Intelligent water conservancy safety detection system based on AI edge computing

Abstract

The application discloses an intelligent water conservancy safety detection system based on AI edge calculation, belongs to the technical field of water conservancy safety detection calculation, and comprises a sensing layer used for collecting real-time water conservancy data arranged at each preset detection point, an edge calculation layer used for determining environmental influence factors after weight distribution calculation based on rainfall, water flow speed, wind speed, shelter probability and wind direction influence factors, determining visible influence factors after weight distribution calculation based on visibility, visible range and environmental influence factors, determining the detection priority of each preset detection point based on the visible influence factors, detection use probability, distance from a preset position and the influence degree of rainfall on each preset detection point, and a management layer used for determining the final preset detection point, wherein the application comprehensively considers environmental factors, introduces energy consumption constraints, fully utilizes historical experience and realizes self-adaptive optimization of the system, and thus the performance and reliability of the intelligent water conservancy safety detection system based on AI edge calculation are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of water conservancy safety monitoring and calculation technology, and in particular to a smart water conservancy safety monitoring system based on AI edge computing. Background Technology

[0002] In recent years, the development of AI edge computing technology has brought new opportunities for smart water conservancy safety monitoring. By deploying a large number of sensors and cameras in the water conservancy system, and using AI edge computing devices to process and analyze the collected data in real time, comprehensive and real-time monitoring of the water environment can be achieved.

[0003] Regarding AI edge computing technology, current technologies often only focus on one or a few environmental factors. Therefore, in strong winds, wind speeds from different directions can blow obstructions onto the detection equipment, blocking its view. Current technologies do not adequately consider this factor, affecting detection accuracy. Secondly, frequent movement of the detection equipment increases energy consumption, shortens its lifespan, and increases operating costs. However, current technologies typically do not consider the energy consumption associated with equipment movement. Furthermore, current technologies do not fully utilize historical data and experience, lacking consideration of historical usage when selecting detection locations. This results in the system being unable to quickly and accurately select the optimal detection location in similar environments. Summary of the Invention

[0004] The technical problem to be solved by this invention is that the existing technology has the shortcomings of not fully considering environmental factors, lacking energy consumption constraints, and not making full use of historical experience. To this end, we propose a smart water conservancy safety detection system based on AI edge computing.

[0005] The technical solution mainly consists of an AI edge computing-based smart water conservancy safety monitoring system, comprising a perception layer, an edge computing layer, a transmission layer, and a management layer, specifically implemented as follows:

[0006] The sensing layer is used to collect real-time water conservancy data, including rainfall, water flow velocity, visibility, visible range and distance from preset locations, deployed at various preset detection points.

[0007] The edge computing layer is used to extract from the management layer the wind speed occlusion probability of the current wind speed, the wind direction influence factor of the current wind direction, and the detection usage probability of the current preset detection point under the same wind direction.

[0008] Environmental impact factors are determined based on the weighted calculations of the rainfall, water flow velocity, wind speed obstruction probability, and wind direction influence factor.

[0009] Based on the weighted calculation of the visibility, the visible range, and the environmental impact factors, the visual impact factor is determined.

[0010] Based on the visual impact factor, the detection usage probability, the distance to the preset location, and the degree of impact of the rainfall on each preset detection point, the detection priority of each preset detection point is determined.

[0011] The transport layer is used for transmissions from the perception layer to the edge computing layer, from the management layer to the edge computing layer, and from the perception layer and the edge computing layer to the management layer.

[0012] The management layer is used for unified access and management of the perception layer and the edge computing layer;

[0013] Based on the ranking of the detection priorities of each preset detection point, the final preset detection points are determined.

[0014] Preferably, the devices used in the sensing layer include rainfall sensors, water flow velocity sensors, visibility sensors, and camera field of view monitoring devices and position sensors;

[0015] The devices used in the edge computing layer include AI edge computing boxes;

[0016] The devices used in the transport layer include wireless routers;

[0017] The equipment used by the management team includes an intelligent management platform.

