Intelligent security early warning method based on big data

Abstract

The invention relates to the technical field of intelligent security early warning based on big data, and discloses an intelligent security early warning method based on big data. The door and window state is accurately identified and timely uploaded by marking and numbering each door and window sensor and acquiring data in sequence, so that misjudgment caused by disordered data sequence is avoided, and the processing efficiency of the system is improved. The infrared motion sensor manages each monitoring area through numbers, effectively tracks family activities and timely discovers abnormal behaviors, so that the security early warning capability is enhanced. The temperature and humidity sensors are arranged in the living room, the bedroom, the kitchen, the toilet and other areas and numbered, independent monitoring and consistent storage of data in each area are ensured, data loss and wrong matching are avoided, and the accuracy of data processing is improved. A unified storage and preprocessing mechanism solves the problems of data redundancy and non-uniform sources, and the overall response efficiency of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent security warning based on big data, and specifically provides an intelligent security warning method based on big data. Background Art

[0002] With the popularization and development of smart home systems, security systems play a crucial role in ensuring home security. Smart home systems achieve real-time monitoring of the home environment through the cooperation of multiple sensors, including temperature and humidity sensors, infrared sensors, door and window sensors, etc. As a part of the intelligent security system, the humidity sensor's main function is to monitor changes in indoor humidity and help identify potential safety hazards such as fires and water leaks. However, in actual applications, humidity sensors are easily affected by factors such as environmental noise, equipment failures, and transmission delays, resulting in false alarms and missed alarms from time to time.

[0003] In the prior art, sensors often have false alarms and missed alarms. The root cause of these problems is that traditional alarm mechanisms mainly rely on fixed thresholds, which do not take into account the dynamic changes in the environment and the stability of the sensors themselves. Therefore, sensors may trigger unnecessary alarms in some cases, or fail to trigger when an alarm is actually needed. Most existing technologies rely on fixed thresholds to trigger alarms. These thresholds are usually set based on historical data or manufacturer recommendations and are difficult to adjust in real time once set. This fixed threshold setting method ignores the influence of environmental factors and fluctuations in sensor performance, resulting in the system being unable to dynamically adjust alarm conditions according to the actual situation, further exacerbating the frequency of false alarms and missed alarms.

[0004] To solve the above problems, this case aims to propose an intelligent security warning method based on big data, design an optimized method for dynamically adjusting thresholds, and combine user feedback and environmental changes to improve the accuracy of the alarm system. The core of this method is to calculate influence factors based on sensor data and real-time user feedback, thereby adjusting the thresholds to achieve adaptive alarms. In this way, the system can continuously optimize the threshold settings, reduce false alarms and missed alarms caused by fixed thresholds, and enhance the accuracy and response ability of the intelligent security system. Summary of the Invention

[0005] The present invention provides an intelligent security warning method based on big data, which helps to solve the problems mentioned in the above background art.

[0006] The present invention provides the following technical solution: An intelligent security warning method based on big data, including:

[0007] Install door and window sensors at the entrances and exits and by the windows of the home, and label each door and window sensor with a number;

[0008] Obtain the opening and closing states of the doors and windows, and record the data of the i-th door and window sensor obtained at time t as S w,i (t);

[0009] Among them, when S w,i (t) = 0, it means that the i-th door and window is closed;

[0010] When S w,i (t) = 1, it means that the i-th door and window is open;

[0011] Set the dataset of the opening and closing states of the doors and windows obtained at time t, denoted as S w (t);

[0012] Add all the opening and closing states of the doors and windows obtained at time t to S w (t) in ascending order according to the corresponding door and window sensor numbers, and obtain where n w is the total number of door and window sensors;

[0013] Set the data acquisition interval of the opening and closing states of the doors and windows as Δt;

[0014] Collect the opening and closing state data of the doors and windows every Δt and upload it to the server;

[0015] Set up infrared motion sensors at the room entrance and the aisle, and label each infrared motion sensor with a number;

[0016] Obtain the areas monitored by each infrared motion sensor, and label the monitored areas with the numbers of the corresponding infrared motion sensors;

[0017] Record the data of the j-th infrared motion sensor obtained at time t as M j (t);

[0018] Among them, when M j (t) = 0, it means that there is no one in area j;

[0019] When M j (t) = 1, it means that there is someone in area j;

[0020] Set the dataset of the infrared motion sensors obtained at time t, denoted as M(t);

[0021] Add all the data of the infrared motion sensors obtained at time t to M(t) in ascending order according to the corresponding infrared motion sensor numbers;

[0022] Obtain where n m is the total number of infrared motion sensors;

[0023] Obtain the central positions of the living room, bedroom, kitchen, and bathroom respectively, set temperature and humidity sensors at each central position, and label each temperature and humidity sensor with a number;

[0024] Among them, the numbers of the temperature and humidity sensors at the same position are the same, and the total numbers of the temperature sensors and humidity sensors are equal;

[0025] Record the temperature data of the k-th temperature sensor obtained at time t as T k (t), and record the humidity data of the k-th humidity sensor obtained at time t as H k (t);

[0026] Set the temperature data dataset of the temperature sensors obtained at time t, denoted as T(t);

[0027] Add the temperature data of all the temperature and humidity sensors obtained at time t to T(t) in ascending order of the numbers, and obtain where n T is the number of temperature sensors;

[0028] Set the humidity data dataset of the humidity sensors obtained at time t, denoted as H(t);

[0029] Add the humidity data of all the humidity sensors obtained at time t to H(t) in ascending order of the numbers, and obtain where n T is the number of humidity sensors;

[0030] Store and preprocess all the obtained data.

[0031] Optionally, the storing and preprocessing all the obtained data specifically includes:

[0032] S21. Classify and store the data:

[0033] Store the door and window switch state data S w (t) and the infrared sensor data M(t) as a time-series boolean dataset;

[0034] Store the temperature and humidity sensor data T(t) and H(t) as a time-series floating-point dataset;

[0035] S22. Preprocess the data:

[0036] Set the size of the sliding window, denoted as n h ;

[0037] Smooth the door and window sensor data by using the sliding window method:

[0038]

[0039] Among them, is the data of the i-th door and window sensor obtained at time t after smoothing;

[0040] The infrared sensor data is smoothed by the sliding window method:

[0041]

[0042] Among them, is the data of the j-th infrared sensor obtained at time t after smoothing.

[0043] Optionally, the storage and preprocessing of all the obtained data specifically include:

[0044] Obtain the number of domain data points for interpolation calculation, denoted as Q;

[0045] If the temperature data of the k-th temperature sensor is missing at time t, use interpolation to fill in the missing value for the temperature data:

[0046]

[0047] Among them, is the temperature data of the k-th temperature sensor obtained at time t after filling in the missing value; T k (t i ) is the temperature value collected by the k-th temperature sensor at time t i ;

[0048] If the temperature data of the k-th humidity sensor is missing at time t, use interpolation to fill in the missing value for the humidity data:

[0049]

[0050] Among them, is the humidity data of the k-th humidity sensor obtained at time t after filling in the missing value; H k (t i ) is the humidity value collected by the k-th humidity sensor at time t i ; is the weight corresponding to time t i ;

[0051]

[0052] Among them, |t - t i | is the current time t and the known data time t iThe time difference between; σ is a hyperparameter that controls the domain range and is used to adjust the influence degree of the most recent moment data on the interpolation result; exp() is an exponential function used to generate a weight decay function that increases with the time difference;

[0053] Intelligent analysis and anomaly detection are performed based on the collected data.

