to cross-coding module C4 is the output of cross-coding module B4. The input to cross-coding module C3 is the output of attention fusion module D4. The input to cross-coding module C2 is the output of attention fusion module D3. The input to cross-coding module C1 is the output of attention fusion module D4. The output of attention fusion module D1 is the input to cross-coding module E1. The output of cross-coding module E1 is the enhanced image.

[0079] Cross-coding modules A1 and E1 each include three coding sub-modules. Cross-coding modules B1, B2, and C1 each include two coding sub-modules and a downsampling network. Cross-coding module B3 includes four coding sub-modules and a downsampling network. Cross-coding modules C4, C2, and C1 each include two upsampling network sub-coding modules. Cross-coding module C3 includes an upsampling network and four sub-coding modules.

[0080] The sub-encoding module includes: one normalization layer, one convolutional layer (with a 1×1 kernel), two multi-head attention mechanism modules, one normalization layer, and one feedforward network. Specifically, the two multi-head attention mechanism modules are: Multi-head attention mechanism module A1 and Multi-head attention mechanism module A2. Multi-head attention mechanism module A1 includes: a multi-scale convolutional layer A... 11 (Multi-scale Convolution) and multi-head attention mechanism layer A 12 (Multi-head Attention). The multi-head attention mechanism module A2 includes: multi-scale convolutional layers A 21 and multi-head attention mechanism layer A 22 Multi-head attention mechanism module A1 focuses on feature changes in the horizontal direction. Multi-head attention mechanism module A2 also focuses on feature changes in the vertical direction. Multi-head attention mechanism modules A1 and A2 are cross-connected, forming a multi-scale convolutional layer A. 11 The output is not only for the multi-head attention mechanism layer A 12 The input is also the input for the multi-head attention mechanism layer A. 22 The input. Multi-scale convolutional layer A 21 The output is not only for the multi-head attention mechanism layer A 22 The input is also the input for the multi-head attention mechanism layer A. 21 Input.

[0081] Each of the attention fusion modules D1 through D4 includes two average pooling layers, two multilayer perceptrons, one sigmoid activation function, and one ReLU activation function. The two average pooling layers and two multilayer perceptrons are symmetrically configured. Taking attention fusion module D1 as an example, the input includes the outputs of cross-encoding modules B1 and C1. Attention fusion module D1 first performs pooling processing on the two outputs in parallel using the two average pooling layers. Then, the pooling results are input into the multilayer perceptrons in parallel. Next, the outputs of the two multilayer perceptrons are summed. Further, the summed features are processed using the sigmoid activation function to generate a weight matrix. Then, the weight matrix is ​​cross-multiplied with the outputs of cross-encoding modules B1 and C1, respectively. Finally, the two cross-multiplication results are summed and input into the ReLU activation function as the output of attention fusion module D1. Attention fusion modules D3 through D4 follow the same principle and will not be described further.

[0082] The aforementioned low-light image enhancement model, a core inventive point of this disclosure, firstly extracts shallow image features through a shallow image feature extraction network. Then, it encodes and decodes these features using an encoder and decoder network with a symmetrical structure. Specifically, it extracts shallow features through a cross-coding module A1 and suppresses noise in the output features to a certain extent through a cross-coding module E1. Furthermore, considering the feature forgetting problem that occurs in conventional convolutional neural networks as the number of layers increases, this disclosure employs an attention fusion module to fuse features between corresponding modules in the symmetrical network structure. In addition, to reduce the amount of feature processing, especially to reduce the deployment difficulty of the Transformer structure on low-power models, the cross-coding module decomposes feature attention into horizontal and vertical directions and achieves the correlation of attention results through cross-design. This approach not only improves the image enhancement capability but also ensures the model's deployment capability on low-power hardware.

[0083] Step 105: Determine the sedimentation status information corresponding to the cable well based on the enhanced image.

[0084] In some embodiments, the aforementioned execution entity can determine the sedimentation status information corresponding to the cable well based on the enhanced image. The sedimentation status information includes: sedimentation type, sedimentation confidence level, and sedimentation risk level. The sedimentation type can include: water accumulation type, mud accumulation type, and other types. The sedimentation confidence level represents the confidence level corresponding to the sedimentation type. The sedimentation risk level represents the degree of impact on the cable well caused by the sedimentation corresponding to the sedimentation type. In practice, a target detection model can be used as the backbone network, with two classifiers connected after the target detection network to output the sedimentation type and sedimentation risk level, respectively. Specifically, the target detection model can be a conventional model such as the YOLO Nano model.