[0018] Preferably, the rainfall sensor is used to acquire the rainfall amount;

[0019] The water flow velocity sensor is used to acquire the water flow velocity;

[0020] The wind direction characteristic influence coefficient is determined based on the product of the wind speed obstruction probability and the wind direction influence factor.

[0021] The environmental impact factors are determined by weighting the wind direction characteristic influence coefficient, the rainfall, and the water flow velocity.

[0022] The weighted values ​​of the wind direction characteristic influence coefficient, the rainfall, and the water flow velocity are all 1.

[0023] Preferably, the management layer counts the number of times the viewpoint of each preset detection point is blocked and the total number of observations under the current wind direction and wind speed in historical data, and transmits the data to the edge computing layer through the transmission layer. The edge computing layer uses the ratio of the number of times the viewpoint is blocked to the total number of observations to determine the wind speed blocking probability.

[0024] The sensing layer collects the angle between the current wind direction and each preset detection point, and transmits it to the edge computing layer through the transmission layer. The edge computing layer determines the wind direction influence factor using a simple angle interval division method, specifically as follows:

[0025] When 0°≤the included angle<45° or 315°≤the included angle<360°, the wind direction influence factor is 0.8. This is because the current wind direction is directly opposite the current preset detection point, and the influence is large.

[0026] When 45° ≤ the included angle < 135° or 225° ≤ the included angle < 315°, the wind direction influence factor is 0.5. This is because the angle between the current wind direction and the current preset detection point is moderate, and the influence is also moderate.

[0027] When 135° ≤ the included angle < 225°, the wind direction influence factor is 0.2, because the current wind direction is opposite to the current preset detection point, and the influence is small.

[0028] Based on the above description of the wind direction influencing factors, the wind direction detected by each preset detection point specifically includes the current wind direction facing directly, diagonally, and away from the current preset detection point.

[0029] Under the same wind direction, the influence of the wind direction on any one of the preset detection points, from largest to smallest, is: facing directly → diagonally → facing away.

[0030] Preferably, the visibility sensor is used to acquire the visibility;

[0031] The visible range monitoring device is used to acquire the visible range;

[0032] The visibility, the visible range, and the environmental impact factors are weighted and assigned to determine the visual impact factors.

[0033] The weights of visibility, visible range, and environmental impact factor are all set to 1.

[0034] Preferably, the maximum probability of the detection is preset to a constant 1;

[0035] Subtracting the detection probability from the constant 1 determines the reverse probability consideration value;

[0036] The position sensor is used to obtain the distance between the current preset detection point and the preset actual detection point that was actually detected under the previous wind speed in the same direction;

[0037] The energy consumption impact coefficient is determined after the weighting of the distance to the preset position plus the weight of the moving distance.

[0038] Based on the product of the probability reverse consideration value, the visual impact factor, and the energy consumption impact coefficient, a comprehensive impact coefficient is determined that reflects the visual detection effect of the current preset detection point and the energy consumption factor of moving to that location on the detection priority FST after excluding the advantages and disadvantages of historical usage experience.

[0039] Based on the product of the detection usage probability and the rainfall, the characteristics of the current preset detection point in responding to water conservancy safety detection are determined, which reflect the comprehensive evaluation characteristics of the current preset detection point in response to water conservancy safety detection, taking into account historical usage experience and current rainfall factors.

[0040] The detection priority is determined by adding the comprehensive influence coefficient to the comprehensive evaluation feature.

[0041] Preferably, the rainfall, water flow velocity, visibility, visible range, and distance from the preset position are all normalized before being incorporated into the calculation by the edge computing layer.

[0042] Preferably, based on the detection priority determined for each preset detection point, the detection priorities of each preset detection point are arranged in order of magnitude to obtain a priority arrangement group;

[0043] Extract the detection priority with the maximum value in the priority arrangement group, and use the corresponding preset detection point as the final preset detection point for actual detection under the current wind speed;

[0044] The preset detection point that is actually detected under any wind speed in the same direction is the preset actual detection point under the next wind speed in the same direction.

