An intelligent terminal monitoring system and method based on edge computing

Abstract

The present invention discloses an intelligent terminal monitoring system and method based on edge computing, which relates to the technical field of terminal monitoring. By obtaining the monitoring area data of traffic monitoring cameras corresponding to each faulty edge computing center, the importance coefficient of the cameras is calculated; and a target edge computing center capable of receiving the data of the faulty edge computing center is screened out from each non-faulty edge computing center; the adaptation coefficient between the target edge computing center and the faulty edge computing center is calculated; according to the adaptation coefficient and the importance coefficient, the final target edge computing center of each faulty edge computing center is determined, and the traffic monitoring camera data is transmitted to this center for analysis; the final target edge computing center analyzes the data in combination with the importance coefficient of the camera to monitor whether the camera is faulty; it ensures that the traffic monitoring camera data can be processed in a timely and efficient manner, and at the same time ensures that the edge computing center can reasonably allocate the load, guaranteeing the stability and accuracy of the monitoring system.

Description

Technical Field

[0001] The present invention relates to the technical field of terminal monitoring, and particularly to an intelligent terminal monitoring system and method based on edge computing. Background Art

[0002] With the acceleration of the urbanization process, real-time monitoring and management become particularly important; traditional traffic flow analysis relies on a centralized data processing architecture, which has a relatively high latency. Especially when dealing with large-scale data and limited bandwidth, it is difficult to meet real-time requirements. To address this challenge, edge computing is introduced, which can push data processing closer to the data source to achieve local analysis, reduce network latency, and improve response speed; in the application field of traffic monitoring, multiple edge computing centers are set up, and an edge computing-based terminal monitoring system is set up in the edge computing center with the maximum computing power. The edge computing-based terminal monitoring system allocates the terminals that need to be calculated and analyzed to each edge computing center. Each edge computing center, according to the terminals allocated by the edge computing center-based terminal monitoring system, such as traffic monitoring cameras, analyzes the allocated traffic monitoring cameras to determine whether there are any faults in the traffic monitoring cameras and issues a reminder in a timely manner to ensure that the traffic monitoring cameras can collect correct image data, and identify key events such as traffic flow and vehicle speed based on the correct image data, so as to monitor the traffic conditions in real time; moreover, the edge computing-based terminal monitoring system also has a redundant design and an emergency mechanism. For example, when a local edge computing center fails, the data of the traffic monitoring cameras in that area is transmitted to other edge computing centers for analysis, thereby effectively balancing computing resources, reducing latency, and improving the response speed and processing capacity of the overall system, realizing the efficient monitoring of traffic monitoring cameras.

[0003] For example, when a certain edge computing center fails, the edge computing-based terminal monitoring system can, through analysis and judgment, select the most suitable edge center among other available edge computing centers to process the data of the terminal devices corresponding to the original faulty edge center; however, when multiple edge computing centers have problems, it is possible that a certain edge computing center is used as a backup by multiple faulty edge computing centers, which may lead to an overloaded load on this edge computing center, a decline in processing capacity, and even delays or overloaded operations. This approach that overly relies on a single backup center may affect the response speed and stability of the system, thereby resulting in problems with the monitoring results of traffic monitoring cameras. Summary of the Invention

[0004] The purpose of the present invention is to solve the above-mentioned problems and provide an intelligent terminal monitoring system and method based on edge computing.

[0005] In the first aspect of the implementation of the present invention, an intelligent terminal monitoring method based on edge computing is first proposed. The method includes:

[0006] Obtain the traffic area data monitored by each traffic monitoring camera, which is used to evaluate the complexity of the traffic area, and calculate the importance coefficient of the camera according to the traffic area data;

[0007] Obtain the resource information of each non-faulty edge computing center. According to the resource information of each non-faulty edge computing center, screen out each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center, which is called the target edge computing center;

[0008] Obtain the distances and historical data transfer times between the target edge computing centers and each faulty edge computing center, and calculate the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the distances and historical data transfer times;

[0009] Determine the final target edge computing center selected by each faulty edge computing center according to the adaptation coefficient and the importance coefficient, and transfer the corresponding traffic monitoring camera data to the selected final target edge computing center for data analysis;

[0010] The final target edge computing center analyzes the transferred traffic monitoring camera data, and monitors whether the corresponding traffic monitoring camera is faulty in combination with the importance coefficient of the transferred traffic monitoring camera.

[0011] Optionally, the steps for calculating the importance coefficient of the camera according to the traffic area data are as follows:

[0012] Take the area photographed by the traffic monitoring camera as the target area, obtain the total number of vehicles that have passed through the target area in the past preset time period, and obtain the driving time spent by each vehicle in the target area. Calculate the variance of the driving time spent according to the driving time spent by all vehicles in the target area as the vehicle speed fluctuation value of the target area;

[0013] Obtain the vehicles that have had traffic accidents in the target area, and divide the vehicles that have had traffic accidents by the total number of vehicles that have passed through to obtain the vehicle accident probability of the target area;

[0014] Obtain the drivable area of the lanes in the target area, and divide the total number of vehicles that have passed through by the drivable area of the lanes to obtain the vehicle density of the target area;

[0015] Calculate the importance coefficient of the camera according to the vehicle speed fluctuation value, vehicle accident probability and vehicle density of the target area.

[0016] Optionally, the steps for calculating the importance coefficient of the camera according to the vehicle speed fluctuation value, vehicle accident probability and vehicle density of the target area are as follows:

[0017] Take the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area as the input items of the fuzzy rules, and take the importance coefficient of the camera as the output item;

[0018] Fuzzify the input items and convert the precise input values into fuzzy sets;

[0019] Define the fuzzy rules according to the fuzzy sets of the input items and map the input items to the output items;

[0020] Infer the input items according to the fuzzy rules to determine the fuzzy values of the output items;

[0021] For each rule, calculate the membership degree of its antecedent and take the minimum value as the activation degree of the rule;

[0022] Synthesize the results of all rules and calculate the fuzzy set of the output item;

[0023] Convert the fuzzy values in the fuzzy set of the output item into precise values, and output the importance coefficient of the camera according to the result of defuzzification.

[0024] Optionally, the steps to screen out each edge computing center that can receive the traffic monitoring camera data of the failed edge computing center, which is called the target edge computing center, are as follows:

[0025] Obtain the number of CPU cores and clock frequency of each CPU of each non-failed edge computing center, calculate the computing power value of the non-failed edge computing center according to the number of CPU cores and clock frequency of each CPU, and the calculation formula is: , where TR is the computing power value C of the non-failed edge computing center i and F i are the number of the i-th CPU core and the clock frequency respectively, and n is the total number of CPUs;

[0026] Obtain the utilization rate of each CPU of each non-failed edge computing center, and calculate the average value of the utilization rates of all CPUs as the load value of the non-failed edge computing center;

[0027] Calculate the computing power coefficient of each non-failed edge computing center according to the computing power value and the load value, and the calculation formula is: , where UG is the computing power coefficient and TE is the load value;

[0028] Compare the computing power coefficients of each non-failed edge computing center with the preset computing power coefficient threshold. If the computing power coefficient is not less than the preset computing power coefficient threshold, each edge computing center that can receive the traffic monitoring camera data of the failed edge computing center is called the target edge computing center.