[0054] Optionally, the intelligent analysis and anomaly detection based on the collected data specifically include:

[0055] S31. Door and window anomaly detection:

[0056] Set the time threshold set for the door and window to open, denoted as S threshold ;

[0057] If Then the i-th door and window at time t is marked as an abnormal state;

[0058] Among them, ∧ is the AND operation;

[0059] S32. Infrared activity anomaly detection:

[0060] Calculate the time series activity density:

[0061]

[0062] Set the j-th activity density threshold, denoted as M j,threshold ;

[0063] If D j (t)>M j,threshold , then the area j at time t is marked as an abnormal state;

[0064] S33. Temperature and humidity anomaly analysis:

[0065] Set the temperature threshold range of the k-th temperature sensor, denoted as [T k,min ,T k,max , where T k,min is the minimum temperature of the k-th temperature sensor, and T k,max is the maximum temperature of the k-th temperature sensor;

[0066] Set the humidity threshold range of the k-th humidity sensor, denoted as [H k,min ,H k,max , where H k,min is the minimum humidity of the k-th humidity sensor, and H k,max is the maximum humidity of the k-th humidity sensor;

[0067] If Then the position temperature obtained by the k-th temperature sensor at time t is marked as abnormal;

[0068] If then record the position humidity obtained by the k-th humidity sensor at time t as abnormal;

[0069] Generate real-time warnings and user notifications based on the results of intelligent analysis and anomaly detection.

[0070] Optionally, the generating of real-time warnings and user notifications according to the results of intelligent analysis and anomaly detection specifically includes:

[0071] Generate an alarm status based on the analysis results of doors and windows, infrared, temperature and humidity, and video:

[0072]

[0073] where V is the OR operation;

[0074] When A(t) = 1, obtain the position and type of the sensors with a status of 1, and send the obtained position and type of the sensors to the user;

[0075] Perform security event recording and feedback optimization.

[0076] Optionally, the performing of security event recording and feedback optimization specifically includes:

[0077] Set the set of unique identifiers of the sensors that trigger the alarm, denoted as SensorID;

[0078] Obtain the unique identifiers of all the sensors that trigger the alarm at time t, and add them to SensorID;

[0079] Obtain the timestamp at the time of the alarm, denoted as Timestamp;

[0080] Record the alarm as Log(t) = {A(t), SensorID, Timestamp};

[0081] Collect the feedback of each alarm from the user and update the threshold:

[0082] Obtain the number of false alarms of the j-th infrared sensor in the user feedback, denoted as

[0083] Obtain the number of valid alarms of the j-th infrared sensor in the user feedback, denoted as Obtain the average value of the measured activity density of the j-th infrared sensor over the historical time, denoted as Calculate the influence factor of the user feedback data of the j-th infrared sensor on the threshold, denoted as

[0084] where M'j,threshold is the optimized j-th activity density threshold; γ1 and γ2 are adjustment factors.

[0085] Optionally, the optimization of safety event recording and feedback specifically includes:

[0086] Obtain the number of false alarms of the minimum temperature of the k-th temperature sensor feedback by the user, denoted as

[0087] Obtain the number of valid alarms of the minimum temperature of the k-th temperature sensor feedback by the user, denoted as

[0088] Obtain the number of false alarms of the maximum temperature of the k-th temperature sensor feedback by the user, denoted as

[0089] Obtain the number of valid alarms of the maximum temperature of the k-th temperature sensor feedback by the user, denoted as

[0090] Obtain the average value of the measured temperature data of the k-th temperature sensor within the historical time, denoted as

[0091] Calculate the influence factor of the data feedback by the k temperature sensor users on the minimum temperature, denoted as

[0092]

[0093] Calculate the influence factor of the data feedback by the k temperature sensor users on the maximum temperature, denoted as

[0094]

[0095] where T′ k,min is the optimized minimum temperature of the k-th temperature sensor; T′ k,max is the optimized maximum temperature of the k-th temperature sensor; γ3 and γ4 are adjustment factors.

[0096] Optionally, the optimization of safety event recording and feedback specifically includes:

[0097] Obtain the number of false alarms of the minimum humidity of the k-th humidity sensor feedback by the user, denoted as

[0098] Obtain the number of valid alarms of the minimum humidity of the γ-th humidity sensor feedback by the user, denoted as

[0099] Obtain the number of false alarms of the maximum humidity of the k-th humidity sensor feedback by the user, denoted as

[0100] Obtain the effective alarm quantity of the maximum humidity value of the k-th humidity sensor for user feedback, denoted as

[0101] Obtain the average value of the measured humidity data of the k-th humidity sensor within the historical time, denoted as

[0102] Calculate the influence factor of the data of user feedback of k humidity sensors on the minimum humidity value, denoted as

[0103]

[0104] Calculate the influence factor of the data of user feedback of k humidity sensors on the maximum humidity value, denoted as

[0105]

[0106] Among them, H′ k,min is the minimum temperature value of the k-th humidity sensor after optimization; H′ k,max is the maximum temperature value of the k-th humidity sensor after optimization; γ5 and γ6 are adjustment factors.

[0107] The present invention has the following beneficial effects:

[0108] 1. By labeling each door and window sensor with a number and collecting and uploading data in the order of the numbers, precise identification and status tracking of each door and window can be achieved. When collecting data on the open / closed status of doors and windows, ensure that the data for each door and window can be recorded and uploaded in a timely and accurate manner. By setting up a dataset for the open / closed status of doors and windows and ensuring that the data is added to the dataset in the order of the door and window sensor numbers, the unity and easy processing of the data are ensured. This not only improves the efficiency of data processing but also avoids system misjudgment problems caused by chaotic data order, ensuring that each collection and upload of the open / closed status can be accurately matched with the corresponding sensor. By setting up infrared motion sensors at the entrance and corridor and labeling each sensor with a number and managing the monitoring areas by numbering, the activities in each area can be effectively identified. The data of each infrared motion sensor is recorded and uploaded in the order of the numbers, ensuring that the system can track the activities in the home in real-time and accurately, promptly detect abnormal behaviors, and enhance the security warning ability. By setting up temperature and humidity sensors in different areas such as the living room, bedroom, kitchen, and bathroom and numbering each sensor, the environmental data for each area can be individually identified and monitored. The sequential storage of the dataset ensures the consistency and accuracy of the data in subsequent processing. By collecting and monitoring temperature and humidity data, abnormal environmental changes such as drastic fluctuations in temperature and humidity can be promptly detected, and corresponding security responses can be made. Through a unified storage and preprocessing mechanism, centralized management of all the sensor data obtained is carried out to ensure the integrity and accuracy of the data. This processing method solves problems such as inconsistent multi-sensor data sources and data redundancy, and improves the processing ability and response efficiency of the system.

