Summary of the invention
Main purpose of the present invention is to provide a kind of vehicle-mounted identification authenticate device, mobile terminal, intelligent vehicle key control system and method, is intended to the recognition of face authentication function realizing remotely controlling vehicle mounted identification authenticate device, improves convenience.
To achieve these objectives, the present invention proposes a total inventive concept, and the mobile terminal remote namely utilizing car owner to carry with controls vehicle-mounted identification authenticate device, and then realizes the object of Long-distance Control recognition of face authentication function.
Therefore, the present invention proposes a kind of vehicle-mounted identification authenticate device, comprises acquisition module, wireless communication module and identification module, wherein:
Acquisition module, for obtaining facial image;
Identification module, for the facial image obtained according to the identification parameter identification certification prestored, if authentification failure, then send described facial image by described wireless communication module to the mobile terminal bound in advance, and draw authentication result according to the feedback information of described mobile terminal.
Preferably, described vehicle-mounted identification authenticate device also comprises online study module, and it is for upgrading described identification parameter according to the feedback information of described mobile terminal.
Preferably, described acquisition module comprises shooting unit, and it is for capture video image, and described vehicle-mounted identification authenticate device also comprises invasion detection module, its for:
Whether the moving region detected in described video image exceedes predetermined threshold value and detects described acquisition module whether get facial image, if described moving region exceedes predetermined threshold value and continues after Preset Time, described acquisition module does not also get facial image, then first send described video image to described mobile terminal, then give the alarm.
The present invention also proposes a kind of mobile terminal, comprises authorization identifying module, and it for receiving and showing the facial image that vehicle-mounted identification authenticate device sends, and sends feedback information to described vehicle-mounted identification authenticate device.
Preferably, described mobile terminal also comprises locking module, and it, for after the lock instruction receiving external device transmission, locks described authorization identifying module.
The present invention proposes a kind of intelligent vehicle key control system simultaneously, comprise the vehicle-mounted identification authenticate device and mobile terminal that carry out information interaction, described vehicle-mounted identification authenticate device comprises acquisition module, wireless communication module and identification module, and described mobile terminal comprises an authorization identifying module, wherein:
Acquisition module, for obtaining facial image;
Identification module, for the facial image obtained according to the identification parameter identification certification prestored, if authentification failure, then send described facial image by described wireless communication module to the mobile terminal bound in advance, and draw authentication result according to the feedback information of described mobile terminal;
Authorization identifying module, for receiving and showing the facial image that vehicle-mounted identification authenticate device sends, and sends feedback information to described vehicle-mounted identification authenticate device.
Preferably, described intelligent vehicle key control system also comprises server end, and described vehicle-mounted identification authenticate device and mobile terminal indirectly carry out information interaction by described server end, wherein:
Vehicle-mounted identification authenticate device, for sending information to described server end, and receives the information that described mobile terminal is sent to described server end;
Mobile terminal, for sending information to described server end, and receives the information that described vehicle-mounted identification authenticate device is sent to described server end;
Server end, for receiving the information of described vehicle-mounted identification authenticate device and mobile terminal transmission and storing.
The present invention proposes a kind of intelligent vehicle key control method simultaneously, comprises step:
After vehicle-mounted identification authenticate device obtains facial image, facial image according to the identification parameter identification certification prestored;
If described vehicle-mounted identification authenticate device authentification failure, then send described facial image to mobile terminal;
Described mobile terminal receives and shows described facial image, and sends feedback information according to car owner's instruction to described vehicle-mounted identification authenticate device;
Described vehicle-mounted identification authenticate device draws authentication result according to described feedback information.
Preferably, also comprise while described vehicle-mounted identification authenticate device draws the step of authentication result according to described feedback information:
Described vehicle-mounted identification authenticate device upgrades described identification parameter according to described feedback information.
Preferably, also comprise before the step of described vehicle-mounted identification authenticate device facial image according to the identification parameter identification certification prestored:
Vehicle-mounted identification authenticate device capture video image;
Whether the described vehicle-mounted identification authenticate device moving region detected in described video image exceedes predetermined threshold value and continues Preset Time, and attempts to obtain facial image from described video image;
If the moving region in described video image exceedes predetermined threshold value and continue Preset Time, and described vehicle-mounted identification authenticate device is unsuccessful gets described facial image, then send described video image to mobile terminal and give the alarm.