[0029] Optionally, the steps for calculating the adaptation coefficients of each target edge computing center and each faulty edge computing center based on the distance and the number of historical data transmissions are as follows:

[0030] Obtain the position coordinates of the target edge computing center and the faulty edge computing center, and calculate the shortest straight-line distance of signal transmission between the target edge computing center and the faulty edge computing center by calculating the Euclidean distance between the two points:

[0031] Calculate the signal transmission attenuation coefficient of the target edge computing center and the faulty edge computing center according to the shortest straight-line distance. The calculation formula is: , where YT is the signal transmission attenuation coefficient, D is the shortest straight-line distance, A is a preset constant representing the basic value of path loss, and m is a preset path loss exponent with a value range of 2 - 3.5;

[0032] Calculate the adaptation coefficients of each target edge computing center and each faulty edge computing center according to the signal transmission attenuation coefficient and the number of historical data transmissions between the target edge computing center and the faulty edge computing center.

[0033] Optionally, the steps for calculating the adaptation coefficients of each target edge computing center and each faulty edge computing center according to the signal transmission attenuation coefficient and the number of historical data transmissions between the target edge computing center and the faulty edge computing center are as follows:

[0034] Obtain the total number of data transmissions and the total number of successful transmissions between the target edge computing center and the faulty edge computing center, and divide the total number of successful transmissions by the total number of data transmissions to obtain the transmission success rate;

[0035] Obtain the time intervals between adjacent transmissions, calculate the mean value of the transmission time intervals as the average transmission time interval, and perform normalization processing on the average transmission time interval to map it to the numerical interval of 0 - 1;

[0036] Calculate the adaptation coefficient of the target edge computing center and the faulty edge computing center according to the signal transmission attenuation coefficient, the transmission success rate, and the normalized average transmission time interval. The calculation formula is: , where ED is the adaptation coefficient, YT, cv, and ku are the signal transmission attenuation coefficient, the transmission success rate, and the normalized average transmission time interval respectively, and a1, a2, and a3 are the preset proportional coefficients of YT, cv, and ku respectively, and a1, a2, and a3 are all greater than 0.

[0037] Optionally, the steps for determining the final target edge computing center selected by each faulty edge computing center are as follows:

[0038] After calculating the adaptation coefficients between each failed edge computing center and each target edge computing center, for each failed edge computing center, sort the adaptation coefficients between the failed edge computing center and each target edge computing center in descending order to obtain the set of corresponding target edge computing centers for each failed edge computing center;

[0039] Take the target edge computing center with the largest set of corresponding target edge computing centers as the corresponding final target edge computing center;

[0040] If a certain target edge computing center is simultaneously used as the final target edge computing center of multiple failed edge computing centers, then take this target edge computing center as the final target edge computing center of the failed edge computing center with the largest importance coefficient.

[0041] Optionally, the steps for monitoring whether the corresponding traffic monitoring camera is faulty by combining the importance coefficient of the transmitted traffic monitoring camera are as follows:

[0042] Extract frames from the video transmitted by the traffic monitoring camera to obtain a number of frames of images. For each frame of image, perform a transformation using the Laplace operator, calculate the variance of the pixel values of the transformed image, and take the variance of the pixel values as the image clarity value of the corresponding image; add up the image clarity values of each frame of image as the video clarity index of the traffic monitoring camera;

[0043] Obtain the actual frame rate of each frame of image, calculate the difference between the actual frame rate and the preset minimum frame rate. When the difference is less than 0, mark the corresponding frame of image as an abnormal image; divide the number of abnormal image frames by the total number of image frames to obtain the image frame rate abnormality ratio;

[0044] Obtain the health assessment index of the traffic monitoring camera based on the video clarity index, the image frame rate abnormality ratio, and the importance coefficient of the corresponding traffic monitoring camera, and monitor whether the corresponding traffic monitoring camera is faulty according to the health assessment index.

[0045] Optionally, the steps for obtaining the health assessment index of the traffic monitoring camera based on the video clarity index, the image frame rate abnormality ratio, and the importance coefficient of the corresponding traffic monitoring camera, and monitoring whether the corresponding traffic monitoring camera is faulty are as follows:

[0046] Obtain the health assessment index of the traffic monitoring camera based on the video clarity index, the image frame rate abnormality ratio, and the importance coefficient of the corresponding traffic monitoring camera. The calculation formula is:

[0047] , where FD is the health assessment index of the traffic monitoring camera, pl, uk, and tu are the video clarity index, the abnormal ratio of the image frame rate, and the importance coefficient of the corresponding traffic monitoring camera respectively, b1 and b2 are the preset proportionality coefficients of the video clarity index and the abnormal ratio of the image frame rate respectively, and both b1 and b2 are greater than 0;

[0048] Compare the health assessment index of the traffic monitoring camera with the preset health assessment index threshold. If the health index is not less than the preset health assessment index threshold, the corresponding traffic monitoring camera has no fault; if it is less, the corresponding traffic monitoring camera has a fault.

[0049] In the second aspect of the implementation of the present invention, an intelligent terminal monitoring system based on edge computing is proposed. The system includes:

[0050] Importance degree module: Obtain the traffic area data monitored by each traffic monitoring camera, used to evaluate the complexity of the traffic area, and calculate the importance coefficient of the camera according to the traffic area data;

[0051] Screening module: Obtain the resource information of each fault-free edge computing center, and screen out each edge computing center that can receive the traffic monitoring camera data of the fault edge computing center from the resource information of each fault-free edge computing center, which is called the target edge computing center;

[0052] Adaptability degree module: Obtain the distance and the historical data transmission times between the target edge computing center and each fault edge computing center, and calculate the adaptation coefficient between each target edge computing center and each fault edge computing center according to the distance and the historical data transmission times;

[0053] Selection module: Determine the final target edge computing center selected by each fault edge computing center according to the adaptation coefficient and the importance coefficient, and transfer the corresponding traffic monitoring camera data to the selected final target edge computing center for data analysis;

[0054] Terminal monitoring module: The final target edge computing center analyzes the transmitted traffic monitoring camera data, and monitors whether the corresponding traffic monitoring camera is faulty in combination with the importance coefficient of the transmitted traffic monitoring camera.