[0109] 2. By storing the door and window switch status data and infrared sensor data as a time-series boolean data set, and storing the temperature and humidity sensor data as a time-series floating-point data set, different types of data can be effectively distinguished and specialized processing can be carried out. The time-series boolean data set helps to quickly determine whether the doors and windows are in the open or closed state, while the time-series floating-point data set can efficiently store and process the fine changes in temperature and humidity. Through this classification storage method, the situation of different types of data being mixed together is avoided, improving the manageability and processing efficiency of the data, and effectively reducing data misreading and processing errors. By applying the sliding window smoothing method to the door and window sensor and infrared sensor data, the outliers generated by signal jitter or temporary interference can be effectively removed. The sliding window method can smooth the data fluctuations and remove unnecessary noise, making the system's response to the changes in the door and window status and infrared motion detection more accurate, and avoiding false alarms and missed alarms. This processing scheme solves the problem of sensor data fluctuations in a dynamic environment, improving the data quality and system reliability. The smoothing process of the key data such as door and window sensors and infrared sensors can better capture the real state changes and avoid misjudgments caused by unstable signals within a short period of time. The data of the temperature and humidity sensors is also smoothed through this processing method, ensuring the stable operation of the system under different environmental conditions. This precise data smoothing process can reduce the interference brought by environmental changes to the system and improve the overall monitoring and response capabilities. By classifying and storing the door and window switch status, infrared sensor data, temperature and humidity data, etc. in different formats and performing smoothing processing, the system can read and process data more quickly in real time. At the same time, the improvement of data quality enables subsequent intelligent analysis to more accurately identify potential security risks, improving the response speed and accuracy of the security system, and timely detecting potential threats and making responses.

[0110] 3. By interpolating and filling the missing data of temperature and humidity sensors, the system can maintain data continuity even when sensor data is missing. This interpolation process ensures that the missing data caused by sensor failures or environmental disturbances no longer affects the overall data analysis, guaranteeing the stability and integrity of the system during real-time data acquisition. The use of interpolation methods can fill the data gaps caused by sensor failures, preventing problems such as incorrect judgments or ineffective analysis due to data loss. When temperature and humidity data are missing, using interpolation methods not only solves the data missing problem but also improves the data accuracy. Interpolation methods fill in the missing data by combining known data points and their weights, making the filled data closer to the true values and thus enhancing the accuracy of subsequent intelligent analysis. For the temperature and humidity monitoring system, accurate data is crucial for anomaly detection and early warning. After interpolation filling, the data quality is significantly improved, enhancing the system's sensitivity and adaptability to environmental changes. In time-series data, missing data from temperature and humidity sensors may affect the temporal continuity of the data, thereby influencing the effectiveness of subsequent analysis. Through interpolation methods, especially those based on time differences and weights, the filling process of missing data can maintain the temporality of the data, ensuring data continuity and consistency. This processing method avoids system misjudgments caused by data interruptions or discontinuities, enhancing the system's ability to process time-series data and ensuring the effective execution of subsequent analysis and prediction tasks. Intelligent analysis and anomaly detection rely on high-quality continuous data, and the data after filling missing values is more conducive to subsequent analysis. Using interpolation methods can avoid calculation biases caused by missing data, ensuring that the data is not affected during anomaly detection. Based on more complete data, the system can accurately identify abnormal patterns, such as excessively high environmental temperature or too low humidity, and issue timely warnings, thereby enhancing the system's security and early warning capabilities. Interpolation methods have an adaptive ability through weight adjustment and the influence of time differences. As the time difference increases, the interpolation method can automatically adjust the weights to ensure that data closer to the current moment has a priority impact on the interpolation result. Through this adaptive interpolation method, the system can more flexibly respond to different environmental changes, further optimizing the data processing effect. Especially for missing data with large time differences, it can flexibly adjust the interpolation result according to the actual situation, improving the system's robustness and accuracy.

[0111] 4. By setting a time threshold for the opening of doors and windows, the system can monitor the opening and closing status of doors and windows in real time and detect any abnormalities. For example, when a door or window remains open for a long time, the system will determine it as an abnormal state, thus promptly alerting the user or triggering an alarm. This process effectively prevents potential safety hazards caused by forgetting to close doors and windows or malfunctions, ensuring home security and reducing the occurrence of theft and other accidents caused by abnormal doors and windows. By calculating the activity density of the time series and comparing it with a preset threshold, the system can effectively detect the human activity status in the area. When the activity density exceeds the set threshold, the system will determine it as abnormal activity. This function can detect the intrusion behavior of personnel in real time, especially efficiently capturing abnormal activities in unmanned monitored areas. By promptly detecting abnormal activities in the area, the system can automatically trigger an alarm, enhancing the security of the home or facility. Through real-time analysis of the data from temperature and humidity sensors, the system can set a threshold range for temperature and humidity. Once the temperature or humidity exceeds the set range, it will be immediately marked as abnormal. This function can effectively monitor environmental changes, ensure that the temperature and humidity are maintained within a comfortable range, and prevent damage to household appliances or buildings caused by extreme environmental conditions. For example, when the indoor temperature is too high or the humidity is too large, the system can notify the user to adjust the equipment through early warning, reducing the risk of overheating or moisture absorption of the equipment. By combining multi-dimensional anomaly detection of doors and windows, infrared motion, and temperature and humidity data, the system can comprehensively analyze the feedback information of different sensors. When the data of a certain sensor shows an abnormality, the system can promptly conduct cross-verification and linkage analysis to further improve the accuracy of anomaly detection. For example, the linkage between the anomaly detection of temperature and humidity and infrared motion can help the system more accurately determine whether the anomaly is caused by external weather changes or abnormal human activities. Through this multi-dimensional combination method, the system can reduce false alarms and enhance the overall reliability of monitoring. Based on the results of intelligent analysis and anomaly detection, the system can automatically generate real-time warnings and notify the user to ensure that the user can receive a warning in the first time when an abnormal situation occurs. This timely feedback mechanism can help the user quickly make response decisions and avoid greater risks caused by delayed handling. The intelligence of the warning and notification system improves the response speed of home and environmental management and enhances the ability to respond to emergencies.

[0112] 5. By comprehensively considering the results of door / window sensors, infrared sensors, temperature / humidity sensors, and video analysis and generating an alarm status through an "OR" operation, the system can achieve more accurate alarm judgment. A single sensor may have false alarms or missed alarms. However, through the collaborative work of multiple sensors, false alarms can be effectively eliminated, and the accuracy of alarms can be improved. For example, the system will only generate an alarm when the doors and windows are not closed and the infrared sensor detects human activities, reducing the error of triggering an alarm based solely on the abnormality of a single sensor. When the system alarms, it can obtain relevant information about the sensor location and type and send this information to the user in a timely manner. In this way, the user not only receives an alarm notification but also can accurately know the specific abnormal location and sensor type. This function effectively reduces the time for the user to find the source of the abnormality and improves the processing efficiency. For example, when the door / window sensor detects an abnormality, the system can immediately inform the user which specific door or window is not closed, thus saving the user's investigation time and increasing the response speed. By recording each alarm event and optimizing the feedback mechanism, the system can achieve post-event tracking and analysis of security events. This process not only helps the user understand the specific reasons for the alarm but also provides data support for further optimizing the alarm rules and algorithms. Over time, the system gradually adjusts the alarm threshold and response strategy through the analysis of historical data, thereby improving the accuracy of overall security monitoring and user satisfaction. Through timely alarms and accurate location feedback, users can more intuitively perceive the security status of their homes or workplaces. This real-time feedback mechanism enhances the user's sense of security, increases the user's trust and dependence on the system, and enables a quick response in the event of an emergency, reducing potential security risks.

