The application is the divisional application of Application No. 201510170322.5, and the applying date of original application is 2015-04-11,
Entitled safe biologic feature recognition and image capturing system for intelligent movable equipment.
The content of the invention
The above technical problem of solution, the present invention provide a kind of safe biologic feature recognition for intelligent movable equipment and
Image capturing system.
A kind of method that image for mobile terminal living things feature recognition shows, it is characterised in that including step:
A) biometric image that collects simultaneously exports image;B) displacement treatment is carried out to the image, so that institute
The biometric image of collection is maintained at axle direct-view state;C) the displacement image that the displacement treatment is obtained is shown.
Preferably, the collection biometric image of the step a is Real-time Collection biometric image.
Preferably, the displacement image that the display of the step c displacement treatment is obtained to show at displacement in real time
The image that reason is obtained.
Preferably, the displacement treatment of the image of the step b is comprised the following steps:
The image center that display screen shows adjusts by X-Y coordinate axial displacement amount, described to be adjusted to calculate display screen
The image center X of display, Y-coordinate axle physical location and biometric image collecting device optical centre X, Y-coordinate axle thing
Displacement between reason position;
According to the displacement for being calculated, displacement treatment is carried out to the image.
Preferably, the image center X for being shown according to display screen, Y-coordinate axle physical location, biometric image collection
Equipment optical centre X, Y-coordinate axle physical location, the optical magnification of biometric image collecting device, optical imaging lens
Equivalent focal length, the unit pixel physical size of object distance and image imaging sensor calculates the displacement of X-axis and Y-axis.
Preferably, the displacement of the calculating X-axis and Y-axis is specifically calculated as follows:
X_SHIFT=β * (Xscreen-Ximager)/PS;
Y_SHIFT=β * (Yscreen-Yimager)/PS;
Wherein,
(X_SHIFT, Y_SHIFT) is respectively the image center X that display screen shows, Y-coordinate axle displacement, unit picture
Element, pixels;
(Xscreen, Yscreen) is respectively the image center X that display screen shows, Y-coordinate axle physical location, unit
Centimetre, cm;
(Ximager, Yimager) is respectively biometric image collecting device optical centre X, Y-coordinate axle physical location,
Unit centimetre, cm;
β is the optical magnification of biometric image collecting device, without unit;
EFL is the equivalent focal length of optical imaging lens, unit millimeter, mm;
D is object distance, unit centimetre, cm;
PS is the unit pixel physical size of image imaging sensor, the every pixel of unit micron, um/pixel.
The mobile terminal that a kind of image for mobile terminal living things feature recognition shows, it is characterised in that:Including:
Biometric image collecting device, for gathering biometric image;The safe drive control unit of display screen, for performing in real time
Detection and feedback prompts that image displacement shows, the safe drive control unit of display screen are aobvious including image displacement
Show unit, carry out the display methods of displacement treatment to the image for performing, methods described includes step:A) receive and adopt
The biometric image that collects simultaneously exports image;B) displacement treatment is carried out to the image, so that the life for being gathered
Thing characteristic image is maintained at axle direct-view state;C) the displacement image that the displacement treatment is obtained is shown.
Preferably, the biometric image collecting device is Real-time Collection biometric image.
Preferably, the safe drive control unit of the display screen is to show the image that displacement treatment is obtained in real time.
Preferably, image displacement display unit carries out the step of displacement is processed and specifically includes:Display screen is calculated to show
Image center adjusted by X-Y coordinate axial displacement amount, it is described to be adjusted to calculate the image center that shows of display screen
X, Y-coordinate axle physical location and biometric image collecting device optical centre X, the displacement between Y-coordinate axle physical location;
According to the displacement, displacement treatment is carried out to the Imaging biological characteristic image.
Preferably, the image center X that the display screen shows according to image displacement display unit, Y-coordinate axle
Physical location, biometric image collecting device optical centre X, Y-coordinate axle physical location, biometric image collecting device
The unit pixel physical size meter of optical magnification, the equivalent focal length of optical imaging lens, object distance and image imaging sensor
Calculate the displacement of X-axis and Y-axis.
Preferably, the image displacement display unit calculates X-axis and the displacement of Y-axis is specifically calculated as follows:
X_SHIFT=β * (Xscreen-Ximager)/PS;
Y_SHIFT=β * (Yscreen-Yimager)/PS;
Wherein,
(X_SHIFT, Y_SHIFT) is respectively the image center X that display screen shows, Y-coordinate axle displacement, unit picture
Element, pixels;
(Xscreen, Yscreen) is respectively the image center X that display screen shows, Y-coordinate axle physical location, unit
Centimetre, cm;
(Ximager, Yimager) is respectively biometric image collecting device optical centre X, Y-coordinate axle physical location,
Unit centimetre, cm;
β is the optical magnification of biometric image collecting device, without unit;
EFL is the equivalent focal length of optical imaging lens, unit millimeter, mm;
D is object distance, unit centimetre, cm;
PS is the unit pixel physical size of image imaging sensor, the every pixel of unit micron, um/pixel.
A kind of safe biologic Feature Recognition System for intelligent movable equipment, including application operating system and safety behaviour
Make system;The application operating system includes the application proxy for performing the application operating system end standardized A PI invocations of procedure
Service interface unit;The SOS includes the peace for performing the SOS end standardized A PI invocations of procedure
The safe computing unit of full application proxy service interface unit, the data for performing SOS end and code calculating,
For providing the memory space of SOS end data and code, and ensure that safe computing unit passes through individual secure address
The secure storage unit of bus access, the memory headroom for providing SOS end data and code, and ensure safety
The secure memory unit that computing unit is accessed by individual secure address bus;The safe computing unit includes PKI encryption label
Name security algorithm unit and living things feature recognition algorithm unit;The living things feature recognition algorithm unit includes registration algorithm unit
With identification and self-learning algorithm unit.
