Camera detection method based on wireless signal

Abstract

The invention is applicable to the field of safety detection, and provides a camera detection method based on wireless signals, which comprises the steps of probe equipment connection, data acquisition, data analysis, data threshold calculation, model calling and judgment, label judgment and classification, equipment list display, positioning search, signal lamp strip control and the like. According to the invention, the network traffic in the WiFi environment is obtained through the Bluetooth probe device, illegal wireless camera detection and positioning are identified by performing feature analysis and uplink and downlink calculation on the traffic, and the method is mainly used for detecting and positioning illegal wireless cameras so as to protect security of confidential places and confidential conferences, personal privacy and social security.

Description

Technical Field

[0001] The present invention belongs to the technical field of safety detection, and in particular relates to a camera detection method based on wireless signals. Background Art

[0002] With the rapid development of technology, wireless cameras have been widely used in security monitoring, smart home, hotel business travel, social entertainment and other fields. However, its convenience also brings problems of privacy leakage and security risks. There are many types of hidden cameras on the market with different shapes. Some hidden cameras use pinholes to hide in unknown corners. It is difficult for ordinary users to know whether there is a privacy leakage problem in their environment.

[0003] Since illegal cameras may be hidden in various environments and have limited range, it is difficult to accurately detect camera devices within a specific range. There are several common methods for detecting illegal cameras on the market, such as infrared detection, mobile phone camera detection, route detection, mirror glass detection, thermal imaging detection, etc. However, these methods often have problems such as low detection accuracy and complex operation, which cannot meet users' needs for convenience and reliability. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to provide a camera detection method based on wireless signals, aiming to solve the technical problems of low detection accuracy and complex operability of the existing method.

[0005] The present invention adopts the following technical solution:

[0006] The camera detection method based on wireless signals comprises the following steps:

[0007] Step S1, establishing a Bluetooth connection between the mobile terminal and the Bluetooth probe device;

[0008] Step S2: The Bluetooth probe device collects WiFi traffic data packets in the surrounding environment and sends the collected data packets to the mobile terminal;

[0009] Step S3: The mobile terminal performs data analysis on the data packet, determines the validity of the analysis data result, performs label matching on the valid data through the MAC library of the server, and saves the matching record;

[0010] Step S4: for valid data, determine whether the data meets the threshold condition. If the threshold condition is met, the device is a device with camera features and the model call discrimination process is performed; if the threshold condition is not met, the device is classified as other devices;

[0011] Step S5, model calling and judging using a signal analysis model to judge whether it is a camera;

[0012] Step S6: For the identified devices, including cameras, suspected cameras and other devices, mark the MAC brand label of the identified device according to the matching record of step S3;

[0013] Step S7: Generate a device list and display the classified device information to the user.

[0014] Furthermore, the method further comprises the following steps:

[0015] Step S8, select a device from the device list and specify a channel for the device. The Bluetooth probe device obtains single-channel data in real time. According to the strength of the data signal, the Bluetooth probe device is moved in the environment, and the signal strength of the Bluetooth probe device is displayed in real time on the interface of the mobile terminal. The stronger the signal, the closer it is to the target device, so as to assist the user in locating and finding the target device.

[0016] Furthermore, the method further comprises the following steps:

[0017] Step S9, the mobile terminal parses the single-channel data and displays the data in real time, and calculates the signal strength value at the same time, and sends the signal strength value to the Bluetooth probe device. The Bluetooth probe device controls the corresponding signal strength signal light according to the value, and intuitively displays the signal strength of the device.

[0018] Furthermore, the specific process of step S5 is as follows:

[0019] Create a training data set, initialize StandardScaler, configure random forest classifier parameters, identify the data according to basic traffic characteristics, time characteristics, size characteristics, and comprehensive characteristics, use StandardScaler to standardize the data, create a SMOTE-RandomForest pipeline, combine standardization, SMOTE oversampling, and random forest classifiers, perform data segmentation and model training, evaluate the indicators, set evaluation indicators, use the logistic regression model for prediction, and calculate the probability value. When the probability is greater than 0.5, it is determined to be a camera, between 0.3 and 0.5, it is determined to be a suspected camera, and less than 0.3, it is determined to be a non-camera.