[0055] The beneficial effects of the present invention:

[0056] The present invention provides an intelligent terminal monitoring system and method based on edge computing. By obtaining the monitoring area data of traffic monitoring cameras corresponding to each faulty edge computing center, and calculating the importance coefficient of the cameras based on the area data to evaluate the importance degree of the cameras; obtaining the resource information of each non-faulty edge computing center, and screening out the target edge computing centers that can receive the data of the faulty edge computing centers; obtaining the distance information between the target edge computing centers and the faulty edge computing centers, calculating the adaptation coefficient to evaluate the adaptation degree between them; according to the adaptation coefficient and the importance coefficient, determining the final target edge computing center for each faulty edge computing center, and transmitting the traffic monitoring camera data to this center for analysis; the final target edge computing center analyzes the data in combination with the importance coefficient of the camera to monitor whether the camera is faulty; in this way, when multiple edge computing centers have problems and a certain edge computing center is backed up by multiple faulty edge centers, it can be appropriately allocated and calculated according to the actual situation, reducing the overloading of the backup edge computing centers, ensuring their processing capabilities, and preventing delays or overloading operations, ensuring the response speed and stability of the entire edge computing-based terminal monitoring system, and making the monitoring results of traffic monitoring cameras correct. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The present invention will be further described below with reference to the accompanying drawings.

[0058] Figure 1 is a flowchart of an intelligent terminal monitoring method based on edge computing;

[0059] Figure 2 is a framework diagram of an intelligent terminal monitoring system based on edge computing. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0061] All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] The embodiments of the present invention provide an intelligent terminal monitoring method based on edge computing. Refer to Figure 1 , Figure 1 is a flowchart of an intelligent terminal monitoring method based on edge computing provided by an embodiment of the present invention. The method includes the following steps:

[0063] Obtain the traffic area data monitored by each traffic monitoring camera, which is used to evaluate the complexity of the traffic area, and calculate the importance coefficient of the camera according to the traffic area data;

[0064] Obtain the resource information of each non-faulty edge computing center. According to the resource information of each non-faulty edge computing center, screen out each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center, which is called the target edge computing center;

[0065] Obtain the distance and historical data transmission times between the target edge computing center and each faulty edge computing center, and calculate the adaptation coefficient between each target edge computing center and each faulty edge computing center according to the distance and historical data transmission times;

[0066] Determine the final target edge computing center selected by each faulty edge computing center according to the adaptation coefficient and the importance coefficient, and transfer the corresponding traffic monitoring camera data to the selected final target edge computing center for data analysis;

[0067] The final target edge computing center analyzes the transmitted traffic monitoring camera data, and combines the importance coefficient of the transmitted traffic monitoring camera to monitor whether the corresponding traffic monitoring camera is faulty.

[0068] Based on an intelligent terminal monitoring method based on edge computing provided by an embodiment of the present invention, in the above manner, when multiple edge computing centers have problems and a certain edge computing center is backed up by multiple faulty edge centers, it is possible to perform appropriate allocation calculations according to the actual situation, reduce the overloading of the backup edge computing center, ensure its processing ability, and prevent delays or overloading operations. Ensure the response speed and stability of the entire edge computing-based terminal monitoring system, and make the monitoring results of traffic monitoring cameras correct.

[0069] In one embodiment, obtain the traffic area data monitored by each traffic monitoring camera, which is used to evaluate the complexity of the traffic area, and calculate the importance coefficient of the camera according to the traffic area data; the importance coefficient is used to evaluate the importance of the traffic monitoring cameras of each faulty edge computing center;

[0070] Among them, the steps of calculating the importance coefficient of the camera according to the traffic area data are:

[0071] Take the area photographed by the traffic monitoring camera as the target area, obtain the total number of vehicles that have passed through the target area in the past preset time period, and obtain the driving time of each vehicle in the target area. Calculate the variance of the driving time according to the driving time of all vehicles in the target area, which is used as the vehicle speed fluctuation value of the target area;

[0072] Obtain the vehicles that have been in traffic accidents in the target area, and divide the number of vehicles that have been in traffic accidents by the total number of vehicles that have passed through, to obtain the vehicle accident probability of the target area;

[0073] Obtain the drivable area of the lanes in the target area, and divide the total number of vehicles that have passed through by the drivable area of the lanes, to obtain the vehicle density of the target area;

[0074] Calculate the importance coefficient of the camera according to the vehicle speed fluctuation value, vehicle accident probability and vehicle density in the target area.

[0075] It should be noted that in the above process of calculating the importance coefficient of the camera, the area photographed by the traffic monitoring camera is used as the target area. The number of vehicles passing by and the driving time can be used to identify the passing time of each vehicle through an image processing algorithm. Combining the timestamp of each vehicle, the total number of vehicles passing through this area within a preset time period and their respective driving times can be accurately calculated; these data are usually provided by the traffic monitoring system or the traffic flow analysis platform; the vehicle speed information can be obtained through license plate recognition technology combined with spatio-temporal positioning data, or estimated by combining ground sensors and camera video streams to obtain the real-time speed of the vehicle; the variance of the vehicle speed reflects the fluctuation of the traffic flow smoothness in this area and can effectively reflect the congestion degree and abnormal behavior; the statistics on traffic accidents can be obtained from the local relevant departments; the drivable area of the lanes can be directly obtained through the local traffic system; there can also be other acquisition methods, which are not specifically limited and elaborated here.

[0076] It should be noted that the greater the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area, the greater the importance coefficient of the corresponding camera and the greater the importance of the terminal device. This is because: when the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area are greater, it indicates that the traffic conditions in this area are more complex and dangerous, and thus more accurate and timely monitoring is required to ensure traffic safety and efficiency. A large vehicle speed fluctuation value usually means poor traffic flow in this area, possibly with severe congestion, frequent stops and accelerations, increasing the risk of traffic accidents; a high vehicle accident probability directly reflects a high frequency of traffic accidents in this area, indicating that the monitoring of this area is crucial for accident prevention and emergency response; a large vehicle density means greater traffic pressure in this area, which may lead to problems such as too small vehicle spacing and increased congestion, increasing the possibility of sudden traffic events. Since these factors directly affect traffic safety and flow, the camera needs to have higher monitoring accuracy and response speed to promptly capture abnormal changes, so as to effectively warn of and handle traffic accidents or optimize traffic flow. Therefore, in areas where these factors are more significant, the importance coefficient of the camera will be calculated to be higher, reflecting the significance and urgency of the monitoring task in this area for the entire traffic management system, and thus making the corresponding terminal device more important and having higher monitoring value.

