CN100485712C - Embedded head portrait intelligent recognition system and method - Google Patents
Embedded head portrait intelligent recognition system and method Download PDFInfo
- Publication number
- CN100485712C CN100485712C CNB2004100512773A CN200410051277A CN100485712C CN 100485712 C CN100485712 C CN 100485712C CN B2004100512773 A CNB2004100512773 A CN B2004100512773A CN 200410051277 A CN200410051277 A CN 200410051277A CN 100485712 C CN100485712 C CN 100485712C
- Authority
- CN
- China
- Prior art keywords
- head portrait
- image
- unit
- signal
- personnel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Abstract
This invention relates to an inserted photo intelligent identifying system and a method including: a shoot unit for shooting image data, a photo location and pre-judgment process unit for positioning and pre-judging the data to solve the position deviation of the phone in the picture resulted in completely random place of the identified object to make sure the accurate position to the identified object in the picture as soon as possible, a processor unit for picking up the property parameter of the head photo data after its positioning, pre-judgment process unit and pre-judgment to compare the property parameters of the head photo with the pre-recorded and an output control unit outputting signals based on the judgment result of the processor unit.
Description
Technical field
The present invention relates to intelligent identification technology, be specifically related to a kind of embedded head portrait intelligent identifying system and method.
Background technology
Customs, airport verification system lean on manually entirely at present, but there is the accuracy rate instability in it, and human factor is heavier, can't stay shelves, put on record and be not easy to drawbacks such as retrieval.And there are " steal, change, forget, lose " problem all the time in key, password, badge mode that automobile, hotel, gate control system adopt, can't guarantee " uniqueness ".Along with the development of infotech, obtained application to a certain extent based on the head portrait recognition technology of high speed PC platform, compare the efficient that has improved checking with artificial checking.Yet existing verification system still exists many deficiencies, such as: volume is big, cost is high, automaticity is low, power consumption reaches greatly and uses inconvenience etc.
Summary of the invention
Technical matters to be solved by this invention is, a kind of embedded head portrait intelligent identifying system and method are provided, and this system has that volume is little, cost is low, dependable performance, power consumption is little and characteristics such as easy to use.
Adopted following technical scheme for solving the problems of the technologies described above the present invention: provide a kind of embedded head portrait intelligent identifying system, comprise shooting head portrait sensor unit, shooting certificate sensor unit, the pre-judgement in head portrait location processing unit, processor unit and output control unit, the pre-input end of judging processing unit of this shooting head portrait sensor unit and shooting certificate sensor unit joint picture location, this pre-judgement in head portrait location processing unit connection processing device unit, this processor unit connects output control unit, this shooting head portrait sensor unit and shooting certificate sensor unit absorb the picture of personnel and this personnel's certificate respectively, the head portrait location is pre-to be judged the processing unit location and judges personnel's head portrait and the accurate position of personnel's certificate head portrait on corresponding picture in advance, processor unit extracts through the pre-characteristic parameter of judging the head portrait data that processing unit location and anticipation are had no progeny in head portrait location, and the characteristic parameter of personnel's head portrait and the characteristic parameter of personnel's certificate head portrait compared judgement, output control unit is according to the judged result output control signal of processor unit.
In the embedded head portrait intelligent identifying system of the present invention, described head portrait location is pre-judges that processing unit comprises: 1) " background image " map section, and the nothing of shooting is identified the background image data of object when being used for store initialization; 2) " on-the-spot input picture " map section, the view data that is identified object that is used to store shooting; 3) comparison operation registers group is used for the view data of " background image " map section and " on-the-spot input picture " map section is compared computing; 4) bias vector valuator is used for " permission variance " value of from processor unit is delivered to each the interior register of comparison operation registers group that participates in comparison operation; 5) and deviation region phasor coordinate register, be used for calling in order to processor unit with vector area coordinate form storage comparison operation result; This background image map section, on-the-spot input picture map section, bias vector valuator and deviation region phasor coordinate register all are connected with the comparison operation registers group, and processor unit all is connected with this bias vector valuator and deviation region phasor coordinate register.
