CN110457987A - Face identification method based on unmanned plane - Google Patents
Face identification method based on unmanned plane Download PDFInfo
- Publication number
- CN110457987A CN110457987A CN201910494410.9A CN201910494410A CN110457987A CN 110457987 A CN110457987 A CN 110457987A CN 201910494410 A CN201910494410 A CN 201910494410A CN 110457987 A CN110457987 A CN 110457987A
- Authority
- CN
- China
- Prior art keywords
- face
- module
- video flowing
- unmanned plane
- acquisition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention belongs to face identification system fields, more particularly to a kind of face identification method based on unmanned plane, video flowing receiving module, face recognition module, face acquisition module and the database interactive module of video flowing acquisition/transmission module, earth station including unmanned generator terminal;Video flowing acquisition/transmission module of the unmanned generator terminal is responsible for the image information in the UVC camera photographic subjects region for calling raspberry pie to pass through USB interface carry;Earth station's video flowing receiving module is mainly responsible for the data packet for receiving and transmitting from unmanned plane subsystem, is decoded into the format that face recognition module is capable of identifying processing;The face recognition module is mainly responsible for face identification function;Earth station's face acquisition module provides real-time acquisition mode and batch introduction model.Accuracy of identification of the present invention is high, and facial image is undistorted, strong antijamming capability.
Description
Technical field
The invention belongs to field of face identification more particularly to a kind of face identification methods based on unmanned plane.
Background technique
By practical investigation discovery, recognition of face still has its limitation with upper public safety field: on the one hand, adopting
The object of collection --- face, be in it is uncertain under the conditions of, including the factors such as illumination condition, angle, shelter will directly lead
Cause the precision of recognition of face;On the other hand, due to the equipment of face acquisition, the i.e. characteristic of camera, physical location is once portion
Administration is fixed, and the operating angle range of backstage face identification system is extremely limited, and there are biggish blind areas.In addition various regions are public
Pacify organ for cost consideration, recognition of face camera covering surface is limited, and the demand that many needs obtain live view in time is difficult
To realize.How met for greater flexibility the needs of video monitoring, investigation, patrol, preferably serving public safety practice is pole
Has the project of challenge.
The fast development of unmanned air vehicle technique is to put the axe in the helve to bring dawn.Booming, the novel microcomputer of material science
Electric transducer and the update of minitype inertial navigation controller, the iteration of flight controller and Flight Control Algorithm, all promote with four
Rotor wing unmanned aerial vehicle is that the unmanned plane family of representative starts to show up prominently in security protection industry, as great Jiang company pushes away in late September, 2018
The police version unmanned plane of longitude and latitude M200 is gone out, police refiting scheme etc. is provided to the civilian unmanned plane of Mavic Pro series.Nobody
Machine brings following positive effect in police, safety-security area utilization: one, solve the problems, such as that traditional camera covering surface is insufficient,
High-altitude macroscopic view visual angle is provided for party in request, makes up the short slab that monitoring probe has observation blind area;Two, unmanned plane is since it is flexible
Characteristic can actively adjust itself observation position, achieve the effect that initiative recognition target, tracking target;Three, for the view of acquisition
Frequency stream can promote the efficiency of decision-making with Intelligent treatment;Four, compared with artificial patrol, more efficient, safety effectively reduces cost.
But regrettably, the utilization of some local public security actual combat departments video acquired to unmanned plane is more to it
Carry out macroscopic view, based on manually visualize method acquisition of information, as group disposition in the detecting of landform, to staff size
Determine etc.;To the statistics of vehicle flowrate in traffic administration;When case is investigated and prosecuted to live macro environment take pictures evidence obtaining etc., to video
The case where flowing directly progress recognition of face work is also more rare, and reason has: one, market provides, and carries mature face knowledge
The unmanned plane product of other function is insufficient;Two, face recognition algorithms are in the interference of the unfavorable factors such as larger angle in face, lead to people
Face image distortion, accuracy of identification are poor.Therefore, how recognition of face effectively to be combined with unmanned plane, how improves unmanned plane people
Face identifying system recognition effect, this is all the content that the present invention studies and research significance place.
Summary of the invention
High the present invention is directed to provide a kind of accuracy of identification in place of overcome the deficiencies in the prior art, facial image is undistorted,
The face identification method based on unmanned plane of strong antijamming capability.
