An artificial intelligence-oriented access control monitoring system
By installing cameras at enterprise access control points, establishing a mobile route database, and analyzing personnel attention and eye direction, the problem of internal theft has been solved, achieving high-precision security monitoring and post-event tracking.
Patent Information
- Application Number
- CN202411081167.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-08
AI Technical Summary
Existing intelligent access control and monitoring systems are difficult to effectively detect theft by employees within enterprises, especially during periods when no one is present, due to insufficient accuracy in identification and behavioral analysis capabilities.
By installing cameras at enterprise access control points, a database of common movement routes is established to monitor personnel location and attention in real time. Combined with infrared ranging and eye gaze direction analysis, abnormal behavior is identified, spatial areas are divided and intersection points are recorded, and abnormal personnel are screened out.
It achieves high recognition rate for multi-area, time-segmented security monitoring, improves access control security detection capabilities, provides traceable evidence after the fact, and meets the security needs of enterprise equipment and facilities.
Smart Images

Figure CN118629137B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of access control monitoring data management, and in particular to an access control monitoring system oriented to artificial intelligence. Background Art
[0002] The intelligent access control monitoring system is a new type of modern security management system that integrates automatic identification technology and modern security management measures, involving many new technologies such as electronics, machinery, computer technology, and communication technology.
[0003] In recent years, the application scope of digital image technology has become more and more extensive, and identity authentication using biometric technology has become a research hotspot in the current field of artificial intelligence due to its huge application scenarios in security monitoring systems. In the existing technology, enterprises have introduced face recognition technology to actual access control monitoring to prevent suspicious persons from entering. With the continuous improvement of the monitoring system, the possibility of external personnel entering the enterprise to steal has been greatly reduced, but there may still be internal personnel stealing, especially during meal time, when few people stay in the work area, and the risk of enterprise equipment and facilities being stolen has been greatly increased. It is necessary to design a method to monitor the access control of the work area when the work area is unattended to prevent theft in the enterprise. Therefore, it is necessary to design an access control monitoring system for artificial intelligence with high behavior detection accuracy and strong target behavior analysis capabilities. Summary of the invention
[0004] The purpose of the present invention is to provide an artificial intelligence-oriented access control monitoring system to solve the problems raised in the above-mentioned background technology.
[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: an access control monitoring system for artificial intelligence, including an access control monitoring method for artificial intelligence, wherein the access control monitoring method mainly includes the following steps:
[0006] Step S1: access the access control monitoring method at the processing terminal of the enterprise access control system, and monitor the personnel entering and leaving the enterprise gate access control through two cameras arranged above the access control device;
[0007] Step S2: Establish a common movement route database for enterprise personnel, record and store all common travel routes of enterprise personnel, and when the abnormal monitoring of the access control gate is turned on, retrieve the common movement route database. When the camera unit detects the characteristics of a person in its monitoring direction, it sends a signal, and the movement route database marks the specific location of the current enterprise personnel in the route database according to the location of the signal.
[0008] Step S3: Locate the corresponding position of the enterprise personnel in real time, and determine whether the enterprise personnel have any abnormal behavior when passing through the access control gate by monitoring their attention. When the system identifies that the movement behavior of the monitored target person is abnormal, monitor the time when the target person appears again. When it is identified that the target person passes through the gate access control again within the limit time, the video information is intercepted and stored.
[0009] According to the above technical solution, step S2 includes:
[0010] Step S21: after the access control system detects that at least L enterprise personnel have left the work area, and during the interval period when at least L enterprise personnel have not entered the work area, start abnormal monitoring of the access control gate, where the specific value of L is 60% of the total number of people authorized to obtain the enterprise work area on that day;
[0011] Step S22: Retrieve the commonly used mobile route database. When the camera unit detects the characteristics of a person within its monitoring direction, it sends a signal. The mobile route database marks the specific location of the enterprise personnel in the route database according to the location of the signal.