Preferably, after described mobile terminal receives the lock instruction of external device transmission, stop carrying out information interaction with described vehicle-mounted identification authenticate device.
Vehicle-mounted identification authenticate device provided by the present invention, mobile terminal, intelligent vehicle key control system and method, mobile terminal when vehicle-mounted identification authenticate device authentification failure then to car owner sends facial image, the facial image that car owner is shown by mobile terminal judges, and utilize mobile terminal to send feedback information to vehicle-mounted identification authenticate device, such as new user's registed authorization, registered users mandate or refusal are authorized.Thus achieve the object of the recognition of face authentication function of remotely controlling vehicle mounted identification authenticate device, car owner is registered without the need in person arriving vehicle place, or avoid vehicle-mounted identification authenticate device None-identified car owner in some cases or other lawful registration user and refuse to start the difficult situation of vehicle, improve convenience.
Detailed description of the invention
Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
See Fig. 1, intelligent vehicle key control system one embodiment of the present invention is proposed, described intelligent vehicle key control system comprises vehicle-mounted identification authenticate device 100 and mobile terminal 200, and described vehicle-mounted identification authenticate device 100 and mobile terminal 200 are bound mutually, can carry out information interaction.First described vehicle-mounted identification authenticate device 100 carries out customer identity registration (log-on message and mobile terminal synchronization), then by carrying out the face of chaufeur identifying that certification completes the driving mandate of vehicle, the processing unit of this device can by realizing based on the treater of ARM framework, the embedded system of digital signal processor or the vehicle-mounted computer based on X86-based on hardware.Described mobile terminal 200 is carried with by car owner usually, can be mobile phone or panel computer.
Described vehicle-mounted identification authenticate device 100 as shown in Figure 2 and Figure 3, comprise acquisition module 110, wireless communication module 130, initial study module 120, on-line study module 150, identification module 140 and intrusion detection module 160, described acquisition module 110, for obtaining facial image and video image, comprises shooting unit 111, Face datection unit 112 and pretreatment unit 113.
Shooting unit 111: for taking the video image in car, in order to cost-saving, single network camera can be adopted.Described shooting unit 111 also can be external camera head, is connected with vehicle-mounted identification authenticate device 100 by USB interface or IEEE1394 interface.
Face datection unit 112: for positioning the human face region in video image, detect face.Its testing process is as follows:
First carry out feature extraction, extract the feature of (different regional area sizes) under various yardstick of each position in video image.Comparatively conventional feature has Haar wavelet transform (HAAR) feature, Jia Bai (Gabor) feature etc.; The feature extracted also can be obtained by the method for study, such as, convolutional neural networks (CNN), by using back transfer algorithm, learns multilayer convolution kernel, then the image of these convolution kernels and different levels carries out convolution, and the multilayer feature of synthetic image is expressed.
Then adjudicate, according to the detector trained (this detector also can be considered a kind of binary classifier, for distinguishing face and non-face region), adjudicating which position and comprising face under which yardstick.This detector can use SVMs to realize, and self adaptation also can be used to strengthen sorting algorithms such as (AdaBoost) and realize.Wherein SVMs different classes of by study between the dividing range of proper vector (or linear or Nonlinear Mapping of vector) realize, this dividing range minimizes structure risk while minimizing training error; And self adaptation strengthens sorting algorithm by the multiple Weak Classifier of study and respective weights, then classify after multiple Weak Classifier series connection, the Weak Classifier of series connection front end can not be the region of face in a large number with the removal of very little calculated amount, the Weak Classifier of back segment detects further to remaining region again, thus can significant increase Face datection speed.
Pretreatment unit 113: for carrying out pretreatment and aligning to the facial image detected, to reduce the relative displacement at illumination and each position of face to the impact of recognition of face.Its treating process is as follows:
First Gamma correction (GammaCorrection) is carried out, to strengthen the dynamic range of topography.The gray value supposing certain pixel is γ, and so after Gamma correction, this gray value equals I
γ, the span of γ is 0 to 1.
