Resident community personnel safety management system and method
Technical Field
The invention relates to the field of artificial intelligence-based community personnel safety management, in particular to a community personnel safety management method compatible with vehicle passenger and temporary access personnel management.
Background
Along with the development of artificial intelligence and network communication technology, the face recognition type intelligent access control management system replacing an access control card is more and more widely applied and gradually spread to common residential districts from an initial security type unit, so that personnel for confirming identity can enter and exit through brushing face, and the safety and management convenience of personnel management are improved. Face recognition has many unique advantages over other recognition techniques, such as: 1) the image is collected in a non-contact manner, the concealment is strong, the psychological acceptance of all people is high, and the image is easy to be accepted by users. Since it is non-contact, the arrangement can be hidden if necessary. 2) The face has uniqueness, is difficult to counterfeit, is easy to obtain and identify, and is high in accuracy, safe and reliable. 3) The equipment for collecting the human face is low in price, low in cost, easy to operate and capable of being widely used. The face recognition system commonly used at present is mainly based on an embedded intelligent system, and the embedded system is developed to be generally combined with a network at present. The intelligent terminal of the embedded system is a necessary trend of face recognition development under the comprehensive consideration of multiple aspects such as universality, safety, maturity, cost and the like, is mature and applied to an access control system at present, and related products are produced in recent years.
In the prior art, the basis of face recognition is face detection, and the target of face detection is to acquire a face effective region. In many current algorithms, a human face can be accurately positioned and image acquisition can be performed during the process of human face detection, so as to be used for the subsequent human face feature extraction. Before extracting features, the detected image is generally preprocessed, which includes normalization and other processing methods, then some face feature representation method is adopted to extract the detected face features, and the known face image stored in the database is represented by the same method. Common representation methods include geometric features (such as Euclidean distance, curvature, and angle), algebraic features (such as matrix eigenvectors), fixed-feature templates, and the like. At present, there are several types of face recognition methods in the prior art: (1) the identification method based on the geometric features comprises the following steps: the eyebrow, eye, nose, mouth, ear and other parts of the human face are distributed at different positions of the human face, the geometric relationship is fixed, and the human face can be expressed according to the characteristic difference of the parts on the human face. Specifically, the IZOD curvature, the angle, the Euclidean distance between two points, and the like of the features on the face can be represented by a geometric feature vector, and the feature vectors form a face feature vector matrix. And realizing face recognition, and classifying the face by using a classifier according to the vector differences. (2) Statistical-based face recognition methods, such as principal component analysis, are the most commonly used algorithms in all current algorithms. (3) The method based on the connection mechanism comprises the following steps: such as an elastic graph matching method and an artificial neural network method. In addition, the method also comprises a face recognition method based on data fusion and the like.
The key step of face recognition is face recognition of a detected face image. The identification process is that the input human face features are matched with the human face features registered in the human face data set, the matching score value is compared with a set threshold value, if the similarity is high, the detected person is considered to be in the authority, the identification result can be sent to an access control system, a door lock is opened, and otherwise, the entrance is refused. Because the face recognition process is greatly affected by changes of hairstyle, light, shelters and the like, a standby recognition system is generally needed in the face recognition access control system.
In the existing patent, CN 106952368A discloses a gatekeeper internet management system and a control method thereof, including an artificial intelligence visitor all-in-one machine, a human-computer interaction module, an image acquisition module, an image recognition module, a gate control module, an intelligent terminal device and a network device.
However, in this method, after the visitor needs to install and register as the visitor APP software client user, a mutual trust relationship is established to the visitor APP software client through the internet, and then the visitor APP software client sets the number of times that the visitor enters the door access within a specified time range, so that the visitor has an access right. Therefore, the application management process is tedious, wastes a large amount of time of visitors and community residents, and is not convenient to operate.
CN 201811603146 discloses a novel access control management system and a detection method, comprising a gate module, a camera module, a control module, an infrared induction module, an alarm module and a storage module, wherein the gate module, the camera module and the infrared induction module are connected with the control module; the gate module is used for controlling the access of a user, the camera module is used for shooting face images of the user, the infrared sensing module is used for detecting whether a person arrives, the control module is used for carrying out face recognition on the user and judging the user permission, and the storage module is used for storing resident information of an access control management area. Although the application realizes the automatic management of the residential housing access of the residential housing by setting the authority of the visiting person and matching with the face recognition technology, the visitor management lacks convenience, for example, the number of times of authorized access gate entry and exit of the visiting person in the storage module of the system on the visiting day needs to be set, however, the number of times of access to and exit from the residential housing is often random, and especially for express delivery personnel, the number of times of access is unpredictable. Causing serious inconvenience to the visitors. It only enables management of the resident and, in addition, no management measures are given for the personnel who enter the cell following.
