A regional security monitoring method and system based on building protection doors and windows

Abstract

The present application relates to a regional security monitoring method and system based on building protection doors and windows, the method comprising the steps of: generating regional security monitoring information based on a preset security detection module in response to a trigger operation of detecting and identifying an object in the external area; generating security assessment information based on a preset convolutional neural network according to the regional security monitoring information; judging whether the risk index is greater than a preset risk threshold; if the judgment is yes, obtaining pre-stored access statistics information, and further judging whether there are people in the internal area according to the access statistics information; if the judgment is yes, executing a first security measure to protect the safety of people in the room; if the judgment is no, executing a second security measure to protect the safety of materials stored in the room. Through the above steps, the present invention implements the execution of corresponding security measures to protect the safety of people and materials in the room according to different risk indexes when a risk event is detected.

Description

Technical Field

[0001] The present application relates to the field of security monitoring, and in particular to a method and system for regional security monitoring based on building protective doors and windows. Background Art

[0002] With the rapid development of the Internet of Things technology, a new type of intelligent security door has emerged, which combines the Internet of Things technology with traditional security doors to form a more advanced intelligent security system. This system can achieve remote control and unlocking through a variety of methods such as smartphones, fingerprint recognition, and facial recognition, greatly improving the security of the access control system and the convenience of users. Although the intelligent security system has made significant progress in many aspects, the system still has certain limitations. Existing intelligent security systems mainly focus on monitoring and protection. When faced with extreme situations such as violent intrusions, in addition to issuing alarms, they often lack effective countermeasures, which greatly reduces the overall protection capabilities of the system.

[0003] Therefore, it is necessary to improve the prior art to solve the above problems. Summary of the invention

[0004] Based on this, it is necessary to provide a regional security monitoring method and system based on building protection doors and windows to address the above-mentioned technical problems, so as to solve the technical problem that the existing security system lacks effective response measures when risk events occur.

[0005] The present invention provides a regional security monitoring method based on building protective doors and windows, which is applied to the security monitoring of the internal area of ​​the protective doors and windows and the external area of ​​the protective doors and windows in the building, comprising:

[0006] In response to a trigger operation of detecting and identifying an object in the external area, generating regional security monitoring information based on a preset security detection module;

[0007] Generate safety assessment information based on a preset convolutional neural network according to the regional safety monitoring information, wherein the safety assessment information includes a risk estimation type and a risk index;

[0008] Determining whether the risk index is greater than a preset risk threshold;

[0009] If the judgment is yes, then obtaining pre-stored access statistics information, and further determining whether there are people in the internal area according to the access statistics information;

[0010] If the judgment is yes, the preset first security measure is executed to protect the safety of personnel in the internal area; if the judgment is no, the preset second security measure is executed to protect the safety of materials stored in the internal area.

[0011] As a preferred solution, the external area includes a door area and a window area, and the security detection module is arranged in the door area and the window area;

[0012] The first security measure includes the following steps:

[0013] Determine whether the trigger operation originates from the door area or the window area;

[0014] If it is determined that the trigger operation originates from the door area, a door closing instruction is generated and sent to a preset door switch module to close the room door;

[0015] Generate an alarm instruction and send it to a preset alarm module, and control the alarm module to issue an alarm message;

[0016] If it is determined that the trigger operation originates from the window area, a door opening instruction is generated and sent to a preset door switch module to open the room door;

[0017] Generate a gate opening command and send it to the preset gate switch module to open the protective door;

[0018] Generate an alarm instruction and send it to the alarm module, and control the alarm module to issue an alarm message.

[0019] As a preferred solution, the external area includes a door area and a window area;

[0020] The second security measure comprises the following steps:

[0021] Generate a door closing command and send it to the preset door switch module to close the room door;

[0022] Generate an alarm instruction and send it to a preset alarm module, and control the alarm module to issue an alarm message;

[0023] The security assessment information is sent to the mobile terminal and sent to the cloud for storage.

[0024] As a preferred solution, in response to the triggering operation of detecting and identifying the object in the external area, generating regional security monitoring information based on a preset security detection module specifically includes:

[0025] In response to a trigger operation of detecting and identifying an object in the external area, a camera recognition unit based on a preset safety detection module acquires a range recognition image, and generates a facial recognition information and a gesture recognition information set according to the range recognition image, wherein the safety detection module is pre-set on the protective door and window, and the safety detection module includes a camera recognition unit, a grating detection unit and a pressure detection unit;

[0026] generating a posture recognition data set and preliminary body shape recognition data according to the posture recognition information set;

[0027] Acquiring infrared detection data based on the grating detection unit, and generating a stay time sequence and auxiliary body shape recognition data according to the infrared detection data;

[0028] generating body shape recognition information according to the preliminary body shape recognition data and the auxiliary body shape recognition data;

[0029] Generating identity recognition information according to the facial recognition information and the body shape recognition information, and comparing the identity recognition information with the resident information in a preset database to generate a comparison similarity, wherein the database is preset and stores the resident information of the resident of the house;

[0030] Determine whether the comparison similarity is greater than a preset similarity threshold. If not, obtain pressure detection information based on the pressure detection unit, and generate regional safety monitoring information according to the identity recognition information, the posture recognition data set, the stay time series and the pressure detection information.