[0045] The technical effects and advantages of this invention are as follows:

[0046] In this invention, firstly, by using an edge computing layer to perform weighted calculations on various environmental factors, including rainfall, water flow velocity, wind speed, obstruction probability, and wind direction influence factors, the overall impact of the environment on detection can be more comprehensively reflected. Secondly, the visibility influence factor combines environmental influence factors with visualization-related factors such as visibility and visible range, further considering the impact of the environment on the visualization of detection points. This allows the system to consider not only the overall impact of the environment but also the actual observation effect of the detection equipment when selecting detection locations, thereby improving the accuracy and reliability of detection.

[0047] In this invention, when calculating the detection priority at the edge computing layer, preset position distance and movement distance weights are introduced. These weights take into account the energy consumption caused by camera movement, enabling the system to achieve energy saving while ensuring detection performance.

[0048] In this invention, by considering the probability of using the current preset detection point under the same wind direction in history, historical data and experience are fully utilized, so that when the system selects the preset position, it gives priority to the detection position with high historical usage frequency and high reliability, thereby improving the system's adaptability and stability in similar environments. Attached Figure Description

[0049] Figure 1 A flowchart illustrating the process of determining the final preset detection points for this intelligent water conservancy safety monitoring system;

[0050] Figure 2 This is a schematic diagram of the equipment used in the functional layers of this intelligent water conservancy safety monitoring system.

[0051] Figure 3 This is a schematic diagram illustrating the determination of the wind direction influence factor FX using the simple angle interval division method in this invention. Detailed Implementation

[0052] The present invention will now be described in further detail with reference to the accompanying drawings and preferred embodiments.

[0053] Reference Figures 1-3 As shown, this invention provides a technical solution: a smart water conservancy safety monitoring system based on AI edge computing, comprising a perception layer, an edge computing layer, a transmission layer, and a management layer, specifically implemented as follows:

[0054] The perception layer is used to collect real-time water conservancy data, including rainfall, water flow velocity, visibility, visible range, and distance from the preset location, deployed at various preset detection points.

[0055] The devices used in the perception layer include rainfall sensors, water flow velocity sensors, visibility sensors, and camera field-of-view monitoring devices and position sensors;

[0056] The edge computing layer is used to extract the wind speed occlusion probability, the wind direction influence factor, and the detection usage probability of the current preset detection point under the same wind direction from the management layer.

[0057] Environmental impact factors are determined based on the weighted calculations of rainfall, water flow velocity, wind speed obstruction probability, and wind direction influence factors.

[0058] Based on the weighted calculation of visibility, visible range and environmental impact factors, the visibility impact factor is determined;

[0059] Based on the visual impact factor, the probability of detection use, the distance from the preset location, and the degree of impact of rainfall on each preset detection point, the detection priority of each preset detection point is determined.

[0060] The devices used in the edge computing layer include AI edge computing boxes;

[0061] The transport layer is used for transmission between the perception layer and the edge computing layer, between the management layer and the edge computing layer, and between the perception layer and the edge computing layer and the management layer.

[0062] The devices used in the transport layer include wireless routers;

[0063] The management layer is used for unified access and management of the perception layer and the edge computing layer.

[0064] Based on the ranking of the detection priorities of each preset detection point, the final preset detection points are determined;

[0065] The equipment used by the management team includes an intelligent management platform.

[0066] Based on this embodiment, the system utilizes the collaborative efforts of various functional sensors and cameras in the perception layer, the AI ​​edge computing box in the edge computing layer, the wireless router in the transmission layer, and the intelligent management platform in the management layer to comprehensively consider various environmental factors and energy consumption constraints.