[0113] 6. By setting and collecting the set of unique identifiers of the sensors that trigger alarms, the system can accurately identify and record the specific sensors in each alarm event. This measure avoids the problems of duplicate or missing alarm information caused by the collaborative work of multiple sensors, ensuring the integrity and traceability of alarm data. When the system issues an alarm, it can accurately feedback which sensor triggered the alarm, improving the accuracy of alarm information. By collecting the feedback of each alarm from the user and combining the number of false alarms and valid alarms feedback by the user, the system can dynamically adjust the alarm threshold. For infrared sensors, by combining the average activity density within the historical time, the system can accurately identify the alarm behavior of the sensors and adjust the activity density threshold according to the user feedback. This mechanism can effectively reduce the false alarm frequency, ensure the accuracy of the alarm system after long-term use, and optimize the alarm strategy of the system, thus reducing unnecessary alarms. By calculating the influence factor of the user feedback data of each infrared sensor on the threshold, the system can reasonably adjust the threshold. The adjustment factor is used to control the influence degree of historical data and user feedback on the threshold adjustment, so as to accurately optimize the alarm trigger threshold in actual operation. This approach avoids the threshold adjustment deviation caused by solely relying on historical data or user feedback, making the alarm threshold more flexible and intelligent, and capable of adapting to environmental changes in real time. The system can adjust the alarm threshold according to the user feedback, solving the defect that traditional alarm systems fail to adapt to environmental changes or user needs in a timely manner. By gradually optimizing the alarm rules of the system, the system can continuously improve the accuracy of alarms and enhance its adaptability during long-term operation. This not only reduces false alarms but also enables more efficient alarm responses under different environments and conditions.

[0114] 7. By collecting the number of false alarms and valid alarms for the minimum and maximum temperatures feedback by users, the system can analyze and judge the alarm accuracy of each sensor within a specific temperature range. The number of false alarms feedback by users can provide an accuracy index for the system's alarms. By comparing these data, the system optimizes and adjusts the alarm thresholds of each temperature sensor, avoiding the problem of reduced system trust caused by frequent false alarms. The system obtains the average value of the measured temperature data of each temperature sensor over historical time, thereby being able to understand the typical working range and normal data fluctuations of the sensor. With the help of these data, the system can better identify normal and abnormal temperature fluctuations, avoid false alarms caused by instantaneous changes in the external environment, and enhance the stability and reliability of the alarm system. By calculating the influence factors of the user feedback data of each temperature sensor on the minimum and maximum temperatures, the system can dynamically adjust the temperature thresholds to reflect the users' feedback on alarm accuracy. The introduction of the adjustment factor enables the system to optimize the threshold adjustment not only based on historical data but also in combination with real-time user feedback. This mechanism improves the adaptability of the alarm threshold, enabling the system to flexibly adjust according to user needs and environmental changes, reducing the inadaptability problem caused by fixed thresholds. By optimizing and adjusting the minimum and maximum values of the temperature sensors, the system can more accurately judge abnormal temperature states, avoiding omissions or false alarms caused by overly loose or overly strict alarm ranges. The optimized thresholds can adapt to changes in different environmental conditions during daily operation and continuously adjust precisely according to users' usage feedback, significantly improving the accuracy and intelligent level of the alarm system.

[0115] 8. By collecting the number of false alarms and valid alarms of each humidity sensor within the minimum and maximum humidity ranges, the accuracy of sensor alarms can be accurately evaluated. Through user feedback, the system can determine which humidity ranges have more reliable alarms, thus avoiding the problem of distrust of the alarm system caused by excessive false alarms. The system can adjust the alarm threshold according to this feedback data to make it more in line with the actual environmental conditions. By obtaining the average value of the humidity data measured by the humidity sensor historically, the system can understand the normal fluctuation range of the humidity sensor. These historical data help to identify real humidity anomalies and short-term environmental fluctuations. Therefore, the system can better distinguish normal humidity changes from abnormal fluctuations, avoiding false alarms caused by temporary environmental changes such as wind speed or temperature changes, thereby improving the accuracy and stability of alarms. By analyzing user feedback on humidity sensors, calculating the influence factors of each humidity sensor's feedback on the minimum and maximum humidity values, and dynamically adjusting the humidity alarm threshold. Through this mechanism, the system can quickly respond to user feedback and timely adjust the humidity threshold, avoiding the problem that traditional static threshold settings cannot adapt to changing environments. Through flexible adjustment, the system realizes personalized optimization of the threshold, reducing the phenomena of false alarms and missed alarms. As the threshold optimization progresses, the humidity sensor can more accurately identify environmental changes and issue alarms in a timely manner. The optimized humidity sensor alarm threshold can be adjusted according to different users' feedback to achieve more accurate anomaly detection. In addition, the system can adaptively adjust the threshold according to the changing environment and device status, improving the intelligence and response ability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0116] Figure 1 It is a schematic flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0117] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0118] Embodiment, referring to Figure 1 , a big data-based intelligent security warning method, including:

[0119] Door and window sensors are respectively arranged at the entrances and exits and windows of the home, and each door and window sensor is marked with a number;

[0120] Obtain the opening and closing states of the doors and windows, and record the data of the i-th door and window sensor obtained at time t as S w,i (t);

[0121] Among them, when S w,i (t) = 0, it indicates that the i-th door or window is closed;

[0122] When S w,i (t) = 1, it indicates that the i-th door or window is open;

[0123] Set the dataset of door and window switch states obtained at time t, denoted as S w (t);

[0124] Add all the door and window switch states obtained at time t to S w (t) in ascending order according to the corresponding door and window sensor numbers, and obtain where n w is the total number of door and window sensors;

[0125] Set the data acquisition interval of door and window switch states as Δt;

[0126] Collect the door and window switch state data every Δt and upload it to the server;

[0127] Set infrared motion sensors at the room entrance and the aisle, and mark numbers for each infrared motion sensor;

[0128] Obtain the areas monitored by each infrared motion sensor, and mark the numbers of the corresponding infrared motion sensors for the monitored areas;

[0129] Record the data of the j-th infrared motion sensor obtained at time t as M j (t);

[0130] Among them, when M j (t) = 0, it indicates that there is no one in area j;

[0131] When M j (t) = 1, it indicates that there is someone in area j;

[0132] Set the dataset of infrared motion sensors obtained at time t, denoted as M(t);

[0133] Add all the infrared motion sensor data obtained at time t to M(t) in ascending order according to the corresponding infrared motion sensor numbers;

[0134] Obtain where n m is the total number of infrared motion sensors;

[0135] Respectively obtain the central positions of the living room, bedroom, kitchen, and bathroom, set temperature and humidity sensors at each central position, and mark numbers for each temperature and humidity sensor;

[0136] The temperature and humidity sensors at the same position have the same number, and the total numbers of temperature sensors and humidity sensors are equal;

[0137] Denote the temperature data of the k-th temperature sensor obtained at time t as T k (t), and denote the humidity data of the k-th humidity sensor obtained at time t as H k (t);

[0138] Set the temperature data dataset of the temperature sensors obtained at time t, denoted as T(t);

[0139] Add the temperature data of all the temperature and humidity sensors obtained at time t to T(t) in ascending order of the numbers, to obtain where n T is the number of temperature sensors;

[0140] Set the humidity data dataset of the humidity sensors obtained at time t, denoted as H(t);

[0141] Add the humidity data of all the humidity sensors obtained at time t to H(t) in ascending order of the numbers, to obtain where n T is the number of humidity sensors;

[0142] Store and preprocess all the obtained data.

