WO2017202169A1

WO2017202169A1 - Access control data processing method, access control method, and electronic apparatus

Info

Publication number: WO2017202169A1
Application number: PCT/CN2017/081884
Authority: WO
Inventors: 王时全; 黄飞跃; 吴永坚; 陈志博
Original assignee: 腾讯科技（深圳）有限公司
Priority date: 2016-05-25
Filing date: 2017-04-25
Publication date: 2017-11-30
Also published as: CN105913528B; CN105913528A

Abstract

An access control data processing method comprises: acquiring an image sequence (402); acquiring, according to the image sequence, personal activity information (404); detecting, according to the personal activity information, a person having intension of entering (406); acquiring location information of the person (407); performing, according to the location information of the person, facial recognition (408); and outputting, when a user identifier associated with the recognized face has an entry permission, a door opening command (409).

Description

Access control data processing method, access control method and electronic device

This application claims to be submitted to the Chinese Patent Office on May 25, 2016, the application number is 201610353026.3, and the title of the invention is the priority of the Chinese patent application entitled "Access Control Data Processing Method and Apparatus, Access Control Method and Apparatus", the entire contents of which are The citations are incorporated herein by reference.

Technical field

The present invention relates to the field of access control technologies, and in particular, to an access control data processing method, an access control method, and an electronic device.

Background technique

Access control is a preventive measure against the door. The access control can guarantee the personal safety of human beings. The doors here include various passages that can pass, such as residential doors and passage gates. At present, automatic access control can be realized to a certain extent. For example, a camera is installed at the door. The outside person can actively align the camera in front of the door, and the face image is obtained by the camera and sent to the access control system, and the face is controlled by the access control system. Identification, thereby using the recognition result to control the passage and closing of the door.

However, the current access control mode requires personnel to actively cooperate with the camera to complete face recognition, and waiting for personnel to cooperate to complete the face recognition time consumption is difficult to predict, resulting in inefficient access control.

Summary of the invention

According to various embodiments of the present application, an access control data processing method, an access control method, and an electronic device are provided.

A method for processing access control data includes:

Obtaining a sequence of images;

Obtaining human activity information according to the image sequence;

Detecting a person who meets the intention of entering the door according to the information of the personnel activity;

Obtaining location information of the person;

Performing face recognition based on the location information of the person; and

The door opening instruction is output when the recognized user ID of the face belongs to the door opening authority.

An access control method includes:

Obtaining location information of the person who meets the intention of entering the door; the person who meets the intention of entering the door is detected according to the information of the personnel activity; and the information of the activity of the person is obtained according to the image sequence;

Performing face recognition based on the location information of the person;

Obtaining the user ID to which the recognized face belongs; and

The door opening instruction is output when the user identification has a door opening authority.

An electronic device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor such that the processor performs the following steps:

Obtaining a sequence of images;

Obtaining human activity information according to the image sequence;

Obtaining location information of the person;

Performing face recognition based on the location information of the person;

Obtaining the user ID to which the recognized face belongs; and

Details of one or more embodiments of the present application are set forth in the accompanying drawings and description below. Other features, objects, and advantages of the invention will be apparent from the description and appended claims.

DRAWINGS

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings to be used in the embodiments or the prior art description will be briefly described below. Obviously, the drawings in the following description are only It is a certain embodiment of the present application, and other drawings can be obtained from those skilled in the art without any creative work.

1 is an application environment diagram of an access control system in an embodiment;

2 is a schematic view showing the internal structure of a camera in an embodiment;

3 is a schematic diagram showing the internal structure of an electronic device in an embodiment;

4 is a schematic flow chart of a method for processing access control data in an embodiment;

FIG. 5 is a schematic diagram of a training and prediction process of a prediction model in an embodiment; FIG.

6 is a flow chart showing the steps of an abnormal alarm in one embodiment;

7 is a schematic flow chart of an access control method in an embodiment;

FIG. 8 is a structural block diagram of an electronic device in an embodiment; FIG.

9 is a structural block diagram of an electronic device in another embodiment;

FIG. 10 is a structural block diagram of an electronic device in an embodiment; FIG.

11 is a block diagram showing the structure of an electronic device in another embodiment.

detailed description

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

As shown in FIG. 1, in one embodiment, an access control system is provided, including a camera 110, a face acquisition camera 120, and an access control device 130. The camera 110 may specifically be a normal camera of a depth camera or a non-depth camera, and the depth camera can not only capture images but also capture images. The distance from the target camera in the captured image to the depth camera is obtained. The camera 110 is configured to capture a sequence of images, obtain human activity information according to the image sequence, detect location information of the person who meets the intention of entering the door according to the personnel activity information, and send the location information of the person to the access control device. The face capture camera 120 is configured to collect an image including a face, and the collected image including the face can be sent to the access control device 130. The access control device 130 may be a server or a device dedicated to the access control, and the access control device 130 may perform targeted face recognition on the image including the face according to the location information of the person, and have the user identifier of the recognized face. The door opening command is output when the door is opened.