CN 208521347 provides a villa face recognition entrance guard security protection video telephone integrated system, which comprises a main control chip, a face recognition module, a video telephone module, a door control module, a USB face acquisition camera, a face database module and a monitoring data storage module; the gate control module, the face recognition module, the video telephone module, the face database module and the monitoring data storage module are all connected with the main control chip; the USB face acquisition camera is respectively connected with the face recognition module and the video telephone module; the main control chip is used for controlling the cooperative work among all the modules of the all-in-one machine. Although this application has realized face identification and has opened the door, and accessible USB face acquisition camera carries out the video recording and stores relevant influence data to the condition before the door, has improved the security performance, but this access control function is relatively more single, can not carry out safety control to all ring edge borders and personnel, only is applicable to rural single villa.
CN 109102610 discloses a safety access control system for security of district security, relates to access control system technical field. The invention comprises a camera shooting and photographing device, an image database, an infrared sensing monitoring device, a data input device, a voice alarm device and a gate opening and closing device; the device comprises an image information acquisition unit, an image data analysis unit, a monitoring information acquisition unit, a data analysis and comparison unit and an information processing unit. Although this application can be discerned the personnel of punching the card, avoided non-district personnel to pick up entrance guard's card and get into the district, however, this system need set up the number that infrared sensing monitoring devices monitored and got into the district through the gate, need input the number that gets into the district when punching the card, so operate and just can remind security personnel, though can avoid non-this district personnel to follow the resident family in theory and get into the district, but actual operability is not strong, especially when the time personnel of working at last or next work are intensive, verify through the mode of input number that the traffic efficiency has seriously reduced and can't prevent the situation that the follow sneaks into.
In conclusion, the existing face recognition access control system is lack of tightness and comprehensiveness, and particularly cannot overcome the defect that a stranger enters or leaves along with the system, so that the safety of face recognition management is reduced. In addition, strange passengers in the vehicle entering the community through the vehicle entrance bring about a loophole of community personnel safety management, and the prior art is difficult to realize corresponding face recognition and image acquisition.
Moreover, with frequent visits of express delivery personnel, decoration personnel or residents, visitors are difficult to access, verification procedures are complicated, a large amount of time is wasted, and the management of the express delivery personnel who compete for seconds is particularly disadvantageous.
Therefore, for the existing personnel conditions of residential communities, it is necessary to improve the existing community security management method, and an intelligent community security management system and method with high security, comprehensive management, rapidness, convenience and simplicity are needed.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a community personnel safety management system and a community personnel safety management method compatible with vehicle and external visitor management based on an artificial intelligence system, so as to solve the problems that the efficiency of artificial community personnel safety management is not high, and particularly, the effective identification management cannot be carried out on personnel entering the community along with a vehicle and visitors.
The artificial intelligence mainly designs an intelligent access control system based on face detection and recognition by means of face detection and recognition technology, and the intelligent access control system is used as a main component of community personnel safety management.
Wherein the community security management is also applicable to the situation without entrance guard.
The community personnel safety management system at least comprises subsystem modules such as a vehicle management system, a multi-information acquisition point system, an intelligent access control management system, a face detection and recognition system, a database storage system, a visitor management system and the like, and a central server, wherein all the subsystems are communicated with the central server.
The intelligent access control system comprises an intelligent access control system at the entrance and exit position of a community and an intelligent access control system of a unit door of each unit of a residential building. The intelligent access control system comprises embedded intelligent equipment terminals. The embedded intelligent device terminal comprises a high-definition camera, an electronic control lock device, a face detection identification module and other system components connected with a community center server, and further comprises a standby identification module such as a fingerprint collection identification module. The camera is used for actively collecting face images, the face detection and recognition module is used for detecting, recognizing and comparing the face images in corresponding areas, the electronic control lock is opened after the face images are compared successfully, personnel are allowed to pass through the electronic control lock, the comparison failure reaches a certain number of times, the standby recognition system is started, after the standby recognition system is matched and failed, alarm information is sent to a server of the management system, suspicious personnel information is displayed on a terminal display screen, and community management personnel intervene.
Furthermore, the community personnel safety management system also comprises a standby identification system such as fingerprint identification or two-dimensional code identification containing personal information, so that the defect that the people cannot pass due to identification failure caused by reasons such as shielding, light and the like of a face identification system can be avoided.