[0031] As a preferred solution, the risk index is generated based on the following formula:

[0032] ;

[0033] Among them, R is the risk index, B is the body recognition similarity, F is the face recognition similarity, Wu is the identity weight parameter, Wv is the posture weight parameter, S is the posture matching degree, T is the person's stay time, Tmax is the total detection time, Wx is the time weight parameter, P is the detection pressure value, Pe is the pressure detection threshold, Wy is the pressure weight parameter, and N is the number of pressure detections.

[0034] As a preferred solution, the generating of security assessment information based on the regional security monitoring information based on a preset convolutional neural network specifically includes:

[0035] Extracting features of the identity recognition information through the convolutional neural network to generate an identity recognition feature vector, and generating a posture feature vector through the PoseNet model according to the posture recognition data set;

[0036] Generate a time series feature vector through the convolutional neural network according to the residence time series, and generate a pressure feature vector through the convolutional neural network according to the pressure detection information;

[0037] Generate a fusion feature parameter according to the identity recognition vector, the posture feature vector, the time series feature vector and the pressure feature vector;

[0038] The fusion feature parameters are input into a preset risk assessment model to generate safety assessment information, wherein the safety assessment information includes a risk index and a risk estimation type, wherein the risk assessment model is pre-trained and set.

[0039] As a preferred solution, the access statistics information is obtained through the following steps:

[0040] In response to the door opening operation, the identity recognition information and the infrared detection data are obtained, and the walking direction information of the personnel is generated by an edge detection algorithm according to the infrared detection data; and the personnel change information is generated according to the identity recognition information and the walking direction information of the personnel;

[0041] In response to the door closing operation, acquiring indoor image data of the interior area based on a preset indoor camera unit, and generating verification identification information according to the indoor image data;

[0042] Comparing the verification identification information with the identity identification information to generate a confidence level, and determining whether the confidence level is greater than a preset confidence threshold;

[0043] If the judgment is yes, the personnel change information is marked to generate personnel number information and time record data, and access statistics information is generated according to the personnel change information, the personnel number information and the time record data;

[0044] If the judgment is no, an abnormal event record is generated and sent to the mobile terminal for user confirmation and feedback information is generated. The personnel change information is corrected according to the feedback information to generate corrected change information, and access statistical information is generated according to the corrected change information, the personnel number information and the time record data.

[0045] The present invention also provides a regional security monitoring system based on building protective doors and windows, which is applied to the security monitoring of the internal area of ​​the protective doors and windows in the building and the external area of ​​the protective doors and windows, including:

[0046] A monitoring information generating module, configured to generate regional security monitoring information based on a preset security detection module in response to a triggering operation of detecting and identifying an object in the external region;

[0047] An assessment information generation module, used to generate safety assessment information based on a preset convolutional neural network according to the regional safety monitoring information;

[0048] A risk index determination module is used to determine whether the risk index is greater than a preset risk threshold;

[0049] A personnel situation determination module, for obtaining pre-stored access statistics information if the determination is yes, and further determining whether there are personnel in the internal area according to the access statistics information;

[0050] The security measure execution module is used to execute a preset first security measure to protect the safety of personnel in the internal area if the judgment is yes; and to execute a preset second security measure to protect the safety of materials stored in the internal area if the judgment is no.

[0051] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps described in the above-mentioned regional security monitoring method based on building protective doors and windows are implemented.

[0052] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps described in the above-mentioned regional security monitoring method based on building protective doors and windows are implemented.

[0053] The technical effects achieved by the present invention are as follows:

[0054] The above-mentioned regional security monitoring method and system based on building protection doors and windows, in turn, generates regional security monitoring information based on a preset security detection module by responding to the trigger operation of detecting and identifying objects in the external area; generates security assessment information based on a preset convolutional neural network according to the regional security monitoring information, wherein the security assessment information includes a risk estimation type and a risk index; determines whether the risk index is greater than a preset risk threshold; if the judgment is yes, obtains pre-stored access statistical information, and further determines whether there are people in the internal area according to the access statistical information; if the judgment is yes, executes the preset first security measure to protect the safety of people in the internal area; if the judgment is no, executes the preset second security measure to protect the safety of materials stored in the internal area; thereby achieving that when a risk event is detected, corresponding security measures are executed according to different risk indexes to protect the safety of people and materials in the room. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flowchart of a method for regional security monitoring based on building protection doors and windows in one embodiment;

[0056] Figure 2 The structure block diagram of a regional security monitoring system based on building protection doors and windows in one embodiment. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application 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 application and are not used to limit the present application.

[0058] In one embodiment, a terminal is provided, which is applied to security monitoring of internal areas and external areas of protective doors and windows in buildings, and the terminal is used to generate regional security monitoring information based on a preset security detection module in response to a trigger operation of detecting and identifying an object in the external area; generate security assessment information based on a preset convolutional neural network according to the regional security monitoring information; determine whether the risk index is greater than a preset risk threshold; if the determination is yes, obtain pre-stored access statistics information, and further determine whether there are people in the internal area according to the access statistics information; if the determination is yes, execute a preset first security measure to protect the safety of people in the internal area; if the determination is no, execute a preset second security measure to protect the safety of materials stored in the internal area.

[0059] The terminal may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices.

[0060] In one embodiment, Figure 1 As shown, a regional security monitoring method based on building protective doors and windows is provided, which is applied to the security monitoring of the internal area of ​​the protective doors and windows in the building and the external area of ​​the protective doors and windows. It should be understood that the buildings mentioned in this article include but are not limited to residential buildings, office buildings, factory production buildings, etc.