[0143] By labeling each door and window sensor with a number and collecting and uploading data in the order of the numbers, accurate identification and status tracking of each door and window can be achieved. When collecting data on the opening and closing status of doors and windows, ensure that the data of each door and window can be recorded and uploaded in a timely and accurate manner. By setting up a dataset for the opening and closing status of doors and windows and ensuring that the data is added to the dataset in the order of the door and window sensor numbers, the unity and easy processing of the data are ensured. This not only improves the efficiency of data processing but also avoids system misjudgment problems caused by chaotic data order, ensuring that each collection and upload of the opening and closing status can be accurately matched with the corresponding sensor. By setting up infrared motion sensors at the entrances and walkways, labeling each sensor with a number, and managing the monitoring areas by numbering, the activities in each area can be effectively identified. The data of each infrared motion sensor is recorded and uploaded in the order of the numbers, ensuring that the system can track the activities in the home in real time and accurately, detect abnormal behaviors in a timely manner, and enhance the security warning ability. By setting up temperature and humidity sensors in different areas such as the living room, bedroom, kitchen, and bathroom, and numbering each sensor, the environmental data of each area can be individually identified and monitored. The sequential storage of the dataset ensures the consistency and accuracy of the data in subsequent processing. Through the collection and monitoring of temperature and humidity data, abnormal environmental changes such as drastic fluctuations in temperature and humidity can be detected in a timely manner, and corresponding security responses can be made. Through a unified storage and preprocessing mechanism, centralized management of all the sensor data obtained is carried out to ensure the integrity and accuracy of the data. This processing method solves problems such as inconsistent data sources and data redundancy of multiple sensors, and improves the processing ability and response efficiency of the system.

[0144] The storage and preprocessing of all the data obtained specifically include:

[0145] S21. Classify and store the data:

[0146] Store the door and window opening and closing status data S w (t) and the infrared sensor data M(t) as a time-series boolean dataset;

[0147] Store the temperature and humidity sensor data T(t) and H(t) as a time-series floating-point dataset;

[0148] S22. Preprocess the data:

[0149] Set the size of the sliding window, denoted as n h , which is the data range for smoothing processing;

[0150] For the door and window sensor data, use the sliding window method for smoothing to remove the outliers caused by signal jitter:

[0151]

[0152] Among them, is the data of the i-th door and window sensor obtained at time t after smoothing;

[0153] The infrared sensor data is smoothed using a sliding window method to remove outliers caused by signal jitter:

[0154]

[0155] Among them, is the data of the j-th infrared sensor obtained at time t after smoothing.

[0156] By storing the door and window switch state data and infrared sensor data as a time-series boolean data set, and storing the temperature and humidity sensor data as a time-series floating-point data set, different types of data can be effectively distinguished and specialized processing can be carried out. The time-series boolean data set helps to quickly determine whether the door and window is in the open or closed state, while the time-series floating-point data set can efficiently store and process the fine changes in temperature and humidity. Through this classification storage method, the situation of different types of data being mixed together is avoided, the manageability and processing efficiency of the data are improved, and data misreading and processing errors are effectively reduced. By applying the sliding window smoothing method to the door and window sensor and infrared sensor data, outliers caused by signal jitter or temporary interference can be effectively removed. The sliding window method can smooth data fluctuations and remove unnecessary noise, so that the system's response to door and window state changes and infrared motion detection is more accurate, avoiding false alarms and missed alarms. This processing solution solves the problem of sensor data fluctuations in a dynamic environment, improves data quality and system reliability. The smoothing process of key data such as door and window sensors and infrared sensors can better capture real state changes and avoid misjudgments caused by unstable signals within a short period of time. The data of temperature and humidity sensors is also smoothed through this processing method, ensuring the stable operation of the system under different environmental conditions. This precise data smoothing process can reduce the interference brought by environmental changes to the system and improve the overall monitoring and response capabilities. By classifying and storing the door and window switch state, infrared sensor data, temperature and humidity data, etc. in different formats and performing smoothing processing, the system can read and process data in real time more quickly. At the same time, the improvement of data quality enables subsequent intelligent analysis to more accurately identify potential security risks, improves the response speed and accuracy of the security system, and timely discovers potential threats and makes responses.

[0157] The storage and preprocessing of all the obtained data specifically include:

[0158] Obtain the number of domain data points for interpolation calculation, denoted as Q;

[0159] If the temperature data of the k-th temperature sensor is missing at time t, the interpolation method is used to fill in the missing value for the temperature data:

[0160]

[0161] Among them, is the temperature data of the k-th temperature sensor obtained at time t after filling in the missing value; T k (t i ) is the temperature value collected by the k-th temperature sensor at time t i ;

[0162] If the temperature data of the k-th humidity sensor is missing at time t, the interpolation method is used to fill in the missing value for the humidity data:

[0163]

[0164] Among them, is the humidity data of the k-th humidity sensor obtained at time t after filling in the missing value; H k (t i ) is the humidity value collected by the k-th humidity sensor at time t i ; is the weight corresponding to time t i ;

[0165]

[0166] Among them, |t - t i | is the time difference between the current time t and the known data time t i ; σ is a hyperparameter that controls the domain range and is used to adjust the influence degree of the nearest time data on the interpolation result; exp() is an exponential function used to generate a weight decay function that increases with the increase of the time difference;

[0167] Intelligent analysis and anomaly detection are carried out based on the collected data.

[0168] By interpolating the missing data of the temperature and humidity sensors, the system can maintain data continuity when sensor data is missing. This interpolation process prevents the missing data originally caused by sensor failure or environmental interference from affecting the overall data analysis, ensuring the stability and integrity of the system during real-time data collection. The use of interpolation can make up for the data gaps caused by sensor failure and avoid the problem of the system making wrong judgments or failing to conduct effective analysis due to data loss. The use of interpolation to fill in the missing temperature and humidity data not only solves the problem of missing data, but also improves the accuracy of the data. The interpolation method fills in the missing data by combining known data points and their weights, making the filled data closer to the true value, thereby improving the accuracy of subsequent intelligent analysis. For the temperature and humidity monitoring system, accurate data is crucial for anomaly detection and early warning. After interpolation filling, the quality of the data has been significantly improved, enhancing the sensitivity and adaptability of the system to environmental changes. In time series data, the missing data of the temperature and humidity sensors may affect the time series continuity of the data, thereby affecting the effect of subsequent analysis. Through interpolation, especially the interpolation method generated based on time difference and weight, the filling process of missing data can maintain the time series of data, thus ensuring the continuity and consistency of data. This processing method avoids system misjudgment caused by data interruption or discontinuity, improves the system's ability to process time series data, and ensures the effective execution of subsequent analysis and prediction tasks. Intelligent analysis and anomaly detection rely on high-quality continuous data, and the data after filling the missing values ​​is more conducive to subsequent analysis. The use of interpolation can avoid the calculation bias caused by missing data and ensure that the data is not affected when performing anomaly detection. Based on more complete data, the system can accurately identify abnormal patterns, such as excessively high ambient temperature or low humidity, and issue warnings in time, thereby improving the safety and warning capabilities of the system. The interpolation method makes the filling process adaptive through weight adjustment and the influence of time difference. As the time difference increases, the interpolation method can automatically adjust the weight to ensure that data closer to the current moment has priority in affecting the interpolation results. Through this adaptive interpolation method, the system can respond to different environmental changes more flexibly and further optimize the data processing effect. In particular, for missing data with large time differences, the interpolation results can be flexibly adjusted according to actual conditions, thereby improving the robustness and accuracy of the system.