Then Gaussian difference filtering (DifferenceofGaussianfiltering) is carried out, to remove the brightness step impact of the overall situation, as shade etc.The difference of two not homoscedastic two-dimensional Gaussian functions is carried out convolution to complete filtering as convolution kernel and facial image by Gaussian difference filtering.Gaussian difference filtering can be considered as a kind of bandpass filter, the noise very high by frequency and frequency lower brightness step change filtering.
Then contrast ratio equilibrium is carried out, with the change of the span and standardized images brightness of readjusting brightness of image.The gray value intermediate value of the gray value of each pixel divided by all pixels of view picture facial image can be taked to realize.
Finally carry out living body authentication, whether the two-dimension human face image received with certification is real human face.Can require in verification process that the face of chaufeur is towards there being certain change, according to judge whether different towards facial image carry out living body authentication.Also can allow filming apparatus up and down, side-to-side movement, the facial image according to whether photographing different visual angles when judging that camera head moves carries out certification.Detection and tracking can also be carried out, according to judging that certification is carried out in the variation whether eye has the reasons such as nictation to cause to eyes.
Wireless communication module 130: for directly communicating with the mobile terminal 200 of binding, this wireless communication module 130 can adopt 3G to communicate and interrogate module or WIFI module.
Initial study module 120: the facial image for obtaining according to acquisition module 110 during registration initially learns, and to obtain identification parameter, and is stored in this locality.Comprise extraction and the decision device parameter learning of training set facial image feature.The extraction of described training set facial image feature, namely from training set, in pretreated facial image, extract the feature of specifying, the feature comparatively commonly used for recognition of face has Jia Bai (Gabor) feature, local binary feature (LocalBinaryPattern) etc.The feature extracted also can be obtained by the method for study, can adopt K means clustering method (K-means), limited bohr hereby agate machine (restrictedboltzmannmachine) or auto coder (auto-encoder) carries out the feature that unsupervised learning obtains needing extraction.Described decision device parameter learning, can adopt SVMs, random forest (RandomForest), self adaptation to strengthen the sorting technique such as (Adaboost), artificial neural net (ANN), distance classifier and realize.
Identification module 140: for according to being pre-stored in local identification parameter, identification certification is carried out to the facial image that acquisition module 110 obtains.The facial image that the described identification parameter prestored comprises initial study module 120 couples of registered users initially learns the identification parameter (comprising characteristic parameter and decision device parameter etc.) obtained, and the follow-up identification parameter upgrading upgrading of on-line study module 150.When the confidence level identified is higher, then authorized user carries out vehicle operating; Otherwise, by wireless communication module 130, the facial image obtained and relevant information (as candidate message) thereof are sent to mobile terminal 200, then draw authentication result according to the feedback information of mobile terminal 200.Such as, if car owner carries out new user's registed authorization or registered users mandate by mobile terminal 200 pairs of current driver's, identification module 140 draws the result of authentication success, authorizes current driver's to carry out vehicle operating; If car owner refuses to authorize current driver's by mobile terminal 200, identification module 140 draws the result of authentification failure, refusal current driver's operation vehicle.
The concrete identifying of identification module 140 is: first extract feature, from the facial image obtained, namely extract the feature of the identification parameter that specific characteristic and extraction prestore.Then adjudicate, the feature of more described specific characteristic and identification parameter, obtain the similarity with the face of registered users, obtain the candidate of some.As used SVMs as decision device, by calculating the distance of sample characteristics to be identified to classification interface, estimate similarity and confidence level; As adopted distance classifier as decision device, then the distance by calculating sample characteristics to be identified and training set sample or center of a sample estimates similarity and confidence level.If the similarity of sample to be identified and all registered users all lower than certain threshold value, is then judged to unauthorized chaufeur; When the similarity with a certain registered users is higher than certain threshold value, is then judged to and authorizes chaufeur; When similarity is in time authorizing between unauthorized, then refuse judgement.When being judged to be unauthorized chaufeur or refusal judgement, call the mobile terminal that the facial image obtained and relevant information (candidate message as similar) are sent to binding by wireless communication module 130.The wherein selection of candidate, can sort according to the similarity of sample characteristics to be identified and existing registered users training sample, selects several registered users alternatively people that similarity is the highest.