[0085] In some optional implementations of certain embodiments, the execution entity determines the siltation status information corresponding to the cable well based on the enhanced image, including:

[0086] The first step is to determine the preset pixel value threshold corresponding to the aforementioned cable well.

[0087] The aforementioned preset pixel value threshold represents the pixel value threshold used for foreground and background separation corresponding to cable wells with stable light distribution. In practice, the foreground and background of the enhanced images have clear boundaries, especially considering the stable light distribution environment inside the cable well. Therefore, to reduce the amount of subsequent feature processing and further reduce the data processing pressure and power consumption of the processor, a preset pixel value threshold corresponding to the cable well can be obtained to perform pixel filtering on the enhanced image. When the pixel value of a pixel in the image is within the threshold range centered on the preset pixel value threshold, the pixel value is initialized to 0. By setting the threshold range, the fault tolerance is improved.

[0088] The second step is to separate the foreground and background of the enhanced image based on the preset pixel value threshold to obtain the separated image.

[0089] In the above-mentioned separated image, the pixel value corresponding to the background is initially 0.

[0090] The third step is to generate the aforementioned sedimentation state information based on the separated images and the pre-trained sedimentation state prediction model.

[0091] The sedimentation state prediction model uses the YOLO Nano model as its backbone network, connected to two classifiers that output sedimentation type and sedimentation risk level, respectively. The confidence level corresponding to the sedimentation type is used as the sedimentation confidence level. For example, the classifiers can be implemented using fully connected layers.

[0092] Optionally, the above method further includes:

[0093] The first step is to add the cable well identifier corresponding to the above cable well to the first list in response to the above siltation status information meeting the first preset condition.

[0094] The first preset condition is that the sedimentation type is either water accumulation or mud accumulation, the sedimentation confidence level is greater than the threshold confidence level, and the sedimentation risk level is within the first risk level range.

[0095] The second step is to add the cable well identifier corresponding to the cable well to the second list in response to the above-mentioned siltation status information meeting the second preset condition, and to send a siltation warning to the remote early warning terminal for the above-mentioned cable well.

[0096] Among them, the silt removal priority corresponding to the second list is higher than the silt removal priority corresponding to the first list. The second preset condition is that the silt type is water accumulation or mud accumulation, the silt confidence level is greater than the threshold confidence level, and the silt risk level is within the second risk level range. The remote early warning terminal can be a mobile terminal bound to a user account that maintains the cable well.

[0097] The above-described embodiments of this disclosure have the following beneficial effects: The method for determining the sedimentation state in cable wells under low-light environments, based on some embodiments of this disclosure, achieves accurate determination of the sedimentation state in cable wells, thereby ensuring the service life of cables and reducing the probability of power equipment failure and damage caused by sedimentation in power wells. Specifically, firstly, weather information and a sensor signal group are acquired. The sensor signal group is collected by a sensor assembly installed in the cable well. The sensor assembly includes a mounting frame and at least three fiber Bragg grating sensors. The at least three fiber Bragg grating sensors are parallel to each other and fixed to the mounting frame at a preset interval. The mounting frame includes a fixing groove facing the target side, and the fixing groove is used to fix one of the fiber Bragg grating sensors. Secondly, based on the weather information and the sensor signal group, a trigger configuration information group is generated. The trigger configuration information includes a trigger frequency value and a trigger probability value. The trigger frequency value is used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when triggered. The target camera is installed in the cable well and faces the bottom of the cable well. In practice, siltation is usually caused by the low-lying location of cable wells, especially during adverse weather conditions such as rain and snow, where rainwater and snowmelt carry silt and sand into the wells. Therefore, this disclosure generates trigger configuration information for the target camera by combining weather information and sensor signals. This avoids the problems of large data volume and high power consumption associated with conventional real-time monitoring via cameras. Next, based on the aforementioned trigger configuration information, the target camera is controlled to acquire target images. This disclosure effectively controls the activation and image acquisition of the target camera through trigger frequency and probability values ​​to ensure effective coverage. Furthermore, the target image is enhanced in low light using a pre-trained low-light image enhancement model, resulting in an enhanced image. In practice, cable wells are typical low-light environments, making it difficult to monitor siltation under low illumination. Therefore, this disclosure effectively enhances the image by designing a low-light image enhancement model. Finally, based on the enhanced image, the siltation status information corresponding to the cable well is determined, including: siltation type, siltation confidence level, and siltation risk level. In summary, this method enables accurate determination of the sedimentation status inside cable wells, thereby ensuring the service life of cables and reducing the probability of power equipment failure and damage caused by sedimentation in power wells.