[0077] In one embodiment, the steps of calculating the importance coefficient of the camera according to the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area are as follows:

[0078] Take the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area as the input items of the fuzzy rule, and take the importance coefficient of the camera as the output item;

[0079] Fuzzify the input items and convert the accurate input values into fuzzy sets;

[0080] Define the fuzzy rules according to the fuzzy sets of the input items and map the input items to the output item;

[0081] According to the fuzzy rules, reason about the input items to determine the fuzzy value of the output item;

[0082] For each rule, calculate the membership degree of its antecedent and take the minimum value as the activation degree of the rule;

[0083] Synthesize the results of all rules and calculate the fuzzy set of the output item;

[0084] Convert the fuzzy values in the fuzzy set of the output item into accurate values, and output the importance coefficient of the camera according to the result of defuzzification.

[0085] It should be noted that the input and output items are defined as follows: the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area are used as the input items of the fuzzy rules, while the importance coefficient of the camera is used as the output item. For each input item and output item, it is necessary to clarify the scope of its fuzzification and possible values.

[0086] Fuzzification of input items: Convert the exact numerical values of the input items (i.e., vehicle speed fluctuation value, vehicle accident probability, and vehicle density) into fuzzy sets. In this process, fuzzification methods such as triangular membership functions or Gaussian membership functions can be used to convert specific numerical values (such as vehicle speed fluctuation values, accident probabilities, etc.) into fuzzy sets. For example, the vehicle speed fluctuation value may be fuzzified into three levels: "low", "medium", and "high", and each level will have corresponding membership degree values.

[0087] Define fuzzy rules: Based on traffic management experience and expert opinions, define fuzzy rules to map the input items to the output item. An example of a fuzzy rule can be: "If the vehicle speed fluctuation value is high, the vehicle accident probability is large, and the vehicle density is high, then the importance coefficient of the camera is high." By setting the rules, the relationship between the input items and the output item is determined.

[0088] Fuzzy inference: According to the defined fuzzy rules, perform fuzzy inference. Map the fuzzy sets of the input items to the fuzzy sets of the output item through inference methods. At this time, fuzzy logic inference methods such as the Mamdani fuzzy inference method are used, combined with the fuzzy sets of the input items, to infer the fuzzy set of the importance coefficient of the camera.

[0089] Calculate the activation degree: For each rule, calculate the membership degree of its antecedent (i.e., the input item), and take the minimum value as the activation degree of this rule. This process ensures that the strength of the rule reflects the actual fuzziness of the input item. For example, when the vehicle speed fluctuation value is very high and the accident probability is relatively large, the activation degree of the rule will be stronger.

[0090] Synthesize the rule results: Synthesize the results of all fuzzy rules and calculate the fuzzy set of the output item. The activation degree of each rule will affect the membership degree of the output item. The outputs of multiple rules are combined through methods such as weighted average to obtain a comprehensive fuzzy output.

[0091] Defuzzification: Convert the fuzzy values in the synthesized output item fuzzy set into exact values. The defuzzification process uses a certain defuzzification algorithm (such as the centroid method) to obtain the exact value of the importance coefficient of the camera, thereby determining the final importance coefficient of each camera. This process provides specific numerical outputs for subsequent monitoring system decisions.

[0092] Through this fuzzy inference method, it is possible to comprehensively consider the influence of different factors, flexibly evaluate the importance of the camera, and thus provide a basis for subsequent edge computing and resource allocation.

[0093] In one implementation, the benefits of calculating the importance coefficient of a camera based on the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area are as follows: The vehicle speed fluctuation value reveals the stability of the traffic flow, while the accident probability and vehicle density are directly related to the occurrence probability of traffic accidents and traffic pressure, and can reflect the complexity and danger of the area from multiple dimensions. Secondly, these factors are the most critical indicators in traffic management, with strong practicality and predictive ability. By combining these three factors, a more comprehensive and effective monitoring and evaluation mechanism can be formed to ensure that the traffic monitoring system can prioritize high-risk areas, optimize resource allocation, and improve overall traffic safety and emergency response capabilities.

[0094] In one embodiment, resource information of each non-faulty edge computing center is obtained, and based on the resource information of each non-faulty edge computing center, each edge computing center that can receive traffic monitoring camera data of the faulty edge computing center is selected therefrom, and is called the target edge computing center;

[0095] Among them, the step of selecting, based on the resource information of each non-faulty edge computing center, each edge computing center that can receive traffic monitoring camera data of the faulty edge computing center and calling it the target edge computing center is as follows:

[0096] Obtain the number of cores and clock frequency of each CPU of each non-faulty edge computing center, and calculate the computing power value of the non-faulty edge computing center according to the number of cores and clock frequency of each CPU. The calculation formula is: , where TR is the computing power value of the non-faulty edge computing center, C i and F i are the number of cores and clock frequency of the i-th CPU respectively, and n is the total number of CPUs;

[0097] Obtain the utilization rate of each CPU of each non-faulty edge computing center, and calculate the average value of the utilization rates of all CPUs as the load value of the non-faulty edge computing center;

[0098] Calculate the computing power coefficient of each non-faulty edge computing center according to the computing power value and the load value. The calculation formula is: , where UG is the computing power coefficient and TE is the load value;

[0099] Select, according to the computing power coefficient of each non-faulty edge computing center, each edge computing center that can receive traffic monitoring camera data of the faulty edge computing center, and call it the target edge computing center.

[0100] It should be noted that when calculating the computing power coefficients of each non-faulty edge computing center, the data acquisition methods for the involved data are as follows: The number of cores and the clock frequency can be obtained through hardware monitoring tools (such as the lscpu command, system management interface, etc.), and are usually directly extracted from the CPU specifications; the current usage rate of each CPU can be dynamically obtained through system monitoring tools (such as top, vmstat, htop, etc.) and will be reflected in the system resource management tool in real time. These data can help accurately evaluate the computing power coefficients of non-faulty edge computing centers.

[0101] It should be noted that when the number of cores of the non-faulty edge computing center is larger, the clock frequency is larger, and the CPU usage rate is smaller, the corresponding computing power coefficient of the non-faulty edge computing center is larger, and it can receive the traffic monitoring camera data of the faulty edge computing center for analysis and calculation. The reason is that: when the number of cores of the non-faulty edge computing center is larger, it means that the center has more computing units and can process more tasks simultaneously, improving the parallel computing ability; when the clock frequency is larger, each core can execute more instructions per unit time, and the computing speed is faster, which improves the processing ability of the center. On the other hand, the smaller the CPU usage rate, it means that the resources of this edge computing center have not been fully consumed, and there is still enough computing power and processing ability to carry more tasks, avoiding performance degradation caused by overload. Therefore, when these conditions are met simultaneously, the non-faulty edge computing center can process the traffic monitoring camera data of the faulty edge computing center more efficiently and stably, avoiding response delays or computing performance degradation caused by high loads, and thus being able to better perform data analysis, real-time monitoring, and rapid response to ensure the stability and efficiency of the overall system.