[0020] Furthermore, the feature identification of the basic traffic features includes identification of uplink traffic ratio, relative stability of traffic size, and periodicity of broadband usage; the feature identification of the time features includes time efficiency, time variability, and time skewness features; the feature identification of the size features includes packet size stability, packet size variability, and burst traffic features; the feature identification of the comprehensive features includes time-size correlation features and traffic density features.

[0021] The beneficial effects of the present invention are: the present invention uses a simple and convenient Bluetooth probe device, and the probe device uses Bluetooth to interact with mobile phone applets for data. Users do not need professional knowledge, and can establish a connection with the probe device by turning on the device and scanning the device QR code. The highly sensitive signal receiving module of the probe device can quickly detect the surrounding environment, and the signal analysis model can be used to analyze the identified network traffic, so as to quickly detect the type of traffic device, and promptly help users discover potential privacy leakage risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a flow chart of a camera detection method based on wireless signals provided by an embodiment of the present invention;

[0023] Figure 2 It is a specific flow chart of the camera detection method based on wireless signals provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0025] In order to illustrate the technical solution of the present invention, a specific embodiment is provided below for illustration.

[0026] like Figure 1 , 2 As shown, the camera detection method based on wireless signals provided in this embodiment includes the following steps:

[0027] Step S1, Bluetooth probe device connection: establish a Bluetooth connection between the mobile terminal and the Bluetooth probe device.

[0028] The mobile terminal is generally a mobile phone, and the functions of the present invention can be realized through an APP or a mini program, preferably a mini program. First, turn on the Bluetooth probe device, the user starts the mini program and searches for nearby Bluetooth devices, when the Bluetooth probe device is searched, the user selects the device to establish a Bluetooth connection request, and the mobile phone mini program establishes a communication connection with the Bluetooth probe device.

[0029] Step S2, data collection: The Bluetooth probe device collects WiFi traffic data packets in the surrounding environment and sends the collected data packets to the mobile terminal.

[0030] The applet sends a command to start scanning, and the probe device starts data collection. The Bluetooth probe device starts to scan and collect data at a fixed time (for example, 120 seconds), collects and obtains WiFi traffic data packets in the surrounding environment, and sends the WiFi traffic data packets to the applet of the mobile terminal via Bluetooth.

[0031] Step S3, data analysis: The mobile terminal performs data analysis on the data packet, determines the validity of the analysis data result, performs label matching on the valid data through the MAC library of the server, and saves the matching record.

[0032] For the collected data packets, the data packets are parsed and the validity is first determined according to the data packet rules. For example, in this embodiment, the data packet length is fixed to 20 bytes. The frame header is 0xaa, the check byte is: 0x6a, and the specific data packet format is shown in the table:

[0033]

[0034] For example: aa 0114008a 4c c64c 272a 3180ea 072e 40780b ac 00790000006a

[0035] The frame header is 0xaa, the data type is 0x01, the data length is 0x14, the uplink and downlink type subtype is 0x8a, the device mac is: 4c c64c 272a 31, the AP's mac is: 80ea 072e 4078, the data channel is 0x0b, the signal strength is 0xac, the traffic size is 0x79 bytes, and the check byte is: 0x6a.

[0036] By judging whether the data matches the data packet rules, the validity can be judged. After judging the validity data, the MAC name is matched and tagged through the MAC library on the server. The MAC record devices that can be matched are named for easy display, and the MAC records that cannot be matched are unknown devices. By matching the valid data with the MAC library, the data is labeled and the matching records are saved, and the preliminary classification and identification of the data can be quickly realized. It ensures that only valuable data enters the subsequent processing link, improving the efficiency and accuracy of the entire system.

