Anti-interference optical fiber perimeter signal identification method

Abstract

The invention relates to signal identification, in particular to an anti-interference optical fiber perimeter signal identification method, which comprises the following steps: acquiring disturbance data of all monitoring point positions in a to-be-monitored area in real time, analyzing the disturbance data, and determining a target monitoring point position and a target monitoring area; acquiring disturbance data of all monitoring points in the target monitoring area, and mapping the disturbance data; judging whether effective disturbance occurs or not according to the mapping value; if the effective disturbance occurs, inputting disturbance data of all monitoring point positions of the target monitoring area into a trained signal recognition model, and performing mode recognition on disturbance categories by using the signal recognition model; according to the technical scheme provided by the invention, the defects of relatively low signal identification accuracy and efficiency in the prior art can be effectively overcome.

Description

Technical Field

[0001] The invention relates to signal recognition, and in particular to an interference-resistant optical fiber perimeter signal recognition method. Background Art

[0002] The fiber optic perimeter alarm system is a key technical means to achieve environmental security monitoring. It can sense and monitor disturbance signals in the fiber optic distribution environment. Compared with traditional electronic security detection methods, the fiber optic perimeter alarm system has the advantages of high sensitivity, strong anti-electromagnetic interference ability, full-range passive, corrosion resistance, simple structure, etc. It is widely used in communication security, energy pipeline transmission, bridge and tunnel monitoring and other fields.

[0003] However, in large-scale complex monitoring environments and with different optical cable layouts, existing fiber optic perimeter alarm systems are difficult to accurately capture and identify different types of intrusion behaviors, and are easily affected by interference signals from the external environment (such as wind, rain, snow, and passing vehicles), resulting in low signal recognition accuracy and efficiency, which in turn leads to frequent false alarms, misreporting, and missed reports, which cannot meet actual application requirements. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides an interference-resistant optical fiber perimeter signal recognition method, which can effectively overcome the defects of low signal recognition accuracy and efficiency in the prior art.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] An anti-interference optical fiber perimeter signal recognition method comprises the following steps:

[0009] S1. Real-time collection of disturbance data of all monitoring points in the area to be monitored, and analysis of the disturbance data to determine the target monitoring points and target monitoring areas;

[0010] S2. Obtain disturbance data of all monitoring points in the target monitoring area and map the disturbance data;

[0011] S3, judging whether effective disturbance occurs according to the mapping value;

[0012] S4. If an effective disturbance occurs, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, and the signal recognition model is used to perform pattern recognition on the disturbance category.

[0013] Preferably, in S1, the disturbance data is analyzed to determine the target monitoring point and the target monitoring area, including:

[0014] If the disturbance data of a certain monitoring point in the monitored area is greater than a preset threshold, the monitoring point will be used as the target monitoring point, and the target monitoring area will be delineated based on a preset size range with the target monitoring point as the center.

[0015] Preferably, in S2, the disturbance data of all monitoring points in the target monitoring area are obtained, and the disturbance data are mapped, including:

[0016] Obtain the disturbance data of all monitoring points in the target monitoring area and map the disturbance data to the range of 0 to 255.

[0017] Preferably, judging whether a valid disturbance occurs according to the mapping value in S3 includes:

[0018] Calculate the gray level co-occurrence matrix according to the mapping value, and calculate the statistical parameters of texture features in multiple directions in the gray level co-occurrence matrix;

[0019] If the statistical parameters of texture features in multiple directions are similar, it is considered that the disturbance data is caused by irregular system noise; otherwise, it is considered that effective disturbance occurs.

[0020] Preferably, the calculating of statistical parameters of texture features in multiple directions in the gray level co-occurrence matrix includes:

[0021] Calculate the texture feature statistical parameters including contrast, correlation, energy, entropy and inverse difference moment at 0°, 45°, 90° and 135° in the gray level co-occurrence matrix.

[0022] Preferably, if an effective disturbance occurs in S4, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, including:

[0023] S41, collecting disturbance data of corresponding monitoring points in the monitored area under different effective disturbances, and dividing them into a training set, a validation set and a test set according to a preset ratio;

[0024] S42, setting the loss function and optimizer of the signal recognition model, and inputting the training set into the signal recognition model for model training;

[0025] S43, calculating the loss value based on the loss function, and the optimizer updates the model parameters according to the loss value and the network gradient information;

[0026] S44. If the loss value is less than the preset threshold, the model training is completed, and the current signal recognition model is the trained signal recognition model. Otherwise, return to S42 and continue to use the training set for model training.