[0067] Specifically, the identified environmental impact factors provide the system with an accurate comprehensive environmental impact assessment, the visible impact factors further consider visual factors based on the environmental impact factors, and the detection priority (FST) is determined by combining historical experience and energy consumption factors. This enables the system to adaptively select the optimal detection location in complex and ever-changing water conservancy environments, improve the accuracy and timeliness of detection, reduce energy consumption, extend equipment life, and achieve intelligent, efficient, and energy-saving water conservancy safety detection.

[0068] In addition, rainfall, water flow velocity, visibility, visible range, and distance from the preset location all need to be normalized before being introduced into the edge computing layer.

[0069] The management layer stores the maximum and minimum values ​​of rainfall, water flow velocity, visibility, visible range, and distance from the preset location, and transmits them to the edge computing layer through the transmission layer. Normalization calculation is used to determine the normalized rainfall, water flow velocity, visibility, visible range, and distance from the preset location, thereby ensuring the consistency of the calculated physical quantities.

[0070] Reference Figures 1-3 As shown in this embodiment: the rainfall sensor is used to acquire rainfall;

[0071] A water flow velocity sensor is used to obtain the water flow velocity;

[0072] The wind direction characteristic influence coefficient is determined by multiplying the wind speed obstruction probability and the wind direction influence factor.

[0073] The environmental impact factors are determined by weighting the wind direction characteristic influence coefficient, rainfall, and water flow velocity.

[0074] Among them, the weighted distribution and value of the wind direction characteristic influence coefficient, rainfall, and water flow velocity are 1.

[0075] Based on this embodiment, the calculation formula for the environmental impact factor is as follows:

[0076] ;

[0077] ;

[0078] in:

[0079] HX represents the environmental impact factor;

[0080] FY is the wind direction characteristic influence coefficient, R is the rainfall, and L is the water flow velocity;

[0081] a1, a2, and a3 are all weighting coefficients, and a1 + a2 + a3 = 1;

[0082] ZG represents the probability of wind speed obstruction, and FX represents the wind direction influence factor.

[0083] In actual strong wind conditions, wind speeds from different directions can cause obstructions to block the camera's view. The visibility monitoring device can not only calculate the probability of wind speed obstruction at each preset detection point based on the number of times the view is blocked and the total number of observations, thus considering the degree of obstruction under different wind directions, but also provide timely warnings about the current camera's view obstruction. Specifically, if the visibility monitoring device shows that there is view obstruction at each preset detection point, it will send a warning signal to the management team, facilitating the timely removal of obstructions. Thus, by combining the influence factors of wind direction, rainfall, and water flow speed, the environmental impact factor HX is calculated, enabling the system to comprehensively assess the combined impact of these environmental factors on the detection, thereby ensuring the accuracy of water conservancy safety monitoring.

[0084] Furthermore, the management team compiles statistics on the number of times each preset detection point's view is blocked and the total number of observations under the current wind direction and speed in historical data. This data is then transmitted to the edge computing layer through the transmission layer. The edge computing layer uses the ratio of the number of times the view is blocked to the total number of observations to determine the wind speed blocking probability.

[0085] The perception layer collects the angle between the current wind direction and each preset detection point, and transmits it to the edge computing layer through the transmission layer. The edge computing layer uses a simple angle interval division method to determine the wind direction influencing factor.

[0086] The specific method for determining the wind direction influence factor FX is as follows:

[0087] When 0°≤angle<45° or 315°≤angle<360°, the wind direction influence factor FX is 0.8. This is because the current wind direction is directly opposite the current preset detection point, and the influence is large.

[0088] When 45° ≤ angle < 135° or 225° ≤ angle < 315°, the wind direction influence factor FX is 0.5. This is because the angle between the current wind direction and the current preset detection point is moderate, and the influence is also moderate.

[0089] When 135° ≤ included angle < 225°, the wind direction influence factor FX is 0.2. This is because the current wind direction is opposite to the current preset detection point, and its influence is small.