[0037] This step collects data within 120 seconds through the scanning program, which can quickly collect signal data in the surrounding environment, save time, reduce the difficulty of operation for users, and improve detection efficiency. For example, in scenarios where it is necessary to quickly determine whether there are suspicious cameras around, such as hotels and business trips, it can help users determine whether the network environment they are in has the risk of privacy leakage in a short time.

[0038] Step S4, data threshold calculation: For valid data, determine whether the data meets the threshold condition. If the threshold condition is met, it is a device with camera features and the model call judgment process is performed; if the threshold condition is not met, it is classified as other devices.

[0039] The threshold conditions are judged based on the upstream and downstream traffic, uploaded data, standard deviation of uploaded data, maximum time interval, mean packet size, standard deviation of packet size, average traffic per second, mean signal strength, and variance of signal strength. Data that obviously does not meet the threshold conditions is classified as other devices, and data that meets the threshold conditions is a camera device with camera features.

[0040] The threshold condition is a condition that describes the basic requirements of the camera data. The condition can be determined based on the general characteristics of the camera signal, historical data, and experimental verification. Data that does not meet the threshold condition is definitely not camera data, and can be judged as data of other devices. In this case, different processing paths are performed to classify it as other devices for subsequent distinction and management. If the threshold condition is met (judged by the feature library of the camera signal), it is data of a device with camera characteristics, and further judgment is made through the following step S5.

[0041] Step S5, model calling and discrimination: the model calling and discrimination uses a signal analysis model to determine whether it is a camera.

[0042] This step realizes signal analysis and judgment through statistics, data processing functions, probability theory methods, linear algebra operation feature engineering, threshold judgment logic, result analysis, etc. The specific process of this step is as follows:

[0043] S51. Create a training data set and collect raw data from camera and non-camera devices to ensure data diversity and representativeness.

[0044] S52. Initialize StandardScaler: Use StandardScaler to standardize the features, eliminate the dimensional differences between different features, and improve the stability and convergence speed of the model.

[0045] S53. Configure random forest classifier parameters: Set the parameters of the random forest classifier, such as the number of trees, maximum depth, etc., to optimize model performance.

[0046] S54, Feature Identification: Identify data according to basic flow characteristics, time characteristics, size characteristics, and comprehensive characteristics to extract key information.

[0047] S55. Use StandardScaler to standardize the data: Apply StandardScaler again to standardize the extracted features.

[0048] S56. Create a SMOTE-RandomForest pipeline: Combine normalization, SMOTE oversampling, and random forest classifiers to form a complete data processing and classification pipeline.

[0049] S57. Data segmentation and model training: Divide the dataset into training set and test set, use the training set to train the model, and verify the model performance on the test set.

[0050] S58. Evaluation indicator setting: Set evaluation indicators, such as accuracy, recall, F1 score, etc., to comprehensively evaluate model performance.

[0051] S59, Logistic regression model prediction: Use the logistic regression model to make the final prediction and convert the linear combination into a probability value through the sigmoid function.

[0052] S510, probability threshold determination: when the probability is greater than 0.5, it is determined to be a camera; between 0.3 and 0.5, it is determined to be a suspected camera; and less than 0.3, it is determined to be a non-camera.

[0053] This sophisticated classification method helps users better understand the devices in their surroundings and take targeted measures to protect personal privacy and security.

[0054] Here, the feature identification of basic traffic features includes identifying the uplink traffic ratio, relative stability of traffic size, and periodicity of broadband use; the feature identification of the time feature includes time efficiency, time variability, and time skewness features; the feature identification of the size feature includes packet size stability, packet size variability, and burst traffic features; the feature identification of the comprehensive feature includes time-size correlation features and traffic density features. Correspondingly, for camera data, in the traffic feature judgment, cameras usually have a higher uplink traffic ratio, relatively stable traffic size, and periodic bandwidth use; in the time feature judgment, cameras usually have higher time efficiency (continuous transmission), lower time variability, and time skewness features reflect the transmission mode; in the size feature judgment, cameras usually have high packet size stability, size variability within a specific range, and obvious burst traffic features; in the comprehensive feature judgment, time-size correlation meets the camera feature, and traffic density meets the camera feature.