As shown in FIG. 2, in one embodiment, a camera is provided that can be used as the camera 110 or the face acquisition camera 120. As shown in FIG. 2, the camera includes a processor coupled through a system bus, a non-volatile storage medium, an image sensor, an internal memory, and a communication interface. The non-volatile storage medium of the camera stores an operating system, and further stores computer readable instructions. When the computer readable instructions are executed by the processor, the processor can implement an access control data processing method. The processor is used to provide computing and control capabilities to support the operation of the entire camera. Computer readable instructions may be stored in the internal memory in the camera, the computer readable instructions being executable by the processor to cause the processor to perform an access control data processing method. The communication interface is for communicating with the access control device 130. Image sensors are used to capture images. It will be understood by those skilled in the art that the structure shown in FIG. 2 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation of the camera to which the solution of the present application is applied. The specific camera may include a ratio. More or fewer components are shown in the figures, or some components are combined, or have different component arrangements.

As shown in FIG. 3, in one embodiment, an electronic device is provided that can be used as the access control device 130. The electronic device includes a processor coupled through a system bus, a non-volatile storage medium, an internal memory, and a communication interface. The non-volatile storage medium of the electronic device stores an operating system, a database, and computer readable instructions. The database stores a correspondence between the user identifier and the opening permission. The computer readable instruction can be used to implement an access control method. The processor of the electronic device is used to provide computing and control capabilities to support the operation of the entire electronic device. Computer readable instructions may be stored in the internal memory of the electronic device, the computer readable instructions being The processor performs a method of access control. The communication interface is for communicating with the camera 110 and the face acquisition camera 120. The electronic device can be a server, and the server can be implemented by a separate server or a server cluster composed of multiple servers. A person skilled in the art can understand that the structure shown in FIG. 3 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the electronic device to which the solution of the present application is applied. The specific electronic device may be It includes more or fewer components than those shown in the figures, or some components are combined, or have different component arrangements.

As shown in FIG. 4, in an embodiment, a method for processing access control data is provided. This embodiment is exemplified by the method applied to the camera 110 in the access control system of FIG. The method specifically includes the following steps:

Step 402: Acquire an image sequence.

Specifically, the camera captures a sequence of images. The image sequence is a collection of images in which the images are arranged in chronological order of shooting. The images captured by the camera are arranged in chronological order to form an image sequence. The camera can be a top view camera. The bird's-eye view camera can take a bird's-eye view image to form a bird's-eye view sequence.

The camera can be mounted on the top or above the wall of the person's height or at the top corner, allowing the camera to capture an image of the top view, with the camera's field of view covering at least the door. The overhead image is an image taken down from a position higher than the height of the person, and may be a plan view or a view with an oblique angle.

The overhead image sequence is a sequence of images formed by chronologically arranging the overhead images captured by the camera, and the overhead image sequence may specifically be a video or a sequence of pictures. The time interval of the adjacent overhead image in the overhead image sequence can be set as needed. The frame rate of the overhead image sequence may be lower or higher than the frame rate threshold that the human eye can withstand. The frame rate of the overhead image sequence is the number of overhead images per unit time in the overhead image sequence, and the frame rate is high based on the sequence of the overhead image. The detection result is more accurate, and the frame rate is low, and the camera hardware resource consumption is small.

Step 404, obtaining personnel activity information according to the image sequence.

The personnel activity information is information that can reflect the activity of the personnel, and the camera can specifically detect each of the overhead images in the image sequence to obtain the personnel activity information. Camera can map The image sequence is sent to the server to receive the human activity information detected by the server and fed back by the server; the camera can also extract the feature for detecting the activity information of the person from the image sequence, and send the extracted feature to the server, and receive the server feedback. The activity information of the personnel obtained based on the extracted features.

The human activity information may include at least one of a human body position, a human body orientation, a human body motion direction, and a human head orientation. The camera may specifically perform at least one of human body detection, human body orientation detection, human motion direction detection, and human head orientation detection on each of the top view images in the image sequence, thereby obtaining at least one of a human body position, a human body orientation, a human body motion direction, and a human head orientation. A kind of information about personnel activities. Human body detection, body orientation detection, human motion direction detection, and head orientation detection can all be obtained by using a pre-trained classifier. The classifier can be a Support Vector Machine (SVM) or a neural network. The classifier can be based on The corresponding sample training of the top view angle is obtained.

Step 406: Detecting a person who meets the intention of entering the door according to the personnel activity information.

Step 407: Acquire location information of the person.

Specifically, after the camera obtains the activity information of the personnel, the camera may predict whether the corresponding person meets the intention of entering the door according to the personnel activity information, and if the person who meets the intention of entering the door obtains the position information of the corresponding person; Persons who meet the intent of entering the door continue to obtain information on the activity of the person based on the sequence of images.