[0012] According to the above technical solution, step S3 further includes:
[0013] Step S31: the system performs infrared ranging on the target person, and measures the relative distance between the target person and the target person in the instantaneous image captured when the target person enters the access control gate;
[0014] Step S32: When the relative distance is detected to be less than V meters, the instantaneous picture is locked, and the front and rear endpoints of the two feet and the coordinates of the middle point of the two feet of the target person's feet outline in the instantaneous picture are marked, and the front and rear endpoints of the two feet are connected respectively, and the marked points are connected in sequence with a smooth curve in the monitoring area, and the route is marked as the monitoring person's trend fitting route, where V is the minimum value of the shooting picture distance detection;
[0015] Step S33: The system obtains the eye gaze direction information of the target person through monitoring, and analyzes the correlation between the trend fitting route information of the monitored target person and the eye gaze direction information of the monitored target person.
[0016] According to the above technical solution, step S33 includes:
[0017] Step S331: In the camera's captured image, a spatial rectangular coordinate system is constructed, with the midpoint of the bottom of the enterprise's access control door as the coordinate origin, and a straight line passing through the coordinate origin and coinciding with the bottom of the enterprise door is established as the X-axis. In the plane at the bottom of the access control door, a straight line passing through the coordinate origin and parallel to the X-axis and entering the access control door from the outside is the positive semi-axis of the Y-axis. A spatial rectangular coordinate system perpendicular to the X-axis and the Y-axis is the positive semi-axis of the Z-axis is constructed through the coordinate origin, wherein the coordinates of the midpoint of the bottom of the enterprise's access control door are (0, 0, 0), and the unit side length of the spatial rectangular coordinate system is 1;
[0018] Step S332: The system divides the coverage area that the enterprise access control monitoring camera can capture and identify features into K spatial areas, wherein the coverage area of the enterprise access control monitoring system is a closed area composed of the K system-divided spatial areas, and marks the spatial area where the fitting route of the personnel's movement trend overlaps with the bottom plane in the spatial area with the largest area;
[0019] Step S333: The camera transmits the captured access control gate image information to the access control monitoring processing terminal. The processing terminal extracts the head image of the target person when entering the access control gate based on the image information, obtains the intersection point of the target person's eye gaze direction and the spatial area, and determines whether the movement behavior of the monitored target person is abnormal.
[0020] According to the above technical solution, in step S332, the specific method of dividing the space area is:
[0021] When the access control gate is fully opened and the four corners of the passable area are parallel to each other, K spatial areas are extended in a straight line at different angles to the coverage area of the enterprise access control monitoring system, and then closed after extending to a distance that the access control gate monitoring camera cannot recognize. The 8 boundary points distributed in a single spatial area of the K spatial areas are (A0, 0, 0), (-A0, 0, 0), (A0, B0, 0), (-A0, B1, 0), (A0, 0, C0), (-A0, 0, C0), (A0, B0, C0), (-A0, B1, C0), where A0 is half the width of the door frame of the access control gate, B0 is the farthest Y-axis distance in the picture taken by the access control monitoring toward one side of the gate at the corresponding extension angle of the right side of the area and the feature can be identified, B1 is the farthest Y-axis distance in the picture taken by the access control monitoring toward one side of the gate at the corresponding extension angle of the left side of the area and the feature can be identified, and C0 is the height value of the access control monitoring setting position from the ground. In the coverage area, the area that the straight line cannot extend to is not recorded as the actual detection area.
[0022] According to the above technical solution, step S333 includes:
[0023] Step S3331: In the spatial rectangular coordinate system constructed by the system, the position of the center of the back of the target person's head is obtained as (A, B, C) based on the image feature recognition technology;
[0024] Step S3332: automatically focus on the left eye of the target person according to the eye features, and calibrate the position of the left eye, inner eye corner and outer eye corner of the target person with the first picture information, take the left eye eyeball convergence center point as the origin, the line segment from the inner eye corner to the outer eye corner as the x-axis, rotate the x-axis 90 degrees in the inner and outer eye corner planes as the y-axis, and establish a plane rectangular coordinate system through the coordinate origin perpendicular to the eye corner plane as the z-axis, set the length of the x-axis as P, the length of the y-axis as Q, and the length of the z-axis as U, where the unit side length of the plane rectangular coordinate system is 0.1, and the current left eye eyeball convergence center point position is (0, 0, 0), continuously obtain the left eye eyeball convergence center point position, and after real-time scaling and comparison with the first picture information, obtain the eyeball convergence center point position (x, y, z), then the real-time position of the eyeball convergence center point transmitted by the system is (x1, y1, z1), (x2, y2, z2)... (x n ,y n , z n ), where n is the total number of times the system monitors the left eyeball position of the target person within the monitoring time period T, and the left eyeball position of the target person is monitored at each interval. The average change distance within ;
[0025] Step S3333: Extend the ray from the target person's back head coordinates (A, B, C) and the eyeball convergence center point to intersect with all target space areas, record all intersection points with K target space areas, count the number of intersection points falling in the marked target space area as K1, and the number of intersection points falling in the unmarked target space area as K2, then the current person's attention degree coefficient to the walking route , where δ is the unit conversion coefficient;
[0026] Step S3334: When the system detects that the target person's attention coefficient to the walking route is lower than the limit value J0, it is determined that the current target person's movement behavior is abnormal.