[0102] In one embodiment, the steps of screening out each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center according to the computing power coefficients of each non-faulty edge computing center, which is called the target edge computing center, are as follows:

[0103] Compare the computing power coefficients of each non-faulty edge computing center with a preset computing power coefficient threshold. If the computing power coefficient is not less than the preset computing power coefficient threshold, each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center is called the target edge computing center.

[0104] It should be noted that the preset computing power coefficient threshold is set by professionals according to the actual situation and is not specifically limited or elaborated here.

[0105] In one implementation manner, when screening edge computing centers that can receive traffic monitoring camera data from a faulty edge computing center, first, compare the computing power coefficient of each non-faulty edge computing center with a preset computing power coefficient threshold. If the computing power coefficient of a certain non-faulty edge computing center is not less than the set threshold, it means that this center has sufficient computing power and resources to process additional traffic monitoring camera data. In this way, it can be ensured that the selected target edge computing center has sufficient processing power, avoiding calculation delay or overload caused by insufficient resources, and ensuring the real-time and accuracy of traffic data analysis. Therefore, the selected target edge computing centers are those that not only meet the computing power threshold requirements but also can efficiently process the data of the faulty edge computing center under the current load, thus maintaining the stability and efficient response of the system.

[0106] In one embodiment, obtain the distances and historical data transmission times between the target edge computing center and each faulty edge computing center, and calculate the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the distances and historical data transmission times; the adaptation coefficients are used to evaluate the adaptation degree between each target edge computing center and each faulty edge computing center;

[0107] Among them, the steps of calculating the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the distances and historical data transmission times are as follows:

[0108] Obtain the position coordinates of the target edge computing center and the faulty edge computing center, and calculate the shortest straight-line distance of signal transmission between the target edge computing center and the faulty edge computing center by calculating the Euclidean distance between two points:

[0109] Calculate the signal transmission attenuation coefficient between the target edge computing center and the faulty edge computing center according to the shortest straight-line distance, and the calculation formula is: , where YT is the signal transmission attenuation coefficient, D is the shortest straight-line distance, A is a preset constant representing the basic value of path loss, and m is a preset path loss exponent with a value range of 2 - 3.5;

[0110] Calculate the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the signal transmission attenuation coefficient and historical data transmission times between the target edge computing center and the faulty edge computing center.

[0111] It should be noted that in the process of calculating the signal transmission attenuation coefficient, the data acquisition methods involved include the following aspects: First, the geographical location coordinates (i.e., longitude and latitude) of each target edge computing center and the faulty edge computing center need to be obtained, and these coordinates can be obtained through the GPS positioning system or the map API. The values of the path loss constant A and the path loss exponent m usually depend on the type of wireless signal propagation environment, and the specific value ranges vary according to different environments. For example, in an urban environment, due to the density of buildings and multiple reflections of signals, the path loss exponent nnn usually takes values between 2.7 and 3.5, and the path loss constant A is generally in the range of 50 to 100 dB. For a suburban environment, the signal propagation is relatively open, the value of m is generally 2.0 to 3.0, and the range of the path loss constant A is 40 to 80 dB. In an open area, due to almost no building interference and relatively straight signal propagation, the path loss exponent m usually takes values between 1.8 and 2.2, and the path loss constant A will be lower, generally between 30 and 60 dB. These values are usually set through on-site measured data or through specific wireless communication standards (such as the ITU-R model); it can also be other acquisition methods, which are not specifically limited and elaborated here,

[0112] It should be noted that when the signal transmission attenuation coefficient between the target edge computing center and the faulty edge computing center is larger, the corresponding adaptation degree between the target edge computing center and the faulty edge computing center is lower. At this time, if the traffic monitoring camera data of the faulty edge computing center is transmitted to the target edge computing center, it may cause monitoring errors at the target edge computing center. The reason is that when the signal transmission attenuation coefficient between the target edge computing center and the faulty edge computing center is larger, it means that the signal loss is more serious during the transmission process, and the signal intensity will drop significantly. This is usually related to a longer transmission distance, complex environmental interference (such as buildings, terrain, etc.) or poor communication conditions (such as signal interference). Therefore, the data transmitted to the target edge computing center may experience delays, data loss or errors, resulting in inaccurate data analysis results. If the attenuation coefficient is large, it means that the target edge computing center may not be able to effectively receive and process the monitoring data from the faulty edge computing center, thereby affecting the accuracy and reliability of the monitoring, and may lead to incorrect monitoring results at the target edge computing center and even make wrong decisions.

[0113] In one embodiment, the steps of calculating the adaptation coefficient of each target edge computing center and each faulty edge computing center according to the signal transmission attenuation coefficient and the historical data transmission times between the target edge computing center and the faulty edge computing center are as follows:

[0114] Obtain the total number of data transmissions and the total number of successful transmissions between the target edge computing center and the faulty edge computing center, and divide the total number of successful transmissions by the total number of data transmissions to obtain the transmission success rate;

[0115] Obtain the transmission time intervals between two adjacent times, calculate the mean value of the transmission time intervals as the average transmission time interval, and perform normalization processing on the average transmission time interval to map it to the numerical interval of 0-1;

[0116] Calculate the adaptation coefficient between the target edge computing center and the faulty edge computing center according to the signal transmission attenuation coefficient, the transmission success rate, and the normalized average transmission time interval. The calculation formula is: , where ED is the adaptation coefficient, YT, cv, and ku are the signal transmission attenuation coefficient, the transmission success rate, and the normalized average transmission time interval respectively, and a1, a2, and a3 are the preset proportional coefficients of YT, cv, and ku respectively, and a1, a2, and a3 are all greater than 0.

[0117] It should be noted that a1, a2, and a3 are set by professionals according to the actual situation. Generally, the sum of a1, a2, and a3 is 1. For example, a1, a2, and a3 can be 0.3, 0.3, and 0.4 respectively, or other numbers, which are not specifically limited; in addition, before calculating the adaptation coefficient, it is necessary to calculate after removing the units of the signal transmission attenuation coefficient, the transmission success rate, and the normalized average transmission time interval.

[0118] It should be noted that all data transmission histories between the target edge computing center and the faulty edge computing center are obtained by monitoring and recording system logs, including the success and failure information of each transmission, so as to calculate the total number of data transmissions and the number of successful transmissions; by obtaining the timestamps of each data transmission, calculate the time intervals between adjacent transmissions, and then obtain the mean value of the transmission time intervals. Then, according to these time interval data, use the normalization method to map it to the 0-1 interval.