On-line study module 150: for upgrading identification parameter according to the feedback information of mobile terminal 200.If car owner selects registered users to authorize current driver's by mobile terminal 200, the feature that the facial image (can be multiple images collected) of registered user's identity that on-line study module 150 is authorized according to car owner and current driver's extracts upgrades identification parameter, as classifier parameters and/or characteristic parameter.If car owner selects to carry out new user's registed authorization to current chaufeur, then illustrate it is new registered user, on-line study module 150 just sets up new user identity identification number, extracts the identification parameter of new user.
Characteristic parameter is upgraded, if the feature that initially study module 120 extracts is obtained by the mode of study, then needs the facial image upgrading to adapt to new mandate or new registration to the feature learnt.If what initial study module 120 adopted is that K means clustering method carries out clustering learning, then online K means clustering method (online-kmeans) can be adopted to carry out the renewal of feature.Wherein, online K means clustering method, the distance according to new samples and original cluster centre upgrades.If the distance of new samples and cluster centre is less than certain threshold value, then recalculate the average of corresponding cluster centre according to this sample; If be greater than certain threshold value, then newly-increased cluster centre.
Decision device parameter is upgraded, namely according to the feature of current face's image zooming-out, decision device parameter is upgraded.If the decision device of initial study module 120 adopts SVMs or artificial neural net (ANN), then can upgrade by stochastic gradient descent method (StochasticGradientDecent) or with the classifying face parameter of loom gradient descent method (StochasticSub-gradientDecent) to SVMs; If the decision device of initial study module 120 adopts random forest, then online random forest (On-lineRandomForest) can be adopted to upgrade.If wherein online random forest carries out according to newly-increased number of samples and to existing decision tree the entropy production that expansion may bring, the decision tree in random forest is upgraded.If newly-increased number of samples is greater than certain threshold value, if carry out expanding the entropy production that can bring to decision tree to be greater than certain threshold value simultaneously, then decision tree is expanded, increase the node of decision tree.
Intrusion detection module 160: whether exceed predetermined threshold value for the moving region detected in video image that described acquisition module 110 collects, and detect described acquisition module 110 and whether get facial image, if described moving region exceedes predetermined threshold value and continues after Preset Time, described acquisition module 110 does not also get facial image, then first send described video image to described mobile terminal 200, then give the alarm.
Concrete, described moving region can adopt the foreground extracting method based on background modeling to obtain, and first uses the span of mixed Gauss model (weighted sum of Gaussian function) to pixel point value each in video image to carry out modeling; Then, in foreground extraction process, if current pixel point can be described by the mixed Gauss model of background, be then judged to be background dot, otherwise be judged to be foreground point.Described moving region also can adopt frame difference method to detect and obtain, and namely by front and back frame subtract to video image, finds out the pixel that front and back frame changes greatly, and judges moving region after continuous multiple frames carries out adding up.After moving region being detected, whether the area and the full images area ratio that calculate moving region are greater than predetermined threshold value, if be greater than predetermined threshold value and continued after Preset Time (as 30 seconds), still do not receive the information that acquisition module 110 obtains facial image, after then sending the video image of this time period by wireless communication module 130 to the mobile terminal 200 of car owner immediately, then give the alarm.Thus destroy filming apparatus after can preventing invasion personnel from finding the game is over, cause cannot collecting evidence afterwards.
Described mobile terminal 200 as shown in Figure 4, Figure 5, can be mobile phone or panel computer etc., comprises authorization identifying module 210 and locking module 220, wherein:
Authorization identifying module 210: for receiving and showing the facial image and video information that vehicle-mounted identification authenticate device 100 sends, and send the feedback informations such as registered users mandate, new user's registed authorization or refusal mandate to described vehicle-mounted identification authenticate device 100.Described mobile terminal 200 can based on Mobile operating system developing applications such as iOS, Android, WindowsPhone, and described authorization identifying module 210 provides clear, maneuverable user interface by described application program to car owner.As shown in Figure 5, illustrate the fundamental element of user interface, wherein region 211 is for showing image or video, checks and reference for user; Region 212 is " registration " new user button, and user can pass through the new user of this button remote login, immediately licenses to current driver's steering vehicle, and gives vehicle-mounted identification authenticate device 100 by information feed back; Multiple rectangular areas in region 213 show facial image and the refusal icon of registered users by phase knowledge and magnanimity, car owner confirms to authorize the registered but chaufeur of vehicle-mounted identification authenticate device 100 None-identified by clicking these facial images, or click refusal icon refusal mandate current driver's, and give vehicle-mounted identification authenticate device 100 by information feed back; Region 214 display application program status, such as, the coupled condition etc. of vehicle-mounted identification authenticate device 100 and mobile terminal 200 or server.