[0027] According to the above technical solution, the access control monitoring system includes an access control monitoring module, a target behavior analysis module and an abnormal data collection module. The access control monitoring module is used to monitor people passing through the enterprise access control gate through monitoring equipment and capture picture information within the monitoring time period; the target behavior analysis module is used to analyze the target person in the captured picture information to determine whether the target person may exhibit abnormal behavior; the abnormal data collection module is used to collect the video data of monitoring abnormal behavior.
[0028] According to the above technical solution, the access control monitoring module includes an access control system module and a video monitoring module. The access control system module includes a host unit and a switch unit. The multiple host units are connected to the switch unit; the video monitoring module includes multiple camera units, video recorder units, and TV screen units. The multiple camera units installed in the same work area are connected to the video recorder unit through a network. The video recorder unit is connected to the TV screen unit, and the video recorder unit is used to record video information and transmit it to the TV screen unit; the host units are respectively installed on the enterprise gate to control the opening and closing of the enterprise gate, and the host units in the enterprise work area are connected to the switch unit arranged in the work area; the camera unit and the host unit are matched, including a camera installed outside the enterprise gate to monitor the situation of people entering and leaving the access control gate.
[0029] According to the above technical solution, the target behavior analysis module includes a trend route fitting module and an abnormality monitoring module. The trend route fitting module is used to mark the trend fitting route of the monitored personnel and record the route information; the abnormality monitoring module is used to determine whether there is an abnormal situation when the target detection personnel passes through the access control gate through the analysis results of the correlation between the trend fitting route information of the monitored target personnel and the eye gaze direction information of the monitored target personnel.
[0030] According to the above technical solution, the abnormality monitoring module includes a spatial area division submodule, an eye focus point detection submodule and an attention degree coefficient analysis submodule. The spatial area division submodule is used to divide the coverage area that the enterprise access control surveillance camera can capture and identify features; the eye focus point detection submodule is used to monitor the specific coordinates of the left eye focus position of the target person in the coordinate system; the attention degree coefficient analysis submodule is used to obtain the intersection point of the target person's eye gaze direction and the spatial area, detect the target person's attention degree coefficient, and determine whether the monitored target person's movement behavior is abnormal.
[0031] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention, by calling up a database of commonly used mobile routes, monitors personnel characteristics and marks the specific locations of current enterprise personnel, and by monitoring the attention of enterprise personnel when passing through the access control gate, determines whether the target personnel have abnormal behavior, and screens out personnel with abnormal situations in the process of different monitoring targets passing through the access control, so that the image information captured by the monitoring camera and the processing terminal cooperate with each other and are comprehensively used in the access control monitoring system to achieve the goals of high recognition rate and intuitive tracking afterwards, realize multi-area and time-divided security monitoring, greatly improve the ability of gate access control entry and exit security detection, and meet the high requirements of enterprises for the safety of enterprise equipment and facilities. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:
[0033] Figure 1 It is a flow chart of an artificial intelligence-oriented access control monitoring system operation method provided by the present invention;
[0034] Figure 2 It is a schematic diagram of the module composition of an artificial intelligence-oriented access control monitoring system provided by the present invention. DETAILED DESCRIPTION
[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0036] See also Figure 1 The present invention provides a technical solution: an artificial intelligence-oriented access control monitoring system, comprising:
[0037] An artificial intelligence-oriented access control monitoring method, the access control monitoring method mainly includes the following steps:
[0038] Step S1: access the access control monitoring method at the processing terminal of the enterprise access control system, and monitor the personnel entering and leaving the enterprise gate access control through two cameras arranged above the access control device;
[0039] Step S2: Establish a common movement route database for enterprise personnel, record and store all common travel routes of enterprise personnel, and when the abnormal monitoring of the access control gate is turned on, retrieve the common movement route database. When the camera unit detects the characteristics of a person in its monitoring direction, it sends a signal, and the movement route database marks the specific location of the current enterprise personnel in the route database according to the location of the signal.