In the embedded head portrait intelligent identifying system of the present invention, described picture picked-up unit is the cmos sensor unit, and it comprises: 1) cmos sensor module is used for photographic images; 2) signal shaping noise reduction pretreatment module is used for the picture signal that the cmos sensor module is taken is carried out shaping, noise reduction; 3) metadata cache is used for the picture signal behind pretreatment module shaping of signal shaping noise reduction and the noise reduction is put in order according to image or video format; 4) image or video output module are used for metadata cache arrangement back with image or video format output; This cmos sensor module, signal shaping noise reduction pretreatment module, metadata cache and image or video output module connect in turn.Described head portrait picked-up unit also can be the ccd sensor unit, and it comprises: 1) ccd sensor module is used for photographic images; 2) signal shaping pretreatment module, the picture signal that is used for the ccd sensor module is obtained is carried out shaping; 3) RGB A/D modular converter is used for the analog picture signal after the shaping of signal shaping pretreatment module is converted to data image signal; 4) metadata cache is used for the data image signal of RGB A/D modular converter conversion is put in order according to image or video format; 5) image or video output module are used for metadata cache arrangement back with image or video format output; This ccd sensor module, signal shaping pretreatment module, RGB A/D modular converter, metadata cache and image or video output module connect in turn.
And a kind of embedded head portrait intelligent identification Method is provided, it comprises the steps:
1) image of personnel and personnel's certificate;
2) judge the processing unit location and judge personnel's head portrait and the accurate position of personnel's certificate head portrait in correspondence image by the head portrait location is pre-;
3) processor unit extracts through processor unit and extracts through the pre-characteristic parameter of judging the head portrait data that processing unit location and anticipation are had no progeny in head portrait location, and the characteristic parameter of personnel's head portrait and the characteristic parameter of personnel's certificate head portrait are compared judgement;
4) export control signal according to the judged result of processor unit by output control unit.
In the embedded head portrait intelligent identification Method of the present invention, head portrait location in the step 2, when judging in advance, processor unit is worth to " vector deviation valuator " transmission " permission variance " earlier, " vector deviation valuator " delivers to " permission variance " value in each register in the comparison operation registers group that participates in comparison operation, registers group compares computing with the data that receive valuator, and the comparison operation result deposits register in the vector area coordinate form and calls in order to processor unit.Described comparison operation process is: on-the-spot camera part is taken unmanned background image earlier during initialization, and data are stored in it " background image " map section of internal memory; The view data of shooting personnel image input is stored in it " on-the-spot input picture " map section of internal memory, the comparison operation registers group compares computing to the data of " background image " map section and " on-the-spot input picture " map section, and head portrait point area data is scaled phasor coordinate.
In the embedded head portrait intelligent identification Method of the present invention, obtaining of image can be by the cmos sensor unit photographs in the step 1), in the shooting process, the cmos sensor signal is done the signal de-noising pre-service earlier and is placed on data buffer area, put in order according to image or video format, be transformed into image or video format output by first in first out.Obtaining of image also can be by the ccd sensor unit photographs in the step 1), in the shooting process, the ccd sensor signal is done pre-service such as signal shaping, level translation earlier, convert digital signal to through A/D converter afterwards and be placed on data buffer area, put in order according to image or video format, be transformed into image or video format output by " first in first out " principle again.Obtaining also of head portrait can be imported by the computer image system IO interface in step 1) and the step 4).
The present invention compared with prior art, utilize embedded computer system design head portrait Intelligent Recognition and automated verification system, flexible good, can embed automobile, customs, airport, hotel, gate inhibition's application scenario easily, volume is little, cost is low, the convenient use, but not Embedded PC system bulk is big, cost is high, is difficult to be directly used in the application scenario.
Description of drawings
Fig. 1 is a system architecture diagram of the present invention.
Fig. 2 is the processor unit block diagram among the present invention.
Fig. 3 is that head portrait location determination of the present invention is handled block diagram.
Fig. 4 is a cmos sensor of the present invention unit block diagram.
Fig. 5 is a ccd sensor of the present invention unit block diagram.
Fig. 6 is applied to process flow diagram in customs, the airport verification system for the present invention.
Fig. 7 is applied to process flow diagram in hotel, the gate control system for the present invention.