As the improvement to the safe biologic Feature Recognition System for intelligent movable equipment of the present invention:Safety is raw
The register flow path of thing feature recognition is as follows:A, when outside initiation apply registration request;B, by application proxy service interface unit
Transmit the registration request and notify safety applications agency service interface unit;Standardized API processes under c, SOS end
Call execution registration algorithm unit;D, PKI encryption signature security algorithm unit perform registering result ciphering signature, should by safety
Ciphering signature registering result is sent with agency service interface unit notify application proxy service interface unit;E, application proxy service
Interface unit returning to external receives ciphering signature registering result.
As the further improvement to the safe biologic Feature Recognition System for intelligent movable equipment of the present invention:
The identifying procedure of safe biologic feature recognition is as follows:1. when application authorization request is initiated in outside;2. connect by application proxy service
Mouth unit transmits the certification request and notifies safety applications agency service interface unit;3. it is standardized under SOS end
The API invocations of procedure perform identification and self-learning algorithm unit;4. PKI encryption signature security algorithm unit performs authentication result encryption
Signature, sends ciphering signature authentication result and notifies application proxy service interface list by safety applications agency service interface unit
Unit;5. application proxy service interface unit returning to external receives ciphering signature authentication result.
As the further improvement to the safe biologic Feature Recognition System for intelligent movable equipment of the present invention:
The registration algorithm unit is used to perform following rate-determining steps:
S200, initialization definitions cycle count COUNT=0, template counts N=0, template number TN, loop number TC;Its
In:COUNT is defined as the cycle count in algorithm register flow path, and N is defined as the template counts in algorithm register flow path;TN is defined
It is template number predetermined in algorithm register flow path, preferred TN>=3;TC is defined as predetermined circulation in algorithm register flow path
Quantity, preferred TC>=3;S201, collection meet the real-time biological characteristic image of adaptive quality standard;S202, extraction image
Biological information produces feature templates, template counts to add up N=N+1 from increasing;S203, judge templet count N >=TN, are to perform
S204, no return performs S201;S204, cycle count add up COUNT=COUNT+1 from increasing;S205, judge the TN mould of quantity
Whether combination comparison result set C (TN, 2) formed between plate all meets predetermined registration threshold denominators, is to perform S208, no
Perform S206;Wherein:Described combination comparison result set C (TN, 2)=TN!/(2!*(TN-2)!) S206, judge circulation meter
Number COUNT>TC, is to perform S207, and no reset template counts N=0 simultaneously returns to execution S201;S207, registration failure;S208, note
Volume success, the template of global optimum is that enrollment and PKI encryption are stored in safe storage in selection form assembly comparison result
Unit.
As the improvement to the safe biologic Feature Recognition System for intelligent movable equipment of the present invention:The knowledge
It is not used to perform following rate-determining steps with self-learning algorithm unit:S301, collection meet the real-time biological of adaptive quality standard
Characteristic image;S302, extraction image biological characteristic information produce current signature template;S303, judge current template with registration mould
Whether plate or self study template matching result meet predetermined identification threshold denominators, are to perform S304, no execution S306;S304、
Whether judge templet comparison result meets predetermined self study threshold denominators, is to perform S305, no execution S307;S305, judgement
Whether image biological characteristic mass meets predetermined self study quality standard, is to perform S308, no execution S307;S306, identification
Fail with template self study;S307, recognize successfully, the failure of self study template;S308, recognize successfully and PKI encryption preserves current
Template is self study template in secure storage unit.
As to the described safe biologic Feature Recognition System for intelligent movable equipment of the present invention, its feature
It is:It is described to gather the real-time biological characteristic image for meeting adaptive quality standard, comprise the following steps:S400, initialization definitions:
High-grade quality standard Qh, inferior grade quality standard Ql, adaptive quality standard Q=Qh, quality standard fine setting step-length △ Q;
S401, collection real-time biological characteristic image;S402, judge adaptive quality standard Q >=Ql, be to perform S403, no Q=Ql is simultaneously returned
Receipt row S401;S403, judge whether present image biological characteristic quality meets adaptive quality standard Q, be to perform S404, it is no
Q=Q- △ Q simultaneously return to execution S401;S404, return meet the real-time biological characteristic image of adaptive quality standard.
A kind of biometric image acquisition system for intelligent movable equipment, including for gathering biometric image
Biometric image collecting device, for perform biometric image collecting device Real-time Feedback drive control output image
The safe drive control unit of biometric image collecting device, shown for performing real time imagery picture displacement, image
The safe drive control unit of the display screen of detection and feedback prompts that biological characteristic is offset and mirror-reflection is disturbed.
As the improvement to biometric image acquisition system of the present invention:The biometric image collecting device
Including the range sensor for measuring object distance information, the near-infrared LED light source for radiant image near infrared light, for driving
The current driver of near-infrared LED radiation of light source intensity adjustment, for filter visible light-transmissive for be imaged near infrared light
Preposition optical filter and/or rearmounted optical filter, the optical imaging lens of near infrared light are focused on for physics, for driving
The auto-focusing driver of optical imaging lens auto-focusing, the image for opto-electronic conversion output image are imaged sensing
Device, for showing displacement image and being shown as the feedback prompts information disturbed as image biological characteristic offset and mirror-reflection
Display screen.
As the further improvement to biometric image acquisition system of the present invention:The biometric image is adopted
Collection equipment safety drive control unit includes object distance information measurement unit, radiation intensity adjustment unit, auto-focusing unit, imaging
Mode controlling unit.
As the further improvement to biometric image acquisition system of the present invention:The display screen drives safely
Control unit includes image displacement display unit, image biological characteristic offset detection unit, the interference inspection of image mirror-reflection
Survey unit, feedback prompts unit.
