Video monitoring system detection method and device, terminal equipment and storage medium

Abstract

The invention discloses a video monitoring system detection method and device, terminal equipment and a storage medium, and relates to the technical field of detection, and the method comprises the steps: collecting equipment operation and maintenance data of video monitoring equipment in a target video monitoring system through an edge computing node; determining a preliminary detection result according to the equipment operation and maintenance data, wherein the preliminary detection result comprises a network flow analysis result, a video anomaly detection result and an equipment power consumption detection result; constructing a global fault knowledge base according to the preliminary detection result, and performing iterative training on the fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model; and inputting collected real-time to-be-detected data into the optimized fault diagnosis model to obtain a fault diagnosis result. According to the invention, multi-angle fault diagnosis results can be output to deal with various fault conditions in a complex network environment, and the generalization ability and adaptability of detection are improved, so that the detection accuracy of a video monitoring system can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of detection technology, and in particular to a video surveillance system detection method, device, terminal equipment and storage medium. Background Art

[0002] With the rapid development of internet technology, video surveillance systems have been widely adopted by businesses, institutions, and individuals, becoming critical infrastructure for ensuring safety and order in public places, commercial areas, residential areas, factories, and other settings. They play a significant role in urban security, helping relevant departments effectively prevent and address various security incidents. However, video surveillance systems face numerous challenges in their actual operation. Existing video surveillance system detection methods typically rely on limited human resources patrolling video surveillance points. This manual detection method is difficult to adapt to complex network environments, resulting in poor detection results. Summary of the Invention

[0003] The present invention provides a video surveillance system detection method, device, terminal equipment and storage medium, which can solve the technical problem in the existing technology that video surveillance system detection relies on limited manpower to patrol video surveillance points, and the manual detection method is difficult to adapt to complex network environments, resulting in poor detection effect.

[0004] The present invention provides a video surveillance system detection method, comprising:

[0005] Collecting device operation and maintenance data of video surveillance devices in the target video surveillance system through edge computing nodes, wherein the device operation and maintenance data includes network traffic data, video data, and device power consumption data;

[0006] Determine preliminary detection results based on the device operation and maintenance data, wherein the preliminary detection results include network traffic analysis results, video anomaly detection results, and device power consumption detection results;

[0007] Building a global fault knowledge base based on the preliminary detection results, and iteratively training a fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model;

[0008] The collected real-time data to be detected is input into the optimized fault diagnosis model to obtain a fault diagnosis result.

[0009] Furthermore, determining the preliminary detection result based on the equipment operation and maintenance data includes:

[0010] Inputting the video data into an artificial intelligence model based on a convolutional neural network, extracting video frame features from the video data according to a convolution operation in the artificial intelligence model, determining whether there is an image anomaly based on the video frame features, and generating a video anomaly detection result;

[0011] Using a sliding window algorithm to analyze the network traffic data to obtain a network traffic analysis result;

[0012] The exponential smoothing method is used to predict the current power consumption forecast value, which is compared with the current power consumption actual value, and the device power consumption detection result is determined based on the comparison result.

[0013] Furthermore, extracting video frame features from the video data according to the convolution operation in the artificial intelligence model includes:

[0014] The convolution operation is performed according to the following formula:

[0015]

[0016] Among them, Q j is the video frame feature, I(x,y) is the input video frame, K j (x,y) represents the value of the j-th convolution kernel at coordinate (x,y), It is an accumulation operation of the multiplication results of the elements at the corresponding positions of the video frame and the convolution kernel.

[0017] Furthermore, the network traffic data is analyzed using a sliding window algorithm to obtain network traffic analysis results, including:

[0018] Using a sliding window algorithm and the network traffic data, calculate the average network traffic within a preset window;

[0019] Calculating a traffic variance based on the traffic mean, and determining whether there is a sudden packet loss anomaly based on the traffic variance;

[0020] Generate network traffic analysis results based on whether there is a sudden packet loss anomaly.

[0021] Furthermore, the expression of the exponential smoothing method is as follows:

[0022]

[0023] in, is the power consumption prediction value at the current moment, α is the smoothing coefficient, P(t-1) is the actual power consumption value at the previous moment, Predict the power consumption for the previous moment.

[0024] Furthermore, the real-time data to be detected is input into the optimized fault diagnosis model to obtain the fault diagnosis result, including:

[0025] Collect real-time network performance indicator data, input the real-time network performance indicator data into the optimized fault diagnosis model, evaluate the network status of the target video surveillance system based on the optimized fault diagnosis model and the real-time network performance indicator data, and obtain a fault diagnosis result corresponding to the network status.