[0169] The intelligent analysis and anomaly detection based on the collected data specifically includes:

[0170] S31, Door and Window Abnormal Detection:

[0171] Set the time threshold set for doors and windows to be opened, denoted as S threshold ;

[0172] like Then, the i-th door or window at time t is marked as an abnormal state;

[0173] Among them, ∧ is the AND operation;

[0174] S32. Infrared activity anomaly detection:

[0175] Calculate the activity density of the time series:

[0176]

[0177] Set the j-th activity density threshold, denoted as M j,threshold ;

[0178] If D j (t) > M j,threshold , then the area j at time t is marked as an abnormal state;

[0179] S33. Temperature and humidity anomaly analysis:

[0180] Set the temperature threshold range of the k-th temperature sensor, denoted as [T k,min , T k,max , where T k,min is the minimum temperature of the k-th temperature sensor, and T k,max is the maximum temperature of the k-th temperature sensor;

[0181] Set the humidity threshold range of the k-th humidity sensor, denoted as [H k,min , H k,max , where H k,min is the minimum humidity of the k-th humidity sensor, and H k,max is the maximum humidity of the k-th humidity sensor;

[0182] If then the position temperature obtained by the k-th temperature sensor at time t is marked as abnormal;

[0183] If then the position humidity obtained by the k-th humidity sensor at time t is marked as abnormal;

[0184] Generate real-time warnings and user notifications based on the results of intelligent analysis and anomaly detection.

[0185] By setting a time threshold for the opening and closing of doors and windows, the system can monitor the opening and closing status of doors and windows in real time and detect any abnormalities. For example, when a door or window remains open for an extended period, the system will determine it as an abnormal state, thus promptly alerting the user or triggering an alarm. This process effectively prevents potential safety hazards caused by forgetting to close doors and windows or malfunctions, ensuring home security and reducing the occurrence of theft and other accidents resulting from abnormal door and window conditions. By calculating the activity density of the time series and comparing it with a preset threshold, the system can effectively detect the human activity status within a region. When the activity density exceeds the set threshold, the system will determine it as abnormal activity. This function can detect intrusion behavior of personnel in real time, especially efficiently capturing abnormal activities in unmanned monitored areas. By promptly detecting abnormal activities within a region, the system can automatically trigger an alarm, enhancing the security of homes or facilities. Through real-time analysis of the data from temperature and humidity sensors, the system can set a threshold range for temperature and humidity. Once the temperature or humidity exceeds the set range, it will be immediately marked as abnormal. This function can effectively monitor environmental changes, ensure that the temperature and humidity are maintained within a comfortable range, and prevent damage to household appliances or buildings caused by extreme environmental conditions. For example, when the indoor temperature is too high or the humidity is too large, the system can notify the user to adjust the equipment through early warning, reducing the risk of overheating or moisture absorption of the equipment. By combining multi-dimensional anomaly detection of doors and windows, infrared motion, and temperature and humidity data, the system can comprehensively analyze the feedback information from different sensors. When the data of a certain sensor shows an abnormality, the system can promptly conduct cross-verification and linkage analysis to further improve the accuracy of anomaly detection. For example, the linkage between temperature and humidity anomalies and infrared motion anomaly detection can help the system more accurately determine whether the anomaly is caused by external weather changes or abnormal human activities. Through this multi-dimensional combination method, the system can reduce false alarms and enhance the overall reliability of monitoring. Based on the results of intelligent analysis and anomaly detection, the system can automatically generate real-time warnings and notify users, ensuring that users can receive alerts immediately when abnormal situations occur. This timely feedback mechanism can help users quickly make response decisions and avoid greater risks caused by delayed handling. The intelligence of the warning and notification system improves the response speed of home and environmental management and enhances the ability to respond to emergencies.

[0186] Generating real-time warnings and user notifications according to the results of intelligent analysis and anomaly detection specifically includes:

[0187] Generating an alarm status based on the analysis results of doors and windows, infrared, temperature and humidity, and video:

[0188]

[0189] Among them, V is the OR operation;

[0190] When A(t) = 1, obtain the position and type of the sensor when the status is 1, and send the obtained position and type of the sensor to the user;

[0191] Conduct security event recording and feedback optimization.

[0192] By comprehensively considering the results of door and window, infrared, temperature and humidity, and video analysis and generating an alarm status through "OR" operation, the system can achieve more accurate alarm judgment. A single sensor may have false alarms or missed alarms, while through the collaborative work of multiple sensors, false alarms can be effectively eliminated and the accuracy of alarms can be improved. For example, the system will generate an alarm only when the doors and windows are not closed and the infrared sensor detects human activities, reducing the error of triggering an alarm based on the abnormality of a single sensor. When the system alarms, it can obtain relevant information about the sensor position and type and send this information to the user in a timely manner. In this way, the user not only receives an alarm notification but also can accurately know the specific abnormal position and sensor type. This function effectively reduces the time for the user to find the source of the abnormality and improves the processing efficiency. For example, when the door and window sensor detects an abnormality, the system can immediately inform the user which specific door or window is not closed, thus saving the user's investigation time and improving the response speed. By recording each alarm event and optimizing the feedback mechanism, the system can achieve post-event tracking and analysis of security events. This process not only helps the user understand the specific reasons for the alarm but also provides data support for further optimizing the alarm rules and algorithms. Over time, the system gradually adjusts the alarm threshold and response strategy through the analysis of historical data, thereby improving the overall accuracy of security monitoring and user satisfaction. Through timely alarms and accurate position feedback, users can more intuitively perceive the security status of their homes or workplaces. This real-time feedback mechanism enhances the user's sense of security, increases the user's trust and dependence on the system, and especially enables a quick response in the event of an emergency, reducing potential security risks.

[0193] The conduct of security event recording and feedback optimization specifically includes:

[0194] Set a set of unique identifiers for the sensors that trigger alarms, denoted as SensorID;

[0195] Obtain the unique identifiers of all sensors that trigger alarms at time t and add them to SensorID;

[0196] Obtain the timestamp at the time of the alarm, denoted as Timestamp;

[0197] Record the alarm as Log(t) = {A(t), SensorID, Timestamp};

[0198] Collect the feedback of each user alarm and update the threshold:

[0199] Obtain the number of false alarms of the j-th infrared sensor in the user feedback, denoted as

[0200] Obtain the number of valid alarms of the j-th infrared sensor in the user feedback, denoted as

[0201] Obtain the average value of the measured activity density of the j-th infrared sensor within the historical time, denoted as

[0202] Calculate the influence factor of the user feedback data of the j-th infrared sensor on the threshold, denoted as

[0203] where M′ j,threshold is the optimized activity density threshold of the j-th; γ1 and γ2 are adjustment factors to control the influence degree of historical data and user feedback on the threshold adjustment.

[0204] By setting and collecting the set of unique identifiers of the sensors that trigger alarms, the system can accurately identify and record the specific sensors in each alarm event. This measure avoids the problems of duplicate or missing alarm information caused by the collaborative work of multiple sensors, ensuring the integrity and traceability of alarm data. When the system issues an alarm, it can accurately feedback which sensor triggered the alarm, improving the accuracy of alarm information. By collecting the feedback of each user alarm and combining the number of false alarms and valid alarms in the user feedback, the system can dynamically adjust the alarm threshold. For infrared sensors, combined with the average value of the activity density within the historical time, the system can accurately identify the alarm behavior of the sensors and adjust the activity density threshold according to the user feedback. This mechanism can effectively reduce the false alarm frequency, ensure the accuracy of the alarm system after long-term use, and optimize the alarm strategy of the system, thereby reducing unnecessary alarms. By calculating the influence factor of the user feedback data of each infrared sensor on the threshold, the system can reasonably adjust the threshold. The adjustment factors are used to control the influence degree of historical data and user feedback on the threshold adjustment, so as to precisely optimize the alarm trigger threshold in actual operation. This approach avoids the threshold adjustment deviation caused by solely relying on historical data or user feedback, making the alarm threshold more flexible and intelligent, and capable of adapting to environmental changes in real time. The system can adjust the alarm threshold according to the user feedback, solving the defect that traditional alarm systems fail to adapt to environmental changes or user needs in a timely manner. By gradually optimizing the alarm rules of the system, the system can continuously improve the alarm accuracy and enhance its adaptability during long-term operation. This not only reduces false alarms but also enables more efficient alarm response under different environments and conditions.