The main flow of the face recognition is that each access control terminal image acquisition assembly and equipment of other image acquisition points acquire a face image of a user through an image acquisition module, and the face image is sent to a face feature extraction module to extract a face feature part after image processing; then the central server processing end calls a face recognition module to recognize and compare the extracted face features. Furthermore, a storage module is arranged for storing the collected face data, and a display module displays the face recognition result; and the early warning module carries out alarm prompt when finding an illegal face image.
Specifically, the main steps of face recognition in the community personnel safety management method of the invention are as follows:
1) establishing a face information database of the community permanent people: the face image of the community permanent person is collected, the face characteristic information is extracted, and the face characteristic information is stored as a permanent face information database.
2) After the camera of the face detection acquisition module acquires a face image, the face image is preprocessed by a processor of the face recognition module, face characteristic information is extracted and is compared and recognized with face information stored in a face information database.
The face acquisition equipment is at least arranged at a community entrance/exit entrance guard end, each building unit entrance guard end and a garage personnel entrance/exit (such as an elevator or a stair entrance/exit), and respectively acquires face information of a community main entrance/exit, each unit door and garage entrance/exit personnel.
Preferably, the face information database includes a temporary face information database in addition to the permanent face information database. The permanent face information database stores face information of community residents and community property, security, cleaning and other working personnel; the temporary face information database stores face information of temporary personnel such as relatives and friends of residents in the community, express delivery and the like, can be set to have timeliness, and is deleted after being stored for a period of time.
Further, the face information database is associated with spare identification information such as fingerprint information or personal two-dimensional codes and corresponding personal information such as identity numbers, contact ways, accurate addresses and the like.
Furthermore, the invention also comprises the steps of intelligent video monitoring of public areas such as high altitude and ground of the outer vertical surface of the floor to monitor the abnormal behaviors of the high altitude parabolic and ground personnel.
Specifically, in order to achieve the above object, the present invention provides a community personnel safety management method compatible with vehicle management and visitor management based on an artificial intelligence system, which comprises the following steps.
Step S1: and managing the personnel passing at the community entrance.
Firstly, collecting face images of community personnel, extracting face characteristic information, and establishing a community face information database. Secondly, the intelligent access control equipment at the entrance and exit position of the community is used as a first information acquisition and identification point, the intelligent access control equipment comprises a face image acquisition module and a face detection and identification module, and the information of the face image acquired on site can be identified and compared with the information in the community face information database.
Specifically, this step S1 includes the following specific steps.
Step S1-1, establishing a community face information database: the face image of community personnel is collected, the face characteristic information is extracted, and the face characteristic information is stored as a face information database. The community face information database comprises a permanent face information sub-database and a temporary face information sub-database; the frequent living face information is face information of community residents and community property, security, cleaning and other workers; the temporary face information is face information of temporary personnel such as relatives and friends of community residents, express delivery and the like.
Preferably, the face information storage time of the temporary person may be set by itself, for example, a day, a month, or a year.
Step S1-2, when the approach of the face is detected, the intelligent access control equipment at the entrance and exit position of the community collects the current face information on site through the image collection module and carries out face feature extraction through the face recognition module, then carries out comparison matching recognition with the face information stored in the face information database, and judges whether the faces are the same or not according to the comparison of the matching similarity and the set threshold value; and after the comparison is successful, automatically opening the entrance guard to allow the entrance.
Step S1-3, after the matching fails for a plurality of times, the standby identification system is automatically started, and when the standby identification system can not pass, the intelligent access control device sends alarm information to a central server of the management system and displays suspicious personnel information on a terminal display screen, and community management personnel intervene in management such as security protection and the like; meanwhile, the obtained information of the face, the fingerprint and the like of the suspicious person is stored in a suspicious person information database.
Furthermore, the suspicious personnel information database is not only connected with a central server of the community management system, but also networked with a public security management system of a public security organization or a police system related to various suspicious personnel for information interaction.
Furthermore, before the suspicious personnel information database is transmitted to the public security management system, community personnel firstly carry out auditing confirmation, and the identification failure condition caused by external factors such as facial occlusion and the like is eliminated.
Step S2: and a face acquisition intelligent terminal device is arranged in the community entrance to acquire all face data entering the community and perform identification comparison. Face acquisition intelligent terminal equipment sets up in the entry direction towards community's entrance guard, sets up the light source in order to provide sufficient shooting ambient light at entrance guard's entry direction simultaneously.
The face collection intelligent terminal device is a second information collection identification point.
Specifically, the intelligent terminal device for face acquisition in this step is active face detection acquisition, and includes a face detection module capable of detecting a face, a face image acquisition module, and a network transmission module, or further includes a face recognition module. The intelligent face acquisition terminal device utilizes the high-definition camera assembly to capture and acquire detected face images, and simultaneously sends one or more acquired images to the server for feature extraction, identification, comparison and processing, so that stranger information entering a community along with the acquired images is obtained through screening, warning information is sent to the management system, and the face data is stored or displayed.