As the further improvement to biometric image acquisition system of the present invention:The biometric image is adopted
Collection equipment safety drive control unit is used to perform biometric image collecting device Real-time Feedback drive control output image
As including following feedback control step:S101. the object distance information of object distance information measurement unit real time dynamic measurement range sensor
D;S102. whether object distance information D being judged in working range, being to perform S103, no feedback prompts adjust distance and return is held
Row S101;S103. according to object distance information D, real-time dynamic feedback control current driver drives near red radiation intensity adjustment unit
Outer LED light source radiation intensity I changes;S104. auto-focusing unit is according to object distance information D, and real-time dynamic feedback control is automatically right
Burnt driver drives optical imaging lens auto-focusing;S105. the real-time dynamic control synchronous sequence arteries and veins of imaging pattern control unit
Rush frame radiation intensity I, frame time T and frame rate F the output image of imaging pattern.
As the further improvement to biometric image acquisition system of the present invention:The auto-focusing unit pair
Currently the corresponding focal position regional areas of object distance information D are further finely tuned, i.e., corresponding poly- with current object distance information D
Centered on burnt position, define regional area carries out the scanning of fine step to obtain the focusing of optimal rear focus position for scope
Effect;Frame time and the frame rate output of the real-time dynamic control synchronous sequence Pulse Imageing pattern of imaging pattern control unit
Image method is specifically by synchronous sequence control image imaging sensor exposure and the pulse of near-infrared LED radiation of light source
Frame time T and frame rate the F output of imaging pattern are realized;The Pulse Imageing pattern realizes that frame is radiated using pulse amplitude modulation
Intensity I controlled output;The Pulse Imageing pattern realizes frame time T controlled outputs using pulse width duty ratio modulation;It is described
Pulse Imageing pattern realizes frame rate F controlled outputs using unit pulse periodic frequency modulation.
As the further improvement to biometric image acquisition system of the present invention:The display screen drives safely
Control unit is used to perform real time imagery picture displacement and show, the detection that the skew of image biological characteristic and mirror-reflection are disturbed
And feedback prompts, including following rate-determining steps:S106, the displacement of image displacement display unit Real time dynamic display treatment into
As image;The position of biological characteristic and yardstick are inclined in S107, image biological characteristic offset detection unit Real-time and Dynamic Detection image
Move;S108, judge whether biological characteristic position and yardstick offset in image, be feedback prompts unit perform biological characteristic position and
Yardstick offset feedback is pointed out and returns to execution S106, no execution S109;S109, image mirror-reflection Interference Detection unit are moved in real time
The position of mirror-reflection and yardstick are disturbed in state detection image;S110, judge whether image mirror-reflection position and yardstick disturb,
It is that feedback prompts unit performs mirror-reflection position and yardstick interference feedback is pointed out and returns to execution S106, no return is performed
S106。
As the further improvement to biometric image acquisition system of the present invention:The displacement of the image
Treatment includes:The image center that display screen shows adjusts (X_SHIFT, Y_SHIFT) by X-Y coordinate axial displacement amount;Institute
State specifically being calculated as follows for adjustment X-Y axial displacements amount (X_SHIFT, Y_SHIFT):
X_SHIFT=β * (Xscreen-Ximager)/PS
Y_SHIFT=β * (Yscreen-Yimager)/PS
β=EFL/ (D-EFL)
Wherein:(X_SHIFT, Y_SHIFT) is respectively the image center X that display screen shows, Y-coordinate axle displacement,
Unit pixel, pixels;(Xscreen, Yscreen) is respectively the image center X that display screen shows, Y-coordinate axle physics
Position, unit centimetre, cm;(Ximager, Yimager) is respectively biometric image collecting device optical centre X, Y-coordinate axle
Physical location, unit centimetre, cm;β is the optical magnification of biometric image collecting device, without unit;EFL is studied for light
As the equivalent focal length of lens, unit millimeter, mm;D is object distance, unit centimetre, cm;;PS is the unit picture of image imaging sensor
Plain physical size, the every pixel of unit micron, um/pixel.
As the further improvement to biometric image acquisition system of the present invention:Described feedback prompts unit
Perform display screen and show dynamic adjustment in real time biological characteristic position and yardstick offset information;The feedback prompts unit performs display
Screen display dynamic adjustment in real time mirror-reflection position and yardstick interference information.
One kind is entered for realizing biometric image acquisition system biological characteristic individuality raying total amount in unit period
The combined optimization control method of row security restriction and the acquisition glitch-free biological characteristic image of high-quality, including:(1) in thing
Near-infrared LED radiation of light source is completely closed when exceeding the closely limit away from information D;(2) biometric image acquisition system is defined
The biological characteristic individuality raying energy J0 within per frame period;J0=E*T;Biometric image acquisition system is defined in unit week
Biological characteristic individuality raying energy JE in phase;JE=J0*F<Jlimit;(3) according to the definition relation of JE in (2), by connection
Close optimal control parameter:Biological characteristic individuality raying energy J0 in per the frame period, and frame number F in unit period, realize to list
Biological characteristic individuality raying energy limits JE in bit period<Jlimit;(4) according to the definition relation of J0 in (2), combined optimization
Control parameter E, T inverse relation, i.e. E=J0/T, T=J0/E, improve the individual per frame period raying illumination E of biological characteristic, drop
Low biological characteristic is individual per frame period raying and time for exposure T, i.e., short per the frame period using radiant illumination E high of per frame period
Radiation and time for exposure T are used to obtain the glitch-free biological characteristic image of high-quality;Wherein:Jlimit is in the unit cycle
The safety margins energy of biological characteristic individuality raying;T is biological characteristic individual per raying time in frame period, i.e. image imaging
The frame time of the synchronous sequence Pulse Imageing pattern of exposure sensor and near-infrared LED radiation of light source;F is frame in the unit cycle
The frame rate of the synchronous sequence Pulse Imageing pattern of the exposure of number, i.e. image imaging sensor and near-infrared LED radiation of light source;E is
Biological characteristic is individual per frame period raying illumination;D is object distance information.