[0205] The optimization of security event recording and feedback specifically includes:

[0206] Obtain the number of false alarms of the minimum temperature of the k-th temperature sensor as feedback by the user, denoted as

[0207] Obtain the number of valid alarms of the minimum temperature of the k-th temperature sensor as feedback by the user, denoted as

[0208] Obtain the number of false alarms of the maximum temperature of the k-th temperature sensor as feedback by the user, denoted as

[0209] Obtain the number of valid alarms of the maximum temperature of the k-th temperature sensor as feedback by the user, denoted as

[0210] Obtain the average value of the measured temperature data of the k-th temperature sensor within the historical time, denoted as

[0211] Calculate the influence factor of the data feedback by the user of the k temperature sensors on the minimum temperature, denoted as

[0212]

[0213] Calculate the influence factor of the data feedback by the user of the k temperature sensors on the maximum temperature, denoted as

[0214]

[0215] Among them, T′ k,min is the minimum temperature of the k-th temperature sensor after optimization; T′ k,max is the maximum temperature of the k-th temperature sensor after optimization; γ3 and γ4 are adjustment factors, controlling the influence degree of temperature and user feedback on the threshold adjustment.

[0216] By collecting the number of false alarms and valid alarms for the minimum and maximum temperatures reported by users, the system can analyze and determine the alarm accuracy of each sensor within a specific temperature range. The number of false alarms reported by users can provide an accuracy indicator for the alarm system. By comparing this data, the system optimally adjusts the alarm thresholds for each temperature sensor, avoiding the problem of reduced system trust caused by frequent false alarms. The system obtains the average value of the measured temperature data of each temperature sensor over historical time, thereby enabling it to understand the typical operating range and normal data fluctuations of the sensor. With this data, the system can better identify normal and abnormal temperature fluctuations, avoid false alarms caused by instantaneous changes in the external environment, and enhance the stability and reliability of the alarm system. By calculating the influence factors of the user feedback data of each temperature sensor on the minimum and maximum temperatures, the system can dynamically adjust the temperature thresholds to reflect the users' feedback on alarm accuracy. The introduction of the adjustment factor enables the system to optimize the threshold adjustment not only based on historical data but also in combination with real-time user feedback. This mechanism improves the adaptability of the alarm threshold, enabling the system to flexibly adjust according to user needs and environmental changes, reducing the inadaptability problems caused by fixed thresholds. By optimally adjusting the minimum and maximum values of the temperature sensor, the system can more accurately determine abnormal temperature states, avoiding omissions or false alarms caused by overly loose or overly strict alarm ranges. The optimized thresholds can adapt to changes in different environmental conditions during daily operation and continuously and precisely adjust according to users' usage feedback, significantly improving the accuracy and intelligence level of the alarm system.

[0217] The safety event recording and feedback optimization are specifically as follows:

[0218] Obtain the number of false alarms for the minimum humidity of the k-th humidity sensor reported by the user, denoted as

[0219] Obtain the number of valid alarms for the minimum humidity of the k-th humidity sensor reported by the user, denoted as

[0220] Obtain the number of false alarms for the maximum humidity of the k-th humidity sensor reported by the user, denoted as

[0221] Obtain the number of valid alarms for the maximum humidity of the k-th humidity sensor reported by the user, denoted as

[0222] Obtain the average value of the measured humidity data of the k-th humidity sensor over historical time, denoted as

[0223] Calculate the influence factor of the data reported by the users of the k humidity sensors on the minimum humidity, denoted as

[0224]

[0225] Calculate the influence factor of the data fed back by k humidity sensors on the maximum humidity value, denoted as

[0226]

[0227] where H′ k,min is the minimum temperature of the k-th humidity sensor after optimization; H′ k,max is the maximum temperature of the k-th humidity sensor after optimization; γ5 and γ6 are adjustment factors that control the influence degree of humidity and user feedback on the threshold adjustment.

[0228] By collecting the number of false alarms and valid alarms of each humidity sensor within the range of minimum and maximum humidity values, the accuracy of the sensor alarm can be accurately evaluated. Through user feedback, the system can determine which humidity ranges have more reliable alarms, thus avoiding the problem of distrust of the alarm system caused by excessive false alarms. The system can adjust the alarm threshold according to this feedback data to make it more in line with the actual environmental conditions. By obtaining the average value of the humidity data measured by the humidity sensor historically, the system can understand the normal fluctuation range of the humidity sensor. These historical data help to identify real humidity anomalies and short-term environmental fluctuations. Therefore, the system can better distinguish normal humidity changes from abnormal fluctuations, avoiding false alarms caused by temporary environmental changes such as wind speed or temperature changes, thereby improving the accuracy and stability of the alarm. By analyzing the user feedback on the humidity sensor, calculate the influence factor of each humidity sensor's feedback on the minimum and maximum humidity values, and dynamically adjust the humidity alarm threshold. Through this mechanism, the system can quickly respond to user feedback and adjust the humidity threshold in a timely manner, avoiding the problem that traditional static threshold settings cannot adapt to changing environments. Through flexible adjustment, the system realizes the personalized optimization of the threshold, reducing the phenomena of false alarms and missed alarms. As the threshold optimization progresses, the humidity sensor can more accurately identify environmental changes and issue alarms in a timely manner. The optimized humidity sensor alarm threshold can be adjusted according to different users' feedback to achieve more accurate anomaly detection. In addition, the system can adaptively adjust the threshold according to the changing environment and device status, improving the intelligence and response ability of the system.

[0229] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0230] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A big data-based intelligent security warning method, characterized in that, Including: Install door and window sensors respectively at the entrance and exit of the house and by the windows, and label each door and window sensor with a number; Obtain the opening and closing status of the doors and windows, and record the data of the i-th door and window sensor obtained at time t as S w,i (t); Among them, when S w,i (t) = 0, it means that the i-th door or window is closed; When S w,i (t) = 1, it indicates that the i-th door or window is open; Set the dataset of the door and window switch states obtained at time t, denoted as S w (t); Add all the door and window switch states obtained at time t to S(t) in ascending order of the corresponding door and window sensor numbers, where n is the total number of door and window sensors; w (t), to obtain where n w is the total number of door and window sensors; Set the data acquisition interval of the door and window switch status as Δt; Collect the door and window switch status data every Δt and upload it to the server; Install infrared motion sensors at the entrance of the room and in the aisle, and label each infrared motion sensor with a number; Obtain the monitored area of each infrared motion sensor, and label the monitored area with the number of the corresponding infrared motion sensor; Denote the data of the j-th infrared motion sensor obtained at time t as M j (t); Among them, when M j (t) = 0, it means that there is no human activity in area j; When M j (t) = 1, it indicates that there is human activity in area j; Set the infrared motion sensor data set obtained at time t, denoted as M(t); Add all the infrared motion sensor data obtained at time t to M(t) in ascending order according to the corresponding infrared motion sensor number; Obtained where n m is the total number of infrared motion sensors; Respectively obtain the central positions of the living room, bedroom, kitchen and bathroom, install temperature and humidity sensors at each central position, and label each temperature and humidity sensor with a number; Among them, the numbers of the temperature and humidity sensors at the same position are the same, and the total numbers of the temperature sensors and the humidity sensors are equal; Denote the temperature data of the k-th temperature sensor obtained at time t as T k (t), and denote the humidity data of the k-th humidity sensor obtained at time t as H k (t); Set the temperature data set of the temperature sensors obtained at time t, denoted as T(t); Add the temperature data of all temperature and humidity sensors obtained at time t to T(t) in ascending order of the numbers, and obtain where n T is the number of temperature sensors; Set the humidity data set of the humidity sensors obtained at time t, denoted as H(t); Add the humidity data of all humidity sensors obtained at time t to H(t) in ascending order of the numbers, to obtain where n T is the number of humidity sensors; Store and preprocess all the obtained data.