[0098] Further reference Figure 4 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a device for determining the sedimentation state in cable wells under low-light conditions. These device embodiments are similar to... Figure 1Corresponding to the method embodiments shown, this device for determining the sedimentation status in cable wells under low-light conditions can be specifically applied to various electronic devices.

[0099] like Figure 4 As shown in some embodiments, a device 400 for determining the siltation status in cable wells under low-light conditions includes: an acquisition unit 401, a generation unit 402, a control unit 403, a low-light image enhancement unit 404, and a determination unit 405. The acquisition unit 401 is configured to acquire weather information and a set of sensor signals. The sensor signals are acquired by a sensor assembly installed in the cable well. The sensor assembly includes a mounting frame and at least three fiber Bragg grating sensors. The at least three fiber Bragg grating sensors are parallel to each other and fixed to the mounting frame at preset intervals. The mounting frame includes a fixing groove facing the target side, used to fix one of the fiber Bragg grating sensors. The generation unit 402 is configured to generate a trigger configuration information set based on the weather information and the sensor signal set. The trigger configuration information includes a trigger frequency value and a trigger probability value. The trigger frequency value is used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when triggered. The target camera is located inside the cable well and faces the bottom of the cable well. The control unit 403 is configured to control the target camera to acquire target images according to the trigger configuration information group. The low-light image enhancement unit 404 is configured to enhance the target image in low light using a pre-trained low-light image enhancement model to obtain an enhanced image. The determination unit 405 is configured to determine the sedimentation status information corresponding to the cable well based on the enhanced image. The sedimentation status information includes sedimentation type, sedimentation confidence level, and sedimentation risk level.

[0100] It is understandable that the units described in the cable well sedimentation state determination device 400 applied in low-light environments are similar to the reference units. Figure 1 The steps in the described method correspond accordingly. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the device 400 for determining the sedimentation state in cable wells in low-light environments and the units contained therein, and will not be repeated here.

[0101] The following is for reference. Figure 5 It illustrates a schematic diagram of the structure of an electronic device (e.g., a computing device) suitable for implementing some embodiments of the present disclosure. Figure 5 The electronic device shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments of this disclosure. Figure 5As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The memory may include a non-volatile storage medium and internal memory. The non-volatile storage medium may store an operating system and a computer program. The computer program includes program instructions that, when executed, cause the processor to perform any of the methods described above. The processor provides computational and control capabilities to support the operation of the entire computer device. The internal memory provides an environment for the execution of the computer program in the non-volatile storage medium; when executed by the processor, the computer program causes the processor to perform any of the methods described above. The network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the computer device to which the present disclosure is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0102] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0103] In one embodiment, the processor is configured to run a computer program stored in a memory to perform the following steps: acquiring weather information and a set of sensor signals, wherein the set of sensor signals is acquired by a sensor assembly disposed in a cable well, the sensor assembly comprising: a mounting bracket and at least three fiber Bragg grating sensors, the at least three fiber Bragg grating sensors being parallel to each other and fixed to the mounting bracket at a predetermined interval, the mounting bracket comprising: a fixing groove, wherein the fixing groove faces the target side, the fixing groove being used to fix one of the fiber Bragg grating sensors; generating a trigger configuration signal based on the weather information and the set of sensor signals. The system comprises a trigger configuration information group, wherein the trigger frequency value and the trigger probability value are used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when triggered. The target camera is located inside the cable well and faces the bottom of the cable well. According to the trigger configuration information group, the system controls the target camera to acquire target images. The target image is then enhanced in low light using a pre-trained low-light image enhancement model to obtain an enhanced image. Based on the enhanced image, the system determines the sedimentation status information corresponding to the cable well, wherein the sedimentation status information includes sedimentation type, sedimentation confidence level, and sedimentation risk level.