[0055] Therefore, through the training and use of the model, it is possible to determine whether the device is a camera, a suspected camera, or not a camera.

[0056] Step S6, label identification and classification: For the identified devices, including cameras, suspected cameras and other devices, identify the device MAC brand label according to the matching record mark of the MAC library in step S3.

[0057] The MAC library is a database that stores a large number of device MAC addresses and related information (such as brand, model, etc.). Each device has a unique MAC address, just like the device's ID card. By collecting and organizing the MAC address information of camera devices of different brands and models and other types of devices, a MAC library is constructed so that relevant matching information can be quickly obtained from it after the device MAC address is identified.

[0058] Since the above steps have determined the category of the device (camera, suspected camera, and non-camera device), it is then necessary to match the MAC address of the identified device with the MAC library. Specifically, use the MAC address of the device as a keyword to search and query in the MAC library to see if there is a matching record. If a matching record is found, it means that the MAC address of the device has been filed in the library, and its corresponding brand, model and other detailed information can be obtained. If no matching record is found, it may be a new device or the MAC library has not yet included the information of the device. In this case, collect the relevant information of the device and add it to the MAC library for subsequent analysis.

[0059] When a matching record is found in the MAC database, the device is labeled with its name based on the brand, model, and other information. For example, if the matched device is a camera of a certain brand, it is labeled as a camera of that brand and model; if it is another type of device, it is labeled as its corresponding brand and model name.

[0060] Based on the brand name and other information, the brand feature display classification can be further realized. Grouping devices of the same brand together and analyzing the commonalities and differences in data indicators of devices of different brands will help to gain a deeper understanding of the performance characteristics of various brands of camera devices. At the same time, in subsequent analysis, the devices can be more accurately classified and judged based on brand features, thereby improving the accuracy and reliability of the entire classification process.

[0061] Step S7, device list display: Generate a device list and display the classified device information to the user.

[0062] A device list is generated based on the aforementioned identification and classification results, and the device information is displayed to the user. The device list includes detailed information such as device name, brand, signal strength, device type (whether it is a camera), etc., so that the user can quickly understand the device status in the surrounding environment.

[0063] Step S8, positioning search: select a device from the device list and specify a channel for the device. The Bluetooth probe device obtains single-channel data in real time. According to the strength of the data signal, the Bluetooth probe device is moved in the environment. The signal strength of the Bluetooth probe device is displayed in real time on the interface of the mobile terminal. The stronger the signal, the closer it is to the target device, so as to assist the user in locating and finding the target device.

[0064] Users can select the target device to be located on the mini program, such as a camera, specify the channel of the device, and obtain single-channel data in real time through channel data scanning. The mini program receives the data and can confirm the strength of the data signal based on the acquired data signal strength, data flow, signal variance, etc. The mini program will display the signal strength of the selected positioning device in real time. Users move the Bluetooth probe device in the surrounding environment according to the signal strength. The stronger the signal, the closer it is to the positioning device, and the approximate location of the target device can be located, helping users detect illegal cameras so that they can take corresponding measures in time to protect personal privacy and security.

[0065] Step S9, signal light strip control: the mobile terminal parses the single-channel data and displays the data in real time, and calculates the signal strength value at the same time, and sends the signal strength value to the Bluetooth probe device. The Bluetooth probe device controls the signal strength signal light accordingly according to the value, and intuitively displays the signal strength of the device.

[0066] This step is the priority method. The Bluetooth probe device turns on the device positioning function. The applet performs a single-channel scan and analysis of the located target device, and displays the signal strength of the data in real time. At the same time, the applet calculates the signal strength value and sends the light control command to the Bluetooth probe device. The Bluetooth probe device controls the lighting and extinguishing of the device signal light according to the light control threshold, and can also control the signal light strength of the Bluetooth probe device according to the strength value. Through the signal light control, the signal strength of different devices is intuitively displayed to the user. This step uses the signal light control to intuitively display the signal strength of different devices, allowing users to quickly understand the signal status of the device. At the same time, signal strength channel selection is performed to optimize the data collection and analysis process and improve the accuracy and efficiency of detection.