Foregoing description is summarized, there is advantages below by the present invention:
1st, complete secure registration identifying procedure system is realized, it is ensured that do not attacked by external security.
2nd, realize in authentication registration porcess system that biometric templates are optimal with stability and uniformity, improve registration and
Success rate during identification, while improving registration and recognition speed;.
3rd, realize in authentication registration porcess system when different acquisition environment and biological characteristic itself change
When further improve registration and recognize when success rate, while improve registration and recognition speed;
4th, image biological characteristic mass standard self adaptation in authentication registration porcess system is realized, registration is improved and is recognized successfully
Rate, while improving registration and recognition speed;
5th, the biological characteristic image of acquisition brightness constancy in IMAQ porcess system is realized.
6th, the biological characteristic image that quick obtaining focus are stablized in 100ms in IMAQ porcess system is realized
Picture.
7th, the image biological feature locations and yardstick skew for realizing collection in IMAQ porcess system are in predetermined model
In enclosing.
8th, the image mirror-reflection position and yardstick interference for realizing collection in IMAQ porcess system are in predetermined model
Enclose outer.
9th, realize that the biometric image gathered in IMAQ porcess system is maintained at axle direct-view state.
10th, the biological characteristic individuality raying total amount security restriction in unit period is realized, high-quality is obtained glitch-free
Biological characteristic image.
Embodiment 1, the present invention provide a kind of safe biologic feature recognition and IMAQ system for intelligent movable equipment
System and the correlation technique based on the system.
Safe biologic feature recognition and image capturing system for intelligent movable equipment of the invention is included for moving
The safe biologic Feature Recognition System of smart machine and the biometric image acquisition system for intelligent movable equipment.
As shown in figure 1, the safe biologic Feature Recognition System to be used for intelligent movable equipment in the present embodiment, including application
Operating system and SOS;Wherein application operating system includes application proxy service interface unit;SOS
Including security proxy service interface unit, safe computing unit, secure storage unit, secure memory unit;It is above-described to answer
It is used to perform the standardized API invocations of procedure in application operating system end with agency service interface unit;Above-described safety should
It is used to perform the standardized API invocations of procedure in SOS end with agency service interface unit;Safe computing unit is used for
Perform the calculating of the data and code at SOS end;Secure storage unit is used to provide the data at SOS end
With the memory space of code, and ensure that safe computing unit is accessed by individual secure address bus;Secure memory unit is used for
The data at SOS end and the memory headroom of code are provided, and ensure that safe computing unit is total by individual secure address
Line is accessed.Wherein, safe computing unit includes PKI encryption signature security algorithm unit and living things feature recognition algorithm unit;It is raw
Thing feature recognition algorithms unit includes registration algorithm unit and identification and self-learning algorithm unit.
The embodiment of the present invention 1 is preferably used in the register flow path of the safe biologic feature recognition of intelligent movable equipment, including
Following steps:
A, when outside initiation apply registration request;
B, transmitted by application proxy service interface unit the registration request notify safety applications agency service interface unit;
Specifically, the registration request performs the standardized API mistakes under application operating system end by application proxy service interface unit
Journey call notification safety applications agency service interface unit performs the standardized API invocations of procedure under SOS end;
The standardized API invocations of procedure perform registration algorithm unit under c, SOS end;
D, PKI encryption signature security algorithm unit perform registering result ciphering signature, are connect by safety applications agency service
Mouth unit sends ciphering signature registering result and notifies application proxy service interface unit;
E, application proxy service interface unit returning to external receive ciphering signature registering result.
The embodiment of the present invention 1 is preferably used in the identifying procedure of the safe biologic feature recognition of intelligent movable equipment, including
Following steps:
1. when application authorization request is initiated in outside;
2. the certification request is transmitted by application proxy service interface unit and notifies safety applications agency service interface unit;
Specifically, the certification request performs the standardized API mistakes under application operating system end by application proxy service interface unit
Journey call notification safety applications agency service interface unit performs the standardized API invocations of procedure under SOS end;
3. the standardized API invocations of procedure perform identification and self-learning algorithm unit under SOS end;
4. PKI encryption signature security algorithm unit performs authentication result ciphering signature, is connect by safety applications agency service
Mouth unit sends ciphering signature authentication result and notifies application proxy service interface unit;
5. application proxy service interface unit returning to external receives ciphering signature authentication result.
In Fig. 1, the preferred application operating system of the embodiment of the present invention 1 is android OS, and SOS is portion
The TrustZone OS of administration TEE (credible performing environment).The embodiment of the present invention 1 is used for the safe biologic feature of intelligent movable equipment
Identifying system, can realize the secure registration identifying procedure system of complete living things feature recognition, it is ensured that do not attacked by external security.
As shown in figure 3, registration algorithm unit of the present invention is used to perform following rate-determining steps, including:
S200, initialization definitions cycle count COUNT=0, template counts N=0, template number TN, loop number TC;
Wherein:
COUNT is defined as the cycle count in algorithm register flow path,
N is defined as the template counts in algorithm register flow path;
TN is defined as predetermined template number in algorithm register flow path, preferred TN>=3;
TC is defined as predetermined loop number in algorithm register flow path, preferred TC>=3;
S201, collection meet the real-time biological characteristic image of adaptive quality standard;
S202, extraction image biological characteristic information produce feature templates, template counts to add up N=N+1 from increasing;
S203, judge templet count N >=TN, are to perform S204, and no return performs S201;
S204, cycle count add up COUNT=COUNT+1 from increasing;
S205, judge whether combination comparison result set C (TN, 2) formed between the TN template of quantity all meets pre-
Fixed registration threshold denominators, are to perform S208, no execution S206;
Wherein:Described combination comparison result set C (TN, 2)=TN!/(2!*(TN-2)!), embodiment 1 is illustrated
TN=4 is that the combination comparison result collection formed between template { 1,2,3,4 }, template is combined into C (4,2)=4!/(2!*(4-2)!)=6,
Combination comparison result set is { R (1,2), R (1,3), R (Isosorbide-5-Nitrae), R (2,3), R (2,4), R (3,4) } specific as follows.