2. The intelligent security warning method based on big data according to claim 1, characterized in that The storing and preprocessing of all the obtained data specifically include: S21. Classify and store the data: Store the door and window switch state data S w (t) and the infrared sensor data M(t) as a time-series boolean data set; Store the temperature and humidity sensor data T(t) and H(t) as a time series floating-point data set; S22. Preprocess the data: Set the sliding window size, denoted as n h ; Smooth the door and window sensor data by using the sliding window method; Among them, is the data of the i-th door and window sensor obtained at time t after smoothing; Smooth the infrared sensor data by using the sliding window method; Among them, is the data of the j-th infrared sensor obtained at time t after smoothing.

3. The intelligent security warning method based on big data according to claim 2, wherein The storing and preprocessing of all the obtained data specifically include: Obtain the number of domain data points for interpolation calculation, denoted as Q; If the temperature data of the kth temperature sensor is missing at time t, use the interpolation method to fill in the missing value for the temperature data; Among them, is the temperature data of the k-th temperature sensor obtained at time t after filling in the missing values; T k (t i ) is the temperature value collected by the k-th temperature sensor at time t i ; If the temperature data of the kth humidity sensor is missing at time t, use the interpolation method to fill in the missing value for the humidity data; Among them, is the humidity data of the k-th humidity sensor obtained at time t after filling in the missing values; H k (t i ) is the humidity value collected by the k-th humidity sensor at time t i ; is the weight corresponding to time t i ; where |t - t i | is the time difference between the current moment t and the known data moment t i ; σ is a hyperparameter that controls the domain range and is used to adjust the influence degree of the most recent moment data on the interpolation result; exp() is an exponential function used to generate a weight decay function that increases with the increase of the time difference; Perform intelligent analysis and anomaly detection based on the collected data.

4. The intelligent security warning method based on big data according to claim 3, wherein The intelligent analysis and anomaly detection based on the collected data specifically include: S31. Door and window anomaly detection: Set the set of time thresholds for opening doors and windows, denoted as S threshold ; If then the i-th door or window at time t will be marked as in an abnormal state; Wherein, ∧ represents the AND operation; S32. Infrared activity anomaly detection: Calculate the time series activity density; Set the j-th active density threshold, denoted as M j,threshold ; If D j (t) > M j,threshold , then the time region j at time t will be marked as an abnormal state; S33. Temperature and humidity anomaly analysis: Set the temperature threshold range of the k-th temperature sensor, denoted as [T k,min , T k,max , where T k,min is the minimum temperature of the k-th temperature sensor, and T k,max is the maximum temperature of the k-th temperature sensor; Set the humidity threshold range of the k-th humidity sensor, denoted as [H k,min , H k,max , where H k,min is the minimum humidity of the k-th humidity sensor, and H k,max is the maximum humidity of the k-th humidity sensor; If then the position temperature obtained by the k-th temperature sensor at time t is recorded as abnormal; If then record the position humidity obtained by the k-th humidity sensor at time t as abnormal; Generate real-time warnings and user notifications according to the results of the intelligent analysis and anomaly detection.

5. The intelligent security warning method based on big data according to claim 4, characterized in that The generating of real-time warnings and user notifications according to the results of the intelligent analysis and anomaly detection specifically includes: Generate an alarm status according to the analysis results of the doors and windows, infrared, temperature and humidity, and video; Wherein, ∨ represents the OR operation; When A(t) = 1, obtain the positions and types of the sensors with the status of 1, and send the obtained positions and types of the sensors to the user; Perform security event recording and feedback optimization.

6. The intelligent security warning method based on big data according to claim 5, characterized in that The performing of security event recording and feedback optimization specifically includes: Set the set of unique identifiers of the sensors that trigger the alarm, denoted as SensorID; Obtain the unique identifiers of all sensors that trigger an alarm at time t and add them to SensorID; Obtain the timestamp at the time of the alarm and denote it as Timestamp; Record the alarm as Log(t) = {A(t), SensorID, Timestamp}; Collect the feedback from the user for each alarm and update the threshold: Obtain the number of false alarms of the j-th infrared sensor for which user feedback is obtained, denoted as Obtain the effective alarm count of the j-th infrared sensor for user feedback, denoted as Obtain the average value of the measured activity density of the j-th infrared sensor within the historical time, denoted as Calculate the influence factor of the user feedback data of the j-th infrared sensor on the threshold, denoted as where M′ j,threshold is the optimized j-th activity density threshold; γ1 and γ2 are adjustment factors.

7. A method for intelligent security warning based on big data according to claim 6, characterized in that, The safety event recording and feedback optimization is carried out, specifically including: Obtain the number of false alarms of the minimum temperature of the k-th temperature sensor for which user feedback is obtained, denoted as Obtain the effective alarm count of the minimum temperature of the k-th temperature sensor for user feedback, denoted as Obtain the number of false alarms of the maximum temperature of the k-th temperature sensor for user feedback, denoted as Obtain the effective alarm count of the maximum temperature of the k-th temperature sensor for which user feedback is obtained, denoted as Obtain the average value of the temperature data measured by the k-th temperature sensor within the historical time, denoted as Calculate the influence factor of the data feedback from k temperature sensors on the minimum temperature, denoted as Calculate the influence factor of the data feedback from k temperature sensors on the maximum temperature, denoted as where T′ k,min is the minimum temperature of the k-th temperature sensor after optimization; T′ k,max is the maximum temperature of the k-th temperature sensor after optimization; γ3 and γ4 are adjustment factors.

8. A method for intelligent security warning based on big data according to claim 7, characterized in that, The safety event recording and feedback optimization is carried out, specifically including: Obtain the number of false alarms of the minimum humidity value of the k-th humidity sensor for user feedback, denoted as Obtain the effective alarm count of the minimum humidity value of the k-th humidity sensor for which user feedback is obtained, denoted as Obtain the number of false alarms of the maximum humidity value of the k-th humidity sensor for user feedback, denoted as The effective alarm count of the maximum humidity value of the k-th humidity sensor for obtaining user feedback is denoted as Obtain the average value of the humidity data measured by the k-th humidity sensor within the historical time, denoted as Calculate the influence factor of the data feedback from k humidity sensors on the minimum humidity, denoted as Calculate the influence factor of the data feedback from k humidity sensors on the maximum humidity value, denoted as Among them, H' k,min is the minimum temperature of the k-th humidity sensor after optimization; H' k,max is the maximum temperature of the k-th humidity sensor after optimization; γ5 and γ6 are adjustment factors.

Images