[0027] Preferably, the signal recognition model adopts a wavelet neural network model, and the wavelet neural network model is expressed by the following formula:

[0028]

[0029] Among them, x is the model input, y i is the output of the i-th node in the hidden layer, P i , Q i are the wavelet basis functions φ of the i-th node in the hidden layer. i The expansion factor and translation factor of i is the weight of the i-th node in the hidden layer, is the normalization factor, and m is the number of nodes in the hidden layer.

[0030] Preferably, setting the loss function and optimizer of the signal recognition model in S42 includes:

[0031] The loss function L of the signal recognition model is expressed as follows:

[0032]

[0033] Among them, y i is the output of the i-th node in the hidden layer, is the actual output of the i-th node in the hidden layer, and m is the number of nodes in the hidden layer.

[0034] Preferably, in S42, the training set is input into the signal recognition model for model training, including:

[0035] During model training, the weight ω of the i-th node in the hidden layer is i The iteration is performed using the following formula:

[0036]

[0037] Among them, ω i (t) is the weight of the i-th node in the hidden layer at the t-th iteration, ω i (t+1) is the weight of the i-th node in the hidden layer at the t+1th iteration, L is the loss function of the recognition model, η is the learning rate, Δω i (t) is the weight change, Δω i (t) = ω i (t)-ω i (t-1), α is the weight change coefficient.

[0038] After obtaining the trained signal recognition model, it includes:

[0039] S45. Input the validation set into the trained signal recognition model, evaluate the generalization ability of the model by observing the performance on the validation set, and tune the hyperparameters and structure of the model;

[0040] S46. Input the test set into the tuned signal recognition model to evaluate the model performance.

[0041] (III) Beneficial effects

[0042] Compared with the prior art, the anti-interference optical fiber perimeter signal identification method provided by the present invention has the following beneficial effects:

[0043] 1) Real-time collection of disturbance data of all monitoring points in the monitored area, and analysis of the disturbance data to determine the target monitoring points and target monitoring areas, obtain disturbance data of all monitoring points in the target monitoring area, and map the disturbance data to the range of 0 to 255, calculate the gray-level co-occurrence matrix according to the mapping value, and calculate the texture feature statistical parameters of multiple directions in the gray-level co-occurrence matrix. If the texture feature statistical parameters of multiple directions are similar, it is considered that the disturbance data is caused by irregular system noise, otherwise it is considered that effective disturbance occurs, and the disturbance data is mapped, and the gray-level co-occurrence matrix and texture feature statistical parameters are calculated for the mapping value. By comparing the texture feature statistical parameters, the invalid disturbance caused by irregular system noise can be accurately identified, providing support for subsequent accurate signal recognition;

[0044] 2) If an effective disturbance occurs, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, and the signal recognition model is used to perform pattern recognition on the disturbance category. Thus, the signal recognition model can be used to perform pattern recognition on the disturbance data of all monitoring points in the target monitoring area, and various disturbance categories including interference disturbances caused by the external environment and intrusion disturbances caused by different types of intrusion behaviors can be accurately and efficiently identified. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 It is a schematic diagram of the process of the present invention;

[0047] Figure 2 It is a schematic diagram of a flow chart for determining whether effective disturbance occurs in the present invention;

[0048] Figure 3It is a schematic diagram of the process of model training for the signal recognition model in the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] An anti-interference optical fiber perimeter signal recognition method, such as Figure 1 and Figure 2 As shown, S1, real-time collection of disturbance data of all monitoring points in the area to be monitored, and analysis of the disturbance data to determine target monitoring points and target monitoring areas.

[0051] Specifically, the disturbance data is analyzed to determine the target monitoring points and target monitoring areas, including:

[0052] If the disturbance data of a certain monitoring point in the monitored area is greater than a preset threshold, the monitoring point will be used as the target monitoring point, and the target monitoring area will be delineated based on a preset size range with the target monitoring point as the center.

[0053] S2. Obtain disturbance data of all monitoring points in the target monitoring area and map the disturbance data, including:

[0054] Obtain the disturbance data of all monitoring points in the target monitoring area and map the disturbance data to the range of 0 to 255.

[0055] S3. Judging whether effective disturbance (including disturbance disturbance caused by the external environment and intrusion disturbance caused by different types of intrusion behaviors) occurs according to the mapping value, specifically including:

[0056] Calculate the gray level co-occurrence matrix according to the mapping value, and calculate the statistical parameters of texture features in multiple directions in the gray level co-occurrence matrix;

[0057] If the statistical parameters of texture features in multiple directions are similar, it is considered that the disturbance data is caused by irregular system noise; otherwise, it is considered that effective disturbance occurs.