The embodiment of the present invention 1 all meets the mesh of predetermined registration threshold denominators using comparison result set is combined between template
Be be ensure template between image biological feature stability and uniformity, improve template recognize when success rate;
S206, judge cycle count COUNT>TC, is to perform S207, and no reset template counts N=0 simultaneously returns to execution
S201;
S207, registration failure
S208, succeed in registration, the template of global optimum is enrollment and PKI encryption in selection form assembly comparison result
It is stored in secure storage unit.
The embodiment of the present invention 1 is preferred, and the template for using global optimum in selection form assembly comparison result is registration mould
Plate, it is template { 1,2,3,4 } to illustrate TN=4, and it is normalization numerical value [0,1] to set comparison result, and 0 representative differs completely
Cause, 1 represents completely the same, and the template of global optimum is the maximum corresponding mould of comparison result overall situation accumulated value between other templates
Plate, max { S1, S2, S3, S4 };
S1=R (1,2)+R (1,3)+R (Isosorbide-5-Nitrae);
S2=R (1,2)+R (2,3)+R (2,4);
S3=R (1,3)+R (2,3)+R (3,4);
S4=R (Isosorbide-5-Nitrae)+R (2,4)+R (3,4);
Wherein:
S1 is comparison result overall situation accumulated value between template 1 and other templates, and S1 is maximum, and template 1 is the template of global optimum;
S2 is comparison result overall situation accumulated value between template 2 and other templates, and S2 is maximum, and template 2 is the template of global optimum;
S3 is comparison result overall situation accumulated value between template 3 and other templates, and S3 is maximum, and template 3 is the template of global optimum;
S4 is comparison result overall situation accumulated value between template 4 and other templates, and S4 is maximum, and template 4 is the template of global optimum.
It is enrollment that the embodiment of the present invention 1 uses the template of global optimum in form assembly comparison result, it is therefore an objective to be
Selection combination comparison result set in image biological feature stability and the optimal template of uniformity, improve registration and recognize when into
Power, while improving registration and recognition speed.
As shown in figure 4, identification of the present invention performs following rate-determining steps with self-learning algorithm unit, including:
S301, collection meet the real-time biological characteristic image of adaptive quality standard;
S302, extraction image biological characteristic information produce current signature template;
S303, judge whether current template and enrollment or self study template matching result meet predetermined identification thresholding
Standard, is to perform S304, no execution S306;
Whether S304, judge templet comparison result meet predetermined self study threshold denominators, are to perform S305, no execution
S307;
S305, judge whether image biological characteristic mass meets predetermined self study quality standard, be to perform S308, it is no to hold
Row S307;
S306, identification and template self study failure;
S307, recognize successfully, the failure of self study template;
S308, recognize successfully and PKI encryption preserve current template be self study template in secure storage unit;
The embodiment of the present invention 1 recognizes threshold denominators and self study threshold denominators using predetermined registration threshold denominators, it is therefore an objective to
Success rate during further to improve registration and recognize, while improving registration and recognition speed;The embodiment of the present invention 1 uses template
Self-learning method purpose is more entered when being for when different acquisition environment and biological characteristic itself (physiological) change
Success rate when one step improves registration and recognizes, while improving registration and recognition speed;The embodiment of the present invention 1 is using setting template ratio
To the self study threshold denominators and self study quality standard of result, it is therefore an objective to further to improve self study template quality, carry
Success rate when height registration and identification, while improving registration and recognition speed;In across large-scale crowd application, biological characteristic is individual
(the biological characteristic quality that physiological is formed), from high-grade quality to inferior grade quality, becomes in itself to there is image biological characteristic mass
Change span scope greatly, if when going to judge with single high-grade quality standard part biological characteristic individual biological characteristic quality
Cannot meet the requirements, influence is registered and recognition success rate, while influenceing registration and recognition speed;If with single inferior grade quality
The biological characteristic quality reduction of part biological characteristic individual, influence registration and recognition success rate when standard goes to judge, while influence
Registration and recognition speed.
As shown in figure 5, the embodiment of the present invention 1 uses image biological characteristic mass standard adaptive approach, including following step
Suddenly:S400, initialization definitions:High-grade quality standard Qh, inferior grade quality standard Ql, adaptive quality standard Q=Qh, quality
Standard fine setting step-length △ Q;
S401, collection real-time biological characteristic image;
S402, judge adaptive quality standard Q >=Ql, be to perform S403, no Q=Ql simultaneously returns to execution S401;
S403, judge whether present image biological characteristic quality meets adaptive quality standard Q, be to perform S404, no Q=
Q- △ Q simultaneously return to execution S401;
S404, return meet the real-time biological characteristic image of adaptive quality standard.
The embodiment of the present invention 1 uses image biological characteristic mass standard adaptive approach, the life of all biological characteristic individual
Thing characteristic mass all meets high-grade quality standard in itself will to the adaptive quality standard of inferior grade quality standard dynamic change
Ask, further improve registration and recognition success rate, while improving registration and recognition speed.