Locking module 220: for after the lock instruction receiving external device transmission, lock described authorization identifying module 210, thus the authorization identifying function of turning-off mobile terminal 200, avoid unauthorized person to walk around the recognition of face certification of vehicle-mounted identification authenticate device 100 by the mobile terminal 200 stealing car owner.Described external device can be final terminal or the server of any transmission lock instruction, and described lock instruction can be self-defined by car owner.
See Fig. 6, intelligent vehicle key control system second embodiment of the present invention is proposed, described intelligent vehicle key control system is compared with above-described embodiment, be to add server end 300, make described vehicle-mounted identification authenticate device 100 and mobile terminal 200 indirectly carry out information interaction by described server end 300.Wherein, the account that described vehicle-mounted identification authenticate device 100 is corresponding with described server end 300 carries out binding that (station server can process the information of multiple different user simultaneously, the server account that different users is corresponding different), send information to described server end 300, and receive the information that described mobile terminal 200 is sent to described server end 300; The account that described mobile terminal 200 is corresponding with described server end 300 is bound, and sends information to described server end 300, and receives the information that described vehicle-mounted identification authenticate device 100 is sent to described server end 300; Described server end 300 is for receiving the information of described vehicle-mounted identification authenticate device 100 and mobile terminal 200 transmission and storing.Described server end 300 can adopt desktop computer, and on desktop computer, loads corresponding application program as server platform; Or use the network server platform that Web server business provides.
For the network server platform that google provides, the application program of server end 300 can use Google's application engine to realize, and it is for the image that receives vehicle-mounted identification authenticate device 100 and upload or video, the simultaneously feedback information of mobile terminal receive 200.Mobile terminal 200 and vehicle-mounted identification authenticate device 100 are then regularly to server end 300 obtaining information.Concrete, when server end 300 receive vehicle-mounted identification authenticate device 100 with the posting request of image after, use Google application engine by image or video storage in the data warehouse of Google's application engine, corresponding state variable is set as very, representative has new renewal (PYTHON language can be adopted to realize this step in Google's application engine) simultaneously.Whether the state variable that mobile terminal 200 listens to server end 100 is in real time true, if be true, the image then just stored by server end 300 li or video download to mobile terminal 200, allow car owner select whether to authorize current driver's steering vehicle, the feedback information that car owner is sent by mobile terminal 200 then to be uploaded onto the server end 300 by posting request.The whole device 100 of vehicle-mounted identification certification also carrys out the feedback information of mobile terminal receive 200 by the state variable of the user instruction listening to server end 300 in real time, and determines whether that authorizing current driver's drives power and upgrade identification parameter according to feedback information.
If the mobile terminal of car owner 200 is lost, car owner can send lock instruction by server end 300 to described mobile terminal 200, close the authorization identifying function of described mobile terminal 200, avoid unauthorized person to pass through to steal the recognition of face certification that mobile terminal 200 walks around vehicle-mounted identification authenticate device 100.
Accordingly, intelligent vehicle key control system of the present invention, has the following advantages:
(1) when vehicle-mounted identification authenticate device 100 None-identified non-registered users, car owner can by mobile terminal 200 long-range carry out new user's registration, car owner being registered without the need in person arriving vehicle scene, improve convenience.
(2) when vehicle-mounted identification authenticate device 100 pairs of registered users None-identifieds, car owner only need select active user according to facial image by mobile terminal 200 from alternative registered user, easily can complete mandate.Avoid in some cases, vehicle-mounted identification authenticate device 100 None-identified car owner or other lawful registration user and refuse the difficult situation starting vehicle.