[0040] Step S3: Locate the corresponding position of the enterprise personnel in real time, and determine whether the enterprise personnel have abnormal behavior when passing through the access control gate by monitoring their attention. When the system identifies that the movement behavior of the monitored target person is abnormal, monitor the time when the target person appears again. When it is identified that the target person passes through the gate access control again within the limit time, the video information is intercepted and stored, and the specific value of the limit time is extracted through the monitoring database.
[0041] The embodiments of the present invention monitor the characteristics of personnel and mark the specific locations of current enterprise personnel by calling up a database of commonly used mobile routes, determine whether the target personnel have abnormal behavior by monitoring the attention of the enterprise personnel when passing through the access control gate, and screen out personnel with abnormal behavior in the process of different monitoring targets passing through the access control, so that the image information captured by the monitoring camera and the processing terminal cooperate with each other and are comprehensively used in the access control monitoring system to achieve the goals of high recognition rate and intuitive tracking afterwards, realize multi-area and time-divided security monitoring, greatly improve the ability of gate access control entry and exit security detection, and meet the high requirements of enterprises for the safety of enterprise equipment and facilities.
[0042] In a preferred embodiment, step S2 comprises:
[0043] Step S21: after the access control system detects that at least L enterprise personnel have left the work area, and during the interval period when at least L enterprise personnel have not entered the work area, start abnormal monitoring of the access control gate, where the specific value of L is 60% of the total number of people authorized to obtain the enterprise work area on that day;
[0044] Step S22: Retrieve the commonly used mobile route database. When the camera unit detects the characteristics of a person within its monitoring direction, it sends a signal. The mobile route database marks the specific location of the enterprise personnel in the route database according to the location of the signal.
[0045] This technical solution solves the problem of high load caused by long-term uninterrupted monitoring. By building a mobile route database, when a person is detected entering the work area, monitoring of the target person is started, and the time period for theft monitoring is determined. This can greatly reduce the operating load of the monitoring system while ensuring the accuracy of monitoring security.
[0046] In a preferred embodiment, step S3 further comprises:
[0047] Step S31: the system performs infrared ranging on the target person, and measures the relative distance between the target person and the target person in the instantaneous image captured when the target person enters the access control gate;
[0048] Step S32: When the relative distance is detected to be less than V meters, the instantaneous picture is locked, and the front and rear endpoints of the two feet and the coordinates of the middle point of the two feet of the target person's feet outline in the instantaneous picture are marked, and the front and rear endpoints of the two feet are connected respectively, and the marked points are connected in sequence with a smooth curve in the monitoring area, and the route is marked as the monitoring person's trend fitting route, where V is the minimum value of the shooting picture distance detection;
[0049] Step S33: The system obtains the eye gaze direction information of the target person through monitoring, and analyzes the correlation between the trend fitting route information of the monitored target person and the eye gaze direction information of the monitored target person.
[0050] When the target person is detected outside the corporate work area facing the access control gate, the timing unit monitors the time the target person continues to appear in the monitoring screen. When the time exceeds the safety detection time, the video data on this side is recorded and stored. If the target person appears in the monitored area when no one is there, and has at least one round trip and lingers in the access control monitoring area for a long time, he / she will be marked as a target person who needs to be paid special attention.
[0051] In this embodiment, step S33 includes:
[0052] Step S331: In the camera's captured image, a spatial rectangular coordinate system is constructed, with the midpoint of the bottom of the enterprise's access control door as the coordinate origin, and a straight line passing through the coordinate origin and coinciding with the bottom of the enterprise door is established as the X-axis. In the plane at the bottom of the access control door, a straight line passing through the coordinate origin and parallel to the X-axis and entering the access control door from the outside is the positive semi-axis of the Y-axis. A spatial rectangular coordinate system perpendicular to the X-axis and the Y-axis is the positive semi-axis of the Z-axis is constructed through the coordinate origin, wherein the coordinates of the midpoint of the bottom of the enterprise's access control door are (0, 0, 0), and the unit side length of the spatial rectangular coordinate system is 1;
[0053] Step S332: The system divides the coverage area that the enterprise access control monitoring camera can capture and identify features into K spatial areas, wherein the coverage area of the enterprise access control monitoring system is a closed area composed of the K system-divided spatial areas, and marks the spatial area where the fitting route of the personnel's movement trend overlaps with the bottom plane in the spatial area with the largest area;
[0054] Step S333: The camera transmits the captured access control gate image information to the access control monitoring processing terminal. The processing terminal extracts the head image of the target person when entering the access control gate based on the image information, obtains the intersection point of the target person's eye gaze direction and the spatial area, and determines whether the movement behavior of the monitored target person is abnormal.