[0061] In this embodiment, the protective doors and windows include protective doors, room doors and protective windows. The external area is defined as a specific area outside the protective door and outside the protective window, and the internal area is defined as a closed space surrounded by the outer wall of the house, the inner side of the protective door and the inner side of the protective window. The external area is divided into a doorway area and a window area according to the different locations where the security detection module is set. The internal area of ​​the house usually includes a public area and multiple rooms. The room doors are used to separate each room from the public area to form multiple independent room areas. A protective window is set in each room area, and the rest of the internal area is defined as a public area.

[0062] The regional security monitoring method based on building protection doors and windows includes:

[0063] Step S100: In response to a triggering operation of detecting and identifying an object in the external area, generating regional security monitoring information based on a preset security detection module.

[0064] In this step, in order to monitor the safety of each house in the building, a security detection module is set on the protective door and the protective window. This solution monitors risk events in the external area through the security detection module on the protective door and the protective window, and implements corresponding security measures to ensure the safety of personnel and materials in the internal area. When an action detection and identification object appears in the external area and a trigger operation occurs, the security detection module detects the detection and identification object and summarizes the detected information to generate regional security monitoring information. The detection and identification object can be a person or an object such as a stone, a hammer, or a stick.

[0065] Step S200: Generate safety assessment information based on a preset convolutional neural network according to the regional safety monitoring information, wherein the safety assessment information includes a risk assessment type and a risk index.

[0066] In this step, a convolutional neural network is pre-set, which can adapt to the various input types and data structures contained in the regional security monitoring information. By carefully designing the network structure, including adjusting the configuration of the convolution layer, the convolutional neural network can handle different types of tasks. Using its hierarchical feature extraction mechanism, the regional security monitoring information can be deeply abstracted and semantically encoded, thereby effectively outputting the risk estimation type and risk index, providing guidance for the implementation of subsequent security measures.

[0067] Step S300: Determine whether the risk index is greater than a preset risk threshold.

[0068] In this step, in many daily scenarios, such as pedestrians occasionally passing by the protective door in a hurry, couriers leaving quickly after dropping off packages outside the door, or birds occasionally flying over the protective window, these situations often do not pose a substantial security threat, so their corresponding risk indexes are usually low. By pre-setting a risk threshold, these low-risk events can be effectively filtered to avoid frequent and unnecessary alarm triggering. When the risk index is lower than this preset risk threshold, only the regional security assessment information will be sent to the user's pre-bound mobile terminal to inform the user of the current security status, and the alarm module will not be activated. The above settings help reduce false alarms while maintaining user trust in the security monitoring system. When it is detected that the risk index is greater than the preset risk threshold, subsequent corresponding security measures are executed.

[0069] Step S400: if the judgment is yes, obtaining pre-stored access statistical information, and further determining whether there are people in the internal area according to the access statistical information;

[0070] In this step, when the risk index is detected to be greater than the preset risk threshold, it indicates that the detected and identified object in the external area is trying to invade the internal area. It is necessary to further determine whether there are people in the internal area based on the pre-stored access statistics information, and choose to implement different security measures based on the presence of people in the internal area.

[0071] Step S500: If the judgment is yes, execute the first security measure to protect the safety of people in the room; if the judgment is no, execute the second security measure to protect the safety of materials stored in the room, wherein the first security measure and the second security measure are pre-set.

[0072] In this embodiment, by setting a security detection module on the protective door and the protective window, the security situation of each house in the building is monitored. When the detection and identification object in the external area, whether it is a person or an object, performs a triggering operation, such as approaching, touching or otherwise interfering with the protective door or the protective window, the security detection module is immediately started, the detection and identification object is identified and detected, and the information is aggregated to generate regional security monitoring information. Then, the above-generated regional security monitoring information is processed using a preset convolutional neural network to output the risk estimation type and risk index. Based on the preset risk threshold and the generated risk index, the risk event that occurs is judged, and different types of security measures are performed according to the severity of the risk. For some low-risk events, only the security assessment information is sent to the bound mobile terminal to notify the residents of the house. For some high-risk events, it is necessary to further determine whether there are people in the internal area. If there are people, the safety of the people is prioritized. If there are no people, the main focus is on protecting the materials in the house.

[0073] In one embodiment, the step S100: in response to a trigger operation of detecting and identifying an object in the external area, generating regional security monitoring information based on a preset security detection module specifically includes:

[0074] Step S110: in response to a trigger operation of detecting and identifying an object in the external area, a camera recognition unit based on a preset safety detection module acquires a range recognition image, and generates facial recognition information and a gesture recognition information set according to the range recognition image, wherein the safety detection module is pre-set on the protective door and window, and the safety detection module includes a camera recognition unit, a grating detection unit and a pressure detection unit;

[0075] In this step, facial recognition information is generated through the MobileNet model based on the range recognition image, and a posture recognition information set is generated through the OpenPose Light model based on the range recognition image.

[0076] Step S120: generating a posture recognition data set and preliminary body shape recognition data according to the posture recognition information set;

[0077] In this step, the suspicious posture data is identified through the K-means clustering algorithm according to the posture recognition information set, the posture matching degree is generated according to the type and proportion of the suspicious posture data, and the preliminary body shape recognition data is generated according to the posture recognition data set through the PoseNet model.