In order to solve the above technical problems, the present invention is implemented as follows:
Face identification method based on unmanned plane, including unmanned generator terminal video flowing acquisition/transmission module, earth station's video flowing receive
Module, face recognition module, face acquisition module and database interactive module;
The unmanned generator terminal video flowing acquisition/transmission module is responsible for that raspberry pie is called to clap by the UVC camera of USB interface carry
The image information in target area is taken the photograph, and two different operating modes: TCP mode and UDP mould are provided in video flowing transmission
Formula;
Earth station's video flowing receiving module is mainly responsible for the data packet for receiving and transmitting from unmanned plane subsystem, decodes it
It is capable of the format of identifying processing at face recognition module;
The face recognition module is mainly responsible for face identification function, i.e., by the target face occurred in video flowing in interactive interface
In outline, alarm;
Earth station's face acquisition module provides real-time acquisition mode and batch introduction model;
The database interactive module includes increase, deletion, the modification of (1) system user and password pair;(2) name and identity are known
Alias matching is mutually looked into;(3) batch of human face data imports.
As a preferred embodiment, video flowing acquisition/transmission module of unmanned generator terminal of the present invention is responsible for calling raspberry
The image information in UVC camera photographic subjects region that group passes through USB interface carry, is embodied as Opencv computer vision library
The instantiation of VideoCapture class under the videoIO module of middle offer.
Further, earth station's video flowing receiving module of the present invention is under TCP operating mode, the specific work of earth station
Make step are as follows: (1) establish socket with server end and connect;(2) negotiation data packet is sent;(3) it receives from server end
TCP data packet;(4) decoding data packet is transmitted to face recognition module.
Further, the specific steps of face recognition module of the present invention are as follows: (1) receive and connect from earth station's video flowing
Receive the Video stream information that module transmits;(2) it initializes face recognition module and loads target face characteristic value information;(3) it detects
Face present in stream is simultaneously marked with box;(4) whether the face that judgement detects is target face, if being then marked simultaneously
System alarm records time of occurrence.
Further, face acquisition module of the present invention mainly provides real-time acquisition mode and batch introduction model;It is real
When acquisition mode workflow are as follows: (1) call local camera, captured in real-time acquisition target face information;(2) it is being suitable for
Under facial angle, suitable illumination condition, shooting still photo acquires image as face;(3) face characteristic value, mark are extracted
Name and identity code;(4) face database is written.
Further, database interactive module of the present invention includes (1) user management: the increasing of system user and password pair
Add, delete, modify;(2) ID inquiring: name is matched with identity recognition number mutually to be looked into;(3) face is put in storage: the batch of human face data
It imports.
Further, user management of the present invention is mainly interacted by the user table with backstage MySQL database
To realize this system user and password to newly-built, increase, deletion.
Further, ID inquiring of the present invention is mainly interacted by the face table with backstage MySQL database,
Realize the mutual inquiry of identity recognition number and name.
Further, face storage of the present invention is provided in face of the individual human face storage of individual facial image and in face of more
Open the batch face storage of facial image.
Further, individual human face of the present invention enters library facility main working process are as follows:
(1) individual portrait is read, the deep learning model extraction face characteristic value of Dlib is used;
(2) characteristic value, identity recognition number, name are inserted into jointly in background data base face table;
The batch face storage enters library facility, workflow using individual facial image of recursive call are as follows:
(1) characteristic value for first face picture stored under storage folder is extracted;
(2) name and identity code to mark in filename is inserted into face database face table with characteristic value jointly;
(3) the lower next picture stored is pressed from both sides to this document and repeats aforesaid operations.
In the test of unmanned plane face tracking function, different several groups of tissue of the present invention height, distance, facial angle
Face recognition experiment.By experiment, this system possesses higher face when 5 meters of height, facial angle are less than 45 degree of work and knows
Other precision, and under closer distance (within 1.8 meters), the operating condition of smaller facial angle, face is blocked and (wears sunglasses)
Situation have higher recognition of face robustness.
Earth station's face acquires function after tested, can provide earth station and acquire in real time to target face, adds identification
Code and name, the function of extracting characteristic value and database is added, time overhead and stability meet design requirement.Database interaction
Tool after tested, mutually look into function time overhead and meet design requirement with stability by name and identification code.
Detailed description of the invention
Present invention will be further explained below with reference to the attached drawings and specific embodiments.Protection scope of the present invention not only office
It is limited to the statement of following content.
Fig. 1 is present system entirety hardware frame figure.
Fig. 2 is invention software part flow diagram.
Fig. 3 is the unmanned generator terminal video flowing acquisition module workflow block diagram of this hair.
Fig. 4 is the unmanned generator terminal video flowing transmission module workflow block diagram of the present invention.
Fig. 5 is earth station's video flowing receiving module workflow block diagram of the present invention.
Fig. 6 is earth station's face recognition module workflow block diagram of the present invention.
Fig. 7 is earth station's face acquisition module workflow block diagram of the present invention.