Preferably, in this step, the face acquisition intelligent terminal device may be a high-definition intelligent monitoring camera including a face detection snapshot module.
Step S3: and arranging an intelligent entrance guard comprising an embedded intelligent equipment terminal at each unit door to collect the face data entering the unit door and carry out recognition and comparison.
The unit door embedded intelligent equipment terminal is a third information acquisition identification point.
The entrance guard terminal image acquisition assembly acquires faces appearing in corresponding areas, and sends the faces to a face feature extraction module to extract face feature parts after image processing; and then the face recognition module carries out recognition comparison on the extracted face features and face feature information stored in a community face database, and determines whether the face entering the cell door is a suspicious person or not according to the similarity. And if the comparison is passed, executing automatic opening of the entrance guard of the unit door, and allowing the entrance.
Further, if the matching failure of the face image obtained by the information acquisition point and the face image stored in the face database exceeds a certain number of times, the standby recognition system is automatically started.
When the standby system can not be identified, the intelligent access control terminal sends prompt information to the server, and stores and displays the face data.
Step S4: a face collection intelligent terminal device is arranged at a garage personnel entrance and exit so as to collect all face data entering a community elevator or a stair from a garage and carry out recognition and comparison. The intelligent face acquisition terminal device is arranged in the garage entrance direction, and a light source is arranged in the direction to provide sufficient shooting environment light.
The face acquisition intelligent terminal equipment is a fourth information acquisition identification point.
In the step, the intelligent face acquisition terminal equipment is used for active face detection and acquisition, and acquires all face data entering a community from a garage and performs recognition ratio. The face acquisition intelligent terminal equipment utilizes the high-definition camera assembly to capture detected face images, and simultaneously sends one or more acquired images to the server for feature extraction, identification, comparison and processing so as to screen stranger information of vehicle personnel entering the community from the garage. And if the stranger exists, sending warning information to a management system, and storing or displaying the face data.
In this step, since a multi-person scene may appear at the entrance, when performing face recognition on the face image, a low recognition threshold may be set or a face recognition algorithm with a fast recognition speed may be used to obtain a relatively high recognition speed, for example, the threshold may be set to 60-70%.
Step S5: managing visitor personnel: the safety management system of the invention also comprises a visitor management subsystem used for managing visitors. The steps are as follows.
Step S5-1: when a visitor passes through a community entrance guard for the first time, an intelligent terminal device at the community entrance enters a visitor management interface, information of a person to be visited is input, and a visit application is submitted; the access is approved after the verification of the manager; then, the image acquisition module of the entrance guard intelligent terminal device at the entrance and exit of the community acquires the face image information of the visitor on site, and stores the face image information acquired on site into the storage module of the visitor management subsystem, so as to grant the visitor with the permission in a certain period of time.
Step S5-2: in this predetermined time quantum, when the visitor passes through the inside unit entrance guard of community or passes through the entrance guard of community once more, by the image acquisition module field acquisition visitor's face image information of entrance guard terminal equipment, the identification module draws face characteristic information and discerns the matching with the face information that prestores in the visitor's subsystem, opens access control after matching qualified, the permission gets into.
Preferably, when visitor personnel enter the community entrance guard, the intelligent terminal device collects identity card and other certificate information and stores the identity card and the face image in an associated mode.
The management system improves the access convenience and the entrance guard safety through the visitor subsystem, and solves the problem that visitors temporarily enter a cell, particularly a unit entrance guard is difficult. The system is particularly suitable for the in-and-out management of express delivery and other personnel.
Step S6: a ground-high altitude combined three-dimensional video monitoring system is arranged for monitoring abnormal behaviors: the video monitoring system preferably comprises an intelligent video monitoring system of a video analysis module. The ground video monitoring system is mainly used for monitoring public areas such as community squares, roads and the like; the high-altitude video monitoring system is provided with a plurality of fixed or rotatable monitoring camera sets for monitoring high-altitude areas covering the outer facades of all floors of the community.
Step S6-1: the process of monitoring the abnormal behavior of the ground personnel by the ground video is as follows.
1) The video analysis module of the community ground video monitoring system is integrated with a plurality of personnel motion detection algorithms and is pre-stored with a motion mathematical model database module;
2) carrying out video monitoring on the community ground and acquiring video data; detecting a moving target, particularly a personnel target, according to a motion recognition algorithm, and continuously tracking to obtain a motion track characteristic;
3) comparing the motion track characteristics with track templates of abnormal behaviors in a model database; the abnormal behaviors comprise behavior characteristics such as falling, lying, personnel gathering, pursuing running and the like;
4) and sending a warning prompt to the control center according to the comparison result, and carrying out manual intervention management.