[0026] Furthermore, after obtaining the fault diagnosis result corresponding to the network status, the method further includes:

[0027] Generate a real-time network status profile based on the fault diagnosis result;

[0028] An optimal transmission path is determined according to the real-time network status portrait, and the current transmission path is switched to the optimal transmission path.

[0029] The present invention provides a video surveillance system detection device, comprising:

[0030] The device operation and maintenance data collection module is used to collect device operation and maintenance data of the video surveillance devices in the target video surveillance system through the edge computing node. The device operation and maintenance data includes network traffic data, video data, and device power consumption data;

[0031] A preliminary detection result determination module is used to determine preliminary detection results based on the device operation and maintenance data, wherein the preliminary detection results include network traffic analysis results, video anomaly detection results, and device power consumption detection results;

[0032] a fault diagnosis model optimization module, configured to construct a global fault knowledge base based on the preliminary detection results, and iteratively train the fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model;

[0033] The fault diagnosis result determination module is used to input the collected real-time data to be detected into the optimized fault diagnosis model to obtain the fault diagnosis result.

[0034] The present invention also provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the video surveillance system detection method as described above is implemented.

[0035] The present invention also provides a computer-readable storage medium, comprising: a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the video surveillance system detection method as described above.

[0036] The following beneficial effects are achieved by implementing the present invention:

[0037] The present invention constructs a global fault knowledge base by collecting equipment operation and maintenance data and preliminary detection results to optimize the fault diagnosis model. The optimized diagnosis model can output multi-angle fault diagnosis results to cope with various fault conditions in complex network environments, improve the generalization ability and adaptability of detection, and thus effectively improve the detection accuracy of the video surveillance system.

[0038] Furthermore, the present invention collects network performance index data in real time and uses an optimized fault diagnosis model to perform automated evaluation and diagnosis, thereby enabling real-time, accurate, and predictive monitoring of the network status of the video surveillance system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the implementation. Obviously, the drawings described below are only some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0040] Figure 1 1 is a flow chart of a video surveillance system detection method provided by an embodiment of the present invention;

[0041] Figure 2 It is a structural diagram of a video surveillance system detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions in this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0044] In the description of the embodiments of this application, the technical terms "first" and "second" are used only to distinguish different objects and should not be understood to indicate or imply relative importance or implicitly specify the quantity, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "plurality" is more than two, unless otherwise clearly and specifically defined.

[0045] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0046] In the description of the embodiments of this application, the term "and / or" is simply a description of the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent the following three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally indicates that the associated objects are in an "or" relationship.

[0047] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0048] In the description of the embodiments of the present application, unless otherwise expressly specified or limited, technical terms such as "installed," "connected," "connected," and "fixed" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integration; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; internal connections between two components or interactions between two components. Those skilled in the art can understand the specific meanings of the above terms in the embodiments of the present application based on specific circumstances.

[0049] See also Figure 1 In order to solve the technical problem that the existing technology relies on limited manpower to patrol video surveillance points to achieve video surveillance system detection, and the manual detection method is difficult to adapt to complex network environments, resulting in poor detection effect, the present invention provides a video surveillance system detection method according to one embodiment of the present invention, including:

[0050] S1. Collecting device operation and maintenance data of video surveillance devices in the target video surveillance system through edge computing nodes. The device operation and maintenance data includes network traffic data, video data, and device power consumption data.

[0051] In the embodiment of the present invention, by collecting equipment operation and maintenance data of the video surveillance equipment, reliable original data can be provided for subsequent detection.

[0052] S2. Determine preliminary test results based on device operation and maintenance data. The preliminary test results include network traffic analysis results, video anomaly detection results, and device power consumption detection results.

[0053] S3. Build a global fault knowledge base based on the preliminary detection results, and iteratively train the fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model;

[0054] In an embodiment of the present invention, the preliminary detection results and the corresponding equipment operation and maintenance data are transmitted to the cloud. After data cleaning and standardization, a global fault knowledge base is constructed. The global fault knowledge base can cover standardized data sets of historical fault cases and characteristic patterns.