[0119] It should be noted that when the transmission success rate between the target edge computing center and the faulty edge computing center is smaller and the average transmission time interval is larger, the degree of adaptation between the corresponding target edge computing center and the faulty edge computing center is lower. At this time, if the traffic monitoring camera data of the faulty edge computing center is transmitted to the target edge computing center, it may cause monitoring errors in the target edge computing center. The reason is that when the transmission success rate between the target edge computing center and the faulty edge computing center is low and the average transmission time interval is large, it means that the transmission stability of the data is poor, and the network communication may be unstable or have a high delay, which directly affects the accuracy and timeliness of the data. A low data transmission success rate indicates that the probability of data loss or error during transmission is large, resulting in incomplete or unreliable data received by the target edge computing center; while a large average transmission time interval means that the data transmission between the target edge computing center and the faulty edge computing center is not frequent, and there is less experience in data transmission. Therefore, in this case, transmitting the camera data of the faulty edge computing center to the target edge computing center may lead to mistakes or inaccuracies in the monitoring results, thus affecting the security and effectiveness of the overall system.

[0120] In one embodiment, the final target edge computing center selected by each faulty edge computing center is determined according to the adaptation coefficient and the importance coefficient, and the corresponding traffic monitoring camera data is transmitted to the selected target edge computing center for data analysis;

[0121] Among them, the steps for determining the final target edge computing center selected by each faulty edge computing center are as follows:

[0122] After calculating the adaptation coefficients between each faulty edge computing center and each target edge computing center, for each faulty edge computing center, the adaptation coefficients between the faulty edge computing center and each target edge computing center are sorted in descending order to obtain the set of adaptation target edge computing centers corresponding to each faulty edge computing center;

[0123] The target edge computing center with the largest set of adaptation target edge computing centers is used as the corresponding final target edge computing center;

[0124] If a certain target edge computing center serves as the final target edge computing center of multiple faulty edge computing centers at the same time, this target edge computing center is used as the final target edge computing center of the faulty edge computing center with the largest importance coefficient.

[0125] Note that there are three faulty edge computing centers: Faulty Center A, Faulty Center B, and Faulty Center C, and three target edge computing centers: Target Center X, Target Center Y, and Target Center Z. After calculation, the adaptation coefficients between each pair of faulty centers and target centers are as follows: The adaptation coefficient between Faulty Center A and Target Center X is 0.85, the adaptation coefficient between Faulty Center A and Target Center Y is 0.90, and the adaptation coefficient between Faulty Center A and Target Center Z is 0.80. The adaptation coefficient between Faulty Center B and Target Center X is 0.75, the adaptation coefficient between Faulty Center B and Target Center Y is 0.92, and the adaptation coefficient between Faulty Center B and Target Center Z is 0.88. The adaptation coefficient between Faulty Center C and Target Center X is 0.88, the adaptation coefficient between Faulty Center C and Target Center Y is 0.91, and the adaptation coefficient between Faulty Center C and Target Center Z is 0.85. According to the sorting of the adaptation coefficients, the following sets of adapted target edge computing centers for each faulty edge computing center are obtained: For Faulty Center A, the sorted set of adapted target edge computing centers is: [Target Center Y, Target Center X, Target Center Z], and the maximum adaptation coefficient is 0.90. Therefore, Faulty Center A selects Target Center Y as the final target edge computing center. For Faulty Center B, the sorted set of adapted target edge computing centers is: [Target Center Y, Target Center Z, Target Center X], and the maximum adaptation coefficient is 0.92. Therefore, Faulty Center B selects Target Center Y as the final target edge computing center. For Faulty Center C, the sorted set of adapted target edge computing centers is: [Target Center Y, Target Center X, Target Center Z], and the maximum adaptation coefficient is 0.91. Therefore, Faulty Center C selects Target Center Y as the final target edge computing center. At this time, it can be seen that Target Center Y is selected as the final target edge computing center by Faulty Centers A, B, and C. However, Target Center Y cannot process the data transmission of all faulty edge computing centers simultaneously. Therefore, Target Center Y needs to select the faulty edge computing center with the largest adaptation coefficient for data analysis. Assuming that in practical applications, the importance coefficient of Faulty Center B is the largest, then Target Center Y will select Faulty Center B as its final target edge computing center and receive the traffic monitoring camera data of Faulty Center B for analysis.

[0126] In summary, the final data transmission arrangement is: Faulty Center A selects Target Center X for data transmission; Faulty Center B selects Target Center Y for data transmission; Faulty Center C selects Target Center Z for data transmission.

[0127] In one implementation method, through sorting and selection based on the adaptation coefficient, it is possible to ensure the maximization of the adaptation degree between each faulty edge computing center and its corresponding target edge computing center, thereby improving the data transmission efficiency and accuracy. By preferentially selecting the target edge computing center with a larger adaptation coefficient, it is possible to reduce the errors and delays that may occur during data transmission, optimize the use of resources, and ensure that traffic monitoring camera data can be analyzed and processed more quickly and accurately. At the same time, when multiple faulty edge computing centers select the same target edge computing center, by considering the maximization selection of the importance coefficient, the priority of key faulty centers is further guaranteed, potential resource conflicts are avoided, and the overall stability and efficiency of the system are improved.

[0128] In one embodiment, the steps for the final target edge computing center to analyze the traffic monitoring camera data transmitted and monitor whether the corresponding traffic monitoring camera is faulty in combination with the importance coefficient of the transmitted traffic monitoring camera are as follows:

[0129] Extract frames from the video transmitted by the traffic monitoring camera to obtain several frame images. For each frame image, perform a transformation through the Laplace operator, calculate the pixel value variance of the transformed image, and use the pixel value variance as the image clarity value of the corresponding image; add the image clarity values of each frame image as the video clarity index of the traffic monitoring camera;

[0130] Obtain the actual frame rate of each frame image, calculate the difference between the actual frame rate and the preset minimum frame rate. When the difference is less than 0, mark the corresponding frame image as an abnormal image; divide the number of abnormal image frames by the total number of image frames to obtain the image frame rate abnormality ratio; obtain the health assessment index of the traffic monitoring camera based on the video clarity index, the image frame rate abnormality ratio, and the importance coefficient of the corresponding traffic monitoring camera, and monitor whether the corresponding traffic monitoring camera is faulty based on the health assessment index; the formula for calculating the health assessment index is:

[0131] , where FD is the health assessment index of the traffic monitoring camera, pl, uk, and tu are the video clarity index, the image frame rate abnormality ratio, and the importance coefficient of the corresponding traffic monitoring camera respectively, and b1 and b2 are the preset proportionality coefficients of the video clarity index and the image frame rate abnormality ratio respectively, and both b1 and b2 are greater than 0;

[0132] Compare the health assessment index of the traffic monitoring camera with the preset health assessment index threshold. If the health index is not less than the preset health assessment index threshold, the corresponding traffic monitoring camera is not faulty; if it is less, the corresponding traffic monitoring camera is faulty.