[0090] Reference Figure 1 and Figure 2 As shown in this embodiment: the visibility sensor is used to acquire visibility;

[0091] Visibility range monitoring equipment is used to obtain the visible range;

[0092] The visibility impact factors are determined by assigning weights to visibility, visible range, and environmental impact factors.

[0093] The weights of visibility, visible range, and environmental impact factors are all set to 1.

[0094] Based on this embodiment, the formula for calculating the visual impact factor is as follows:

[0095] ;

[0096] in:

[0097] KX is the visible impact factor;

[0098] KL stands for visibility, and KF stands for visible range.

[0099] a4, a5, and a6 are all weighting coefficients, and a4+a5+a6=1.

[0100] In actual water conservancy safety inspections:

[0101] If visibility is low in KL, the camera may not be able to clearly capture floating objects or illegal vessels on the water.

[0102] If the visible range KF is small, some areas will not be able to be monitored;

[0103] Therefore, by using the calculation formula of the visible impact factor KX, the Linfen system can comprehensively evaluate the impact of environmental and visual factors on detection, providing a more comprehensive reference for selecting the best detection location. This helps to improve the accuracy and timeliness of detection and promptly identify potential safety hazards in the water conservancy system.

[0104] Reference Figure 1 and Figure 2 As shown in this implementation scheme: the maximum probability of detecting the probability of use is preset to a constant of 1;

[0105] Subtracting the probability of detection from the constant 1 determines the reverse probability consideration value;

[0106] The position sensor is used to obtain the preset position distance between the current preset detection point and the preset actual detection point under the previous wind speed in the same direction;

[0107] After calculating the weight of the preset distance plus the moving distance weight, the energy consumption impact coefficient is determined;

[0108] Based on the product of probability inverse consideration value, visual impact factor and energy consumption impact coefficient, the comprehensive impact coefficient is determined, which reflects the visual detection effect of the current preset detection point and the energy consumption factor of moving to the location on the detection priority FST after excluding the advantages and disadvantages of historical usage experience.

[0109] Based on the product of the probability of detection use and the amount of rainfall, the characteristics of the current preset detection points in response to water conservancy safety detection are determined, which reflect the comprehensive evaluation characteristics of the current preset detection points in response to water conservancy safety detection, taking into account historical usage experience and current rainfall factors.

[0110] The overall impact coefficient and the overall evaluation characteristics are added together to determine the detection priority.

[0111] Based on this embodiment, the calculation formula for the detection priority is as follows:

[0112] ;

[0113] in:

[0114] FST stands for detection priority.

[0115] FG represents the probability of detection and use, QL represents the distance to the preset position, and q represents the weight of the moving distance.

[0116] The calculation result is the energy consumption impact coefficient;

[0117] The calculation result is a comprehensive impact coefficient. This part of the calculation is a quantitative contribution value to the detection priority FST after considering historical usage, visual impact, and camera movement energy consumption factors. The specific content reflected is as follows:

[0118] Impact of historical usage: The calculation results reflect a reverse consideration of the probability of using this angle under the same historical wind direction. If the detection probability FG is small, it means that the current preset detection point has a low frequency of use under the same historical wind direction. The result is large, at which point the visible influencing factors KX and mobile energy consumption factors are significant. The impact weight on detection priority (FST) is relatively large. This indicates that when evaluating detection priority (FST), the system will place more emphasis on current visibility conditions and mobility energy consumption factors when historical usage experience is insufficient. Conversely, if the detection usage probability (FG) is high, it means that the current preset detection point has been frequently used historically under the same wind direction, indicating high reliability. The result is smaller, with visible influencing factors KX and mobile energy consumption factors. The impact weight on detection priority FST is relatively reduced, meaning the system is more inclined to refer to historical experience.