The intention of entering the door reflects the probability of the corresponding person entering the door. If the intention of entering the door is met, it indicates that the probability of the corresponding person entering the door is high according to the information of the personnel activity; if the intention of entering the door is not met, the probability of the corresponding person entering the door is predicted according to the information of the personnel activity. . Whether or not the intent to enter the door can be obtained by inputting the personnel activity information into the pre-trained prediction model, and outputting whether or not the prediction result conforms to the intent of entering the door, and if so, obtaining the position information of the corresponding person.

Step 408: Perform face recognition according to the location information of the person.

Step 409: Output a door opening instruction when the identified user identifier of the face belongs to the door opening authority.

Specifically, the camera transmits location information of a person who meets the intention of entering the door to the access control device. The access control device acquires the location information of the person sent by the camera, and acquires the face acquisition camera The image captured by the head.

The access control device can use the unified coordinate system to map the position information of the person to the collected image, and perform face recognition on the face image matching the position information of the person in the image collected by the face acquisition camera to obtain a face. Identify the results.

The access control device determines the user identifier to which the face belongs according to the face recognition result, so as to query whether the user identifier has the right to open the door, and if the query has the permission to open the door, the door opening instruction is output. The door opening command can be output to a device that directly controls the passage and closing of the door, such as a door lock or a gate lock, so that the door can pass.

In an embodiment, the face recognition may employ a face recognition algorithm based on template matching, a subspace analysis face recognition algorithm, a Locality Preserving Projections (LPP) face recognition, a principal component analysis (PCA) algorithm, Eigenface method (based on KL transform), artificial neural network face recognition algorithm, and support vector machine face recognition algorithm. The core idea of face recognition algorithm based on template matching uses a person's facial feature law to establish a stereoscopic model frame. After positioning the face position of the person, the model frame is used to locate and adjust the facial features of the person. Factors such as observation angle, occlusion, and expression changes during face recognition.

The above-mentioned access control data processing method obtains the personnel activity information by using the image sequence, and uses the personnel activity information to detect the position information of the person who meets the intention of entering the door. The detected position information of the person is the position information of the person who meets the intention of entering the door, and then the targeted face recognition is performed according to the position information of the person, and the recognized face is used for the access control. In this way, the person who meets the intention of entering the door can be automatically judged, and the targeted face recognition can be performed, which can reduce the time consumption for the waiting person to actively cooperate with the face recognition, and improve the access control efficiency.

In one embodiment, the personnel activity information includes a human motion feature, an image sequence feature, and an image appearance feature. Step 406 specifically includes: detecting, by the trained predictive model, the human motion feature, the image sequence feature, and the image appearance feature. A person who meets the intention of entering the door.

Specifically, the human motion feature may include at least one of a human body position feature, a human body orientation feature, and a human head orientation feature, reflecting a feature of the human body in a bird's-eye view image. An image sequence feature is a feature between consecutive top-view images in an image sequence, including detection and follow-up between successive top-view images. Tracer changes in characteristics, such as human motion direction characteristics or human motion trajectory characteristics. The image appearance feature is a shape feature and a shape distribution feature in a top view image.

The prediction model can adopt a model such as a support vector machine or a neural network. Referring to FIG. 5, the corresponding human motion feature, image sequence feature and image apparent feature can be extracted from the training image sequence, and the specific training image sequence can be extracted by the human body detection and the human body to obtain the corresponding human motion feature, image sequence feature and The apparent features of the image, and then the prediction model can be obtained according to the motion characteristics of the person extracted from the training image sequence, the image sequence features and the apparent features of the image, and whether or not the marker data of the door is entered. When step 406 is performed, the access control device may input the human motion feature, the image sequence feature, and the image appearance feature obtained from the image sequence into the trained prediction model to obtain a prediction result that conforms to the intent of entering the door.

In this embodiment, the personnel motion state, the image sequence feature, and the image appearance feature can more accurately describe the personnel activity status, so that the prediction model can accurately determine whether the corresponding person has the door intention and improve the accuracy of detecting the door intention. Sex, which in turn increases the efficiency of access control.

In one embodiment, the image sequence is acquired by the depth camera; the access control data processing method further includes: acquiring the personnel distance information corresponding to the personnel activity information; and the step 406 includes: detecting the intent to enter the door according to the personnel activity information and the personnel distance information; Personnel.

Specifically, the depth camera is a camera that can acquire not only the image but also the distance between the target and the depth camera in the image, such as TOF (Time offlight), which refers to atoms, molecules or sound waves, electromagnetic waves, etc., which propagate in the medium. Distance required time) camera or 3D laser camera. The camera can not only obtain the information of the personnel activity, but also obtain the corresponding personnel distance information, and the distance information of the person can reflect the distance of the corresponding person from the camera.