[0055] In an embodiment of the present invention, a federated learning algorithm can be used to iteratively train a pre-trained fault diagnosis model. During the training process, a secure multi-party computing protocol can be used to encrypt and transmit model parameters to ensure the privacy of edge node data. At the same time, the fault feature library is updated in combination with historical cases and real-time data to improve the model's recognition accuracy and generalization ability for new faults.

[0056] S4. Input the collected real-time data to be detected into the optimized fault diagnosis model to obtain the fault diagnosis result.

[0057] In an embodiment of the present invention, the real-time data to be detected can be real-time equipment operation and maintenance data, including real-time network traffic data, real-time video data and real-time equipment power consumption data. The real-time data to be detected is input into the optimized fault diagnosis model, and the fault diagnosis results corresponding to each equipment operation and maintenance data are output.

[0058] The embodiment of the present invention constructs a global fault knowledge base by collecting equipment operation and maintenance data and preliminary detection results to optimize the fault diagnosis model. The optimized diagnosis model can output multi-angle fault diagnosis results to cope with various fault conditions in complex network environments, improve the generalization ability and adaptability of detection, and thus effectively improve the detection accuracy of the video surveillance system.

[0059] In one embodiment, step S2, determining preliminary detection results based on equipment operation and maintenance data, includes:

[0060] S21. Input the video data into an artificial intelligence model based on a convolutional neural network, extract video frame features from the video data according to the convolution operation in the artificial intelligence model, determine whether there is an image abnormality based on the video frame features, and generate a video anomaly detection result;

[0061] In an embodiment of the present invention, an artificial intelligence model can be trained based on pre-labeled video data. In an embodiment of the present invention, the acquired video frame features can be similarly calculated with normal video frames, and when the similarity is lower than a preset threshold, it is determined that there is an abnormality in the picture.

[0062] S22. Analyze the network traffic data using a sliding window algorithm to obtain a network traffic analysis result;

[0063] S23. Use the exponential smoothing method to predict the current power consumption prediction value, compare the current power consumption prediction value with the current power consumption actual value, and determine the device power consumption detection result according to the comparison result.

[0064] This embodiment of the present invention automatically extracts video frames from video data using a convolution operation, reducing the amount of data required for processing while retaining key information, thereby effectively improving the detection efficiency of the video surveillance system. Furthermore, this embodiment of the present invention uses a sliding window algorithm to analyze network traffic data, enabling real-time capture of changing trends in network traffic and timely detection of abnormal traffic.

[0065] In one embodiment, step S21, extracting video frame features from the video data according to the convolution operation in the artificial intelligence model, includes:

[0066] The convolution operation is performed according to the following formula:

[0067]

[0068] Among them, Q j is the video frame feature, I(x,y) is the input video frame, K j (x,y) represents the value of the j-th convolution kernel at coordinate (x,y), It is an accumulation operation of the multiplication results of the elements at the corresponding positions of the video frame and the convolution kernel.

[0069] In one embodiment, step S22 uses a sliding window algorithm to analyze the network traffic data to obtain network traffic analysis results, including:

[0070] Using the sliding window algorithm and network traffic data, the average network traffic within the preset window is calculated;

[0071] Calculate the traffic variance based on the traffic mean and determine whether there is a sudden packet loss anomaly based on the traffic variance;

[0072] Generate network traffic analysis results based on whether there is a sudden packet loss anomaly.

[0073] In an embodiment of the present invention, the sliding window algorithm is as follows:

[0074] Calculate the mean flow rate within the preset window on the time series t1, t2, ..., tn Among them, N is the sliding window size, F(t i ) indicates that at time t i The network traffic value at time For t k to t k+N-1 The network traffic value at each moment is accumulated and then divided by the window size N to obtain the average traffic value within the window.

[0075] The variance is in, is the calculated mean flow rate within the window, It is the square of the difference between the flow value and the mean at each moment. By accumulating these square values ​​and dividing them by the window size N, the variance σ of the flow is obtained. 2 , variance is used to determine whether there is an abnormal situation of sudden packet loss in network traffic. When σ 2 When the set threshold is exceeded, it is determined as a burst packet loss anomaly.

[0076] The embodiment of the present invention adopts a sliding window algorithm to analyze network traffic data, which can capture the changing trend of network traffic in real time and promptly determine whether there is a sudden packet loss anomaly.

[0077] In one embodiment, the expression of the exponential smoothing method in step S23 is as follows:

[0078]

[0079] in, is the power consumption prediction value at the current moment, α is the smoothing coefficient, P(t-1) is the actual power consumption value at the previous moment, Predict the power consumption for the previous moment.