Specific embodiment
As shown, the face identification method based on unmanned plane, including unmanned generator terminal video flowing acquisition/transmission module,
Face station video flowing receiving module, face recognition module, face acquisition module and database interactive module;
The unmanned generator terminal video flowing acquisition/transmission module is responsible for that raspberry pie is called to clap by the UVC camera of USB interface carry
The image information in target area is taken the photograph, and two different operating modes: TCP mode and UDP mould are provided in video flowing transmission
Formula;
Earth station's video flowing receiving module is mainly responsible for the data packet for receiving and transmitting from unmanned plane subsystem, decodes it
It is capable of the format of identifying processing at face recognition module;
The face recognition module is mainly responsible for face identification function, i.e., by the target face occurred in video flowing in interactive interface
In outline, alarm;
Earth station's face acquisition module provides real-time acquisition mode and batch introduction model;
The database interactive module includes increase, deletion, the modification of (1) system user and password pair;(2) name and identity are known
Alias matching is mutually looked into;(3) batch of human face data imports.
Video flowing acquisition/transmission module of unmanned generator terminal of the present invention is responsible for that raspberry pie is called to pass through USB interface carry
UVC camera photographic subjects region in image information, the videoIO mould provided in Opencv computer vision library is provided
The instantiation of VideoCapture class under block.
Earth station's video flowing receiving module of the present invention is under TCP operating mode, the specific work steps of earth station are as follows:
(1) socket is established with server end to connect;(2) negotiation data packet is sent;(3) the TCP data packet from server end is received;
(4) decoding data packet is transmitted to face recognition module.
The specific steps of face recognition module of the present invention are as follows: (1) receive and passed from earth station's video flowing receiving module
The Video stream information come;(2) it initializes face recognition module and loads target face characteristic value information;(3) exist in detection stream
Face and marked with box;(4) whether the face that judgement detects is target face, if simultaneously system report is then marked
It is alert, record time of occurrence.
Face acquisition module of the present invention mainly provides real-time acquisition mode and batch introduction model;Real-time acquisition mode
Workflow are as follows: (1) call local camera, captured in real-time acquisition target face information;(2) suitable facial angle,
Under suitable illumination condition, shooting still photo acquires image as face;(3) face characteristic value is extracted, name and body are marked
Part identification code;(4) face database is written.
Database interactive module of the present invention includes (1) user management: the increase of system user and password pair, deletion,
Modification;(2) ID inquiring: name is matched with identity recognition number mutually to be looked into;(3) face is put in storage: the batch of human face data imports.
User management of the present invention mainly interacts to realize that this is by the user table with backstage MySQL database
System user and password are to newly-built, increase, deletion.
ID inquiring of the present invention is mainly interacted by the face table with backstage MySQL database, realizes identity
The mutual inquiry of identifier and name.
Face storage of the present invention provides the individual human face storage for facing individual facial image and faces multiple face figures
The batch face of picture is put in storage.
Individual human face of the present invention enters library facility main working process are as follows:
(1) individual portrait is read, the deep learning model extraction face characteristic value of Dlib is used;
(2) characteristic value, identity recognition number, name are inserted into jointly in background data base face table;
The batch face storage enters library facility, workflow using individual facial image of recursive call are as follows:
(1) characteristic value for first face picture stored under storage folder is extracted;
(2) name and identity code to mark in filename is inserted into face database face table with characteristic value jointly;
(3) the lower next picture stored is pressed from both sides to this document and repeats aforesaid operations.
The Hardware Design and realization
As shown in Figure 1, this system is divided into unmanned plane subsystem and earth station's subsystem in hardware view.Earth station's subsystem is by pen
Remember that this computer is constituted, the management of operation, face database in terms of responsible recognition of face.Unmanned plane subsystem, which removes, to be provided substantially
Outside stabilized flight, should also have a following function: the acquisition of video flowing, the processing of video flowing: video standard can not directly pass through
WiFi signal transmission, it is therefore desirable to further compacting, coded treatment, the transmission of video flowing, security guarantee, earth station property
The functions such as energy guarantee.
The design and realization of this system hardware are carried out based on modular theory.Hardware platform is divided into following several moulds
Block: data acquisition and transmission module, control module, power plant module.Its hardware block diagram is illustrated in fig. 1 shown below.
The major function of this module is in target area, and the face of ground staff shoots video flowing and passes after being handled
It returns.This module comes carry camera, wireless network card using microcomputer as core, using USB port.Supplier of electricity
Face selects a set of independent power module to be responsible for powering to microcomputer, and solving the stability that module itself is run with this needs
It asks.The communication with earth station is carried out by carry high-gain aerial wireless network card, is guaranteed centainly with this in physical hardware level
Functional stabilization.Installation point of one three axis holder as camera is installed, facilitates the control of camera shooting angle, reaches bat
Take the photograph the controllable purpose of content.In addition, the glissandoes such as gyroscope in holder can also play the effect of control picture vibration.