[0119] The role of visibility impact: The visibility impact factor KX comprehensively considers the impact of environmental impact factor HX, visibility KL, and visible range KF on the visualization of the detection point. It reflects the quality of the visualization detection effect of the current preset detection point under the current environment. The higher the KX value, the more beneficial the current preset detection point is to water conservancy safety detection in terms of visualization. This is achieved through comparison with... and The more they are multiplied, the more significant the improvement in detection priority (FST) becomes.

[0120] Considerations for mobile energy consumption: This part considers the energy consumption impact of the camera moving from its current position to a preset position. The larger the distance QL from the preset position, the higher the energy consumption. By multiplying by the movement distance weight q and adding 1, the movement distance has a negative impact on the detection priority FST. That is, the longer the movement distance, the larger this calculation result becomes. Multiplying by KX reduces the boost to detection priority FST, thus lowering the priority of preset positions that consume too much energy due to excessive movement distance. This demonstrates that when selecting preset positions, the system considers not only detection performance but also energy consumption costs to optimize resource utilization.

[0121] The calculation results are comprehensive evaluation features. If the detection usage probability FG is high and the rainfall R is also high, it indicates that the preset location has performed well in the same wind direction in history. At the same time, the current rainfall R is high and may bring higher risks. Therefore, the current preset detection point may be of great significance for dealing with water conservancy safety detection related to rainfall under the current situation, and has a positive effect on improving the detection priority FST.

[0122] Conversely, if the detection usage probability FG is low and the rainfall R is also low, it indicates that the current preset detection point has insufficient historical usage experience and the current rainfall R has little impact on the detection, so its contribution to the detection priority FST is relatively small.

[0123] Reference Figure 1 and Figure 2 As shown in this implementation scheme: based on the detection priority determined by each preset detection point, the detection priorities of each preset detection point are arranged in order of magnitude to obtain a priority arrangement group;

[0124] Extract the detection priority of the maximum value in the priority group and use its corresponding preset detection point as the final preset detection point for actual detection under the current wind speed.

[0125] In this context, the preset detection point that is actually tested under any wind speed in the same direction will become the preset actual detection point under the next wind speed in the same direction.

[0126] Based on this embodiment, firstly, the obtained priority arrangement group can quickly determine and rotate the camera displacement to the best preset detection point among each preset detection point, and the dynamic setting of the preset actual detection point is also more conducive to real-time changes according to recent changes.

[0127] In addition, when the detection priority FST value is high, it indicates that the current preset detection point has a high priority. This may be because one of the environmental factors among rainfall R, water flow velocity L, and wind direction characteristic influence coefficient FY is beneficial to the detection. In this case, the weight of a certain factor in the calculation of environmental impact factor HX can be appropriately increased.

[0128] Conversely, when the detection priority FST value is low, it indicates that the current preset detection point has a low priority. This may be because one of the environmental factors among rainfall R, water flow velocity L, and wind direction characteristic influence coefficient FY has a significant negative impact on the detection. In this case, the weight of a certain factor in the calculation of the environmental impact factor HX can be appropriately reduced.

[0129] In this way, the detection priority FST is cyclically affected when calculating the environmental impact factor HX, enabling the system to dynamically adjust the degree of attention to different environmental factors according to the actual situation, thereby improving the rationality of the selection of preset detection points and the detection efficiency of the system.

[0130] It should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should also be within the scope of protection of this invention.

Claims

1. A smart water conservancy safety monitoring system based on AI edge computing, characterized in that: It includes a perception layer, an edge computing layer, a transport layer, and a management layer, which are implemented as follows: The sensing layer is used to collect real-time water conservancy data, including rainfall, water flow velocity, visibility, visible range and distance from preset locations, deployed at various preset detection points. The edge computing layer is used to extract from the management layer the wind speed occlusion probability of the current wind speed, the wind direction influence factor of the current wind direction, and the detection usage probability of the current preset detection point under the same wind direction. Environmental impact factors are determined based on the weighted calculations of the rainfall, water flow velocity, wind speed obstruction probability, and wind direction influence factor. Based on the weighted calculation of the visibility, the visible range, and the environmental impact factors, the visual impact factor is determined. Based on the visual impact factor, the detection usage probability, the distance to the preset location, and the degree of impact of the rainfall on each preset detection point, the detection priority of each preset detection point is determined. The transport layer is used for transmissions from the perception layer to the edge computing layer, from the management layer to the edge computing layer, and from the perception layer and the edge computing layer to the management layer. The management layer is used for unified access and management of the perception layer and the edge computing layer; Based on the ranking of the detection priorities of each preset detection point, the final preset detection points are determined.