Further, the camera specifically inputs the personnel activity information and the personnel distance information into the prediction model, and outputs whether the prediction result conforms to the intention of entering the door. If the intent of entering the door is met, the camera acquires the location information of the person of the corresponding person. The prediction model is also trained by the personnel activity information obtained in the training image sequence and the corresponding personnel distance information during training.

In the embodiment, when the door intent detection is performed, the personnel distance information is further considered, and the accuracy of detecting the intention of entering the door can be further improved, and the efficiency of the access control is further improved.

As shown in FIG. 6, in an embodiment, the access control data processing method further includes the step of abnormal alarm, which specifically includes the following steps:

Step 602: Obtain a human activity trajectory according to the image sequence.

Among them, the human activity track is data describing the change of the position of the corresponding person over time. Specifically, the camera can perform human body detection on the bird's-eye view image, and track the detected human body in the image sequence according to the human body detection result, thereby obtaining the human activity track of the corresponding person. The human activity track can be represented by a number of position points arranged in chronological order.

In step 604, the result of the person entering the door is obtained according to the trajectory of the person activity.

Among them, the detection result of the personnel entering the door is the detection result of the fact that the person enters the door, and may also include the trajectory of the personnel movement of the corresponding person. Specifically, the camera can detect whether the person's activity trajectory enters the door inner area from the outer door area, thereby obtaining a person entrance detection result of whether the person corresponding to the person activity track enters the door. The area inside the door refers to the area protected by the door controlled by the access control, generally the inner area of the room where the door is located; the area outside the door is the area within the field of view of the camera that does not belong to the area inside the door, generally the area outside the room where the door is located. .

Step 606: Output an alarm message when the recognized face does not match the personnel entry detection result.

Specifically, the camera sends the personnel door detection result to the access control device, and specifically may only be sent when detecting the entry result of the person entering the door by the corresponding person. The access control device receives the door detection result, and determines whether the person corresponding to the recognized face matches the detection result of the person entering the door. If the match is matched, the detection may continue, and if not, the alarm information is output.

Whether the recognized face and the person entering the door detection result match can be determined by judging whether the human activity track of the recognized face matches the person activity track of the entered person. If the identified human activity trajectory of the face is inconsistent with the trajectory of the person who has entered the door, or the position information of the person corresponding to the recognized face is inconsistent with the trajectory of the person who has entered the door, the recognized face and The identity of the entered personnel does not match. At this time, it is determined that the person corresponding to the recognized face does not match the detection result of the person entering the door, indicating that there may be a situation in which the face of another person is used for face recognition, and the alarm information may be output.

Whether the recognized face and the person entering the door detection result match, and can also be identified by judging Judging whether the number of faces matches the number of people who have entered the door. For example, if the recognized face is one, but the personnel entry detection result indicates that there are two entrance personnel, indicating that there may be trailing malicious personnel, and the alarm information can be output.

When the alarm information is output, the access control device may specifically output an alarm sound, control the alarm light to flash, send the alarm information to the designated mobile terminal, or combine two or more of the manners of outputting the alarm information, and may also adopt other The mode outputs alarm information.

In this embodiment, the access control device outputs an alarm message according to the matching result of the recognized face and the person entering the door detection result, and can output an alarm message when the matching result is not matched, and can perform an alarm reminder when the safety condition is met, but the security alarm is obtained. The person who goes to the alarm information can take corresponding measures against the access control scene with potential safety hazards to ensure the security of the access control.

In one embodiment, step 604 includes obtaining a human door detection result that the corresponding person has entered the door when the human activity trajectory passes from the outer door region through a line segment representing the door or more than one parallel line segment reaches the inner door region.

Specifically, in the overhead view image, the inner door area and the inner door area may be represented by corresponding line segments representing the door. The shape of the line segment is related to the projected shape of the door in the overhead image, and the line segment can be a straight line segment, a curved segment or a broken line segment.

The number of doors in the overhead view image may be one or more, one of which may be represented by one line segment or by more than one parallel line segment. When a door corresponds to a line segment, the two sides of the line segment are the inner door area and the outer door area respectively; when one door corresponds to more than one parallel line segment, the two sides of the parallel line segment are respectively the inner door area and the outer door area. The side-by-side line segment is a fault-tolerant area. The parallel line segments are a plurality of line segments that are substantially parallel, such as two or three parallel line segments.

When the personnel activity track passes from the outer door area to a door segment corresponding to a door to reach the door inner area, it can be determined that the corresponding person has entered the door. When the trajectory of the personnel moves from the outside area to the area inside the door through the parallel line indicating the extra one corresponding to one door, it means that the corresponding person is not at the door, but does enter the door area, which can improve the door detection. The accuracy of the results, in turn, can improve the accuracy of the alarm and prevent misjudgment.

In one embodiment, the predictive model may update the predictive model based on personnel information that meets the intent to enter, a face recognition result, and a person entry detection result that meets the entry intention of the person entering the door. The information of the person who meets the intention of entering the door includes the location information of the person who meets the intention of entering the door.