Step S6-2: the high-altitude video monitoring system is used for monitoring abnormal behaviors such as high-altitude parabolic behavior. The high-altitude video monitoring system comprises a video analysis module integrated with a moving object identification algorithm, and a motion mathematical model database module is prestored; the video analysis module can be used for intelligently analyzing the movement behaviors of the videos acquired by the high-definition camera and extracting corresponding pictures, filtering the pictures by objects such as birds and fallen leaves according to models prestored in the database, and then sending the data to the management terminal and giving warning prompts.
Step S7: management of children's entrance and exit to and from community: the child face information needs to be associated with the family member face information, and an access control system at a community entrance and exit is set to have no independent passing permission. That is, when the face recognition module of the community access control equipment acquires the face information of a child who leaves the door independently through the camera, the access control permission is not granted in the face database even if the face information is in the face database; at the moment, the management system sends out a warning prompt and informs or contacts the relevant guardian through telephone, short message and other modes through manual intervention.
Preferably, the face information of children under six years old is not stored in the community face database.
The management method of the present invention further includes a vehicle management step S8, as follows.
1) When a vehicle enters a community area, relevant equipment of a community vehicle management system at an entrance of a community vehicle takes a picture to collect vehicle picture information and uploads the vehicle picture information to a system server, the vehicle management system compares the collected vehicle picture with a community vehicle stored in a database, and if the comparison is successful, a control channel gate is started to release; if there is the garage in the community and when there is remaining parking stall in the garage, the passageway floodgate machine that can open the garage entry need not to compare through the secondary when the floodgate machine equipment of garage entry senses the vehicle and is close.
Wherein, the floodgate machine equipment of garage entry replaces vehicle identification unit module with vehicle response control module, avoids carrying out secondary discernment to the vehicle and compares, improves the current efficiency of vehicle.
The vehicle management system comprises a vehicle picture acquisition unit, a vehicle information identification unit and a channel gate control unit, wherein the vehicle information identification unit compares vehicle information captured by a camera with vehicle information in a storage database.
2) When a strange vehicle enters a community area, the reserved vehicle and the non-reserved vehicle are managed separately.
When a vehicle is ready to enter a cell for reservation, the vehicle information is registered in advance through a community management system or a telephone mode, wherein the vehicle information at least comprises a license plate number and can further comprise information such as vehicle type, vehicle color and the like.
When the reserved vehicle enters the community area, the system related equipment at the entrance of the community vehicle captures the vehicle information and compares the vehicle information with the vehicle information in the storage database, after the comparison is successful, the passage gate is controlled to release, and if the comparison fails for a certain number of times, prompt information is automatically uploaded to a manager.
The comparison of the vehicle information at least comprises license plate number comparison, and preferably further comprises vehicle type comparison and vehicle color comparison.
3) When an unreserved vehicle (such as a taxi and a license plate-free vehicle) temporarily enters a community, a vehicle management system at the entrance of a community vehicle identifies an unconserved strange vehicle through comparison, sends out a warning prompt that the strange vehicle is about to enter the community, and determines whether the vehicle is released through the intervention of management personnel.
In conclusion, the community personnel safety management system based on simple, quick and low-cost face recognition improves the community personnel safety management efficiency with complex personnel and fussy safety management, carries out high-precision detection in key areas of the entrance guard, and combines with the video monitoring of public areas of the community to realize comprehensive community personnel safety management compatible with visitor management and with higher concealment.
The invention also has the beneficial effect of overcoming the defect that the prior face recognition access control system can not overcome the defect that a stranger enters or leaves along with the system. In addition, the blind area of community personnel safety management brought by the strange passengers in the vehicle entering the community through the vehicle entrance is effectively overcome, and all strange passengers entering the community can be comprehensively brought into the face recognition and image acquisition of the management system. Meanwhile, for visitor management such as express delivery personnel, decoration personnel or resident relatives and friends, verification procedures are greatly reduced, and a large amount of time in the visitor verification process is saved.
Detailed Description
The present invention is described below with reference to specific examples, but these examples do not limit the actual scope of the present invention in any way, and do not limit the scope of the present invention.
Implementation mode one
A community personnel safety management method based on an artificial intelligence system comprises a multi-information acquisition point system, an intelligent access control management system, a face detection and recognition system, a database storage system, a visitor management system and other subsystem modules, and a central server. The method comprises the following steps.
Step S1: managing the passage of personnel at the community entrance and exit: the specific flow is as follows.