[0058] Specifically, the texture feature statistical parameters in multiple directions in the gray level co-occurrence matrix are calculated, including:

[0059] Calculate the texture feature statistical parameters including contrast, correlation, energy, entropy and inverse difference moment at 0°, 45°, 90° and 135° in the gray level co-occurrence matrix.

[0060] The above technical scheme collects the disturbance data of all monitoring points in the monitored area in real time, analyzes the disturbance data, determines the target monitoring points and the target monitoring area, obtains the disturbance data of all monitoring points in the target monitoring area, and maps the disturbance data to the range of 0 to 255, calculates the grayscale co-occurrence matrix according to the mapping value, and calculates the texture feature statistical parameters of multiple directions in the grayscale co-occurrence matrix. If the texture feature statistical parameters of multiple directions are similar, it is considered that the disturbance data is caused by irregular system noise, otherwise it is considered that effective disturbance occurs, maps the disturbance data, and calculates the grayscale co-occurrence matrix and texture feature statistical parameters of the mapping value. By comparing the texture feature statistical parameters, the invalid disturbance caused by irregular system noise can be accurately identified, providing support for subsequent accurate signal recognition.

[0061] like Figure 1 As shown, S4, if an effective disturbance occurs, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, and the signal recognition model is used to perform pattern recognition on the disturbance category.

[0062] Specifically, if an effective disturbance occurs, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, such as Figure 3 As shown, including:

[0063] S41, collecting disturbance data of corresponding monitoring points in the monitored area under different effective disturbances, and dividing them into a training set, a validation set and a test set according to a preset ratio;

[0064] S42, setting the loss function and optimizer of the signal recognition model, and inputting the training set into the signal recognition model for model training;

[0065] S43, calculating the loss value based on the loss function, and the optimizer updates the model parameters according to the loss value and the network gradient information;

[0066] S44. If the loss value is less than the preset threshold, the model training is completed, and the current signal recognition model is the trained signal recognition model. Otherwise, return to S42 and continue to use the training set for model training.

[0067] In the technical solution of the present application, the signal recognition model adopts a wavelet neural network model, and the wavelet neural network model is expressed by the following formula:

[0068]

[0069] Among them, x is the model input, y i is the output of the i-th node in the hidden layer, P i , Q iare the wavelet basis functions φ of the i-th node in the hidden layer. i The expansion factor and translation factor of i is the weight of the i-th node in the hidden layer, is the normalization factor, and m is the number of nodes in the hidden layer.

[0070] 1) In S42, the loss function and optimizer of the signal recognition model are set, including:

[0071] The loss function L of the signal recognition model is expressed as follows:

[0072]

[0073] Among them, y i is the output of the i-th node in the hidden layer, is the actual output of the i-th node in the hidden layer, and m is the number of nodes in the hidden layer.

[0074] 2) In S42, the training set is input into the signal recognition model for model training, including:

[0075] During model training, the weight ω of the i-th node in the hidden layer is i The iteration is performed using the following formula:

[0076]

[0077] Among them, ω i (t) is the weight of the i-th node in the hidden layer at the t-th iteration, ω i (t+1) is the weight of the i-th node in the hidden layer at the t+1th iteration, L is the loss function of the recognition model, η is the learning rate, Δω i (t) is the weight change, Δω i (t) = ω i (t)-ω i (t-1), α is the weight change coefficient.

[0078] 3) After obtaining the trained signal recognition model in S44, it includes:

[0079] S45. Input the validation set into the trained signal recognition model, evaluate the generalization ability of the model by observing the performance on the validation set, and tune the hyperparameters and structure of the model;

[0080] S46. Input the test set into the tuned signal recognition model to evaluate the model performance.

[0081] In the above technical scheme, if an effective disturbance occurs, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, and the signal recognition model is used to perform pattern recognition on the disturbance category, so that the signal recognition model can be used to perform pattern recognition on the disturbance data of all monitoring points in the target monitoring area, and various disturbance categories including interference disturbances formed by the external environment and intrusion disturbances formed by different categories of intrusion behaviors can be accurately and efficiently identified.