And, the mobile terminal 200 of car owner is with lock function, when the mobile terminal 200 of car owner is stolen, car owner can send lock instruction by other external devices such as mobile terminal or server to stolen mobile terminal 200, the authorization identifying function of locking stolen mobile terminal 200, thus avoid other people to avoid the face authentication of vehicle by the mobile terminal 200 stealing car owner, improve safety.
(3) simultaneously, based on the feedback information such as new user's registed authorization, registered users mandate that car owner is fed back by mobile terminal 200, the on-line study module 150 of vehicle-mounted identification authenticate device 100 can be carried out upgrading to identification parameter and upgrade.By completing the on-line study of vehicle-mounted identification authenticate device 100 alternately with car owner, avoid the issuable wrong replacement problem of full automaticity on-line study method, decrease unnecessary renewal rewards theory and calculated amount, make the comformability of system to same face changing features under various conditions stronger, the performance automatically identified is become better and better.Along with the increase of system period of service, the number of times needing car owner to intervene will significantly reduce, and improve the accuracy of identification.
(4) the intrusion detection module 160 of vehicle-mounted identification authenticate device 100 possesses the function monitored vehicle interior.When there being people to enter vehicle, vehicle-mounted identification authenticate device 100 can remind the personnel of entering to carry out face authentication.When the personnel face entering vehicle deliberately avoids filming apparatus, can not complete certification at the appointed time, video image can be sent to the mobile terminal 200 of server end 300 or car owner by intrusion detection module 160 immediately, then gives the alarm.For evidence obtaining is in the future produced evidence, invasion personnel are avoided deliberately to destroy Evidence loss after camera head.
See Fig. 7, intelligent vehicle key control method one embodiment of above-mentioned intelligent vehicle key control system is proposed, when the intelligent vehicle key control system that described intelligent vehicle key control system is the first embodiment as shown in Figure 1, when vehicle-mounted identification authenticate device and mobile terminal directly carry out information interaction, described intelligent vehicle key control method comprises the following steps:
Step S101, vehicle-mounted identification authenticate device obtain facial image.
After user enters operator's compartment, vehicle-mounted identification authenticate device obtains the facial image of user immediately.
Step S102, vehicle-mounted identification authenticate device judge whether facial image mates with the identification parameter prestored.
After obtaining facial image, vehicle-mounted identification authenticate device, then according to the identification parameter identification certification facial image prestored, if facial image matches with the identification parameter prestored, then enters step S103; If facial image does not mate with the identification parameter prestored, then enter step S104.
Step S103, authentication success.
After authentication success, vehicle-mounted identification authenticate device then authorized user carries out vehicle operating.
Step S104, vehicle-mounted identification authenticate device send facial image to mobile terminal.
After authentification failure, the facial image obtained and relevant information (as candidate message) thereof are then sent to mobile terminal by vehicle-mounted identification authenticate device, and this mobile terminal is carried with by car owner.
Step S105, mobile terminal receive and show facial image, and send feedback information according to car owner's instruction to vehicle-mounted identification authenticate device.
After mobile terminal receives facial image, show this facial image to car owner, car owner carries out corresponding operation immediately and sends corresponding instruction to mobile terminal, as sent registered users authorized order, new user's registed authorization instruction, refusal instruction etc.Mobile terminal sends corresponding feedback information according to car owner's instruction to vehicle-mounted identification authenticate device, as sent the feedback informations such as registered users mandate, new user's registed authorization or refusal mandate.
Step S106, vehicle-mounted identification authenticate device draw authentication result according to feedback information.
If vehicle-mounted identification authenticate device receives the feedback information of new user's registed authorization or registered users mandate, then draw the result of authentication success, authorize current driver's to carry out vehicle operating; If when vehicle-mounted identification authenticate device receives the feedback information of refusal mandate, then draw the result of authentification failure, refusal current driver's operation vehicle.
Step S107, vehicle-mounted identification authenticate device upgrade identification parameter according to feedback information.