[0133] It should be noted that b1 and b2 are set by professionals according to the actual situation. Generally, the sum of b1 and b2 is 1. For example, b1 and b2 can be 0.4 and 0.6 respectively, or other numbers, and there is no specific limitation; in addition, before calculating the health assessment index, it is necessary to calculate the video clarity index and the abnormal ratio of the image frame rate after removing the unit; in addition, the preset health assessment index threshold is set by professionals according to the actual situation, and there is no specific limitation and elaboration.

[0134] It should be noted that the larger the video clarity index and the lower the abnormal ratio of the image frame rate of the traffic monitoring camera, the smaller the possibility of the traffic monitoring camera malfunctioning, indicating that the camera can stably output clear and detailed images, and the continuity and stability of the images are good. A high video clarity index indicates that the imaging quality of the camera is very good, and the lens and sensor are working properly; a low abnormal ratio of the image frame rate means that the camera continuously captures images at the normal working frequency without freezing or abnormal frame loss. Generally speaking, both of these two indicators are within the normal range, indicating that the camera can efficiently and accurately complete its functions when performing monitoring tasks, so the possibility of malfunction is small.

[0135] In one implementation manner, through the above method, the health assessment is combined with the importance coefficient of the traffic monitoring camera. When the importance coefficient is larger, the corresponding health assessment index decreases because the importance of the camera in the monitoring network is relatively high. In order to ensure the monitoring stability and data accuracy of key areas, more strict evaluations are required for these high-importance cameras; by improving the evaluation criteria for important cameras, the overall security and monitoring quality of the system can be better guaranteed, and the traffic management and safety guarantee of key areas can be prevented from being affected due to camera malfunctions.

[0136] Based on the same inventive concept, the embodiment of the present invention also provides an intelligent terminal monitoring system based on edge computing. See Figure 2 , Figure 2 is a framework diagram of an intelligent terminal monitoring system based on edge computing provided by the embodiment of the present invention. The system includes:

[0137] Importance degree module: obtaining the traffic area data monitored by each traffic monitoring camera, used to evaluate the complexity of the traffic area, and calculating the importance coefficient of the camera according to the traffic area data;

[0138] Screening module: obtaining the resource information of each non-faulty edge computing center, and screening out each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center from the resource information of each non-faulty edge computing center, which is called the target edge computing center;

[0139] Adaptation degree module: Obtain the distances and historical data transmission times between the target edge computing center and each faulty edge computing center, and calculate the adaptation coefficients between each target edge computing center and each faulty edge computing center based on the distances and historical data transmission times;

[0140] Selection module: Determine the final target edge computing centers selected by each faulty edge computing center according to the adaptation coefficients and importance coefficients, and transfer the corresponding traffic monitoring camera data to the selected final target edge computing centers for data analysis;

[0141] Terminal monitoring module: The final target edge computing center analyzes the transferred traffic monitoring camera data, and monitors whether the corresponding traffic monitoring camera is faulty in combination with the importance coefficient of the transferred traffic monitoring camera.

[0142] Based on the intelligent terminal monitoring system based on edge computing provided by the embodiment of the present invention, in the above manner, when problems occur in multiple edge computing centers and a certain edge computing center is backed up by multiple faulty edge centers, it is possible to appropriately allocate calculations according to the actual situation, reduce the overloading of the backup edge computing center, ensure its processing capacity, and prevent delays or overloading operations. Ensure the response speed and stability of the entire edge computing-based terminal monitoring system, and make the monitoring results of traffic monitoring cameras correct.

[0143] The above has described an embodiment of the present invention in detail, but the content described is only a preferred embodiment of the present invention and cannot be artificially used to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the application of the present invention shall still fall within the scope covered by the patent of the present invention.

Claims

1. An intelligent terminal monitoring method based on edge computing, characterized in that It includes the following steps: Obtain the traffic area data monitored by each traffic monitoring camera for evaluating the complexity of the traffic area, and calculate the importance coefficient of the camera according to the traffic area data; Obtain the resource information of each non-faulty edge computing center. According to the resource information of each non-faulty edge computing center, screen out each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center, which is called the target edge computing center; Obtain the distances and historical data transmission times between the target edge computing centers and each faulty edge computing center, and calculate the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the distances and historical data transmission times; Determine the final target edge computing center selected by each faulty edge computing center according to the adaptation coefficient and the importance coefficient, and transfer the corresponding traffic monitoring camera data to the selected final target edge computing center for data analysis; The final target edge computing center analyzes the transferred traffic monitoring camera data, and monitors whether the corresponding traffic monitoring camera is faulty in combination with the importance coefficient of the transferred traffic monitoring camera; The steps of calculating the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the distances and historical data transmission times are as follows: Obtain the position coordinates of the target edge computing center and the faulty edge computing center, and calculate the shortest straight-line distance of signal transmission between the target edge computing center and the faulty edge computing center by calculating the Euclidean distance between two points: Calculate the signal transmission attenuation coefficient between the target edge computing center and the fault edge computing center according to the shortest straight-line distance. The calculation formula is as follows: , where YT is the signal transmission attenuation coefficient, D is the shortest straight-line distance, A is a preset constant representing the basic value of path loss, and m is a preset path loss exponent with a value range of 2 - 3.5; Calculate the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the signal transmission attenuation coefficient and historical data transmission times between the target edge computing center and the faulty edge computing center; The steps of calculating the adaptation coefficients between each target edge computing center and each faulty edge computing center according to the signal transmission attenuation coefficient and historical data transmission times between the target edge computing center and the faulty edge computing center are as follows: Obtain the total number of data transmissions and the total number of successful transmissions between the target edge computing center and the faulty edge computing center, and divide the total number of successful transmissions by the total number of data transmissions to obtain the transmission success rate; Obtain the time intervals between adjacent transmissions, calculate the average value of the transmission time intervals as the average transmission time interval, and normalize the average transmission time interval to map it to the numerical interval of 0-1; Calculate the adaptation coefficient between the target edge computing center and the faulty edge computing center according to the signal transmission attenuation coefficient, transmission success rate, and the average transmission time interval after normalization. The calculation formula is as follows: , where ED is the adaptation coefficient, YT, cv, and ku are the signal transmission attenuation coefficient, transmission success rate, and the average transmission time interval after normalization respectively, and a1, a2, and a3 are the preset proportionality coefficients of YT, cv, and ku respectively, and a1, a2, and a3 are all greater than 0.