Flight control modules are responsible for receiving the telecommand from ground remote control device, pass through the floating-point meter to flying quality
It calculates, manages electron speed regulator output power, unmanned plane is finally made to adjust flight attitude by the instruction that ground remote control device inputs.The mould
Block is connect by the interface on flight controller with GPS electronic compass, remote control signal receiver using flight controller as core.
Three axis holders are connect with remote control signal receiver.Flight controller receives the instruction from ground remote control device by receiver, to
Power plant module issues flight directive.In addition, control module is by the battery-powered of power plant module.
The major function of power plant module is that the lift of flight is provided for unmanned aerial vehicle platform, realizes the flight of unmanned plane.Therefore
Power plant module eliminates circuit (BEC) module by electron speed regulator (ECS), motor, propeller, battery, lithium battery group forms.Wherein
BEC module be mainly responsible for lithium battery and take electricity, voltage is fallen to after the level of 5V~6V as the electronics other than motor circuit
Equipment power supply, such as flight controller, electronic compass, remote control receiver.In order to mitigate Aircraft Quality and volume, some ECS
In be integrated with BEC circuit, can be directly from wherein taking electricity, the fairly simple circuit of setting.
The main object of the present invention is to realize system components module design, carries out flight survey after attaching it to rack
Examination.Structure design by each main modular is selected with crucial equipment, it is contemplated that unmanned plane main screw lift is in 3kg or so, consideration
It is larger with transmission module equipment volume to needing video flowing to be mounted to acquire, therefore select up to sub- 680 models as unmanned plane machine
Frame.The model frame main body material is 3K carbon fiber board, and CNC manufactures folded piece, and wheelbase is standard x type 680, rack weight 476
Gram.Flexibility is high and drop resistant performance is strong, supports the folding of horn, is able to satisfy certain portable demand.Inner space is divided into up and down
Two layers, the installation of convenient device, so far, the selection of component needed for this system finish, and each primary clustering model see the table below.
Component | Model | Component | Model |
Rack | Daya680 | Flight controller | PixHawk V1 |
Motor | X4110s * 4 | ECS | Carefree 40A * 4 |
Blade | Tarrot 1555 *4 | Battery pack | Poly- sea 5400mah6S |
Remote controler/receiver | Fs-i6s / Fs-iA6B | Electronic compass | M8N |
Microcomputer | Raspberry pie 3b | Camera | Kingcent KS2A242 |
High-gain wireless network card | Tp-Link WN722N | Holder | Three axis holders |
The realization of system hardware is broadly divided into: fly control and the connection of remote control receiver, the connection of holder and receiver,
The installation and debugging of GPS compass, electricity adjust with the connection of motor, rack assembling, wiring, recommend handbook to install according to official
Bi Hou is only A4 paper size under folded state.
The realization main task of video flowing transmission module is the connection of raspberry pie and camera, wireless network card, power module
With test, can the purpose is to detect the module carry out the transmission of video flowing with earth station.Camera, network interface card pass through general serial
Bus port is connect with raspberry pie.Power module then passes through the port PowerInput and connect with raspberry pie.After tested, raspberry
After group's installation AR9271 driving, raspberry pie can normally identify that wn722n wireless network card, hardware are No. 5 equipment of No. 1 bus.Pass through
UVC agreement, raspberry pie can normally identify that camera, hardware location are No. 6 equipment of No. 1 bus.
Shown in Figure 3, the video flowing acquisition module of unmanned generator terminal is responsible for that raspberry pie is called to pass through USB interface carry
Image information in UVC camera photographic subjects region is the main data source of whole system.The module major function
The instantiation of VideoCapture class under the videoIO module provided in Opencv computer vision library is provided.Such is main
For capturing video from video file, image sequence or video camera.
Under Python, the creation mode of VideoCapture object be "<VideoCapture object>=
cv.VideoCapture(index[, apiPreference])".Index value therein is the equipment index number of camera,
As the default camera head of FrameCaptured=cv2.VideoCapture (0), as calling computer carry is regarded in real time
The acquisition of frequency stream.After VideoCapture instantiation, using .isOpened () function under such, from the boolean of its return
Value can detecte whether video capture initializes.