Fig. 8 is applied to process flow diagram in the automotive theft proof system for the present invention.
Embodiment
Below in conjunction with drawings and the specific embodiments the present invention is described in further detail: the present invention has disclosed a kind of embedded head portrait intelligent identifying system that can embed application scenarios such as automobile, customs, airport, hotel, gate inhibition easily, but the total system integrated setting on circuit board, about 6 * 10 * 2 (cm of small product size
3).
Figure 1 shows that the system chart in a preferred embodiment of the invention, as shown in the figure, on-the-spot head portrait is taken sensor unit, sensor unit taken by on-the-spot certificate and computer image system IO interface connector picture location is judged the processing unit input end in advance, the pre-judgement in head portrait location processing unit connection processing device unit, processor unit connects output control unit, and output control unit connects pilot lamp, door lock, automobile control or other unit.Processor unit can adopt " SOC (system on a chip) " (SOC) mode realize, also can " system on the plate " (SOB) mode realize, be provided with the application software that is used for head portrait characteristic parameter extraction, systematic comparison and judges output in it.Groundwork process of the present invention is: with CMOS or ccd image sensor, the picture signal of computing machine input with dedicated head as the Intelligent Recognition algorithm, (interocular distance as everyone is to change parameter to extract the head portrait characteristic parameter, angle between left side pupil-right pupil-nose-ear-face exterior feature-eyebrow-mouth type etc., corresponding relations such as position are unique), automatically detect, differentiate detected personnel's head portrait characteristic parameter and certificate image or archival image (passport, I.D., the pass, home return permits etc.) whether the head portrait characteristic parameter conforms to, and by output control unit output corresponding control signal.
Figure 2 shows that the circuit block diagram of processor unit of the present invention, the core of processor unit is 8～64 bit processors, it is communicated by letter with extraneous by IO, peripheral bus, and processor unit also comprises logic control driver module, electrification reset module, ROM/FLASH module, RAM module and the bus control module that is connected with processor by bus.
Among the present invention, the computer image system IO interface can be directly in the following manner or conversion: 1) communication, network input and output (wired, wave point, for example internet, LAN (Local Area Network)); 2) peripheral hardware input is as scanner, video camera, digital camera, video frequency collection card etc.; And 3) other input-output systems are as floppy disk, hard disk, CD, USB flash disk USB, UART, 1394, serial ports and printing, parallel port etc.The computer image system IO interface is mainly used in and identifies special personnel and with its information filing, output from be identified personnel.
Among the present invention, an important step of head portrait identification is to judge the head portrait position, being identified object (being subjected to reference object) institute's station location in shooting picture is at random fully, cause the upper and lower, left and right deviation in the picture by " height the is fat or thin " deviation that is subjected to reference object and " left side of standing, the right side of standing " deviation, must judge as soon as possible and be identified the accurate position of object in picture, could implement head portrait identification computing.Figure 3 shows that head portrait of the present invention location, judge the processing unit block diagram in advance, head portrait locatees, judges in advance that the effect of processing unit is: solve owing to be identified the upper and lower, left and right deviation of head portrait position in the object institute station location shooting picture that completely random causes, judge as soon as possible and be identified the accurate position of object in picture.In the course of work, processor unit is worth to " vector deviation valuator " transmission " permission variance " earlier, " vector deviation valuator " delivers to " permission variance " value in each register in the comparison operation registers group that participates in comparison operation, registers group compares computing with the data that receive valuator, and the comparison operation result deposits register in the vector area coordinate form and calls in order to processor unit.The process of comparison operation is: on-the-spot camera part is taken unmanned background image earlier during initialization, and view data is stored in it " background image " map section of internal memory; The view data of shooting personnel image input is stored in it " on-the-spot input picture " map section of internal memory, the comparison operation registers group compares computing to the data of " background image " map section and " on-the-spot input picture " map section, and head portrait point area data is scaled phasor coordinate.The setting of processing unit is located, judged in advance to head portrait among the present invention, can directly reduce the operand of processor unit " head portrait location " more than 95%, can improve system running speed 60%.