[0104] This disclosure also provides a computer-readable storage medium storing a computer program, the computer program including program instructions, and the method implemented when the program instructions are executed can be referred to the various embodiments of the methods described above.

[0105] The aforementioned computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. Alternatively, the aforementioned computer-readable storage medium may be an external storage device of the computer device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device.

[0106] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0107] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for determining the sedimentation state in cable wells under low-light conditions, characterized in that, include: The system acquires weather information and sensor signal groups, wherein the sensor signal groups are acquired by a sensor assembly installed in a cable well. The sensor assembly includes a mounting frame and at least three fiber Bragg grating sensors, which are parallel to each other and fixed to the mounting frame at a preset interval. The mounting frame includes a fixing groove facing the target side, which is used to fix one of the fiber Bragg grating sensors. Based on the weather information and the sensor signal group, a trigger configuration information group is generated. The trigger configuration information includes a trigger frequency value and a trigger probability value. The trigger frequency value is used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when it is triggered. The target camera is set inside the cable well and faces the bottom of the cable well. According to the trigger configuration information group, control the target camera to acquire target images; The target image is enhanced in low light using a pre-trained low-light image enhancement model to obtain the enhanced image. Based on the enhanced image, the sedimentation status information corresponding to the cable well is determined, wherein the sedimentation status information includes: sedimentation type, sedimentation confidence level, and sedimentation risk level.

2. The method according to claim 1, characterized in that, When the weather changes in the area where the cable well is located, the weather information is actively distributed and synchronized by a remote server. This weather information includes: weather type, predicted precipitation, predicted precipitation probability, predicted temperature, and predicted duration. The step of generating a trigger configuration information group based on the weather information and the sensor signal group includes: The weather type, the predicted precipitation, the predicted precipitation probability, the predicted temperature, and the predicted duration are respectively encoded to obtain weather type features, precipitation features, precipitation probability features, temperature features, and duration features; Based on the pre-trained first trigger frequency classification model, the weather type feature, the precipitation feature, the precipitation probability feature, the temperature feature, and the duration feature, trigger configuration information corresponding to the weather information is generated in the trigger configuration information group. The first trigger frequency classification model includes: 1 input layer, 5 hidden layers, and 1 output layer. The input layer includes: 5×N neurons. The hidden layers include: a first hidden layer, a second hidden layer, a third hidden layer, a fourth hidden layer, and a fifth hidden layer. The first hidden layer and the fifth hidden layer each include 5×N neurons. The second hidden layer and the fourth hidden layer each include 7×N neurons. The third hidden layer includes 9×N neurons. The output layer includes 5×N neurons. The 5 neurons in the output layer correspond to 5 preset trigger frequency values ​​at different time scales.

3. The method according to claim 2, characterized in that, The method further includes: In response to the siltation status information meeting the first preset condition, the cable well identifier corresponding to the cable well is added to the first list; In response to the siltation status information meeting the second preset condition, the cable well identifier corresponding to the cable well is added to the second list, and a siltation warning is sent to the remote early warning terminal for the cable well, wherein the siltation cleaning priority corresponding to the second list is greater than the siltation cleaning priority corresponding to the first list.

4. The method according to claim 3, characterized in that, The step of generating a trigger configuration information group based on the weather information and the sensor signal group further includes: For each sensor signal in the sensor signal group other than the target sensor signal, a reference strain change is determined based on the wavelength offset corresponding to the sensor signal, the sensitivity coefficient group corresponding to the sensor signal, the wavelength offset corresponding to the target sensor signal, and the sensitivity coefficient group corresponding to the target sensor signal. The target sensor signal is a sensor signal collected by a fiber optic grating sensor fixed in the fixing groove. The sensitivity coefficient group includes a temperature sensitivity coefficient and a strain sensitivity coefficient. All at least three fiber optic grating sensors are calibrated for sensitivity coefficients before use. The average value of the changes in each reference strain change in the obtained set of reference strain changes is determined as the target strain change. Based on the target strain change and the preset second trigger frequency classification model, trigger configuration information corresponding to the sensor signal group is generated in the trigger configuration information group, wherein the second trigger frequency classification model is a pre-constructed decision tree model.