After FrameCaptured is by successful examples, get (propID) function and set (propID, value) are used
Function is available or the parameter of frame is arranged.The width of the parameter that can be set such as cv.CAP_PROP_FRAME_WIDTH(frame),
Cv.CAP_PROP_FRAME_HEIGHT, cv.CAP_PROP_FPS etc..Video flowing can be easily read using read () function
Next frame, and return to the frame grabbed in real time.The function is to grab the next frame of camera shooting by grab () function, with
Retrieve () function returns to this frame grabbed in real time, the overload function being jointly formed.Therefore the function can be returned when calling
Two values are returned, shaped like " retval, image=FrameCaptured.read () ".If camera break link or error,
Retval will return to false, and image will return to null value.It, can using release () function after the completion of video flowing acquisition tasks
To close the camera of corresponding equipment call number.
Shown in Figure 4, in this system, the video flowing transmission module of unmanned generator terminal is needed according to module design above
There is provided two different operating modes: TCP mode and UDP mode, the two are differing principally in that and earth station's transmitting video-frequency flow
It is based on which kind of agreement when data packet.Under TCP operating mode, workflow under the mode are as follows: (1) unmanned generator terminal is as client
End, earth station is as server end, and after the two sets up connection, unmanned plane is terminated by the configuration parameter from earth station, parameter
Include frame number and resolution ratio etc.;(2) video frame that video flowing acquisition module constantly transmits is received, by parameter received in the first step
Adjust image;(3) each frame adjusted is encoded, is converted into the form of matrix, be packaged after generating the character string for facilitating transmission;
(4) earth station is sent to by the TCP connection having built up.What is faced due to UDP operating mode is that radio signal conditions are good, dry
Therefore workflow is more succinct in the case where disturbing less, and the configuration of client and server synchronization parameter, the Working mould is omitted
Execution efficiency is higher under formula.
In the module, realize that the mode of TCP connection is to realize that socket is also known as using the library socket that python is provided
Socket, the application program at network both ends issue request to other side by network by establishing socket or provide response,
So that different application, different hosts realize the function of communication.
Shown in Figure 5, earth station's video flowing receiving module is mainly responsible for the number for receiving and transmitting from unmanned plane subsystem
According to packet, it is decoded into the format that face recognition module is capable of identifying processing.Under TCP operating mode, earth station is as client
Work, workflow are as follows: (1) establish socket with server end and connect;(2) negotiation data packet is sent;(3) it receives from service
The TCP data packet at device end;(4) decoding data packet is transmitted to face recognition module.
Socket is established with server end, and the library the socket realization for relying primarily on python and providing is provided.Come in logic from executing
It says, unmanned generator terminal has instantiated socket object and has been bundled with corresponding ports and begun listening at this time, therefore the video of earth station
Receiving module, which is flowed, as client initiates socket connection.
Shown in Figure 6, earth station's face recognition module is the main functional modules of this system.Module chief leading cadre
The target face occurred in video flowing is outlined in interactive interface, is alarmed by face identification function.
The workflow of recognition of face are as follows: (1) receive the Video stream information transmitted from earth station's video flowing receiving module;
(2) it initializes face recognition module and loads target face characteristic value information;(3) face present in detection stream and with box mark
Out;(4) whether the face that judgement detects is target face, if simultaneously system alarm is then marked, records time of occurrence.
Since earth station's video flowing receiving module parses the two different protocol packages of TCP, UDP, in the module, people
Face tracking function equally corresponds to two kinds of different operating modes, but principle is consistent by and large.Therefore work is not distinguished in this part
Operation mode does comprehensive discussion.
Face identification functions need to read the characteristic value data of target face in advance.The link is known by specified target identities
The form of alias, read in background data base in corresponding list item face characteristic value (database interactive module provide name with
The function that identity recognition number is mutually looked into).After video flowing is passed to this module with numpy matrix (numpy arry), Dlib is used first
Face datection model (the dlib.cnn_face_detection_ based on convolutional neural networks (cnn) in computer vision library
Model_v1) confirm in video flowing with the presence or absence of face, realize that the key code of the function is " cnn_face_
Detector=dlib.cnn_face_detection_ model_v1 (model_path) ", herein model_path parameter
It is set as the dlib tissue good recognition of face template of precondition, the text of mmod_human_face_ detector.dat
Part path, determine in video flowing detect face after, using Opencv provide rectangle () function by face frame,
Label.
When detected in video flowing face there are after, use Dlib provide " dlib.face_ recognition_
Model_v1(path) " function configures path parameter, loads the resnet model of dlib, extracts in video flowing and face occurs
128 dimensional feature vectors.The target feature vector is by using mumpy's between loaded target face characteristic
Linalg.norm() function seeks the Euclidean distance of both feature vectors stored in characteristic value and database, if the distance is less than
The error threshold of one setting, then assert the face and target face is same face.The putText provided using Opencv
The name of target face is placed in the lower section of indicia framing by () function.