[0055] In step S332 of this embodiment, the specific method of dividing the space area is:
[0056] When the access control gate is fully open and the four corners of the passable area are parallel to each other, K spatial areas are extended in a straight line at different angles in the area covered by the enterprise access control monitoring system. They are closed after being extended to a distance that the access control gate monitoring camera cannot recognize. The 8 boundary points distributed in a single spatial area of the K spatial areas are (A0, 0, 0), (-A0, 0, 0), (A0, B0, 0), (-A0, B1, 0), (A0, 0, C0), (-A0, 0, C0), (A0, B0, C0), ( -A0, B1, C0), where A0 is half the width of the door frame of the access control gate, B0 is the farthest Y-axis distance that can identify features in the image taken by the access control monitoring on the right side of the gate with the corresponding extension angle, B1 is the farthest Y-axis distance that can identify features in the image taken by the access control monitoring on the left side of the gate with the corresponding extension angle, and C0 is the height value of the access control monitoring setting position from the ground. In the coverage area, the area that cannot be extended by a straight line is not recorded as the actual detection area. The actual area of the actual detection area is small and has little impact on the monitoring results. By dividing the spatial area, the attention direction of the target person can be further accurately determined, providing a strong basis for the subsequent detection of the direction of eye attention.
[0057] In this embodiment, step S333 includes:
[0058] Step S3331: In the spatial rectangular coordinate system constructed by the system, the position of the center of the back of the target person's head is obtained as (A, B, C) based on the image feature recognition technology;
[0059] Step S3332: automatically focus on the left eye of the target person according to the eye features, and calibrate the position of the left eye, inner eye corner and outer eye corner of the target person with the first picture information, take the left eye eyeball convergence center point as the origin, the line segment from the inner eye corner to the outer eye corner as the x-axis, rotate the x-axis 90 degrees in the inner and outer eye corner planes as the y-axis, and establish a plane rectangular coordinate system through the coordinate origin perpendicular to the eye corner plane as the z-axis, set the length of the x-axis as P, the length of the y-axis as Q, and the length of the z-axis as U, where the unit side length of the plane rectangular coordinate system is 0.1, and the current left eye eyeball convergence center point position is (0, 0, 0), continuously obtain the left eye eyeball convergence center point position, and after real-time scaling and comparison with the first picture information, obtain the eyeball convergence center point position (x, y, z), then the real-time position of the eyeball convergence center point transmitted by the system is (x1, y1, z1), (x2, y2, z2)... (x n ,y n , z n ), where n is the total number of times the system monitors the left eyeball position of the target person within the monitoring time period T, and the left eyeball position of the target person is monitored at each interval. The average change distance within ;
[0060] Step S3333: Extend the ray from the target person's back head coordinates (A, B, C) and the eyeball convergence center point to intersect with all target space areas, record all intersection points with K target space areas, count the number of intersection points falling in the marked target space area as K1, and the number of intersection points falling in the unmarked target space area as K2, then the current person's attention degree coefficient to the walking route , where δ is the unit conversion coefficient, and the same eyeball convergence direction can intersect with multiple spatial regions at multiple intersection points;
[0061] Step S3334: When the system detects that the target person's attention coefficient to the walking route is lower than the limit value J0, it is determined that the current target person's movement behavior is abnormal.