[0067] The present invention provides a technical solution that can identify illegal video recording based on wifi traffic analysis, effectively detect WIFI traffic devices within a 10-meter space range, including camera devices, other network devices, etc., to solve the problem that hidden cameras within a specific range are difficult to be discovered. The camera Bluetooth probe device must have the function of automatically scanning the wifi channel after powering on, combining factors such as uplink and downlink traffic, uploaded data, uploaded data standard deviation, maximum time interval, packet size mean, packet size standard deviation, average traffic per second, signal strength mean, signal strength variance, etc., through real-time data capture and signal analysis model deep algorithm analysis, the traffic usage device in the environment can be identified within 120 seconds. The camera Bluetooth probe device interacts with the mobile phone applet through Bluetooth, and provides the analysis results to the user so that the user can promptly understand the privacy and security situation around him.

[0068] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A camera detection method based on wireless signals, characterized in that: The method comprises the following steps: Step S1, establishing a Bluetooth connection between the mobile terminal and the Bluetooth probe device; Step S2: The Bluetooth probe device collects WiFi traffic data packets in the surrounding environment and sends the collected data packets to the mobile terminal; Step S3: The mobile terminal performs data analysis on the data packet, determines the validity of the analysis data result, performs label matching on the valid data through the MAC library of the server, and saves the matching record; Step S4: for valid data, determine whether the data meets the threshold condition. If the threshold condition is met, it is a device with camera features, and a model call discrimination process is performed; If the threshold conditions are not met, it is classified as other devices; Step S5, model calling and judging using a signal analysis model to judge whether it is a camera; Step S6: For the identified devices, including cameras, suspected cameras and other devices, mark the MAC brand label of the identified device according to the matching record of step S3; Step S7: Generate a device list and display the classified device information to the user.

2. The camera detection method based on wireless signals as claimed in claim 1, characterized in that: The method further comprises the steps of: Step S8, select a device from the device list and specify a channel for the device. The Bluetooth probe device obtains single-channel data in real time. According to the strength of the data signal, the Bluetooth probe device is moved in the environment, and the signal strength of the Bluetooth probe device is displayed in real time on the interface of the mobile terminal. The stronger the signal, the closer it is to the target device, so as to assist the user in locating and finding the target device.

3. The camera detection method based on wireless signals as claimed in claim 2, characterized in that: The method further comprises the steps of: Step S9, the mobile terminal parses the single-channel data and displays the data in real time, and calculates the signal strength value at the same time, and sends the signal strength value to the Bluetooth probe device. The Bluetooth probe device controls the corresponding signal strength signal light according to the value, and intuitively displays the signal strength of the device.

4. The camera detection method based on wireless signals as claimed in claim 1, characterized in that: The specific process of step S5 is as follows: Create a training data set, initialize StandardScaler, configure random forest classifier parameters, identify the data according to basic traffic characteristics, time characteristics, size characteristics, and comprehensive characteristics, use StandardScaler to standardize the data, create a SMOTE-RandomForest pipeline, combine standardization, SMOTE oversampling, and random forest classifiers, perform data segmentation and model training, evaluate the indicators, set evaluation indicators, use the logistic regression model for prediction, and calculate the probability value. When the probability is greater than 0.5, it is determined to be a camera, between 0.3 and 0.5, it is determined to be a suspected camera, and less than 0.3, it is determined to be a non-camera.

5. The camera detection method based on wireless signals as claimed in claim 4, characterized in that: The feature identification of the basic traffic characteristics includes identifying the uplink traffic ratio, relative stability of traffic size, and periodicity of broadband usage; the feature identification of the time characteristics includes time efficiency, time variability, and time skewness characteristics; the feature identification of the size characteristics includes packet size stability, packet size variability, and burst traffic characteristics; the feature identification of the comprehensive characteristics includes time-size correlation characteristics and traffic density characteristics.

Images