2. The intelligent terminal monitoring method based on edge computing according to claim 1, wherein, The steps of calculating the importance coefficient of the camera according to the traffic area data are as follows: Take the area photographed by the traffic monitoring camera as the target area, obtain the total number of vehicles that have passed through the target area in the past preset time period, and obtain the driving time of each vehicle in the target area. Calculate the variance of the driving time of all vehicles in the target area as the vehicle speed fluctuation value of the target area; Obtain the vehicles that have had traffic accidents in the target area, and divide the number of vehicles that have had traffic accidents by the total number of vehicles that have passed through to obtain the vehicle accident probability of the target area; Obtain the drivable area of the lanes in the target area, and divide the total number of vehicles that have passed through by the drivable area of the lanes to obtain the vehicle density of the target area; Calculate the importance coefficient of the camera based on the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area.

3. The intelligent terminal monitoring method based on edge computing according to claim 2, wherein, The steps to calculate the importance coefficient of the camera based on the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area are as follows: Take the vehicle speed fluctuation value, vehicle accident probability, and vehicle density in the target area as the input items of the fuzzy rule, and take the importance coefficient of the camera as the output item; Fuzzify the input items and convert the precise input values into fuzzy sets; Define the fuzzy rules according to the fuzzy sets of the input items and map the input items to the output item; Infer the input items according to the fuzzy rules to determine the fuzzy value of the output item; For each rule, calculate the membership degree of its antecedent and take the minimum value as the activation degree of the rule; Synthesize the results of all rules to calculate the fuzzy set of the output item; Convert the fuzzy values in the fuzzy set of the output item into precise values, and output the importance coefficient of the camera according to the result of defuzzification.

4. The intelligent terminal monitoring method based on edge computing according to claim 1, characterized in that, The steps to screen out each edge computing center that can receive the traffic monitoring camera data of the failed edge computing center, which is called the target edge computing center, are as follows: Obtain the number of cores and clock frequencies of each CPU of each non-faulty edge computing center, and calculate the computing power value of the non-faulty edge computing center according to the number of cores and clock frequencies of each CPU. The calculation formula is: , where TR is the computing power value of the non-faulty edge computing center, C i and F i are the number of cores and clock frequency of the i-th CPU respectively, and n is the total number of CPUs; Obtain the usage rate of each CPU of each non-failed edge computing center, and calculate the average value of the usage rates of all CPUs as the load value of the non-failed edge computing center; Calculate the computing power coefficients of each non-faulty edge computing center according to the computing power value and the load value. The calculation formula is as follows: , where UG is the computing power coefficient and TE is the load value; Compare the computing power coefficient of each non-failed edge computing center with the preset computing power coefficient threshold. If the computing power coefficient is not less than the preset computing power coefficient threshold, each edge computing center that receives the traffic monitoring camera data of the failed edge computing center is called the target edge computing center.

5. The intelligent terminal monitoring method based on edge computing according to claim 1, wherein The steps to determine the final target edge computing center selected by each failed edge computing center are as follows: After calculating the adaptation coefficient between each failed edge computing center and each target edge computing center, for each failed edge computing center, sort the adaptation coefficients between the failed edge computing center and each target edge computing center in descending order to obtain the set of adaptation target edge computing centers corresponding to each failed edge computing center; Take the target edge computing center with the largest set of adaptation target edge computing centers as the corresponding final target edge computing center; If a certain target edge computing center serves as the final target edge computing center of multiple failed edge computing centers at the same time, take this target edge computing center as the final target edge computing center of the failed edge computing center corresponding to the camera with the largest importance coefficient.

6. The intelligent terminal monitoring method based on edge computing according to claim 1, wherein The steps to monitor whether the corresponding traffic monitoring camera is faulty by combining the importance coefficient of the transmitted traffic monitoring camera are as follows: Extract frames from the video transmitted by the traffic monitoring camera to obtain several frames of images. For each frame of image, perform a transformation through the Laplacian operator, calculate the variance of the pixel values of the transformed image, and take the variance of the pixel values as the image clarity value of the corresponding image; Add up the image clarity values of each frame of image as the video clarity index of the traffic monitoring camera; Obtain the actual frame rate of each frame of image, calculate the difference between the actual frame rate and the preset minimum frame rate. When the difference is less than 0, record the corresponding frame of image as an abnormal image; divide the number of abnormal image frames by the total number of image frames to obtain the image frame rate abnormality ratio. Obtain the health assessment index of a traffic monitoring camera based on the video clarity index, the abnormal ratio of the image frame rate, and the importance coefficient of the corresponding traffic monitoring camera, and monitor whether the corresponding traffic monitoring camera is faulty according to the health assessment index.

7. An intelligent terminal monitoring method based on edge computing according to claim 6, characterized in that, The steps to obtain the health assessment index of a traffic monitoring camera based on the video clarity index, the abnormal ratio of the image frame rate, and the importance coefficient of the corresponding traffic monitoring camera, and to monitor whether the corresponding traffic monitoring camera is faulty are as follows: Obtain the health assessment index of a traffic monitoring camera based on the video clarity index, the abnormal ratio of the image frame rate, and the importance coefficient of the corresponding traffic monitoring camera. The calculation formula is: ， In the formula, FD is the health assessment index of the traffic monitoring camera, pl, uk, and tu are the video clarity index, the abnormal ratio of the image frame rate, and the importance coefficient of the corresponding traffic monitoring camera respectively, b1 and b2 are the preset proportionality coefficients of the video clarity index and the abnormal ratio of the image frame rate respectively, and both b1 and b2 are greater than 0; Compare the health assessment index of the traffic monitoring camera with the preset health assessment index threshold. If the health index is not less than the preset health assessment index threshold, the corresponding traffic monitoring camera is not faulty; if it is less, the corresponding traffic monitoring camera is faulty.

8. An intelligent terminal monitoring system based on edge computing, which is used to implement the method for monitoring an intelligent terminal based on edge computing according to any one of claims 1-7 above, and is characterized in that, The system includes: Importance degree module: Obtain the traffic area data monitored by each traffic monitoring camera, used to evaluate the complexity of the traffic area, and calculate the importance coefficient of the camera according to the traffic area data; Screening module: Obtain the resource information of each non-faulty edge computing center, and screen out each edge computing center that can receive the traffic monitoring camera data of the faulty edge computing center from the resource information of each non-faulty edge computing center, which is called the target edge computing center; Adaptability degree module: Obtain the distance and the historical data transmission times between the target edge computing center and each faulty edge computing center, and calculate the adaptation coefficient between each target edge computing center and each faulty edge computing center according to the distance and the historical data transmission times; Selection module: Determine the final target edge computing center selected by each faulty edge computing center according to the adaptation coefficient and the importance coefficient, and transfer the corresponding traffic monitoring camera data to the selected final target edge computing center for data analysis; Terminal monitoring module: The final target edge computing center analyzes the transferred traffic monitoring camera data, and monitors whether the corresponding traffic monitoring camera is faulty in combination with the importance coefficient of the transferred traffic monitoring camera.

Images