[0078] Step S130: acquiring infrared detection data based on the grating detection unit, and generating a stay time sequence and auxiliary body shape recognition data according to the infrared detection data;

[0079] In this step, the grating detection unit includes an infrared emission component and a grating component. When the detection and identification object enters the detection range of the grating detection unit, the infrared light emitted by the infrared emission component is blocked by the detection and identification object, causing the light signal received by the grating component to change, and these changed light signals are converted into electrical signals and infrared detection data are generated. Then, the residence time sequence and auxiliary body shape recognition data are generated based on the infrared detection data, and the residence time of the person can be obtained based on the residence time sequence.

[0080] Step S140: generating body shape recognition information according to the preliminary body shape recognition data and the auxiliary body shape recognition data;

[0081] Step S150: generating identity recognition information according to the facial recognition information and the body shape recognition information, and comparing the identity recognition information with the household information in a preset database to generate a comparison similarity, wherein the database is preset and stores the household information of the household in the house;

[0082] Step S160: determining whether the comparison similarity is greater than a preset similarity threshold; if not, obtaining pressure detection information based on the pressure detection unit, and generating regional safety monitoring information according to the identity recognition information, the gesture recognition data set, the stay time sequence and the pressure detection information;

[0083] Step S170: If the judgment is yes, generate the personnel residence time according to the residence time sequence. When the personnel residence time is greater than the preset value and it is confirmed that the triggering operation comes from the door area, generate a door opening instruction and send it to the preset door switch module to open the protective door.

[0084] In this embodiment, the safety detection module includes three parts: a camera recognition unit, a grating detection unit, and a pressure detection unit. When the detection and identification object appears in the external area, the camera recognition unit shoots the detection and identification object to obtain a range recognition image, performs face recognition and posture recognition on the detection and identification object through the range recognition image, and generates facial recognition information and a posture recognition information set. When the detection and identification object enters the recognition range of the grating detection unit, infrared detection data is obtained, and then the body recognition information of the detection and identification object is generated by fusing the posture recognition information set and the infrared detection data. The identity recognition information composed of the facial recognition information and the body recognition information can more accurately reflect the appearance characteristics of the detection and identification object. The identity recognition information is compared with the pre-stored resident information. If the similarity is low, the identity recognition information, posture recognition data set, residence time series and pressure detection information are used to generate regional safety monitoring information. If the similarity is high, it can be directly identified as a resident of the house, and the protective door is opened to allow the resident to enter.

[0085] In this embodiment, the MobileNet model, OpenPose Light model, OpenPose Light model and PoseNet model are preferably used, all of which are lightweight models with small memory usage and are suitable for edge devices in building protection application scenarios.

[0086] Furthermore, the body shape recognition information is generated based on the following steps, specifically including:

[0087] Step S181: generating a key point coordinate set through the PoseNet model according to the posture recognition data set, generating preliminary body shape recognition data based on the convolutional neural network according to the key point coordinate set, and performing feature extraction on the preliminary body shape recognition data to generate first feature data;

[0088] Step S182: generating depth information according to the infrared detection data, generating auxiliary body shape recognition data based on the depth information based on point cloud processing technology, and performing feature extraction on the auxiliary body shape recognition data to generate second feature data;

[0089] Step S183: Generate body shape feature data based on the weighted sum fusion algorithm according to the first feature data and the second feature data, and generate body shape recognition information based on the convolutional neural network according to the body shape feature data.

[0090] In this embodiment, in order to enhance the accuracy of body shape recognition, a comprehensive data processing strategy is adopted. The key point coordinate set is extracted from the posture recognition data set through the PoseNet model, and the coordinate set is processed through the convolutional neural network to form preliminary body shape recognition data. Depth information is constructed using infrared detection data, and auxiliary body shape recognition data is extracted from it with the help of point cloud processing technology. The features of the preliminary body shape recognition data and the auxiliary body shape recognition data are integrated by adopting a weighted sum fusion algorithm to form more comprehensive body shape feature data. Then, the body shape feature data is input into the convolutional neural network for further processing to output accurate body shape recognition information to improve the accuracy and credibility of body shape recognition.

[0091] In one embodiment, the step S200: generating safety assessment information based on a preset convolutional neural network according to the regional safety monitoring information includes:

[0092] Step S210: extracting features of the identity recognition information through the convolutional neural network to generate an identity recognition feature vector, and generating a posture feature vector through the PoseNet model according to the posture recognition data set;

[0093] Step S220: generating a time series feature vector through the convolutional neural network according to the residence time series, and generating a pressure feature vector through the convolutional neural network according to the pressure detection information;

[0094] Step S230: generating fusion feature parameters according to the identity recognition vector, the posture feature vector, the time series feature vector and the pressure feature vector;

[0095] Step S240: inputting the fusion feature parameters into a preset risk assessment model to generate safety assessment information, wherein the safety assessment information includes a risk index and a risk estimation type, wherein the risk assessment model is pre-trained and set.

[0096] In this embodiment, for each piece of information obtained by the camera recognition unit, the grating detection unit and the pressure detection unit, feature extraction is performed based on the convolutional neural network to generate corresponding feature vectors, and the generated multiple feature vectors are fused through the convolutional neural network to generate fusion feature parameters. A trained risk assessment model is pre-set, and the fusion feature parameters are input into the risk assessment model, and the security assessment information including the risk index and the risk assessment type is output to provide guidance for the subsequent implementation of security measures.

[0097] In one embodiment, the risk index is generated based on the following formula:

[0098] ;

[0099] Among them, R is the risk index, B is the body recognition similarity, F is the face recognition similarity, Wu is the identity weight parameter, Wv is the posture weight parameter, S is the posture matching degree, T is the person's stay time, Tmax is the total detection time, Wx is the time weight parameter, P is the detection pressure value, Pe is the pressure detection threshold, Wy is the pressure weight parameter, and N is the number of pressure detections.