Shown in Figure 7, earth station's face acquisition module mainly provides two kinds of operating modes --- real-time acquisition mode with
Batch introduction model.The workflow of real-time acquisition mode are as follows: (1) call local camera, captured in real-time acquisition target face
Information;(2) under suitable facial angle, suitable illumination condition, shooting still photo acquires image as face;(3) it mentions
Face characteristic value is taken, name and identity code are marked;(4) face database is written.
Under real-time acquisition mode, call the operation of local camera by VideoCapture (num) letter in the library Opencv
Number provides, and num parameter is the number of local camera, the camera for being capable of calling local reference numeral by configuring the parameter.It obtains
It is shown on interactive interface after taking video flowing by the transformation of color channel.Determine acquisition present image as human face data
After extracting object, in the corresponding position of interactive interface, identification number, the name of the current acquisition target of input, click are taken pictures
Key, present frame will be saved under specified folder for extracting characteristic value as acquisition image.
The tool mainly provides the interactive function between user and database, specifically includes: (1) system user and password pair
Increase, deletion, modification function;(2) name is matched with identity recognition number mutually looks into function;(3) batch of human face data imports function
Energy.
Subscriber management function mainly interacts to realize this system user by the user table with backstage MySQL database
The purpose of with password to newly-built, increase, deletion.The realization of the function drives mysql-connector by the python of MySQL
Corresponding SQL statement is executed to realize.
Identity lookup function is mainly interacted by the face table with backstage MySQL database, realizes identity recognition number
With the mutual query function of name.The function can provide information branch for the pre-loaded target face characteristic value of face recognition module
It holds.
Face enters library facility and mainly provides two kinds of operating modes: i.e. in face of the individual human face storage of individual facial image and face
To the batch face storage of multiple facial images.Individual human face enters library facility main working process are as follows: (1) individual portrait is read,
Use the deep learning model extraction face characteristic value of Dlib;(2) characteristic value, identity recognition number, name are inserted into backstage jointly
In database face table.Batch introduction model working principle is to enter library facility using individual facial image of recursive call, basic
Workflow are as follows: (1) extract the characteristic value for first face picture stored under storage folder;(2) to be marked in filename
Name and identity code, be inserted into face database face table jointly with characteristic value;(3) the next of lower storage is pressed from both sides to this document
Picture repeats aforesaid operations.
It is to be understood that being merely to illustrate the present invention above with respect to specific descriptions of the invention and being not limited to this
Technical solution described in inventive embodiments, those skilled in the art should understand that, still the present invention can be carried out
Modification or equivalent replacement, to reach identical technical effect;As long as meet use needs, all protection scope of the present invention it
It is interior.
Claims (10)
1. the face identification method based on unmanned plane, which is characterized in that including unmanned generator terminal video flowing acquisition/transmission module,
Face station video flowing receiving module, face recognition module, face acquisition module and database interactive module;
The unmanned generator terminal video flowing acquisition/transmission module is responsible for that raspberry pie is called to clap by the UVC camera of USB interface carry
The image information in target area is taken the photograph, and two different operating modes: TCP mode and UDP mould are provided in video flowing transmission
Formula;
Earth station's video flowing receiving module is mainly responsible for the data packet for receiving and transmitting from unmanned plane subsystem, decodes it
It is capable of the format of identifying processing at face recognition module;
The face recognition module is mainly responsible for face identification function, i.e., by the target face occurred in video flowing in interactive interface
In outline, alarm;
Earth station's face acquisition module provides real-time acquisition mode and batch introduction model;
The database interactive module includes increase, deletion, the modification of (1) system user and password pair;(2) name and identity are known
Alias matching is mutually looked into;(3) batch of human face data imports.
2. according to claim 1 based on the face identification method of unmanned plane, it is characterised in that: the video of the unmanned generator terminal
Stream acquisition/transmission module is responsible for the image letter in the UVC camera photographic subjects region for calling raspberry pie to pass through USB interface carry
Breath, is embodied as the instantiation of VideoCapture class under the videoIO module provided in Opencv computer vision library.
3. according to claim 2 based on the face identification method of unmanned plane, it is characterised in that: earth station's video flowing connects
Module is received under TCP operating mode, the specific work steps of earth station are as follows: (1) establish socket with server end and connect;(2)
Send negotiation data packet;(3) the TCP data packet from server end is received;(4) decoding data packet is transmitted to face recognition module.
4. according to claim 3 based on the face identification method of unmanned plane, it is characterised in that: the face recognition module
Specific steps are as follows: (1) receive the Video stream information transmitted from earth station's video flowing receiving module;(2) recognition of face is initialized
Module simultaneously loads target face characteristic value information;(3) it face present in detection stream and is marked with box;(4) judgement detects
Face whether be target face, if being then marked and system alarm, record time of occurrence.