Step S1-1, establishing a community face information database: the face image of community personnel is collected, the face characteristic information is extracted, and the face characteristic information is stored as a face information database. The community face information database comprises a permanent face information sub-database and a temporary face information sub-database; the frequent living face information is face information of community residents and community property, security, cleaning and other workers; the temporary face information is face information of temporary personnel such as relatives and friends of community residents, express delivery and the like.
Preferably, the face information storage time of the temporary person may be set to three months.
Step S1-2, when the approach of the face is detected, the intelligent access control equipment at the entrance and exit position of the community collects the current face information on site through the image collection module and carries out face feature extraction through the face recognition module, then carries out comparison matching recognition with the face information stored in the face information database, and judges whether the faces are the same or not according to the comparison of the matching similarity and the set threshold value; and after the comparison is successful, automatically opening the entrance guard to allow the entrance.
Specifically, during face detection, face detection is performed by using an AdaBoost algorithm based on Haar features, and face positioning of an acquired image is realized by installing and calling a library function and an xml file in OpenCV (open circuit interface 2. xml file and related functions), so that a preprocessed gray face image is acquired. The face recognition is to compare the face to be recognized with a trained face database, and can adopt Euclidean distance to calculate the similarity, if the similarity is greater than a set threshold, the corresponding face identity is displayed and the door is opened.
In the step, the similarity of the two is analyzed according to the characteristics so as to carry out face recognition comparison, and the threshold value can be set to be 0.7, namely, the recognition success can be confirmed when the similarity reaches 70%.
In this step, the comparison of the face images may be set as follows:
1) after the image normalization processing, the geometric data of the positions and relative distances of the eyes, the nose, the mouth, the chin and the like which form the main features of the face are used for recognizing different human faces.
For example, the feature points may be right corner of the eye, left corner of the eye, nose tip, right corner of the mouth, left corner of the mouth, outer brow endpoint, inner brow endpoint, outer brow endpoint, upper orbital endpoint, lower orbital endpoint, nasion endpoint, alar endpoint, lip endpoint.
Wherein, the eye center point can be determined according to the external end point, the upper end point, the lower end point and the internal end point of the orbit; for example, the left eye center point is determined according to the left eye extraorbital end point, the left eye supraorbital end point, the left eye infraorbital end point and the left eye intraorbital end point, and the right eye center point is determined in the same way.
And determining the central point of the face according to the midpoint of the connecting line of the central point of the right eye and the central point of the left eye.
The method specifically comprises the following steps: distance feature vectors are constructed based on the feature points, for example, feature vector parameters as shown in table 1 below are constructed.
TABLE 1 feature point-based feature vector parameters
Parameter name
|
Meaning of feature vector
|
d1
|
Distance between center point of face and nose tip point
|
d 2
|
Distance between center point of face and left mouth angle
|
d 3
|
Distance between center point of face and right mouth angle
|
d 4
|
Distance between the center of the left and right corners of the mouth and the nose tip
|
d 5
|
Distance between left and right mouth corners
|
d 6
|
Distance between center points of left and right eyes
|
d 7
|
Distance between left canthus and nose tip
|
d 8
|
Distance between right canthus and nose tip
|
d 9
|
Distance between upper and lower endpoints of right orbit
|
d 10
|
Distance between upper and lower endpoints of left orbit |
2) And obtaining the histogram component of each parameter according to the extracted feature vector so as to construct a histogram and calculate the similarity.
Furthermore, 3-5 images are taken to extract features respectively, and the average value of similarity is taken for use.
Step S1-3, automatically starting a fingerprint standby identification system after face matching fails for 3 times, and when the standby identification system can not pass, sending alarm information to a central server of a management system by the intelligent access control device, displaying suspicious personnel information on a terminal display screen, and performing intervention management by community management personnel such as security protection; meanwhile, the acquired information of the face, the fingerprint and the like of the suspicious person is stored in a suspicious person information database for data information filing; if necessary, the information of the suspicious personnel is transmitted to the public security management system of the public security organization after the verification and confirmation of the community personnel.
Step S2: a face collection intelligent terminal device with a high-definition camera is arranged in a community entrance to collect all face data entering a community and perform recognition and comparison. Face acquisition intelligent terminal equipment sets up in the entry direction towards community's entrance guard, sets up the light source in order to provide sufficient shooting ambient light at entrance guard's entry direction simultaneously. The face collection intelligent terminal device is a second information collection identification point. The intelligent face acquisition terminal equipment can be video monitoring equipment with a face detection and recognition module.