Existing head portrait recognition technology all is to adopt video camera (or camera) to obtain the head portrait signal usually, utilize video camera (or camera) rgb signal " and string conversion " that will walk abreast earlier to become serial signal, " going here and there and also change " again behind the input computing machine becomes parallel signal to carry out algorithm process, but having cost, this design increases, power consumption increases, volume increases, process is complicated (does " and string conversion " as first consume system resources, consume system resources is done " string and conversion " more afterwards), low and the automaticity of reliability is low to be waited not enoughly, and this point influence weight especially for embedded system.For this reason, the present invention adopts sensor unit to replace traditional video camera (or camera) and absorbs the head portrait signal, and directly the parallel rgb signal with the CCD/CMOS sensor is sent to processor unit.Figure 4 shows that cmos sensor unit block diagram among the present invention, this sensor unit is made up of cmos sensor module, signal shaping noise reduction pretreatment module, metadata cache pixel sorting module and image or video format output module.In the course of work, the cmos sensor signal is done the signal de-noising pre-service earlier and is placed on data buffer area, puts in order according to image or video format, is transformed into image or video format output by FIFO (" first in first out ") principle.
The present invention also can select ccd sensor unit shown in Figure 5 for use, and this sensor unit is made up of ccd sensor module, signal shaping pretreatment module, RGB A/D modular converter, metadata cache pixel sorting module and image or video format output module.In the course of work, the ccd sensor signal is done pre-service such as signal shaping, level translation earlier, carrying out analog-digital converter (A/D) afterwards converts digital signal to and is placed on data buffer area, put in order according to image or video format, be transformed into image or video format output by " first in first out " principle again.
Figure 6 shows that the present invention is applied to the process flow diagram in customs, the airport verification system, the process flow diagram brief description is as follows: certificate is placed the appointed area, and the automatic shooting certificate photograph extracts the parameter of certificate photograph and parameter is filed; During checking, take personnel's head portrait, extract personnel's head portrait parameter, the personnel's head portrait parameter and the parameter of certificate photograph of shooting compared, differentiates both whether conform to.Conform to and then control lamp A mode and show, do not conform to and then control lamp B mode and show, need artificial cognition or the personnel of noting abnormalities to wait other other states then to control lamp C mode and show.
Fig. 7 is applied to process flow diagram in hotel, the gate control system for the present invention, and the process flow diagram brief description is as follows: take head portrait when using (or the guest moves in) and be stored in system's (or network) for the first time; Start the gate inhibition during use, shooting user head portrait compares judgement with head portrait parameter and the parameter that is stored in system's (or network); When parameter meets in the parameter of the head portrait of shooting and the system, open the gate inhibition; When parameter is not inconsistent in the parameter of the head portrait of shooting and the system, image is sent to related system further differentiates, open the gate inhibition when conforming to, start warning system or other when not conforming to.
Figure 8 shows that the present invention is applied to the process flow diagram in the automotive theft proof system, take the headstock picture when car owner or driver use for the first time and be stored in the system, the car owner is identical when car other drivers for public use are driven, system can file driver's head portrait of car owner's mandate, take driver's head portrait during each later on the startup, with driver's head portrait and the car owner of file or stayed shelves the comparison of driver's head portrait, judge whether to conform to, start the normal driving control system when conforming to, start warning device or other when not conforming to.
Be understandable that; for the person of ordinary skill of the art; can make other various corresponding changes and distortion according to technical scheme of the present invention and technical conceive, and all these changes and distortion should be covered by within the protection domain of claim of the present invention all.