As shown in Fig. 2 the embodiment of the present invention 1 provides a kind of biometric image for intelligent movable equipment gathers system
System, including biometric image collecting device, the safe drive control unit of biometric image collecting device, display screen drive safely
Dynamic control unit.Biometric image collecting device described in the embodiment of the present invention 1 is used to gather real-time biological characteristic image, carries
For the high quality graphic needed for safe biologic Feature Recognition System.Biometric image collection described in the embodiment of the present invention 1 sets
It is ready for use on collection biometric image, including range sensor, near-infrared LED light source, current driver, preposition optical filter
And/or rearmounted optical filter, optical imaging lens, auto-focusing driver, image imaging sensor, display screen;Wherein:Away from
It is used to measure object distance information from sensor;Near-infrared LED light source is used for radiant image near infrared light;Current driver is used to drive
Near-infrared LED radiation of light source intensity adjustment;Preposition optical filter and/or rearmounted optical filter are used to filter visible light-transmissive
For the near infrared light being imaged;Optical imaging lens are used for physics and focus near infrared light;Auto-focusing driver is used to drive light
Learn imaging len auto-focusing;Image imaging sensor is used for opto-electronic conversion and exports image;Display screen is used to show displacement
Image and the feedback prompts information for being shown as picture image biological characteristic offset and mirror-reflection interference.Above-described biology
The safe drive control unit of characteristic image collecting device is used to perform biometric image collecting device Real-time Feedback drive control
Output image;The above-described safe drive control unit of display screen is used to perform real time imagery picture displacement and show, into
The detection disturbed as image biological characteristic offset and mirror-reflection and feedback prompts.
As shown in fig. 6, the safe drive control unit of biometric image collecting device described in the embodiment of the present invention 1 includes
Object distance information measurement unit, radiation intensity adjustment unit, auto-focusing unit, imaging pattern control unit;Display screen drives safely
Dynamic control unit includes image displacement display unit, image biological characteristic offset detection unit, the interference of image mirror-reflection
Detection unit, feedback prompts unit.
As shown in fig. 7, the safe drive control unit of biometric image collecting device of the present invention is used to perform life
Thing characteristic image collecting device Real-time Feedback drive control output image includes following feedback control step:
S101, object distance information measurement unit, the object distance information D of real time dynamic measurement range sensor, (unit cm, li
Rice);
S102, whether object distance information D being judged in the working range, being to perform S103, no feedback prompts adjustment distance is simultaneously
Return and perform S101;
, according to object distance information D, real-time dynamic feedback control current driver drives near red for S103, radiation intensity adjustment unit
Outer LED light source radiation intensity I changes;
The embodiment of the present invention 1 is preferred, and near-infrared LED radiation of light source intensity (unit mW/sr, milliwatt is per surface of sphere) has
I=E*D<2>, E is biological characteristic individual raying illumination (unit mW/cm<2>, milliwatt is every square centimeter), radiation intensity regulation
The real-time dynamic regulation near-infrared LED radiation of light source intensity of unit is realized keeping E constant ranges, is not influenceed by object distance image-forming range,
Image luminance constant scope, security restriction is carried out when closely to biological characteristic individual raying total amount.Such as thing
When exceeding closely 10cm away from information D near-infrared LED radiation of light source is completely closed during the limit;
The method that the present invention is used can obtain the biological characteristic image of brightness constancy, while avoiding when closely
Biological characteristic is individual to receive excess radiation;
The present invention is prevented effectively from using brightness of image analysis and assessment method because picture material change causes control process unreal
When, it is unstable;
S104, auto-focusing unit are according to object distance information D, real-time dynamic feedback control auto-focusing driver drives optics
Imaging len auto-focusing;
The embodiment of the present invention 1 is preferred, and auto-focusing driver includes VCM motor, closed-loop closed loop motors, in
Put motor, liquid crystal driver, MEMS actuator etc.;
The embodiment of the present invention 1 is preferred, auto-focusing drive parameter P=L (D), and L is object distance information D and optical imagery
The corresponding physics focal position relation function of lens forming, auto-focusing drive parameter P corresponds to rear focus position;
As auto-focusing driver uses VCM motor, closed-loop closed loop motors, in put motor etc., auto-focusing drives
Dynamic device parameter P is image distance, and P=L (D)=EFL*D/ (D-EFL), wherein EFL are the equivalent focal length of optical imaging lens, the image distance
P corresponds to rear focus position;
As auto-focusing driver uses liquid crystal driver, MEMS actuator, auto-focusing drive parameter P to be bent for optics
Luminosity, P=L (D)=1/D, optical diopter P corresponds to rear focus position;
Further, auto-focusing driver is PDAF, i.e. phase-detection auto-focusing, is detected by closed loop phase and fed back
Control the auto-focusing of optical imaging lens;
It is currently preferred, it is contemplated that stability and accuracy of the actual optical system under different image-forming conditions, to working as
The corresponding focal position regional area of preceding object distance information D is further finely tuned, i.e., with the corresponding focusing of current object distance information D
Centered on position, define regional area and imitated for scope carries out focusing of the scanning of fine step to obtain optimal rear focus position
Really;
The present invention use Atomatic focusing method can in 100ms quick obtaining focus stabilization biological characteristic into
As image;
The frame radiation intensity I of the real-time dynamic control synchronous sequence Pulse Imageing pattern of S105, imaging pattern control unit, frame
Time T and frame rate F output images;
The embodiment of the present invention 1 is preferred, the real-time dynamic control synchronous sequence Pulse Imageing pattern of imaging pattern control unit
Frame time T (unit, millisecond per frame) and frame rate F (unit, frame per unit cycle) export image method be specially it is logical
During the frame of the Pulse Imageing pattern for crossing synchronous sequence control image imaging sensor exposure (integration) and near-infrared LED radiation of light source
Between T and frame rate F output realize;That is, the frame time T and frame rate F and near-infrared LED of image imaging sensor exposure (integration)
The frame time T and frame rate F of radiation of light source keep the sequential of Pulse Imageing pattern identical;
Further, the embodiment of the present invention 1 is preferred, and described Pulse Imageing pattern realizes frame using pulse amplitude modulation
Radiation intensity I controlled outputs.
Further, the embodiment of the present invention 1 is preferred, and described Pulse Imageing pattern uses pulse width duty ratio modulation
Realize frame time T controlled outputs.
Further, the embodiment of the present invention 1 is preferred, and described Pulse Imageing pattern is adjusted using unit pulse period frequency
System realizes frame rate F controlled outputs.