In this embodiment, the prediction model is iteratively trained by using the personnel information, the face recognition result and the personnel entry detection result that meet the intent of entering the door, so that the prediction result of the prediction model is more accurate, and the accuracy of the access control is improved.

In one embodiment, the various processes performed by the camera, the face capture camera, and the access control device described above may be performed locally or through remote server processing.

As shown in FIG. 7, in an embodiment, an access control method is provided. This embodiment is applied to the access control device 130 in FIG. 1 to illustrate. The method specifically includes the following steps:

Step 702: Obtain location information of the person who meets the intention of entering the door; the person who meets the intention of entering the door is detected according to the information of the personnel activity; and the information of the personnel activity is obtained according to the image sequence.

Specifically, the camera can acquire the image sequence, obtain the personnel activity information according to the image sequence, detect the position information of the person who meets the intention of entering the door according to the personnel activity information, and send the position information of the person to the access control device, and the access control device receives the entry. Location information of the intended person.

In other embodiments, the camera can acquire a sequence of images and send it to the access control device, which can obtain the activity information of the person according to the image sequence, and detect the position information of the person who meets the intention of entering the door according to the activity information of the person.

Among them, the camera captures a sequence of images. The camera can be installed at the top or above the wall of the person's height or at the top corner, so that the camera can capture the image from the top view, and the camera's field of view covers at least the position of the door.

The image sequence is a sequence of images formed by chronologically arranging the overhead images captured by the camera, and the image sequence may specifically be a video or a sequence of pictures. The time interval of adjacent overhead images in the image sequence can be set as desired. The frame rate of the image sequence can be lower or higher than the frame rate threshold that the human eye can withstand. The frame rate of the image sequence is the number of overhead images in the image sequence per unit time, and the frame rate is higher based on the detection result of the image sequence. The frame rate is low, and the camera hardware resource consumption is small.

The personnel activity information is information that can reflect the activity of the personnel, and the camera can specifically detect each of the overhead images in the image sequence to obtain the personnel activity information. The camera can send the image sequence to the server to receive the human activity information detected by the server and fed back by the server; the camera can also extract the feature for detecting the activity information of the person from the image sequence, and send the extracted feature to the server. Receiving the personnel activity information detected by the server based on the extracted feature detection.

After the camera obtains the activity information of the personnel, it can predict whether the corresponding personnel meets the intention of entering the door through the trained prediction model according to the information of the personnel activity, and if the person who meets the intention of entering the door obtains the position information of the corresponding person; if it does not meet the intention of entering the door The personnel continue to obtain personnel activity information based on the image sequence.

Step 704: Perform face recognition according to the location information of the person.

Specifically, the access control device acquires the location information of the person sent by the camera, and acquires the image collected by the face acquisition camera. The access control device can use the unified coordinate system to map the position information of the person to the captured image, and the image collected by the face acquisition camera and the person The face image matched by the location information is subjected to face recognition to obtain a face recognition result.

Step 706: Acquire a user identifier to which the recognized face belongs.

Step 708, outputting a door opening instruction when the user identifier has the right to open the door.

The above-mentioned access control method uses the image sequence to obtain the personnel activity information, thereby using the personnel activity information to detect the position information of the person who meets the intention of entering the door. The detected position information of the person is the position information of the person who meets the intention of entering the door, and then the targeted face recognition is performed according to the position information of the person, and the recognized face is used for the access control. In this way, the person who meets the intention of entering the door can be automatically judged, and the targeted face recognition can be performed, which can reduce the time consumption for the waiting person to actively cooperate with the face recognition, and improve the access control efficiency.

In an embodiment, the access control method further includes: acquiring a door detection result of the person; and outputting the alarm information when the recognized face does not match the detection result of the person entering the door.

Specifically, the access control device may obtain the detection result of the person entering the door sent by the camera, or may obtain the result of the door entry detection according to the image sequence sent by the camera. The results of the personnel entry detection can be obtained according to the trajectory of the personnel activity, and the trajectory of the human activity can be obtained according to the image sequence.

In one embodiment, the personnel activity information includes a person motion feature, an image sequence feature, and an image appearance feature; and obtaining location information of the person who meets the door intention, including: The sign, the image sequence feature and the image appearance feature are detected by the trained prediction model, and the position information of the corresponding person is obtained.

In one embodiment, obtaining location information of the person who meets the intent of entering the door includes: acquiring personnel distance information corresponding to the personnel activity information; and detecting location information of the person who meets the intention of entering the door according to the personnel activity information and the personnel distance information.

In one embodiment, the camera can perform human body detection on the top view image, and track the detected human body in the image sequence according to the human body detection result, thereby obtaining the human activity track of the corresponding person. The human activity track can be represented by a number of position points arranged in chronological order.

The camera sends the personnel entry detection result to the access control device, which can be sent only when the detection result of the person entering the door of the corresponding person is detected. The access control device receives the door detection result, and determines whether the person corresponding to the recognized face matches the detection result of the person entering the door. If the match is matched, the detection may continue, and if not, the alarm information is output.