[0100] In this embodiment, a risk index is generated by comprehensively evaluating from multiple dimensions according to various parameters collected by the security detection module. The facial recognition information and body recognition information are compared with the pre-stored resident information to generate facial recognition similarity and body recognition similarity, and the identity dimension risk value is generated according to the facial recognition similarity and body recognition similarity. Suspicious posture data is obtained according to the posture recognition information, and the posture matching degree is generated according to the type of suspicious posture data and the proportion in the posture recognition information, and the posture dimension risk value is generated according to the posture matching degree. The time dimension risk value is generated according to the residence time of the detected identification object within the detection range of the grating detection unit, and the total detection time from the detection to the trigger operation until the end of the risk event. When calculating the pressure dimension risk value, a pressure detection threshold is set. When the pressure does not exceed the pressure detection threshold, this part of the value is not included in the total risk index. The pressure dimension risk value is generated according to the detection pressure value exceeding the pressure detection threshold and the number of pressure detections. The risk value of each dimension will generate a corresponding weight parameter through a convolutional neural network, and the identity dimension risk value, posture dimension risk value, time dimension risk value and pressure dimension risk value are multiplied by their respective corresponding weight parameters and added to generate the final risk index.

[0101] In this embodiment, based on the various detection information collected by the security detection module, the risk type can be preliminarily assessed, and then the risk type can be further verified through the risk values ​​of each dimension in the risk index. If the verification result does not match, the risk type of the preliminary assessment is corrected. Through the above operation, a more accurate risk assessment type is generated for the user. In some embodiments, the risk assessment type can also be presented in the form of a percentage, and the risk assessment type includes the proportion of each risk type assessed.

[0102] In one embodiment, the access statistics information is obtained by the following steps:

[0103] Step S410: in response to the door opening operation, obtaining the identity recognition information and the infrared detection data, generating personnel walking direction information through an edge detection algorithm according to the infrared detection data; generating personnel change information according to the identity recognition information and the personnel walking direction information;

[0104] In this step, when the detection and identification object appears in front of the protective door, the detection and identification object is monitored through the aforementioned steps S110-S150 to obtain its identity information. Whether the detection and identification object enters the internal area by itself through the verification mechanism of the protective door, or the residents in the internal area actively open the protective door for the detection and identification object to allow the detection and identification object to enter the internal area, once the walking direction information of the person generated by the infrared detection data confirms that the detection and identification object has passed through the protective door, the number of people in the internal area is increased by 1. If n detection and identification objects enter the protective door at the same time, the number of people in the internal area is increased by n.

[0105] Step S420: in response to the door closing operation, obtaining indoor image data of the interior area based on a preset indoor camera unit, and generating verification identification information according to the indoor image data;

[0106] In this step, when all the detection and identification objects that have passed the verification or are allowed to enter the internal area, the protective door is closed. An indoor camera unit is set in the public area of ​​the internal area, and the indoor camera unit is used to capture the detection and identification objects that have entered to obtain indoor image data, and the indoor image data is recognized to generate verification recognition information.

[0107] Step S430: Compare the review identification information with the identity identification information to generate a confidence level, and determine whether the confidence level is greater than a preset confidence threshold;

[0108] Step S440: if the judgment is yes, marking the personnel change information to generate personnel number information and time record data, and generating access statistics information according to the personnel change information, the personnel number information and the time record data;

[0109] If the confidence level is greater than the preset confidence threshold, it indicates that the detection information of the detected object before and after passing through the protective door is highly similar, and there is no misjudgment. The personnel change information is marked, and the unique personnel number of the detected object and the timestamp of entry are added. This information is saved and access statistics are generated to achieve accurate records of entering personnel.

[0110] Step S450: If the judgment is no, an abnormal event record is generated and sent to the mobile terminal for user confirmation and feedback information is generated. The personnel change information is corrected according to the feedback information to generate corrected change information. Access statistical information is generated according to the corrected change information, the personnel number information and the time record data.

[0111] In this embodiment, after the detection and identification object is identified by the camera recognition unit, it is then reviewed by the indoor camera unit, and the results of the two detections are compared and confirmed to improve the accuracy of recognition. Avoid recognition errors in some special cases, such as when the detection and identification object is blocked by carrying large furniture or appliances, or when a large detection and identification object blocks a small detection and identification object, etc., which is prone to errors and affects the decision-making of subsequent security measures. Preferably, the camera recognition unit and the indoor camera unit shoot the detection and identification object from different angles to improve the accuracy of detection.

[0112] In one embodiment, the first security measure comprises the following steps:

[0113] Step S511: determining whether the triggering operation originates from the door area or the window area;

[0114] Step S512: if it is determined that the triggering operation originates from the door area, a door closing instruction is generated and sent to a preset door switch module to close the room door;

[0115] Step S513: Generate an alarm instruction and send it to a preset alarm module to control the alarm module to issue an alarm message;

[0116] Step S514: if it is determined that the triggering operation originates from the window area, a door opening instruction is generated and sent to a preset door switch module to open the room door;

[0117] Step S515: Generate an alarm instruction and send it to the alarm module, and control the alarm module to issue an alarm message;

[0118] Step S516: Generate a door opening instruction and send it to a preset door switch module to open the protective door.