The embodiment of the present invention 1, the safe drive control unit of biometric image collecting device passes through feedback control step
S101-S105 final outputs image is simultaneously further used for the safe drive control unit of display screen;
As shown in figure 8, the safe drive control unit of display screen of the present invention is used to perform real time imagery picture displacement
It has been shown that, the detection and feedback prompts of the skew of image biological characteristic and mirror-reflection interference, including following rate-determining steps:
S106, the image of image displacement display unit Real time dynamic display displacement treatment;
Due to physical installation structure limitation reason, the physical centre of the image that display screen shows and biometric image
The physical optics center of collecting device is cannot to realize same optic axis (coaxial state);
The embodiment of the present invention 1 citing such as mobile phone application, the biometric image collecting device of mobile phone front camera position
Actual displayed screen can not possibly be arranged on behind, biometric image collecting device optical centre physical location and actual mobile phone show
The physical centre position for shielding the image of display is that cannot realize same optic axis certainly;Same optics cannot be realized this
In the case of axle, when user observes the image of mobile phone display screen, by the life installed in mobile phone front camera position
The biometric image of thing characteristic image collecting device collection can be resulted in be looked side ways and non-coaxial direct-view state off axis;It is off-axis oblique
There is biological characteristic tissue deformation in three-dimensional off-axis state depending on directly resulting in biological characteristic tissue under state, make biological spy
The distortion that image produces non-geometric is levied, greatly registration and recognition success rate and registration and recognition speed is reduced;The embodiment of the present invention 1
Physical centre and the physical light of biometric image collecting device of the image that display screen shows are adjusted by dynamic in real time
Displacement between center, realizes that the biometric image of collection is maintained at axle direct-view state;The present invention is further especially explained
Point out:In axle direct-view state confronting coaxial direct-view state, difference is only in that generation optic axis center (X-Y coordinate axle) displacement,
Reflect that actual imaging image effect is exactly that picture centre (X-Y coordinate axle) is subjected to displacement, do not produce biological characteristic group
Knit the deformation of off-axis stravismus;
Further, the embodiment of the present invention 1, the imaging of image displacement display unit Real time dynamic display displacement treatment
Image, the displacement treatment specific method of image:
The image center that display screen shows adjusts (X_SHIFT, Y_SHIFT) by X-Y coordinate axial displacement amount;
Described adjustment X-Y axial displacements amount (X_SHIFT, Y_SHIFT's) is specifically calculated as follows:
X_SHIFT=β * (Xscreen-Ximager)/PS
Y_SHIFT=β * (Yscreen-Yimager)/PS
β=EFL/ (D-EFL)
Wherein:
(X_SHIFT, Y_SHIFT) is respectively the image center X that display screen shows, Y-coordinate axle displacement, unit picture
Element, pixels;
(Xscreen, Yscreen) is respectively the image center X that display screen shows, Y-coordinate axle physical location, unit
Centimetre, cm;
(Ximager, Yimager) is respectively biometric image collecting device optical centre X, Y-coordinate axle physical location,
Unit centimetre, cm;
β is the optical magnification of biometric image collecting device, without unit;
EFL is the equivalent focal length of optical imaging lens, unit millimeter, mm;
D is object distance, unit centimetre, cm;;
PS is the unit pixel physical size of image imaging sensor, the every pixel of unit micron, um/pixel.
The embodiment of the present invention 1, is citing with actual mobile phone application, and design parameter is as follows:
EFL=3mm, D=20cm, PS=1.12um/pixel,
(Xscreen-Ximager)=1cm;
(Yscreen-Yimager)=3cm;
β=0.0152;(X_SHIFT, Y_SHIFT)=(136pixels, 408pixels);
The image center that i.e. display screen shows is adjusted to 136 pixels and 408 pixels by X-Y axial displacement amounts.
The image that the embodiment of the present invention 1 is processed by Real time dynamic display displacement, realizes the biological characteristic figure of collection
As being maintained at axle direct-view state.
The position of biological characteristic and yardstick are inclined in S107, image biological characteristic offset detection unit Real-time and Dynamic Detection image
Move;The embodiment of the present invention 1 is preferred, and image biological characteristic offset detection unit calculates biological characteristic and exists according to real time imagery image
Active position and effective scale information in image, i.e. XY direction of principal axis position offset and the skew of Z axis yardstick, represent left and right upper and lower
Offset to position skew and far and near distance yardstick;
The embodiment of the present invention 1 is preferred, and biological characteristic bias detecting method includes AdaBoost detection algorithms, active profile
Detection algorithm, LOG edge detection operators, Canny detective operators, Moravec corner detection operators and Harris Corner Detections are calculated
Son etc.;
Further, the embodiment of the present invention 1 is preferred, and biological characteristic bias detecting method is also included using biological characteristic point
Cloth statistical function;
S108, judge whether biological characteristic position and yardstick offset in image, be that feedback prompts unit performs biological characteristic
Position and yardstick offset feedback are pointed out and return to execution S106, no execution S109;
The embodiment of the present invention 1 is preferred,
Whether position offsets biological characteristic in judging image, more than predetermined image boundary regional extent;
Whether yardstick offsets biological characteristic in judging image, more than predetermined graphical rule magnitude range;
The embodiment of the present invention 1 is preferred, and feedback prompts unit performs display screen and shows the biological characteristic of dynamic adjustment in real time position
Put and yardstick offset information;
The position of mirror-reflection and yardstick are dry in S109, image mirror-reflection Interference Detection unit Real-time and Dynamic Detection image
Disturb;The embodiment of the present invention 1 is preferred, and image mirror-reflection Interference Detection unit calculates mirror-reflection and exists according to real time imagery image
Active position interference and effective scale interference information in image;
The embodiment of the present invention 1 is preferred, and image mirror-reflection interference detection method includes AdaBoost detection algorithms, actively
Contour detecting algorithm, LOG edge detection operators, Canny detective operators, Moravec corner detection operators and Harris angle points are examined
Measuring and calculating etc.;