The method comprises the steps of utilizing a high-definition camera assembly to capture and collect detected face images, sending one or more collected images to a server for feature extraction and identification comparison processing, obtaining stranger information which enters a community along with the images through screening, sending warning information to a management system, and storing or displaying the face data.
In the step, a commercial application third-party class library OpenCV in the field of face recognition is used to realize efficient face recognition.
Exemplarily, the flow of extracting the face features by using the Haar feature classifier in the OpenCV library is as follows.
1) Loading a function library, setting a Haar feature classifier, and completing feature extraction and classification through an xml file corresponding to the classifier;
2) detecting a face in a picture using a face detection function haarddetect:
CvSeq @ haarddetect (cvar @, cvhaarclasssiccacade @, CvSizemin _ size). Cas is a classifier for face recognition.
3) After a feature data set of five sense organs is obtained according to the classifier, the data is processed to extract position data information of feature points, and then a feature matrix of distance values of a plurality of human face feature points is established, so that human face feature vector data are obtained. Further, in step S1, a histogram is constructed based on the histogram component of each parameter of the feature vector, and the similarity is calculated.
Wherein, the threshold value of the similarity of the second information acquisition points can be set to be medium to low, so as to improve the identification efficiency. For example, set to 0.6 or 0.7, so that strangers who are not in the community face information database are quickly screened.
Step S3: and arranging an intelligent entrance guard comprising an embedded intelligent equipment terminal at each unit door to collect the face data entering the unit door and carry out recognition and comparison. The unit door embedded intelligent equipment terminal is a third information acquisition identification point.
The entrance guard terminal image acquisition assembly acquires faces appearing in corresponding areas, and sends the faces to a face feature extraction module to extract face feature parts after image processing; and then the face recognition module carries out recognition comparison on the extracted face features and face feature information stored in a community face database, and determines whether the face entering the cell door is a suspicious person or not according to the similarity. And if the comparison is passed, executing automatic opening of the entrance guard of the unit door, and allowing the entrance. Further, if the matching failure of the face image obtained by the information acquisition point and the face image stored in the face database exceeds 3 times, the standby recognition system is automatically started. When the standby system can not be identified, the intelligent access control terminal sends prompt information to the server, and stores and displays the face data.
The face recognition operation in this step is the same as the above step.
The threshold value of the similarity comparison is set to be medium or high, for example, 0.8 to 0.9.
Step S4: a face collection intelligent terminal device is arranged at a garage personnel entrance and exit so as to collect all face data entering a community elevator or a stair from a garage and carry out recognition and comparison. The intelligent face acquisition terminal device is arranged in the garage entrance direction, and a light source is arranged in the direction to provide sufficient shooting environment light. The face acquisition intelligent terminal equipment is a fourth information acquisition identification point.
In the step, the face acquisition intelligent terminal equipment utilizes the high-definition camera assembly to capture and acquire detected face images, and simultaneously sends one or more acquired images to the server to perform feature extraction and identification comparison processing so as to screen stranger information of vehicle passengers entering the community from the garage. If the stranger exists, warning information is sent to the management system, and the face data is stored or displayed for data recording.
In this step, since a multi-person scene may appear at the entrance, when performing face recognition on the face image, a low recognition threshold is set or a face recognition algorithm with a fast recognition speed is used to obtain a relatively high recognition speed, for example, the threshold may be set to 0.6 or 0.7.
When the intelligent terminal device for face acquisition adopts a high-definition video monitoring system, the intelligent terminal device at least comprises a face detection module and an image acquisition module. The main flow is as follows.
1) Acquiring a video frame from a camera image; obtaining a real-time picture through function conversion in OpenCV;
2) when a face is detected, acquiring a face area and generating a target face;
3) extracting the face information, comparing the face information with the face information in the face database, identifying strangers, sending warning information to a management system, and storing or displaying the face data for data recording.
Step S5: and managing the visitors entering the community based on the visitor management subsystem. The steps are as follows.
Step S5-1: when a visitor passes through a community entrance guard for the first time, an intelligent terminal device at the community entrance enters a visitor management interface, information of a person to be visited is input, and a visit application is submitted; the access is approved after the verification of the manager; then, the image acquisition module of the entrance guard intelligent terminal device at the entrance and exit of the community acquires the face image information of the visitor on site, and stores the face image information acquired on site into the storage module of the visitor management subsystem, so as to grant the visitor with the permission in a certain period of time.
Step S5-2: in this predetermined time quantum, when the visitor passes through the inside unit entrance guard of community or passes through the entrance guard of community once more, by the image acquisition module field acquisition visitor's face image information of entrance guard terminal equipment, the identification module draws face characteristic information and discerns the matching with the face information that prestores in the visitor's subsystem, opens access control after matching qualified, the permission gets into.