5. according to claim 4 based on the face identification method of unmanned plane, it is characterised in that: the face acquisition module master
Real-time acquisition mode and batch introduction model are provided;The workflow of real-time acquisition mode are as follows: (1) local camera is called,
Captured in real-time acquisition target face information;(2) under suitable facial angle, suitable illumination condition, shooting still photo is made
Image is acquired for face;(3) face characteristic value is extracted, name and identity code are marked;(4) face database is written.
6. according to claim 5 based on the face identification method of unmanned plane, it is characterised in that: the database interactive module
Including (1) user management: increase, deletion, the modification of system user and password pair;(2) ID inquiring: name and identity recognition number
Matching is mutually looked into;(3) face is put in storage: the batch of human face data imports.
7. according to claim 6 based on the face identification method of unmanned plane, it is characterised in that: the user management is mainly led to
It crosses and interacts to realize this system user and password to newly-built, increase, deletion with the user table of backstage MySQL database.
8. according to claim 7 based on the face identification method of unmanned plane, it is characterised in that: the ID inquiring mainly leads to
It crosses and is interacted with the face table of backstage MySQL database, realize the mutual inquiry of identity recognition number and name.
9. according to claim 8 based on the face identification method of unmanned plane, it is characterised in that: the face is put in storage offer face
The individual human face storage and the batch face in face of multiple facial images of individual facial image are put in storage.
10. according to claim 9 based on the face identification method of unmanned plane, it is characterised in that: the individual human face storage
Function main working process are as follows:
(1) individual portrait is read, the deep learning model extraction face characteristic value of Dlib is used;
(2) characteristic value, identity recognition number, name are inserted into jointly in background data base face table;
The batch face storage enters library facility, workflow using individual facial image of recursive call are as follows:
(1) characteristic value for first face picture stored under storage folder is extracted;
(2) name and identity code to mark in filename is inserted into face database face table with characteristic value jointly;
(3) the lower next picture stored is pressed from both sides to this document and repeats aforesaid operations.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910494410.9A CN110457987A (en) | 2019-06-10 | 2019-06-10 | Face identification method based on unmanned plane |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910494410.9A CN110457987A (en) | 2019-06-10 | 2019-06-10 | Face identification method based on unmanned plane |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110457987A true CN110457987A (en) | 2019-11-15 |
Family
ID=68480654
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910494410.9A Pending CN110457987A (en) | 2019-06-10 | 2019-06-10 | Face identification method based on unmanned plane |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110457987A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111552828A (en) * | 2020-04-26 | 2020-08-18 | 上海天诚比集科技有限公司 | 1-to-N face comparison method |
CN112969024A (en) * | 2020-06-30 | 2021-06-15 | 华为技术有限公司 | Camera calling method, electronic equipment and camera |
CN113518218A (en) * | 2021-05-25 | 2021-10-19 | 上海商汤智能科技有限公司 | Camera equipment serial number determining method and device, electronic equipment and storage medium |
CN116878468A (en) * | 2023-09-06 | 2023-10-13 | 山东省国土测绘院 | Information acquisition system for mapping |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20030061047A (en) * | 2002-01-07 | 2003-07-18 | 이철희 | Face recognition algorithm |
KR20130131719A (en) * | 2012-05-24 | 2013-12-04 | 현대모비스 주식회사 | Face authentication apparatus and method for vehicle |
CN105447459A (en) * | 2015-11-18 | 2016-03-30 | 上海海事大学 | Unmanned plane automation detection target and tracking method |
CN106056075A (en) * | 2016-05-27 | 2016-10-26 | 广东亿迅科技有限公司 | Important person identification and tracking system in community meshing based on unmanned aerial vehicle |
CN107197165A (en) * | 2017-07-28 | 2017-09-22 | 哈尔滨市舍科技有限公司 | Unmanned plane self-heterodyne system and method |
CN107451527A (en) * | 2017-06-29 | 2017-12-08 | 广东容祺智能科技有限公司 | A kind of large-scale public place hunting system based on unmanned plane |
CN108141512A (en) * | 2015-09-30 | 2018-06-08 | 株式会社尼康 | Flight instruments, mobile device, server and program |
CN108200168A (en) * | 2017-12-29 | 2018-06-22 | 南京奇蛙智能科技有限公司 | A kind of unmanned plane during flying method for remote management |