Whether the recognized face and the person entering the door detection result match can also be judged by judging whether the number of recognized faces and the number of people who have entered the door match. For example, if the recognized face is one However, the result of the personnel entry detection indicates that there are two people entering the door, indicating that there may be trailing malicious personnel, and the alarm information can be output.

In one embodiment, the personnel door detection result is a result of a door entrance detection result obtained when a person activity track passes from a door outer area through a line segment representing the door or more than one side line segment reaches the door inner area, indicating that the corresponding person has entered the door. The human activity track is obtained from the image sequence.

Specifically, in the overhead view image, the in-door area and the in-door area may be represented by corresponding line segments representing the door. The shape of the line segment is related to the projected shape of the door in the overhead image, and the line segment can be a straight line segment, a curved segment or a broken line segment.

As shown in FIG. 8, in one embodiment, the present application further provides an electronic device 800, electronic The internal structure of device 800 may correspond to the structure shown in FIG. 2, and each of the modules described below may be implemented in whole or in part by software, hardware, or a combination thereof. The electronic device 800 includes an image sequence acquisition module 801, a door intent detection module 802, and a transmission module 803.

The image sequence obtaining module 801 is configured to acquire a sequence of images.

The portal intention detecting module 802 is configured to obtain the personnel activity information according to the image sequence; and detect the location information of the person who meets the intention of entering the door according to the personnel activity information.

The sending module 803 is configured to send the location information of the person to the access control device, so that the access control device performs face recognition according to the location information of the person, and outputs a door opening command when the identified user identifier of the face belongs to the door opening authority.

The electronic device 800 obtains the person activity information by using the image sequence, and uses the person activity information to detect the position information of the person who meets the intention of entering the door. The detected position information of the person is the position information of the person who meets the intention of entering the door, and then the targeted face recognition is performed according to the position information of the person, and the recognized face is used for the access control. In this way, the person who meets the intention of entering the door can be automatically judged, and the targeted face recognition can be performed, which can reduce the time consumption for the waiting person to actively cooperate with the face recognition, and improve the access control efficiency.

In one embodiment, the human activity information includes a human motion feature, an image sequence feature, and an image appearance feature; the portal intent detection module 802 is configured to pass the trained predictive model based on the human motion feature, the image sequence feature, and the image appearance feature. A person who meets the intention of entering the door is detected and the location information of the corresponding person is obtained.

In one embodiment, the image sequence is acquired by the depth camera; the door intention detection module 802 is further configured to acquire the personnel distance information corresponding to the personnel activity information; and the position information of the person who meets the door opening intention is detected according to the personnel activity information and the personnel distance information. .

As shown in FIG. 9, in an embodiment, the electronic device 800 further includes: a door detection module 804, configured to obtain a human activity trajectory according to the image sequence; and obtain a human door detection result according to the human activity trajectory.

The sending module 803 is further configured to send a personnel door detection result to the access control device, so that the access control device outputs the alarm information when the recognized face does not match the personnel door detection result.

In one embodiment, the door detection module 804 is further configured to: when a person activity track passes from a door outer area through a line segment representing the door or more than one side line segment reaches the door inner area, the person who has entered the door of the corresponding person enters the door. Test results.

As shown in FIG. 10, in an embodiment, an electronic device 1000 is provided, including: a personnel location information acquisition module 1001, a face recognition module 1002, and an access control module 1003.

The person position information obtaining module 1001 is configured to obtain position information of a person who meets the intention of entering the door; the position information of the person is position information of the person who meets the intention of entering the door detected according to the personnel activity information; and the personnel activity information is obtained according to the image sequence.

The face recognition module 1002 is configured to perform face recognition according to the location information of the person.

The access control module 1003 is configured to obtain a user identifier to which the recognized face belongs; and output a door opening command when the user identifier has the right to open the door.

The electronic device 1000 obtains the person activity information using the image sequence, and uses the person activity information to detect the position information of the person who meets the intention of entering the door. The location information of the detected person is The position information of the person who meets the intention of entering the door is used, and the targeted face recognition is performed according to the position information of the person, and the recognized face is used for the access control. In this way, the person who meets the intention of entering the door can be automatically judged, and the targeted face recognition can be performed, which can reduce the time consumption for the waiting person to actively cooperate with the face recognition, and improve the access control efficiency.

As shown in FIG. 11, in an embodiment, the electronic device 1000 further includes: an alarm module 1004, configured to acquire a door detection result of the person; and output an alarm message when the recognized face does not match the door detection result.

In this embodiment, when the human activity trajectory passes from the outer door area to a line segment corresponding to one door to reach the inner door area, it can be determined that the corresponding person has entered the door. When the trajectory of the personnel moves from the outside area to the area inside the door through the parallel line indicating the extra one corresponding to one door, it means that the corresponding person is not at the door, but does enter the door area, which can improve the door detection. The accuracy of the results, in turn, can improve the accuracy of the alarm and prevent misjudgment.