[0062] This technical solution can detect the target person's eye direction and screen whether the target person has abnormal attention when passing through the access control gate normally. If the target person's eye deviation is too many times, it means that the possibility of abnormality has increased. By comparing the target person's eye deviation with the actual walking route, the possibility of the target person being abnormal in the access control area can be judged. When a possible theft is detected, the video is recorded and stored to provide evidence for finding the thief in time when the subsequent theft occurs. The action information of the enterprise personnel in the enterprise work area is analyzed, and the corresponding control signal emission intensity is controlled, so that the system can make different control preparation times according to different signal strengths; the process of monitoring the direction of the target person's eye gaze when passing through the access control monitoring system room is carried out, and the observation swing of the target person's eye gaze direction during the monitoring period is detected, which can greatly increase the monitoring density and improve the monitoring accuracy of the access control monitoring system.
[0063] See also Figure 2 The access control monitoring system includes an access control monitoring module, a target behavior analysis module and an abnormal data collection module. The access control monitoring module is used to monitor the personnel passing through the enterprise access control gate through monitoring equipment and capture the image information within the monitoring time period; the target behavior analysis module is used to analyze the target personnel in the captured image information and determine whether the target personnel may have abnormal behavior; the abnormal data collection module is used to collect the video data of monitoring abnormal behavior.
[0064] The access control monitoring module includes an access control system module and a video monitoring module. The access control system module includes a host unit and a switch unit. Multiple host units are connected to the switch unit; the video monitoring module includes multiple camera units, video recorder units, and TV screen units. Multiple camera units installed in the same work area are connected to the video recorder unit through a network. The video recorder unit is connected to the TV screen unit. The video recorder unit is used to record video information and transmit it to the TV screen unit. Enterprise managers can view video information by calling the TV screen unit; the host units are installed on the enterprise gate respectively to control the opening and closing of the enterprise gate. The host units in the enterprise work area are connected to the switch unit located in the work area; the camera unit and the host unit are set up in conjunction with each other, including a camera installed outside the enterprise gate to monitor the situation of people entering and leaving the access control gate.
[0065] The target behavior analysis module includes a trend route fitting module and an anomaly monitoring module. The trend route fitting module is used to mark the trend fitting route of the monitored personnel and record the route information; the anomaly monitoring module is used to determine whether there is any abnormality in the target detection personnel passing through the access control gate by analyzing the correlation between the trend fitting route information of the monitored target personnel and the eye gaze direction information of the monitored target personnel.
[0066] The abnormality monitoring module includes a spatial area division submodule, an eyeball gathering point detection submodule and an attention degree coefficient analysis submodule. The spatial area division submodule is used to divide the coverage area that the enterprise access control surveillance camera can capture and identify features; the eyeball gathering point detection submodule is used to monitor the specific coordinates of the left eyeball gathering position of the target person in the coordinate system; the attention degree coefficient analysis submodule is used to obtain the intersection point of the target person's eye gaze direction and the spatial area, detect the target person's attention degree coefficient, and determine whether the monitored target person's movement behavior is abnormal.
[0067] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0068] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. An artificial intelligence-oriented access control monitoring system, characterized in that: An artificial intelligence-oriented access control monitoring method is provided, and the access control monitoring method mainly comprises the following steps: Step S1: access the access control monitoring method at the access control monitoring processing terminal of the access control monitoring system, and monitor the personnel entering and leaving the enterprise gate access control through two enterprise access control monitoring cameras arranged above the access control device; Step S2: Establish a database of commonly used mobile routes of target enterprise personnel, record and store all commonly used travel routes of target enterprise personnel, and retrieve the database of commonly used mobile routes of target enterprise personnel when abnormal monitoring of the enterprise access control gate is turned on. When the enterprise access control surveillance camera detects the characteristics of a person within its monitoring direction, it sends a signal, and the commonly used mobile route database of target enterprise personnel marks the specific location of the monitored target enterprise personnel in the database of commonly used mobile routes of target enterprise personnel according to the detected characteristic position of the person; Step S3: Locate the corresponding position of the target enterprise personnel in real time, determine whether the target enterprise personnel have abnormal behavior when passing through the enterprise access control gate by monitoring their attention when passing through the enterprise access control gate, and transmit corresponding monitoring instructions to the access control monitoring system; The step S2 comprises: After the access control monitoring system detects that at least L target enterprise personnel have left the enterprise work area, and during the interval period when at least L target enterprise personnel have not entered the enterprise work area, the abnormal monitoring of the enterprise access control gate is started, where the specific value of L is 60% of the total number of people authorized to obtain the enterprise work area on that day; The step S3 further comprises: Step S31: the access control monitoring system performs infrared ranging on the target enterprise personnel, and measures the relative distance between the target enterprise personnel and the target enterprise personnel in the instantaneous image captured when the target enterprise personnel enters the enterprise access control gate; Step S32: When the relative distance is detected to be less than V meters, the instantaneous picture is locked, the front and rear end points of the two feet and the coordinates of the middle point of the two feet of the target enterprise personnel in the instantaneous picture are marked, the front and rear end points of the two feet are connected respectively, and the marked points are connected in the monitoring area in turn with smooth curves, and the route is marked as the fitting route of the trend of the target enterprise personnel, where V is the minimum value of the distance detection of the shooting picture; Step S33: the access control monitoring system obtains the eye gaze direction information of the target enterprise personnel through monitoring, and analyzes the correlation between the trend fitting route information of the target enterprise personnel and the eye gaze direction information of the target enterprise personnel.