CN109740577A (en) * | 2019-02-28 | 2019-05-10 | 南京信息工程大学 | A kind of real-time face based on raspberry pie identifies camera system and its adjustment method again |
CN109819208A (en) * | 2019-01-02 | 2019-05-28 | 江苏警官学院 | A kind of dense population security monitoring management method based on artificial intelligence dynamic monitoring |
-
2019
- 2019-06-10 CN CN201910494410.9A patent/CN110457987A/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20030061047A (en) * | 2002-01-07 | 2003-07-18 | 이철희 | Face recognition algorithm |
KR20130131719A (en) * | 2012-05-24 | 2013-12-04 | 현대모비스 주식회사 | Face authentication apparatus and method for vehicle |
CN108141512A (en) * | 2015-09-30 | 2018-06-08 | 株式会社尼康 | Flight instruments, mobile device, server and program |
CN105447459A (en) * | 2015-11-18 | 2016-03-30 | 上海海事大学 | Unmanned plane automation detection target and tracking method |
CN106056075A (en) * | 2016-05-27 | 2016-10-26 | 广东亿迅科技有限公司 | Important person identification and tracking system in community meshing based on unmanned aerial vehicle |
CN107451527A (en) * | 2017-06-29 | 2017-12-08 | 广东容祺智能科技有限公司 | A kind of large-scale public place hunting system based on unmanned plane |
CN107197165A (en) * | 2017-07-28 | 2017-09-22 | 哈尔滨市舍科技有限公司 | Unmanned plane self-heterodyne system and method |
CN108200168A (en) * | 2017-12-29 | 2018-06-22 | 南京奇蛙智能科技有限公司 | A kind of unmanned plane during flying method for remote management |
CN109819208A (en) * | 2019-01-02 | 2019-05-28 | 江苏警官学院 | A kind of dense population security monitoring management method based on artificial intelligence dynamic monitoring |
CN109740577A (en) * | 2019-02-28 | 2019-05-10 | 南京信息工程大学 | A kind of real-time face based on raspberry pie identifies camera system and its adjustment method again |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111552828A (en) * | 2020-04-26 | 2020-08-18 | 上海天诚比集科技有限公司 | 1-to-N face comparison method |
CN112969024A (en) * | 2020-06-30 | 2021-06-15 | 华为技术有限公司 | Camera calling method, electronic equipment and camera |
CN112969024B (en) * | 2020-06-30 | 2022-03-11 | 华为技术有限公司 | Camera calling method, electronic equipment and camera |
CN113518218A (en) * | 2021-05-25 | 2021-10-19 | 上海商汤智能科技有限公司 | Camera equipment serial number determining method and device, electronic equipment and storage medium |
CN116878468A (en) * | 2023-09-06 | 2023-10-13 | 山东省国土测绘院 | Information acquisition system for mapping |
CN116878468B (en) * | 2023-09-06 | 2023-12-19 | 山东省国土测绘院 | Information acquisition system for mapping |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110457987A (en) | Face identification method based on unmanned plane | |
CN103839194B (en) | Unmanned aerial vehicle routing inspection image retrieval system and method based on electric transmission line and GIS | |
CN205581646U (en) | Unmanned aerial vehicle electric power inspection image acquisition and processing system | |
CN112270267B (en) | Camera shooting identification system capable of automatically capturing line faults | |
CN105915856A (en) | Unmanned aerial vehicle power line inspection image data acquiring and processing system | |
CN203909857U (en) | Distribution network polling intelligent positioning and management system | |
CN204557510U (en) | Unmanned plane is taken photo by plane combination unit | |
CN105425815A (en) | Intelligent pasture management system and method by using unmanned aerial vehicle | |
CN105741477B (en) | Aircraft with intelligent fire voice assistant | |
CN107357311A (en) | A kind of reconnaissance system with unmanned plane based on mixing control technology | |
CN206224618U (en) | A kind of long-range cruising inspection system in wireless base station based on technology of Internet of things | |
CN206726271U (en) | Unmanned plane delivery control system based on Internet of Things | |
CN110493369A (en) | Method, system and the detection device that a kind of pair of concealment electronic equipment is detected | |
CN206042375U (en) | Management system is patrolled and examined to intelligence based on BLE | |
CN116092199B (en) | Employee working state identification method and identification system | |
Shen et al. | Person tracking and frontal face capture with UAV | |
CN109522996A (en) | Constant mark object automatic identification control system | |
CN108346265B (en) | A kind of risk object identification alarm method | |
CN109376707A (en) | Law-enforcing recorder with independent face identification functions | |
CN104835231B (en) | A kind of recognition of face lockset | |
Saha et al. | Face recognition drone | |
CN106303930A (en) | A kind of intelligent patrol detection based on BLE management system and method | |
CN202121706U (en) | Intelligent personnel monitoring system | |
CN110174904A (en) | A kind of more rotors based on cloud platform are taken photo by plane unmanned plane job scheduling system | |
CN109765925A (en) | A kind of location navigation fraud detection system of view-based access control model image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191115 |