5. The method according to claim 4, characterized in that, The step of controlling the target camera to acquire target images according to the trigger configuration information group includes: In response to the fact that the trigger frequency values ​​included in the various trigger configuration information in the trigger configuration information group are different, the following first processing step is performed: For each trigger configuration information in the trigger configuration information group, a timing clock corresponding to the trigger configuration information is generated according to the trigger frequency value included in the trigger configuration information, wherein the timing clock is used to trigger the target camera at a fixed frequency; In response to the presence of a target timer in the obtained timer group, the system determines whether to activate the target camera based on the trigger probability value included in the trigger configuration information corresponding to the target timer, wherein the target timer is the timer that reaches the corresponding trigger time. In response to determining that the target camera is to be turned on, the system controls the target camera to be turned on and controls the target camera to acquire the target image; Since the trigger frequency values ​​included in each trigger configuration information in the trigger configuration information group are the same, the following second processing step is performed: Generate the timing clock corresponding to the trigger configuration information group; In response to the timing clock reaching the corresponding trigger time, it is determined whether to activate the target camera based on the average trigger probability, wherein the average trigger probability is the average of the trigger probabilities included in each trigger configuration information in the trigger configuration information group; In response to determining that the target camera is to be turned on, the system controls the target camera to be turned on and controls the target camera to acquire the target image.

6. The method according to claim 5, characterized in that, The low-light image enhancement model includes a shallow image feature extraction network and a deep image feature extraction network. The shallow image feature extraction network includes a cross-coding module A1. The deep image feature extraction network includes an encoder network and a decoder network. The encoder network includes cross-coding modules B1, B2, B3, and B4, which are serially connected. The decoder network includes cross-coding modules C1, C2, C3, and C4. Block C2, cross-coding module C3, and cross-coding module C4 are connected in series. An attention fusion module D1 is provided between cross-coding module B1 and cross-coding module C1, an attention fusion module D2 is provided between cross-coding module B2 and cross-coding module C2, an attention fusion module D3 is provided between cross-coding module B3 and cross-coding module C3, and an attention fusion module D4 is provided between cross-coding module B4 and cross-coding module C4. A cross-coding module E1 is connected after the attention fusion module D1. The process of enhancing the target image in low light using a pre-trained low-light image enhancement model to obtain an enhanced image includes: The shallow image feature extraction network is used to extract shallow image features from the target image to generate shallow image features. The shallow image features are encoded using the encoder network to generate encoded image features. The enhanced image is generated through the decoder network, the attention fusion module located between the encoder network and the decoder network, and the cross-coding module E1.

7. The method according to claim 6, characterized in that, The step of determining the sedimentation status information corresponding to the cable well based on the enhanced image includes: A preset pixel value threshold is determined for the cable well, wherein the preset pixel value threshold represents the pixel value threshold for foreground and background separation for cable wells with stable light distribution. Based on the preset pixel value threshold, the enhanced image is separated into foreground and background to obtain a separated image, wherein the pixel value corresponding to the background in the separated image is initially 0; The sedimentation status information is generated based on the separated image and the pre-trained sedimentation status prediction model.

8. A device for determining the sedimentation state in cable wells under low-light conditions, characterized in that, include: The acquisition unit is configured to acquire weather information and a sensor signal group, wherein the sensor signal group is acquired by a sensor assembly installed in a cable well. The sensor assembly includes: a mounting frame and at least three fiber Bragg grating sensors, wherein the at least three fiber Bragg grating sensors are parallel to each other and fixed on the mounting frame at a preset interval. The mounting frame includes: a fixing groove, wherein the fixing groove faces the target side, and the fixing groove is used to fix one of the fiber Bragg grating sensors. The generation unit is configured to generate a trigger configuration information group based on the weather information and the sensor signal group. The trigger configuration information includes a trigger frequency value and a trigger probability value. The trigger frequency value is used to control the image acquisition frequency of the target camera, and the trigger probability value is used to control the activation probability of the target camera when it is triggered. The target camera is located inside the cable well and faces the bottom of the cable well. The control unit is configured to control the target camera to acquire target images according to the trigger configuration information group; The low-light image enhancement unit is configured to perform low-light image enhancement on the target image using a pre-trained low-light image enhancement model to obtain an enhanced image; The determining unit is configured to determine the sedimentation status information corresponding to the cable well based on the enhanced image, wherein the sedimentation status information includes: sedimentation type, sedimentation confidence level, and sedimentation risk level.

9. An electronic device, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 7.

10. A computer-readable medium, characterized in that, It stores a computer program thereon, wherein the computer program, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images