[0119] In this embodiment, when there are people in the internal area, the first security measure mainly for protecting the safety of people in the internal area is implemented. Different schemes are implemented according to the different intrusion locations of the detection and identification object. When the detection and identification object is detected to invade from the door area, the room door is controlled to be closed to prevent the detection and identification object from entering the room area, thereby ensuring the safety of people in the room area. When the detection and identification object is detected to invade from the window area, the room door is controlled to be opened and an alarm is issued to remind the people in the room area, so that the people in the room area can leave quickly and avoid facing the detection and identification object without preparation. At the same time, the protective door is opened so that the people can decide whether to flee the internal area according to the situation. In this embodiment, the three actions of opening the room door, issuing an alarm message and opening the protective door can be performed simultaneously, or they can be performed in the order of steps S514-S516.

[0120] It should be noted that when this method is applied to houses in residential buildings, there is usually an additional balcony area in the house. A protective window can be set between the public area and the balcony area, and a security detection module can be set at the same time. If the detected identification object invades the internal area from the balcony area, it should be regarded as invading from the door area, and the first security measure should be executed.

[0121] In one embodiment, the second security measure comprises the following steps:

[0122] Step S521: Generate a door closing instruction and send it to a preset door switch module to close the room door;

[0123] Step S522: Generate an alarm instruction and send it to a preset alarm module to control the alarm module to issue an alarm message;

[0124] Step S523: Send the security assessment information to the mobile terminal and to the cloud for storage.

[0125] In this embodiment, when there are no people in the internal area, the second security measure is implemented to protect the materials stored in the internal area. The preset door switch module controls the room door to close, prevents the detected and identified object from invading the room area, and issues an alarm message. The alarm message can be in the form of an alarm sound, a warning voice, a flashing light, etc., without limitation. The dangerous event is sent to a pre-bound mobile terminal to notify the occupants of the house, and then the security assessment information is stored in the cloud.

[0126] In one embodiment, Figure 2 As shown, a regional security monitoring system based on building protective doors and windows is provided, which is applied to the security monitoring of the internal areas and external areas of the protective doors and windows in the building, including a monitoring information generation module, an evaluation information generation module, a risk index judgment module, a personnel situation judgment module and a security measure execution module.

[0127] The monitoring information generation module is used to generate regional security monitoring information based on a preset security detection module in response to a triggering operation of detecting and identifying an object in the external area.

[0128] The assessment information generation module is used to generate security assessment information based on the regional security monitoring information based on a preset convolutional neural network.

[0129] The risk index determination module is used to determine whether the risk index is greater than a preset risk threshold.

[0130] The personnel situation determination module is used to obtain pre-stored access statistical information if the determination is yes, and further determine whether there are personnel in the internal area according to the access statistical information.

[0131] The security measure execution module is used to execute a preset first security measure to protect the safety of personnel in the internal area if the judgment is yes; and to execute a preset second security measure to protect the safety of materials stored in the internal area if the judgment is no.

[0132] In one embodiment, the monitoring information generation module is further used to:

[0133] In response to a trigger operation of detecting and identifying an object in the external area, a camera recognition unit based on a preset safety detection module acquires a range recognition image, and generates facial recognition information and a gesture recognition information set according to the range recognition image, wherein the safety detection module is pre-set on the protective door and window, and the safety detection module includes a camera recognition unit, a grating detection unit and a pressure detection unit; generates a gesture recognition data set and preliminary body recognition data according to the gesture recognition information set; acquires infrared detection data based on the grating detection unit, generates a stay time sequence and auxiliary body recognition data according to the infrared detection data; generates body recognition information according to the preliminary body recognition data and the auxiliary body recognition data; generates body recognition information according to the facial recognition information and the body recognition data The information generates identity recognition information, and compares the identity recognition information with the resident information in a preset database to generate a comparative similarity, wherein the database is pre-set and stores the resident information of the house resident; it is determined whether the comparative similarity is greater than a preset similarity threshold, and if it is determined to be negative, the pressure detection information is obtained based on the pressure detection unit, and regional safety monitoring information is generated according to the identity recognition information, the posture recognition data set, the residence time sequence and the pressure detection information; if it is determined to be positive, the personnel residence time is generated according to the residence time sequence, and when the personnel residence time is greater than a preset value and it is confirmed that the triggering operation originates from the door area, a door opening instruction is generated and sent to a preset door switch module to open the protective door.

[0134] In one embodiment, the monitoring information generation module is further used to:

[0135] A key point coordinate set is generated through the PoseNet model according to the posture recognition data set, preliminary body shape recognition data is generated based on the convolutional neural network according to the key point coordinate set, and features are extracted from the preliminary body shape recognition data to generate first feature data; depth information is generated according to the infrared detection data, auxiliary body shape recognition data is generated based on the point cloud processing technology according to the depth information, and features are extracted from the auxiliary body shape recognition data to generate second feature data; body shape feature data is generated based on a weighted sum fusion algorithm according to the first feature data and the second feature data, and body shape recognition information is generated based on the convolutional neural network according to the body shape feature data.

[0136] In one embodiment, the evaluation information generating module is further used to:

[0137] The identity recognition information is feature extracted by the convolutional neural network to generate an identity recognition feature vector, and a posture feature vector is generated by the PoseNet model according to the posture recognition data set; a time series feature vector is generated by the convolutional neural network according to the residence time series, and a pressure feature vector is generated by the convolutional neural network according to the pressure detection information; fusion feature parameters are generated according to the identity recognition vector, the posture feature vector, the time series feature vector and the pressure feature vector; the fusion feature parameters are input into a preset risk assessment model to generate safety assessment information, and the safety assessment information includes a risk index and a risk estimation type, wherein the risk assessment model is pre-trained and set.