[0080] In one embodiment, step S4, inputting the collected real-time data to be detected into the optimized fault diagnosis model to obtain a fault diagnosis result, includes:

[0081] Collect real-time network performance indicator data, input the real-time network performance indicator data into the optimized fault diagnosis model, evaluate the network status of the target video surveillance system based on the optimized fault diagnosis model and the real-time network performance indicator data, and obtain the fault diagnosis results corresponding to the network status.

[0082] In an embodiment of the present invention, real-time network performance indicator data includes bandwidth utilization, delay data and packet loss rate of the transmission link. These data are input into the optimized fault diagnosis model to evaluate the network status and obtain fault diagnosis results corresponding to the network status. The embodiment of the present invention can also identify congested and unstable links.

[0083] The embodiment of the present invention collects network performance index data in real time and uses an optimized fault diagnosis model to perform automated evaluation and diagnosis, thereby achieving real-time, accurate, and predictive monitoring of the network status of a video surveillance system.

[0084] In one embodiment, after obtaining the fault diagnosis result corresponding to the network status in step S5, the method further includes:

[0085] Generate real-time network status portrait based on fault diagnosis results;

[0086] In an embodiment of the present invention, the network status portrait is a visual representation of the network status in the fault diagnosis results, which includes the current status of each node (such as routers, switches, servers, etc.) and link (communication path connecting nodes) in the network.

[0087] Determine the optimal transmission path based on the real-time network status portrait, and switch the current transmission path to the optimal transmission path.

[0088] In the real-time network status portrait, network devices are represented as nodes, and links between them as edges. Edge weights represent link performance metrics, such as latency or bandwidth. Based on a given starting point, an arbitrary endpoint is selected. Among multiple paths between the starting and ending points, the Dijkstra algorithm is used to find the path with the smallest weight, which is the optimal transmission path.

[0089] Switching the current transmission path to the optimal transmission path can avoid abnormal links, thereby ensuring low latency and high reliability of video transmission.

[0090] In one embodiment, blockchain technology can be used to immutably store network status data and path switching records, forming a traceable network history log. Furthermore, for multi-tenant data, virtual network technology can be used to partition each tenant into an independent logical space to isolate network resources and avoid diagnostic errors caused by cross-tenant data interference.

[0091] In one embodiment, the video data may be further processed to quantify the video clarity using peak signal-to-noise ratio and structural similarity index, specifically:

[0092] The peak signal-to-noise ratio is calculated as:

[0093]

[0094] Among them, PSNR is the peak signal-to-noise ratio, MAX i is the maximum value of the image pixel, MSE is the mean square error, and the calculation formula of MSE is:

[0095]

[0096] Where I(i, j) is the pixel value of the original image at the coordinate (i, j), K(i, j) is the pixel value of the processed image at the coordinate (i, j), the processed image is the image after compression, transmission or restoration, m is the number of rows of the image, and n is the number of columns of the image. The embodiment of the present invention calculates the sum of the squares of the corresponding pixel differences between the original image and the processed image, and then divides it by the total number of pixels to obtain the mean square error. The peak signal-to-noise ratio is calculated by MSE and MAX. i Calculated to measure video clarity.

[0097] In the embodiment of the present invention, the calculation formula of the structural similarity index SSIM(x, y) is:

[0098]

[0099] Among them, μ x and μ y are the mean of image x and y respectively, and are the variances of images x and y, σ xy is the covariance of images x and y, c1 and c2 are constants used to maintain calculation stability and prevent the denominator from being zero.

[0100] In one embodiment, video stream clarity, frame rate, device logs, and environmental sensor data can be integrated, and data redundancy can be eliminated through dimensionality reduction techniques such as principal component analysis. A cross-layer correlation analysis model can be established to mine fault association rules in the hidden layer between data.

[0101] In one embodiment, NLP technology can be used to process device logs, and high-frequency feature keywords such as overheating, packet loss, and abnormal interruption can be extracted through the bag-of-words model and TF-IDF algorithm to parse the fault semantic information in the logs.

[0102] In this embodiment of the present invention, the calculation formula of the TF-IDF algorithm is as follows:

[0103] The formula for calculating word frequency TF is:

[0104]

[0105] Among them, n t，d is the number of times word t appears in document d, ∑ t′∈d n t′,d The term frequency TF(t,d) of word t in document d is obtained by dividing the number of occurrences of word t in document d by the sum of the number of occurrences of all words in document d.