Furthermore, when the visitor enters the community entrance guard, the intelligent terminal device collects identity card and other certificate information and stores the identity card and the face image in an associated mode.
Step S6: a ground-high altitude combined three-dimensional video monitoring system is arranged for monitoring abnormal behaviors: the video monitoring system is an intelligent video monitoring system with a video analysis module. The ground video monitoring system is mainly used for monitoring public areas such as community squares, roads and the like; the high-altitude video monitoring system is provided with a plurality of fixed or rotatable monitoring camera sets for monitoring high-altitude areas covering the outer facades of all floors of the community. The details are as follows.
Step S6-1: the process of monitoring the abnormal behavior of the ground personnel by the ground video is as follows.
1) The video analysis module of the community ground video monitoring system is integrated with a plurality of personnel motion detection algorithms and is pre-stored with a motion mathematical model database module;
2) carrying out video monitoring on the community ground and acquiring video data; detecting a moving target, particularly a personnel target, according to a motion recognition algorithm, and continuously tracking to obtain a motion track characteristic;
3) comparing the motion track characteristics with track templates of abnormal behaviors in a model database; the abnormal behaviors comprise behavior characteristics such as falling, lying, personnel gathering, pursuing running and the like;
4) and sending a warning prompt to the control center according to the comparison result, and carrying out manual intervention management.
Step S6-2: the high-altitude video monitoring system is used for monitoring abnormal behaviors such as high-altitude parabolic behavior. The high-altitude video monitoring system comprises a video analysis module integrated with a moving object identification algorithm, and a motion mathematical model database module is prestored; the video analysis module can be used for intelligently analyzing the movement behaviors of the videos acquired by the high-definition camera and extracting corresponding pictures, filtering the pictures by objects such as birds and fallen leaves according to models prestored in the database, and then sending the data to the management terminal and giving warning prompts.
Step S7: management of children's entrance and exit to and from community:
the child face information is associated with the family member face information, and an access control system at the community entrance and exit is set to have no independent passing permission. When a face recognition module of community access control equipment acquires face information of a child who singly exits through a camera, access control access permission is not granted in a face database even if the face information is in the face database; at the moment, the management system sends out a warning prompt and informs or contacts the relevant guardian through telephone, short message and other modes through manual intervention.
Further, the face information of children under six years old is not stored in the community face database.
Second embodiment
The management method according to the present embodiment further includes a vehicle management step in addition to the first embodiment, and the vehicle management step is specifically as follows.
1) When a vehicle enters a community area, relevant equipment of a community vehicle management system at an entrance of a community vehicle takes a picture to collect vehicle picture information and uploads the vehicle picture information to a system server, the vehicle management system compares the collected vehicle picture with a community vehicle stored in a database, and if the comparison is successful, a control channel gate is started to release; if there is the garage in the community and when there is remaining parking stall in the garage, the passageway floodgate machine that can open the garage entry need not to compare through the secondary when the floodgate machine equipment of garage entry senses the vehicle and is close.
Wherein, the floodgate machine equipment of garage entry replaces vehicle identification unit module with vehicle response control module, avoids carrying out secondary discernment to the vehicle and compares, improves the current efficiency of vehicle.
The vehicle management system comprises a vehicle picture acquisition unit, a vehicle information identification unit and a channel gate control unit, wherein the vehicle information identification unit compares vehicle information captured by a camera with vehicle information in a storage database.
2) When an external strange vehicle enters a community area, the reserved vehicle and the non-reserved vehicle are managed separately.
2.1) when an external vehicle is ready to enter a cell in a reservation mode, the vehicle information is required to be registered and stored in advance through a community management system or a telephone mode, wherein the vehicle information at least comprises a license plate number and can further comprise information such as vehicle type, vehicle color and the like.
When the reserved vehicle enters the community area, the system related equipment at the entrance of the community vehicle captures the vehicle information and compares the vehicle information with the vehicle information in the storage database, after the comparison is successful, the passage gate is controlled to release, and if the comparison fails for a certain number of times, prompt information is automatically uploaded to a manager.
The comparison of the vehicle information at least comprises license plate number comparison, and further comprises vehicle type comparison and vehicle color comparison.
2.2) when non-reserved vehicles such as taxis and license plates without numbers enter the community temporarily, identifying the non-reserved strange vehicles by comparing through a vehicle management system at the entrance of the community vehicle, giving a warning prompt that the strange vehicles are about to enter the community, and determining whether to pass through by the intervention of management personnel.
It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Further, various changes, modifications and/or alterations of the present invention may be made by those skilled in the art, all of which fall within the scope of the present invention as defined in the appended claims.