In one embodiment, an electronic device includes a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor, causing the processor to perform the steps of: acquiring an image sequence; Obtaining the activity information of the personnel; detecting the location information of the person who meets the intention of entering the door according to the information of the personnel activity; and transmitting the location information of the person to the access control device, so that the access control device performs face recognition according to the location information of the person, and identifies the When the user ID of the face belongs to the user, the door opening command is output.

In one embodiment, the personnel activity information includes a person motion feature, an image sequence feature, and an image appearance feature; and the location information of the person who meets the door intention is detected according to the personnel activity information, the package Including: according to the characteristics of the personnel movement, the characteristics of the image sequence and the appearance of the image, the trained prediction model is used to detect the person who meets the intention of entering the door, and obtain the position information of the corresponding person.

In one embodiment, the electronic device includes a depth camera that acquires a sequence of images; the computer readable instructions further cause the processor to perform the steps of: acquiring human distance information corresponding to the human activity information; and detecting the person in accordance with the intent to enter the door based on the personnel activity information The location information includes: detecting location information of a person who meets the intention of entering the door according to the personnel activity information and the personnel distance information.

In one embodiment, the computer readable instructions further cause the processor to: obtain a human activity trajectory according to the image sequence; obtain a human door detection result according to the human activity trajectory; and send a human door detection result to the access control device to enable the access control The device outputs an alarm message when the recognized face does not match the personnel entry detection result.

In one embodiment, obtaining a door entry detection result according to a human activity trajectory includes: when a person activity trajectory passes from a door outer area through a line segment representing the door or more than one side line segment reaches the door inner area, obtaining the corresponding person has The person entering the door enters the door to test the result.

The electronic device obtains the person activity information by using the image sequence, and uses the person activity information to detect the position information of the person who meets the intention of entering the door. The detected position information of the person is the position information of the person who meets the intention of entering the door, and then the targeted face recognition is performed according to the position information of the person, and the recognized face is used for the access control. In this way, the person who meets the intention of entering the door can be automatically judged, and the targeted face recognition can be performed, which can reduce the time consumption for the waiting person to actively cooperate with the face recognition, and improve the access control efficiency.

An electronic device comprising a memory and a processor, wherein the memory stores computer readable instructions, wherein when the computer readable instructions are executed by the processor, the processor performs the steps of: obtaining location information of a person who meets the intention of entering the door; The information is the location information of the person who meets the intention of entering the door according to the information of the personnel activity; the information of the personnel activity is obtained according to the image sequence; the face recognition is performed according to the location information of the person; and the user identifier to which the recognized face belongs is obtained; The door open command is output when the user logo has the right to open the door.

In one embodiment, the computer readable instructions further cause the processor to perform the steps of: obtaining a human door detection result; and outputting an alarm message when the recognized face does not match the human door detection result.

In one embodiment, the result of the door entry detection is that when a person activity track passes from a door outer area through a line segment representing the door or more than one side line segment reaches the door inner area, the obtained person indicating that the corresponding person has entered the door enters the door. The detection result; the human activity track is obtained according to the image sequence.

A person skilled in the art can understand that all or part of the process of implementing the above embodiment method can be completed by a computer program to instruct related hardware, and the computer program can be stored in a computer readable storage medium. When executed, the flow of an embodiment of the methods as described above may be included. The storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

The technical features of the above embodiments may be arbitrarily combined. For the sake of brevity of description, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, It is considered to be the range described in this specification.

The above embodiments merely express several embodiments of the present invention, and the description thereof is more specific and detailed. However, it cannot be construed as limiting the scope of the invention patent. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of the invention should be determined by the appended claims.

Claims

A method for processing access control data includes:

Obtaining a sequence of images;

Obtaining human activity information according to the image sequence;

Detecting a person who meets the intention of entering the door according to the information of the personnel activity;

Obtaining location information of the person;

Performing face recognition based on the location information of the person; and

The door opening instruction is output when the recognized user ID of the face belongs to the door opening authority.
The method according to claim 1, wherein said personnel activity information comprises a person motion feature, an image sequence feature, and an image appearance feature;

The person who detects the intent to enter the door according to the personnel activity information includes:

According to the motion characteristics of the person, the characteristics of the image sequence, and the apparent features of the image, the trained predictive model is used to detect the person who meets the intention of entering the door.
The method of claim 1 wherein the sequence of images is obtained by a depth camera; the method further comprising:

Obtaining personnel distance information corresponding to the personnel activity information;

The person who detects the intent to enter the door according to the personnel activity information includes:

A person who meets the intention of entering the door is detected based on the person activity information and the person distance information.
The method of claim 1 further comprising:

Obtaining a human activity trajectory according to the image sequence;

Obtaining a person entry detection result according to the trajectory of the person activity; and