If vehicle-mounted identification authenticate device receives the feedback information of registered users mandate, then the feature extracted according to registered user's identity of car owner's mandate and the facial image (can be multiple images collected) of current driver's upgrades identification parameter.If vehicle-mounted identification authenticate device receives the feedback information of new user's registed authorization, then illustrate it is new registered user, then set up new user identity identification number, extract the identification parameter of new user.
See Fig. 8, intelligent vehicle key control method second embodiment of above-mentioned intelligent vehicle key control system is proposed, when the intelligent vehicle key control system that described intelligent vehicle key control system is the second embodiment as shown in Figure 6, when vehicle-mounted identification authenticate device and mobile terminal indirectly carry out information interaction by server end, described intelligent vehicle key control method comprises the following steps:
Step S201, vehicle-mounted identification authenticate device obtain facial image.
After user enters operator's compartment, vehicle-mounted identification authenticate device obtains the facial image of user immediately.
Step S202, vehicle-mounted identification authenticate device judge whether facial image mates with the identification parameter prestored.
After obtaining facial image, vehicle-mounted identification authenticate device, then according to the identification parameter identification certification facial image prestored, if facial image matches with the identification parameter prestored, then enters step S203; If facial image does not mate with the identification parameter prestored, then enter step S204.
Step S203, authentication success.
After authentication success, vehicle-mounted identification authenticate device then authorized user carries out vehicle operating.
Step S204, vehicle-mounted identification authenticate device send facial image to server end.
After authentification failure, the facial image obtained and relevant information (as candidate message) thereof are sent to server end by vehicle-mounted identification authenticate device, and received server-side also stores facial image and relevant information thereof.
Step S205, mobile terminal are downloaded facial image from server end and are shown, and send feedback information according to car owner's instruction to server end.
After mobile terminal receives facial image, show this facial image to car owner, car owner operates accordingly immediately and sends corresponding instruction to mobile terminal, as sent registered users authorized order, new user's registed authorization instruction, refusal instruction etc.Mobile terminal sends corresponding feedback information according to car owner's instruction to server end, and as sent the feedback informations such as registered users mandate, new user's registed authorization or refusal mandate, received server-side also stores described feedback information.
Step S206, vehicle-mounted identification authenticate device are downloaded feedback information from server end, and are drawn authentication result according to feedback information.
If vehicle-mounted identification authenticate device receives the feedback information of new user's registed authorization or registered users mandate, then draw the result of authentication success, authorize current driver's to carry out vehicle operating; If when vehicle-mounted identification authenticate device receives the feedback information of refusal mandate, then draw the result of authentification failure, refusal current driver's operation vehicle.
Step S207, vehicle-mounted identification authenticate device download feedback information from server end, and upgrade identification parameter according to feedback information.
If vehicle-mounted identification authenticate device receives the feedback information of registered users mandate, then the feature extracted according to registered user's identity of car owner's mandate and the facial image (can be multiple images collected) of current driver's upgrades identification parameter.If vehicle-mounted identification authenticate device receives the feedback information of new user's registed authorization, then illustrate it is new registered user, then set up new user identity identification number, extract the identification parameter of new user.
See Fig. 9, propose intelligent vehicle key control method the 3rd embodiment of the present invention, described intelligent vehicle key control method can also be monitored in real time to vehicle, and prevent lawless people from invading vehicle and implement theft, concrete steps are as follows:
Step S301, vehicle-mounted identification authenticate device capture video image.
After vehicle is out of service, vehicle-mounted identification authenticate device captured in real-time video image, monitors vehicle interior.
Whether step S302, the vehicle-mounted identification authenticate device moving region detected in video image exceedes predetermined threshold value and continues Preset Time.
Vehicle-mounted identification authenticate device attempts to obtain facial image from video image with while carrying out recognition of face certification, also detects the moving region in video image, has judged whether that personnel enter vehicle with this.Described moving region can adopt the foreground extracting method based on background modeling to obtain, and first uses the span of mixed Gauss model (weighted sum of Gaussian function) to pixel point value each in video image to carry out modeling; Then, in foreground extraction process, if current pixel point can be described by the mixed Gauss model of background, be then judged to be background dot, otherwise be judged to be foreground point.Described moving region also can adopt frame difference method to detect and obtain, and namely by front and back frame subtract to video image, finds out the pixel that front and back frame changes greatly, and judges moving region after continuous multiple frames carries out adding up.After moving region being detected, whether the area and the full images area ratio that calculate moving region are greater than predetermined threshold value, if be greater than predetermined threshold value and continue Preset Time, then illustrating has personnel into the car, enters step S303; If be less than predetermined threshold value, then illustrate and do not have personnel into the car, proceed to detect.