[0106] The inverse document frequency IDF calculation formula is:

[0107]

[0108] Where D is the document set, |D| is the total number of documents, and |{d∈D:t∈d}| is the number of documents containing the word t. The inverse document frequency IDF(t,D) of the word t is obtained by taking the logarithm of the ratio of the total number of documents to the number of documents containing the word t.

[0109] In one embodiment, time series data such as power, traffic, and link delay data can be input into a long short-term memory network, and the isolation forest algorithm can be used to detect abnormal points in these time series. Potential failures such as hard drive aging and cooling system failure can be predicted in advance for a preset period of time, which can be 72 hours.

[0110] In the embodiment of the present invention, the long short-term memory network combines the isolation forest algorithm to process time series data as follows: the long short-term memory network uses the input gate i t , forget gate f t , output gate o t and memory unit C t The specific formula for processing time series data is as follows:

[0111] Input gate i t =σ(W ii x t +W hi h t-1 +b i ), where σ is the sigmoid function, which is used to map the input value to between 0 and 1 and control the degree of information passing through. ii is the input x t The weight matrix to the input gate, x t is the input at the current moment, W hi is the hidden state h at the previous moment t-1 The weight matrix to the input gate, b i It is the bias vector of the input gate. Through the linear combination of these parameters and then through the sigmoid function, the input gate i is obtained. t The value of determines how much information of the current input is retained in the memory unit;

[0112] Forget Gate f t =σ(W if x t +W hf h t-1 +b f ), where W if and W hf The input x is t and the previous hidden state h t-1To the weight matrix of the forget gate, b f is the bias vector of the forget gate, the forget gate f t Determines which past information in a memory cell is retained or forgotten;

[0113] Memory unit C t =f t ⊙C t-1 +i t ⊙tanh(W ic x t +W hc h t-1 +b c ), where C t-1 is the memory unit of the previous moment, ⊙ is the element-by-element multiplication, W ic and W hc The input x is t and the previous hidden state h t-1 to the weight matrix of the memory unit, b c is the bias vector of the memory unit, and the tanh function maps the input value to between -1 and 1, passing through the forget gate f t For the previous moment memory unit C t-1 The retained part and input gate i t The new information generated by the current input under control is combined to obtain the memory unit C at the current moment t ;

[0114] Output gate o t =σ(W io x t +W ho h t-1 +b0), by the corresponding weight matrix W io 、W ho and the bias vector b o With input x t and the previous hidden state h t-1 After linear combination, the sigmoid function is used to determine the memory unit C. t How much information is output to the hidden state h at the current moment t .

[0115] h t =o t ⊙tanh(C t ), the current hidden state h t By output gate o t and the memory unit C after tanh function processing t The element-by-element multiplication is used for subsequent calculations and potential fault prediction.

[0116] In one embodiment, the video quality parameters of the peak signal-to-noise ratio and the structural similarity index can be periodically collected at the camera end. With the help of the fast Fourier transform algorithm, the video frame I(x, y) is transformed into a frequency domain representation F(u, v) after FFT transformation, which is used for frequency domain feature extraction for peak signal-to-noise ratio and structural similarity index calculation, and uploaded to the edge node through the LoRaWAN protocol.

[0117] In one embodiment, the blur level can be determined by calculating the gradient information of the video frame, specifically:

[0118] Assume the gradient amplitude of the video frame I(x,y) The degree of image blur is determined by the gradient amplitude, where: Represents the gradient of the video frame I(x,y) in the x direction, that is, the rate of change of the pixel value of the image in the horizontal direction, It represents the gradient of the video frame I(x,y) in the y direction, that is, the rate of change of the pixel value of the image in the vertical direction. The gradient amplitude G(x,y) of the video frame at the coordinate (x,y) is obtained by taking the square root of the sum of the squares of the gradients in the two directions. It is used to determine the degree of image blur.

[0119] The implementation of the embodiments of the present invention has the following beneficial effects:

[0120] The embodiment of the present invention constructs a global fault knowledge base by collecting equipment operation and maintenance data and preliminary detection results to optimize the fault diagnosis model. The optimized diagnosis model can output multi-angle fault diagnosis results to cope with various fault conditions in complex network environments, improve the generalization ability and adaptability of detection, and thus effectively improve the detection accuracy of the video surveillance system.