The alarm information is output when the recognized face does not match the person entry detection result.
The method according to claim 4, wherein the obtaining the door detection result according to the human activity trajectory comprises:

When the human activity trajectory passes from a door outer region through a line segment representing the door or more than one parallel line segment reaches the inner door region, the result of the door entry detection of the person who has entered the door is obtained.
An access control method includes:

Obtaining location information of the person who meets the intention of entering the door; the person who meets the intention of entering the door is detected according to the information of the personnel activity; and the information of the activity of the person is obtained according to the image sequence;

Performing face recognition based on the location information of the person;

Obtaining the user ID to which the recognized face belongs; and

The door opening instruction is output when the user identification has a door opening authority.
The method according to claim 6, wherein the personnel activity information comprises a person motion feature, an image sequence feature, and an image appearance feature;

The obtaining location information of the person who meets the intention of entering the door includes:

According to the motion characteristics of the person, the image sequence feature and the image appearance feature, the trained prediction model is used to detect the person who meets the intention of entering the door, and obtain the position information of the corresponding person.
The method according to claim 6, wherein the obtaining location information of a person who meets the intention of entering the door comprises:

Obtaining personnel distance information corresponding to the personnel activity information; and

The person who meets the intention of entering the door is detected according to the personnel activity information and the person distance information, and the location information of the corresponding person is obtained.
The method of claim 6 further comprising:

Obtaining the results of personnel entry detection; and

The alarm information is output when the recognized face does not match the person entry detection result.
The method according to claim 9, wherein said personnel door detection result is obtained when a person activity trajectory passes from a door outer region through a line segment representing the door or more than one parallel line segment reaches the door inner region The person indicating that the corresponding person has entered the door enters the door detection result; the person activity track is obtained according to the image sequence.
An electronic device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor such that the processor performs the following steps:

Obtaining a sequence of images;

Obtaining human activity information according to the image sequence;

Detecting a person who meets the intention of entering the door according to the information of the personnel activity;

Obtaining location information of the person;

Performing face recognition based on the location information of the person; and

The door opening instruction is output when the recognized user ID of the face belongs to the door opening authority.
The electronic device according to claim 11, wherein the personnel activity information comprises a person motion feature, an image sequence feature, and an image appearance feature;

The person who detects the intent to enter the door according to the personnel activity information includes:

According to the motion characteristics of the person, the characteristics of the image sequence, and the apparent features of the image, the trained predictive model is used to detect the person who meets the intention of entering the door.
The electronic device of claim 11 wherein said electronic device comprises a depth camera that acquires a sequence of images; said computer readable instructions further causing said processor to perform the steps of:

Obtaining personnel distance information corresponding to the personnel activity information;

The person who detects the intent to enter the door according to the personnel activity information includes:

A person who meets the intention of entering the door is detected based on the person activity information and the person distance information.
The electronic device of claim 11 wherein said computer readable instructions further cause said processor to perform the following steps:

Obtaining a human activity trajectory according to the image sequence;

Obtaining a person entry detection result according to the trajectory of the person activity; and

The alarm information is output when the recognized face does not match the person entry detection result.
The electronic device according to claim 14, wherein the obtaining the door detection result according to the human activity trajectory comprises:

When the human activity trajectory passes from a door outer region through a line segment representing the door or more than one parallel line segment reaches the inner door region, the result of the door entry detection of the person who has entered the door is obtained.
An electronic device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor such that the processor performs the following steps:

Obtaining location information of the person who meets the intention of entering the door; the person who meets the intention of entering the door is detected according to the information of the personnel activity; and the information of the activity of the person is obtained according to the image sequence;

Performing face recognition based on the location information of the person;

Obtaining the user ID to which the recognized face belongs; and

The door opening instruction is output when the user identification has a door opening authority.
The electronic device according to claim 16, wherein the personnel activity information comprises a person motion feature, an image sequence feature, and an image appearance feature;

The obtaining location information of the person who meets the intention of entering the door includes:

According to the motion characteristics of the person, the image sequence feature and the image appearance feature, the trained prediction model is used to detect the person who meets the intention of entering the door, and obtain the position information of the corresponding person.
The electronic device according to claim 16, wherein the obtaining location information of the person who meets the intention of entering the door comprises:

Obtaining personnel distance information corresponding to the personnel activity information; and

The person who meets the intention of entering the door is detected according to the personnel activity information and the person distance information, and the location information of the corresponding person is obtained.
The electronic device of claim 16 wherein said computer readable instructions further cause said processor to perform the following steps:

Obtaining the results of personnel entry detection; and

The alarm information is output when the recognized face does not match the person entry detection result.
The electronic device according to claim 19, wherein said personnel door detection result is that when a person activity track passes from a door outer region through a line segment representing the door or more than one parallel line segment reaches the door inner region, The obtained result indicates that the person who has entered the door has entered the door; the human activity track is obtained according to the image sequence.