2. The artificial intelligence-oriented access control monitoring system according to claim 1, characterized in that: The step S33 comprises: Step S331: In the shooting picture of the enterprise access control surveillance camera, a spatial rectangular coordinate system is constructed, with the midpoint of the bottom of the enterprise access control door as the coordinate origin, and a straight line passing through the coordinate origin and coinciding with the bottom of the enterprise door is established as the X-axis. In the plane at the bottom of the enterprise access control door, a straight line passing through the coordinate origin parallel to the X-axis and entering the enterprise access control door from the outside is the positive semi-axis of the Y-axis. A spatial rectangular coordinate system perpendicular to the X-axis and the Y-axis is the positive semi-axis of the Z-axis is made through the coordinate origin, wherein the coordinates of the midpoint of the bottom of the enterprise access control door are (0, 0, 0), and the unit side length of the spatial rectangular coordinate system is 1; Step S332: The access control monitoring system divides the coverage area of the enterprise access control monitoring camera that can capture and identify features into K system-divided spatial areas, wherein the coverage area of the access control monitoring system is a closed area composed of the K system-divided spatial areas, and marks the spatial area where the fitting route of the personnel's movement trend overlaps with the bottom plane in the spatial area with the largest area; Step S333: the enterprise access control monitoring camera transmits the captured enterprise access control gate image information to the access control monitoring processing terminal, and the access control monitoring processing terminal extracts the head image of the target enterprise personnel when entering the enterprise access control gate based on the image information, obtains the intersection point of the target enterprise personnel's eye gaze direction and the spatial area, and determines whether the movement behavior of the target enterprise personnel is abnormal; Step S334: When the system identifies that the movement behavior of the target enterprise personnel is abnormal, the time when the target enterprise personnel appears again is monitored. When it is identified that the target enterprise personnel passes through the gate access control again within the limit time, the video information is intercepted and stored.
3. The artificial intelligence-oriented access control monitoring system according to claim 2, characterized in that: In step S332, the specific method of dividing the space area is: When the enterprise access control door is fully open and the four corners of the passable area are parallel to each other, K system-divided space areas are extended in straight lines at different angles in the access control monitoring system coverage area, and then closed after extending to a distance that the enterprise access control monitoring camera cannot recognize. The 8 boundary points distributed in a single space area of the K system-divided space areas are (A0, 0, 0), (-A0, 0, 0), (A0, B0, 0), (-A0, B1, 0), (A0, 0, C0), (-A0, 0, C0), (A0, B0, C0), (-A0, B1, C0), where A0 is half of the door frame width of the enterprise access control door, B0 is the farthest Y-axis distance in the picture taken by the access control monitoring to one side of the door at the corresponding extension angle of the right side of the area and the feature can be identified, B1 is the farthest Y-axis distance in the picture taken by the access control monitoring to one side of the door at the corresponding extension angle of the left side of the area and the feature can be identified, C0 is the height value of the access control monitoring setting position from the ground, and the area within the coverage area that the straight line cannot extend is not recorded as the actual detection area.