[0121] Furthermore, the embodiments of the present invention collect network performance index data in real time and use the optimized fault diagnosis model to perform automated evaluation and diagnosis, thereby achieving real-time, accurate, and predictive monitoring of the network status of the video surveillance system.

[0122] See also Figure 2 Based on the same inventive concept as the above embodiment, the present invention provides a video surveillance system detection device, comprising:

[0123] The device operation and maintenance data collection module 10 is used to collect device operation and maintenance data of the video surveillance devices in the target video surveillance system through the edge computing node. The device operation and maintenance data includes network traffic data, video data, and device power consumption data;

[0124] A preliminary detection result determination module 20 is used to determine preliminary detection results based on device operation and maintenance data. The preliminary detection results include network traffic analysis results, video anomaly detection results, and device power consumption detection results.

[0125] The fault diagnosis model optimization module 30 is used to build a global fault knowledge base based on the preliminary detection results, and iteratively train the fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model;

[0126] The fault diagnosis result determination module 40 is used to input the collected real-time data to be detected into the optimized fault diagnosis model to obtain the fault diagnosis result.

[0127] In one embodiment, determining preliminary detection results based on device operation and maintenance data includes:

[0128] The video data is input into an artificial intelligence model based on a convolutional neural network. The video frame features in the video data are extracted based on the convolution operation in the artificial intelligence model. The video frame features are used to determine whether there are any image anomalies and generate video anomaly detection results.

[0129] Use sliding window algorithm to analyze network traffic data and obtain network traffic analysis results;

[0130] The exponential smoothing method is used to predict the current power consumption forecast value, which is compared with the current power consumption actual value, and the device power consumption detection result is determined based on the comparison result.

[0131] In one embodiment, extracting video frame features from video data based on a convolution operation in an artificial intelligence model includes:

[0132] The convolution operation is performed according to the following formula:

[0133]

[0134] Among them, Q j is the video frame feature, I(x,y) is the pixel value of the input video frame at the coordinate (x,y), K j (x,y) represents the value of the j-th convolution kernel at coordinate (x,y), It is an accumulation operation of the multiplication results of the elements at the corresponding positions of the video frame and the convolution kernel.

[0135] In one embodiment, a sliding window algorithm is used to analyze network traffic data to obtain network traffic analysis results, including:

[0136] Using the sliding window algorithm and network traffic data, the average network traffic within the preset window is calculated;

[0137] Calculate the traffic variance based on the traffic mean and determine whether there is a sudden packet loss anomaly based on the traffic variance;

[0138] Generate network traffic analysis results based on whether there is a sudden packet loss anomaly.

[0139] In one embodiment, the expression of the exponential smoothing method is as follows:

[0140]

[0141] in, is the power consumption prediction value at the current moment, α is the smoothing coefficient, P(t-1) is the actual power consumption value at the previous moment, Predict the power consumption for the previous moment.

[0142] In one embodiment, the collected real-time data to be detected is input into the optimized fault diagnosis model to obtain a fault diagnosis result, including:

[0143] Collect real-time network performance indicator data, input the real-time network performance indicator data into the optimized fault diagnosis model, evaluate the network status of the target video surveillance system based on the optimized fault diagnosis model and the real-time network performance indicator data, and obtain the fault diagnosis results corresponding to the network status.

[0144] In one embodiment, after obtaining the fault diagnosis result corresponding to the network status, the method further includes:

[0145] Generate real-time network status portrait based on fault diagnosis results;

[0146] Determine the optimal transmission path based on the real-time network status portrait, and switch the current transmission path to the optimal transmission path.

[0147] It can be understood that the above-mentioned device embodiment corresponds to the method embodiment of the present invention, and can implement the video surveillance system detection method provided by any of the above-mentioned method embodiments of the present invention.

[0148] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Furthermore, in the drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that they have a communication connection, which may be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement the present invention without inventive effort.

[0149] Based on the above-mentioned embodiment of the video surveillance system detection method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the video surveillance system detection method of any embodiment of the present invention.

[0150] For example, in this embodiment, the computer program may be divided into one or more modules, one or more of which are stored in a memory and executed by a processor to implement the present invention. One or more module elements may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in a terminal device.

[0151] The terminal device can be a computing device such as a desktop computer, notebook computer, PDA, or cloud server. The terminal device may include, but is not limited to, a processor and memory.

[0152] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting various parts of the entire terminal device using various interfaces and lines.