[0082] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An anti-interference optical fiber perimeter signal recognition method, characterized in that: The following steps are involved: S1. Real-time collection of disturbance data of all monitoring points in the area to be monitored, and analysis of the disturbance data to determine the target monitoring points and target monitoring areas; S2. Obtain disturbance data of all monitoring points in the target monitoring area and map the disturbance data; S3, judging whether effective disturbance occurs according to the mapping value; S4. If an effective disturbance occurs, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, and the signal recognition model is used to perform pattern recognition on the disturbance category.

2. The anti-interference optical fiber perimeter signal identification method according to claim 1, characterized in that: In S1, the disturbance data is analyzed to determine the target monitoring points and target monitoring areas, including: If the disturbance data of a certain monitoring point in the monitored area is greater than a preset threshold, the monitoring point will be used as the target monitoring point, and the target monitoring area will be delineated based on a preset size range with the target monitoring point as the center.

3. The anti-interference optical fiber perimeter signal identification method according to claim 2, characterized in that: S2 obtains the disturbance data of all monitoring points in the target monitoring area and maps the disturbance data, including: Obtain the disturbance data of all monitoring points in the target monitoring area and map the disturbance data to the range of 0 to 255.

4. The anti-interference optical fiber perimeter signal identification method according to claim 3, characterized in that: In S3, whether a valid disturbance occurs is determined according to the mapping value, including: Calculate the gray level co-occurrence matrix according to the mapping value, and calculate the statistical parameters of texture features in multiple directions in the gray level co-occurrence matrix; If the statistical parameters of texture features in multiple directions are similar, it is considered that the disturbance data is caused by irregular system noise; otherwise, it is considered that effective disturbance occurs.

5. The anti-interference optical fiber perimeter signal identification method according to claim 4, characterized in that: The calculating of the statistical parameters of texture features in multiple directions in the gray level co-occurrence matrix includes: Calculate the texture feature statistical parameters including contrast, correlation, energy, entropy and inverse difference moment at 0°, 45°, 90° and 135° in the gray level co-occurrence matrix.

6. The anti-interference optical fiber perimeter signal identification method according to claim 1, characterized in that: If an effective disturbance occurs in S4, the disturbance data of all monitoring points in the target monitoring area are input into the trained signal recognition model, including: S41, collecting disturbance data of corresponding monitoring points in the monitored area under different effective disturbances, and dividing them into a training set, a validation set and a test set according to a preset ratio; S42, setting the loss function and optimizer of the signal recognition model, and inputting the training set into the signal recognition model for model training; S43, calculating the loss value based on the loss function, and the optimizer updates the model parameters according to the loss value and the network gradient information; S44. If the loss value is less than the preset threshold, the model training is completed, and the current signal recognition model is the trained signal recognition model. Otherwise, return to S42 and continue to use the training set for model training.

7. The anti-interference optical fiber perimeter signal identification method according to claim 6, characterized in that: The signal recognition model adopts a wavelet neural network model, and the wavelet neural network model is expressed by the following formula: Among them, x is the model input, y i is the output of the i-th node in the hidden layer, P i , Q i are the wavelet basis functions φ of the i-th node in the hidden layer. i The expansion factor and translation factor of i is the weight of the i-th node in the hidden layer, is the normalization factor, and m is the number of nodes in the hidden layer.

8. The anti-interference optical fiber perimeter signal identification method according to claim 7, characterized in that: The loss function and optimizer of the signal recognition model are set in S42, including: The loss function L of the signal recognition model is expressed as follows: Among them, y i is the output of the i-th node in the hidden layer, is the actual output of the i-th node in the hidden layer, and m is the number of nodes in the hidden layer.

9. The anti-interference optical fiber perimeter signal identification method according to claim 8, characterized in that: In S42, the training set is input into the signal recognition model for model training, including: During model training, the weight ω of the i-th node in the hidden layer is i The iteration is performed using the following formula: Among them, ω i (t) is the weight of the i-th node in the hidden layer at the t-th iteration, ω i (t+1) is the weight of the i-th node in the hidden layer at the t+1th iteration, L is the loss function of the recognition model, η is the learning rate, Δω i (t) is the weight change, Δω i (t) = ω i (t)-ω i (t-1), α is the weight change coefficient.

10. The anti-interference optical fiber perimeter signal identification method according to claim 6, characterized in that: After obtaining the trained signal recognition model, it includes: S45. Input the validation set into the trained signal recognition model, evaluate the generalization ability of the model by observing the performance on the validation set, and tune the hyperparameters and structure of the model; S46. Input the test set into the tuned signal recognition model to evaluate the model performance.

Images