2. The smart water conservancy safety monitoring system based on AI edge computing according to claim 1, characterized in that: The devices used in the perception layer include rainfall sensors, water flow velocity sensors, visibility sensors, and camera field-of-view monitoring devices and position sensors. The devices used in the edge computing layer include AI edge computing boxes; The devices used in the transport layer include wireless routers; The equipment used by the management team includes an intelligent management platform.

3. The smart water conservancy safety monitoring system based on AI edge computing according to claim 2, characterized in that: The rainfall sensor is used to acquire the rainfall amount; The water flow velocity sensor is used to acquire the water flow velocity; The wind direction characteristic influence coefficient is determined based on the product of the wind speed obstruction probability and the wind direction influence factor. The environmental impact factors are determined by weighting the wind direction characteristic influence coefficient, the rainfall, and the water flow velocity. The weighted values ​​of the wind direction characteristic influence coefficient, the rainfall, and the water flow velocity are all 1.

4. The smart water conservancy safety monitoring system based on AI edge computing according to claim 3, characterized in that: The management layer calculates the number of times each preset detection point view is blocked and the total number of observations under the current wind direction and wind speed in historical data, and transmits it to the edge computing layer through the transmission layer. The edge computing layer uses the ratio of the number of times blocked to the total number of observations to determine the wind speed blocking probability. The sensing layer collects the angle between the current wind direction and each preset detection point, and transmits it to the edge computing layer through the transmission layer. The edge computing layer uses a simple angle interval division method to determine the wind direction influence factor.

5. The smart water conservancy safety monitoring system based on AI edge computing according to claim 3, characterized in that: The visibility sensor is used to acquire the visibility. The visible range monitoring device is used to acquire the visible range; The visibility, the visible range, and the environmental impact factors are weighted and assigned to determine the visual impact factors. The weights of visibility, visible range, and environmental impact factor are all set to 1.

6. The smart water conservancy safety monitoring system based on AI edge computing according to claim 5, characterized in that: The maximum probability of the detection usage probability is preset to a constant 1; Subtracting the detection probability from the constant 1 determines the reverse probability consideration value; The position sensor is used to obtain the distance between the current preset detection point and the preset actual detection point that was actually detected under the previous wind speed in the same direction; The energy consumption impact coefficient is determined by weighting the distance to the preset position. The comprehensive impact coefficient is determined based on the product of the probability reverse consideration value, the visible impact factor, and the energy consumption impact coefficient. The comprehensive evaluation characteristics are determined based on the product of the detection usage probability and the rainfall amount; The detection priority is determined by adding the comprehensive influence coefficient to the comprehensive evaluation feature.

7. The smart water conservancy safety monitoring system based on AI edge computing according to claim 5, characterized in that: The rainfall, water flow velocity, visibility, visible range, and distance from the preset position all need to be normalized before being incorporated into the edge computing layer.

8. The smart water conservancy safety monitoring system based on AI edge computing according to claim 6, characterized in that: Based on the detection priority determined for each preset detection point, the detection priorities of each preset detection point are arranged in order of magnitude to obtain a priority arrangement group; Extract the detection priority with the maximum value in the priority arrangement group, and use the corresponding preset detection point as the final preset detection point for actual detection under the current wind speed; The preset detection point that is actually detected under any wind speed in the same direction is the preset actual detection point under the next wind speed in the same direction.

Images