4. The artificial intelligence-oriented access control monitoring system according to claim 3 is characterized in that: The step S333 comprises: Step S3331: In the spatial rectangular coordinate system constructed by the system, the position of the center of the back of the head of the target enterprise personnel is obtained as (A, B, C) based on the image feature recognition technology; Step S3332: automatically focus on the left eye of the target enterprise personnel according to the eye features, and calibrate the positions of the left eye, inner eye corner and outer eye corner of the target enterprise personnel with the first picture information, take the left eye eyeball convergence center point as the origin, the line segment from the inner eye corner to the outer eye corner as the x-axis, rotate the x-axis 90 degrees in the inner and outer eye corner planes as the y-axis, and establish an empty plane rectangular coordinate system through the coordinate origin perpendicular to the eye corner plane as the z-axis, set the length of the x-axis as P, the length of the y-axis as Q, and the length of the z-axis as U, wherein the unit side length of the empty plane rectangular coordinate system is 0.1, and the current left eye eyeball convergence center point position is (0, 0, 0), continuously obtain the left eye eyeball convergence center point position, and after real-time scaling and comparison with the first picture information, obtain the eyeball convergence center point position (x, y, z), then the real-time position of the eyeball convergence center point transmitted by the system is (x1, y1, z1), (x2, y2, z2)... (x n ,y n , z n ), where n is the total number of times the system monitors the left eyeball focus position of the target enterprise personnel within the monitoring time period T, and the left eyeball focus position of the target enterprise personnel is in each interval time. The average change distance within ; Step S3333: Extend the ray from the coordinates (A, B, C) of the back of the head of the target enterprise personnel and the center point of the eyeball convergence to intersect with all target space areas, record all the intersection points with K target space areas, count the number of intersection points falling in the marked target space area as K1, and the number of intersection points falling in the unmarked target space area as K2, then the current personnel's attention degree coefficient to the walking route , where δ is the unit conversion coefficient; Step S3334: When the system detects that the target enterprise personnel's attention coefficient to the walking route is lower than the limit value J0, it is determined that the movement behavior of the current target enterprise personnel is abnormal.
5. The artificial intelligence-oriented access control monitoring system according to claim 4, characterized in that: The access control monitoring system includes an access control monitoring module, a target behavior analysis module and an abnormal data collection module. The access control monitoring module is used to monitor personnel passing through the enterprise access control gate through the enterprise access control monitoring camera and capture image information within the monitoring time period; the target behavior analysis module is used to analyze the monitored target enterprise personnel in the captured image information to determine whether the target enterprise personnel may exhibit abnormal behavior; the abnormal data collection module is used to collect the video data of monitoring abnormal behavior.
6. The artificial intelligence-oriented access control monitoring system according to claim 5, characterized in that: The access control monitoring module includes an access control system module and a video monitoring module. The access control system module includes a host unit and a switch unit. The host unit is connected to the switch unit. The video monitoring module includes multiple camera units, a video recorder unit, and a TV screen unit. The multiple camera units installed in the same enterprise work area are connected to the video recorder unit through a network. The video recorder unit is connected to the TV screen unit. The video recorder unit is used to record video information and transmit it to the TV screen unit. The host units are respectively installed on the enterprise gate to control the opening and closing of the enterprise gate. The host units in the enterprise work area are connected to the switch unit located in the enterprise work area. The camera unit is matched with the host unit, including a camera installed outside the enterprise gate to monitor the situation of people entering and leaving the access control gate.
7. The artificial intelligence-oriented access control monitoring system according to claim 6, characterized in that: The target behavior analysis module includes a trend route fitting module and an anomaly monitoring module. The trend route fitting module is used to mark the trend fitting route of the target enterprise personnel and record the route information; the anomaly monitoring module is used to determine whether there is any abnormality in the target enterprise personnel passing through the enterprise access control gate through the analysis results of the correlation between the trend fitting route information of the target enterprise personnel and the eye gaze direction information of the target enterprise personnel.
8. The artificial intelligence-oriented access control monitoring system according to claim 7, characterized in that: The abnormality monitoring module includes a spatial area division submodule, an eyeball gathering point detection submodule and an attention degree coefficient analysis submodule. The spatial area division submodule is used to divide the coverage area that the enterprise access control monitoring camera can capture and identify features; the eyeball gathering point detection submodule is used to monitor the specific coordinates of the left eyeball gathering position of the target enterprise personnel in the coordinate system; the attention degree coefficient analysis submodule is used to obtain the intersection point of the target enterprise personnel's eye gaze direction and the spatial area, detect the attention degree coefficient of the target enterprise personnel, and determine whether the target enterprise personnel's movement behavior is abnormal.