[0138] In one embodiment, the evaluation information generating module is further used to:

[0139] The risk index is generated based on the following formula:

[0140] ;

[0141] Among them, R is the risk index, B is the body recognition similarity, F is the face recognition similarity, Wu is the identity weight parameter, Wv is the posture weight parameter, S is the posture matching degree, T is the person's stay time, Tmax is the total detection time, Wx is the time weight parameter, P is the detection pressure value, Pe is the pressure detection threshold, Wy is the pressure weight parameter, and N is the number of pressure detections.

[0142] In one embodiment, the personnel situation determination module is further used to:

[0143] In response to the door opening operation, the identity recognition information and the infrared detection data are obtained, and the walking direction information of the personnel is generated through an edge detection algorithm according to the infrared detection data; the personnel change information is generated according to the identity recognition information and the walking direction information of the personnel; in response to the door closing operation, the indoor image data of the internal area is obtained based on the preset indoor camera unit, and the verification recognition information is generated according to the indoor image data; the verification recognition information is compared with the identity recognition information to generate a confidence level, and it is determined whether the confidence level is greater than a preset confidence threshold; if the judgment is yes, the personnel change information is marked to generate personnel number information and time record data, and access statistics information is generated according to the personnel change information, the personnel number information and the time record data; if the judgment is no, an abnormal event record is generated and sent to the mobile terminal for confirmation by the user and feedback information is generated, the personnel change information is corrected according to the feedback information to generate corrected change information, and access statistics information is generated according to the corrected change information, the personnel number information and the time record data.

[0144] In one embodiment, the safety measure execution module is further used to:

[0145] Determine whether the trigger operation originates from the door area or the window area; if it is determined that the trigger operation originates from the door area, generate a door closing instruction and send it to a preset door switch module to close the room door; generate an alarm instruction and send it to a preset alarm module to control the alarm module to issue an alarm message; if it is determined that the trigger operation originates from the window area, generate a door opening instruction and send it to a preset door switch module to open the room door; generate a door opening instruction and send it to a preset door switch module to open the protective door; generate an alarm instruction and send it to the alarm module to control the alarm module to issue an alarm message.

[0146] In one embodiment, the safety measure execution module is further used to:

[0147] Generate a door closing command and send it to a preset door switch module to close the room door; generate an alarm command and send it to a preset alarm module to control the alarm module to issue an alarm message; send the security assessment information to a mobile terminal and to cloud storage.

[0148] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps described in the above-mentioned regional security monitoring method based on building protective doors and windows when executing the computer program.

[0149] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps described in the above-mentioned regional security monitoring method based on building protective doors and windows are implemented.

[0150] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0151] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0152] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A regional security monitoring method based on building protective doors and windows, which is applied to the security monitoring of the internal area and the external area of ​​the protective doors and windows in the building, characterized in that: include: In response to a trigger operation of detecting and identifying an object in the external area, generating regional security monitoring information based on a preset security detection module; Generate safety assessment information based on a preset convolutional neural network according to the regional safety monitoring information, wherein the safety assessment information includes a risk estimation type and a risk index; Determining whether the risk index is greater than a preset risk threshold; If the judgment is yes, then obtaining pre-stored access statistics information, and further determining whether there are people in the internal area according to the access statistics information; If the judgment is yes, the preset first security measure is executed to protect the safety of personnel in the internal area; if the judgment is no, the preset second security measure is executed to protect the safety of materials stored in the internal area; Wherein, in response to the triggering operation of detecting and identifying the object in the external area, generating regional security monitoring information based on a preset security detection module specifically includes: In response to a trigger operation of detecting and identifying an object in the external area, a camera recognition unit based on a preset safety detection module acquires a range recognition image, and generates a facial recognition information and a gesture recognition information set according to the range recognition image, wherein the safety detection module is pre-set on the protective door and window, and the safety detection module includes a camera recognition unit, a grating detection unit and a pressure detection unit; generating a posture recognition data set and preliminary body shape recognition data according to the posture recognition information set; Acquiring infrared detection data based on the grating detection unit, and generating a stay time sequence and auxiliary body shape recognition data according to the infrared detection data; generating body shape recognition information according to the preliminary body shape recognition data and the auxiliary body shape recognition data; Generating identity recognition information according to the facial recognition information and the body shape recognition information, and comparing the identity recognition information with the resident information in a preset database to generate a comparison similarity, wherein the database is preset and stores the resident information of the resident of the house; Determine whether the comparison similarity is greater than a preset similarity threshold, and if not, obtain pressure detection information based on the pressure detection unit, and generate regional safety monitoring information according to the identity recognition information, the posture recognition data set, the stay time sequence and the pressure detection information; Wherein, generating security assessment information based on a preset convolutional neural network according to the regional security monitoring information specifically includes: Extracting features of the identity recognition information through the convolutional neural network to generate an identity recognition feature vector, and generating a posture feature vector through the PoseNet model according to the posture recognition data set; Generate a time series feature vector through the convolutional neural network according to the residence time series, and generate a pressure feature vector through the convolutional neural network according to the pressure detection information; Generate a fusion feature parameter according to the identity recognition vector, the posture feature vector, the time series feature vector and the pressure feature vector; The fused feature parameters are input into a preset risk assessment model to generate safety assessment information, wherein the safety assessment information includes a risk estimation type and a risk index.