Step S303, vehicle-mounted identification authenticate device judge whether to get facial image from video image.
After the personnel of having detected enter vehicle, vehicle-mounted identification authenticate device then points out the personnel of entering to carry out recognition of face certification, judge whether from video image, successfully to have got facial image simultaneously, if do not get facial image, then illustrate that the personnel of entering have a mind to hide face authentication, be likely that lawless people invades in car, then enter step S304; If successfully got facial image, then enter step S305-S310, step S305-S310 is identical with the step S102-S107 in aforementioned intelligent car key control method first embodiment, does not repeat them here.Step S305-S310 also can replace to the step S202-S207 in aforementioned intelligent car key control method second embodiment simultaneously.
Step S304, vehicle-mounted identification authenticate device send video image to mobile terminal and give the alarm.
Vehicle-mounted identification authenticate device defines lawless people and invades, then first send video image to retain evidence to the mobile terminal of car owner, and then give the alarm.Or also can send video image to registered server end.
In above-mentioned intelligent vehicle key control method first, second, third embodiment, if after the mobile terminal of car owner receives the lock instruction that external device sends as other mobile terminal or server end, then stop carrying out information interaction with described vehicle-mounted identification authenticate device, receive the human face image information directly sent by vehicle-mounted identification authenticate device as stopped or stopping downloading from server end the human face image information uploaded by vehicle-mounted identification authenticate device, or stopping directly sending feedback information to vehicle-mounted identification authenticate device or uploading feedback information to server end.The intelligent vehicle key control system of vehicle is controlled after stealing car owner's mobile terminal to prevent lawless people.
In sum, intelligent vehicle key control method of the present invention, has the following advantages:
(1) after vehicle-mounted identification authenticate device is to active user's recognition of face authentification failure, mobile terminal then to car owner sends human face image information, the command operating maybe will not authorized is authorized by long-range the carrying out of car owner, achieve the object of the recognition of face authentication function of remotely controlling vehicle mounted identification authenticate device, improve convenience.
(2) mobile terminal of car owner is with lock function, when the mobile terminal of car owner is stolen, car owner can send lock instruction by other external devices such as mobile terminal or server to stolen mobile terminal, the authorization identifying function of locking stolen mobile terminal, thus avoid other people to avoid the face authentication of vehicle by the mobile terminal stealing car owner, improve safety.
(3) simultaneously, based on the feedback information such as new user's registed authorization, registered users mandate that car owner is fed back by mobile terminal, vehicle-mounted identification authenticate device can carry out upgrading to identification parameter and upgrade.By completing the on-line study of vehicle-mounted identification authenticate device alternately with car owner, avoid the issuable wrong replacement problem of full automaticity on-line study method, decrease unnecessary renewal rewards theory and calculated amount, make the comformability of system to same face changing features under various conditions stronger, the performance automatically identified is become better and better.Along with the increase of system period of service, the number of times needing car owner to intervene will significantly reduce, and improve the accuracy of identification.
(4) vehicle-mounted identification authenticate device also possesses the function monitored vehicle interior.When detecting that someone enters vehicle, vehicle-mounted identification authenticate device can remind the personnel of entering to carry out face authentication.When the personnel face entering vehicle deliberately avoids filming apparatus, can not complete certification at the appointed time, video image can be sent to the mobile terminal of server end or car owner by vehicle-mounted identification authenticate device immediately, then gives the alarm.For evidence obtaining is in the future produced evidence, invasion personnel are avoided deliberately to destroy Evidence loss after camera head.
Should be understood that; these are only the preferred embodiments of the present invention; can not therefore limit the scope of the claims of the present invention; every utilize specification sheets of the present invention and accompanying drawing content to do equivalent structure or equivalent flow process conversion; or be directly or indirectly used in other relevant technical fields, be all in like manner included in scope of patent protection of the present invention.