[0153] Based on the above-mentioned method embodiments, another embodiment of the present invention provides a computer-readable storage medium, including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the video surveillance system detection method of any one of the above-mentioned method embodiments of the present invention.

[0154] Among them, the module / unit integrated in the device / terminal equipment, if implemented in the form of a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, it can implement the steps of the above-mentioned various method embodiments. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer-readable media may include: any entity or device that can carry computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0155] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A video surveillance system detection method, characterized in that: include: Collecting device operation and maintenance data of video surveillance devices in the target video surveillance system through edge computing nodes, wherein the device operation and maintenance data includes network traffic data, video data, and device power consumption data; Determine preliminary detection results based on the device operation and maintenance data, wherein the preliminary detection results include network traffic analysis results, video anomaly detection results, and device power consumption detection results; Building a global fault knowledge base based on the preliminary detection results, and iteratively training a fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model; The collected real-time data to be detected is input into the optimized fault diagnosis model to obtain a fault diagnosis result.

2. The video surveillance system detection method according to claim 1, wherein: Determining the preliminary detection result based on the equipment operation and maintenance data includes: Inputting the video data into an artificial intelligence model based on a convolutional neural network, extracting video frame features from the video data according to a convolution operation in the artificial intelligence model, determining whether there is an image anomaly based on the video frame features, and generating a video anomaly detection result; Using a sliding window algorithm to analyze the network traffic data to obtain a network traffic analysis result; The exponential smoothing method is used to predict the current power consumption forecast value, which is compared with the current power consumption actual value, and the device power consumption detection result is determined based on the comparison result.

3. The video surveillance system detection method according to claim 2, wherein: The extracting video frame features from the video data according to the convolution operation in the artificial intelligence model includes: The convolution operation is performed according to the following formula: Among them, Q j is the video frame feature, I(x,y) is the input video frame, K j (x,y) represents the value of the j-th convolution kernel at coordinate (x,y), It is an accumulation operation of the multiplication results of the elements at the corresponding positions of the video frame and the convolution kernel.

4. The video surveillance system detection method according to claim 2, wherein: The network traffic data is analyzed using a sliding window algorithm to obtain a network traffic analysis result, including: Using a sliding window algorithm and the network traffic data, calculate the average network traffic within a preset window; Calculating a traffic variance based on the traffic mean, and determining whether there is a sudden packet loss anomaly based on the traffic variance; Generate network traffic analysis results based on whether there is a sudden packet loss anomaly.

5. The video surveillance system detection method according to claim 2, wherein: The expression of the exponential smoothing method is as follows: in, is the power consumption prediction value at the current moment, α is the smoothing coefficient, P(t-1) is the actual power consumption value at the previous moment, Predict the power consumption for the previous moment.

6. The video surveillance system detection method according to claim 1, wherein: The method of inputting the collected real-time data to be detected into the optimized fault diagnosis model to obtain a fault diagnosis result includes: Collect real-time network performance indicator data, input the real-time network performance indicator data into the optimized fault diagnosis model, evaluate the network status of the target video surveillance system based on the optimized fault diagnosis model and the real-time network performance indicator data, and obtain a fault diagnosis result corresponding to the network status.

7. The video surveillance system detection method according to claim 6, wherein: After obtaining the fault diagnosis results corresponding to the network status, the following steps are also performed: Generate a real-time network status profile based on the fault diagnosis result; An optimal transmission path is determined according to the real-time network status portrait, and the current transmission path is switched to the optimal transmission path.

8. A video surveillance system detection device, characterized in that: include: The device operation and maintenance data collection module is used to collect device operation and maintenance data of the video surveillance devices in the target video surveillance system through the edge computing node. The device operation and maintenance data includes network traffic data, video data, and device power consumption data; A preliminary detection result determination module is used to determine preliminary detection results based on the device operation and maintenance data, wherein the preliminary detection results include network traffic analysis results, video anomaly detection results, and device power consumption detection results; a fault diagnosis model optimization module, configured to construct a global fault knowledge base based on the preliminary detection results, and iteratively train the fault diagnosis model based on the global fault knowledge base to obtain an optimized fault diagnosis model; The fault diagnosis result determination module is used to input the collected real-time data to be detected into the optimized fault diagnosis model to obtain the fault diagnosis result.

9. A terminal device, characterized in that: The device comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the video surveillance system detection method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that include: A stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the video surveillance system detection method according to any one of claims 1 to 7.

Images