2. The regional security monitoring method based on building protection doors and windows according to claim 1 is characterized in that: The external area includes a door area and a window area, and the safety detection module is arranged in the door area and the window area; The first security measure includes the following steps: Determine whether the trigger operation originates from the door area or the window area; If it is determined that the trigger operation originates from the door area, a door closing instruction is generated and sent to a preset door switch module to close the room door; Generate an alarm instruction and send it to a preset alarm module, and control the alarm module to issue an alarm message; If it is determined that the trigger operation originates from the window area, a door opening instruction is generated and sent to a preset door switch module to open the room door; Generate an alarm instruction and send it to the alarm module, and control the alarm module to issue an alarm message; Generate a gate opening command and send it to the preset gate switch module to open the protective door.

3. The regional security monitoring method based on building protection doors and windows according to claim 1 is characterized in that: The second security measure comprises the following steps: Generate a door closing command and send it to the preset door switch module to close the room door; Generate an alarm instruction and send it to a preset alarm module, and control the alarm module to issue an alarm message; The security assessment information is sent to the mobile terminal and sent to the cloud for storage.

4. The regional security monitoring method based on building protection doors and windows according to any one of claims 1 to 3 is characterized in that: The access statistics information is obtained by following the steps below: In response to the door opening operation, the identity recognition information and the infrared detection data are obtained, and the walking direction information of the personnel is generated by an edge detection algorithm according to the infrared detection data; and the personnel change information is generated according to the identity recognition information and the walking direction information of the personnel; In response to the door closing operation, acquiring indoor image data of the interior area based on a preset indoor camera unit, and generating verification identification information according to the indoor image data; Comparing the verification identification information with the identity identification information to generate a confidence level, and determining whether the confidence level is greater than a preset confidence threshold; If the judgment is yes, the personnel change information is marked to generate personnel number information and time record data, and access statistics information is generated according to the personnel change information, the personnel number information and the time record data; If the judgment is no, an abnormal event record is generated and sent to the mobile terminal for user confirmation and feedback information is generated. The personnel change information is corrected according to the feedback information to generate corrected change information, and access statistical information is generated according to the corrected change information, the personnel number information and the time record data.

5. The regional security monitoring method based on building protection doors and windows according to claim 1 is characterized in that: The risk index is generated based on the following formula: ； Among them, R is the risk index, B is the body recognition similarity, F is the face recognition similarity, Wu is the identity weight parameter, Wv is the posture weight parameter, S is the posture matching degree, T is the person's stay time, Tmax is the total detection time, Wx is the time weight parameter, P is the detection pressure value, Pe is the pressure detection threshold, Wy is the pressure weight parameter, and N is the number of pressure detections.

6. A regional security monitoring system based on building protective doors and windows, which is applied to the security monitoring of the internal area and the external area of ​​the protective doors and windows in the building, characterized in that: include: A monitoring information generating module, configured to generate regional security monitoring information based on a preset security detection module in response to a triggering operation of detecting and identifying an object in the external region; An assessment information generation module, used to generate safety assessment information based on a preset convolutional neural network according to the regional safety monitoring information; A risk index determination module is used to determine whether the risk index is greater than a preset risk threshold; A personnel situation determination module, for obtaining pre-stored access statistics information if the determination is yes, and further determining whether there are personnel in the internal area according to the access statistics information; A security measure execution module, for executing a preset first security measure to protect the safety of personnel in the internal area if the judgment is yes; and for executing a preset second security measure to protect the safety of materials stored in the internal area if the judgment is no; Wherein, the monitoring information generation module is also used for: In response to a trigger operation of detecting and identifying an object in the external area, a camera recognition unit based on a preset safety detection module acquires a range recognition image, and generates a facial recognition information and a gesture recognition information set according to the range recognition image, wherein the safety detection module is pre-set on the protective door and window, and the safety detection module includes a camera recognition unit, a grating detection unit and a pressure detection unit; generating a posture recognition data set and preliminary body shape recognition data according to the posture recognition information set; Acquiring infrared detection data based on the grating detection unit, and generating a stay time sequence and auxiliary body shape recognition data according to the infrared detection data; generating body shape recognition information according to the preliminary body shape recognition data and the auxiliary body shape recognition data; Generating identity recognition information according to the facial recognition information and the body shape recognition information, and comparing the identity recognition information with the resident information in a preset database to generate a comparison similarity, wherein the database is preset and stores the resident information of the resident of the house; Determine whether the comparison similarity is greater than a preset similarity threshold, and if not, obtain pressure detection information based on the pressure detection unit, and generate regional safety monitoring information according to the identity recognition information, the posture recognition data set, the stay time sequence and the pressure detection information; Wherein, the evaluation information generating module is also used for: Extracting features of the identity recognition information through the convolutional neural network to generate an identity recognition feature vector, and generating a posture feature vector through the PoseNet model according to the posture recognition data set; Generate a time series feature vector through the convolutional neural network according to the residence time series, and generate a pressure feature vector through the convolutional neural network according to the pressure detection information; Generate a fusion feature parameter according to the identity recognition vector, the posture feature vector, the time series feature vector and the pressure feature vector; The fused feature parameters are input into a preset risk assessment model to generate safety assessment information, wherein the safety assessment information includes a risk estimation type and a risk index.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images