Further, the embodiment of the present invention 1 is preferred, and image mirror-reflection interference detection method also includes anti-using minute surface
Penetrate feature distribution statistical function;
S110, judge whether image mirror-reflection position and yardstick disturb, be that feedback prompts unit performs mirror-reflection position
Put and pointed out with yardstick interference feedback and return execution S106, no return performs S106;
The embodiment of the present invention 1 is preferred,
Judge in image mirror-reflection whether Position disturbance, more than biological characteristic regional extent in predetermined image;
Whether yardstick is disturbed mirror-reflection in judging image, more than biological characteristic magnitude range in predetermined image;
The embodiment of the present invention 1 is preferred, and feedback prompts unit performs display screen and shows the mirror-reflection of dynamic adjustment in real time position
Put and yardstick interference information;
The safe drive control unit rate-determining steps of display screen that the present invention is used, according to the reality that image imaging sensor is exported
When image, the image of feedback control display screen display real-time displacement treatment, and biological characteristic position in detection image
Adjusted for feedback prompts with yardstick skew and the information such as mirror-reflection position and yardstick interference;
Realize the dynamic adjustment of the skew of image biological feature locations and yardstick skew, it is ensured that biological characteristic position is in predetermined
Image boundary regional extent in, it is ensured that biological characteristic yardstick be in predetermined graphical rule magnitude range in;
Realize the dynamic adjustment of image mirror-reflection Position disturbance and yardstick interference, it is ensured that mirror-reflection position is in predetermined
Image in outside biological characteristic regional extent, it is ensured that mirror-reflection yardstick is in biological characteristic magnitude range in predetermined image
Outward.
The embodiment of the present invention 1 is using a kind of for realizing biometric image acquisition system biological characteristic in unit period
Individual raying total amount carries out security restriction and obtains the combined optimization control of the glitch-free biological characteristic image of high-quality
Method, including:
1st, near-infrared LED radiation of light source is completely closed when object distance information D exceedes the closely limit;
2nd, biometric image acquisition system biological characteristic individuality raying energy J0 within per frame period is defined, (unit,
Joules per cm per frame period, J/cm<2>/ frame),
J0=E*T;
Biometric image acquisition system biological characteristic individuality raying energy JE in unit period is defined, (unit, it is burnt
Ear is every square centimeter, J/cm<2>),
JE=J0*F<Jlimit;
3rd, according to the definition relation of JE in 2, by combined optimization control parameter:Biological characteristic is individual in per the frame period receives spoke
Energy J0, and frame number F in unit period are penetrated, realizes limiting JE to biological characteristic individuality raying energy in unit period<
Jlimit;
Further, the embodiment of the present invention 1 illustrates, according to the definition relation of JE in 2, by predetermined unit period
Interior frame number F, it is determined that biological characteristic individuality raying energy J0 in per the frame period.
4th, according to the definition relation of J0 in 2, combined optimization control parameter E, T inverse relation, i.e. E=J0/T, T=J0/E,
Improve biological characteristic individual per frame period raying illumination E, reduce biological characteristic individual per frame period raying and time for exposure
T, i.e., using radiant illumination E high of per frame period, per the frame period, short radiation and time for exposure T are used to obtain the glitch-free life of high-quality
Thing characteristic imaging image.
Further, the embodiment of the present invention 1 illustrates, excellent using parametric joint optimal control maximum inversely
Point is can to meet radiant illumination high of per frame period, and can be met per frame period short radiation and time for exposure, realizes parameter E, T's
Product meets the biological characteristic individuality raying energy J0 for making a reservation for be limited within per frame period.
It is used to realize the imaging interference such as effective filter background light per frame period radiant illumination high, per frame period short radiation and exposure
The light time be used for realize solve motion defocus wait imaging disturb.
Further, the embodiment of the present invention 1 illustrates, using the spoke high of per frame period of synchronous sequence Pulse Imageing pattern
Penetrating illumination imaging can solve the imaging interference such as effective filter background light by improving the signal to noise ratio of imaging and bias light.
Further, the embodiment of the present invention 1 illustrates, using synchronous sequence Pulse Imageing pattern per frame period in short-term
Between radiation and exposure image can solve the motion of whole X-Y-Z axles and defocus.
Wherein:
Jlimit is the safety margins energy of biological characteristic individuality raying in the unit cycle;
T is biological characteristic individual per raying time in frame period, the i.e. exposure of image imaging sensor (integration) and near-infrared
The frame time of the synchronous sequence Pulse Imageing pattern of LED light source radiation;
F is frame number in the unit cycle, i.e. the synchronization of image imaging sensor exposure (integration) and near-infrared LED radiation of light source
The frame rate of time sequential pulse imaging pattern;
E is biological characteristic individual per frame period raying illumination;
D is object distance information.
The embodiment of the present invention 1 is used to realize that biometric image acquisition system biological characteristic in unit period is individual by spoke
Firing association's amount carries out the combined optimization control method of security restriction and the glitch-free biometric image imaging of high-quality, realizes filtering
Bias light etc. imaging interference and motion defocus wait imaging interference, obtain the glitch-free biological characteristic image of high-quality, and in fact
Biological characteristic individuality raying total amount carries out security restriction in present unit period.
The specific embodiment content and technical characteristic of present invention description, can be in the range of identical or equivalent understanding by reality
Apply, identical purpose also should equally be understood with the replacement of equal value for realizing function.
Finally, in addition it is also necessary to it is noted that listed above is only several specific embodiments of the invention.Obviously, this hair
It is bright to be not limited to above example, there can also be many deformations.One of ordinary skill in the art can be from present disclosure
The all deformations directly derived or associate, are considered as protection scope of the present invention.