Claims (7)
1, a kind of embedded head portrait intelligent identifying system, it is characterized in that: comprise shooting head portrait sensor unit, shooting certificate sensor unit, the pre-judgement in head portrait location processing unit, processor unit and output control unit, the pre-input end of judging processing unit of this shooting head portrait sensor unit and shooting certificate sensor unit joint picture location, this pre-judgement in head portrait location processing unit connection processing device unit, this processor unit connects output control unit, this shooting head portrait sensor unit and shooting certificate sensor unit absorb the picture of personnel and this personnel's certificate respectively, the head portrait location is pre-to be judged the processing unit location and judges personnel's head portrait and the accurate position of personnel's certificate head portrait on corresponding picture in advance, processor unit extracts through the pre-characteristic parameter of judging the head portrait data that processing unit location and anticipation are had no progeny in head portrait location, and the characteristic parameter of personnel's head portrait and the characteristic parameter of personnel's certificate head portrait compared judgement, output control unit is according to the judged result output control signal of processor unit, described head portrait location is pre-judges that processing unit comprises: the nothing of shooting is identified the background image map section of the background image data of object when being used for store initialization, be used to store the on-the-spot input picture map section of the view data that is identified object of shooting, be used for the view data of background image map section and on-the-spot input picture map section is compared the comparison operation device registers group of computing, be used for the permission variance yields of from processor unit being delivered to the bias vector valuator of each register in the comparison operation registers group that participates in comparison operation and being used for preparing against the deviation region phasor coordinate register that processor unit calls with vector area coordinate form storage comparison operation result, this background image map section, on-the-spot input picture map section, bias vector valuator and deviation region phasor coordinate register all are connected with the comparison operation registers group, and processor unit all is connected with this bias vector valuator and deviation region phasor coordinate register.
2, embedded head portrait intelligent identifying system according to claim 1, it is characterized in that: head portrait picked-up unit is the cmos sensor unit, it comprises the cmos sensor module that is used for photographic images, be used for the picture signal that the cmos sensor module is taken is carried out the signal shaping noise reduction pretreatment module of shaping and noise reduction, be used for the metadata cache that the picture signal behind pretreatment module shaping of signal shaping noise reduction and the noise reduction is put in order according to image or video format and be used for the image or the video output module of metadata cache arrangement back with image or video format output, this cmos sensor module, signal shaping noise reduction pretreatment module, metadata cache and image or video output module connect in turn.
3, embedded head portrait intelligent identifying system according to claim 1, it is characterized in that: head portrait picked-up unit is the ccd sensor unit, it comprises the ccd sensor module that is used for photographic images, the picture signal that is used for that the ccd sensor module is obtained is carried out the signal shaping pretreatment module of shaping, be used for the analog picture signal after the shaping of signal shaping pretreatment module is converted to the RGB A/D modular converter of data image signal, be used for the metadata cache that the data image signal of RGB A/D modular converter conversion is put in order according to image or video format and be used for the image or the video output module of metadata cache arrangement back with image or video format output, this ccd sensor module, the signal shaping pretreatment module, RGB A/D modular converter, metadata cache and image or video output module connect in turn.
4, a kind of embedded head portrait intelligent identification Method, it comprises the steps:
1) obtains the image of personnel and personnel's certificate;
2) judge the processing unit location and judge personnel's head portrait and the accurate position of personnel's certificate head portrait in correspondence image by the head portrait location is pre-;
3) processor unit extracts through processor unit and extracts through the pre-characteristic parameter of judging the head portrait data that processing unit location and anticipation are had no progeny in head portrait location, and the characteristic parameter of personnel's head portrait and the characteristic parameter of personnel's certificate head portrait are compared judgement;
4) export control signal according to the judged result of processor unit by output control unit;
Described step 2) head portrait location in, when judging in advance, processor unit is worth to " vector deviation valuator " transmission " permission variance " earlier, " vector deviation valuator " delivers to " permission variance " value in each register in the comparison operation registers group that participates in comparison operation, registers group compares computing with the data that receive valuator, and the comparison operation result deposits register in the vector area coordinate form and calls in order to processor unit.
5, embedded head portrait intelligent identification Method according to claim 4, it is characterized in that: described comparison operation process is: on-the-spot camera part is taken unmanned background image earlier during initialization, and data are stored in it " background image " map section of internal memory; The view data of shooting personnel image input is stored in it " on-the-spot input picture " map section of internal memory, the comparison operation registers group compares computing to the data of " background image " map section and " on-the-spot input picture " map section, and head portrait point area data is scaled phasor coordinate.
6, according to claim 4 or 5 described embedded head portrait intelligent identification Methods, it is characterized in that: obtaining of image can be by the cmos sensor unit photographs in the step 1), in the shooting process, the cmos sensor signal is done the signal de-noising pre-service earlier and is placed on data buffer area, put in order according to image or video format, be transformed into image or video format output by first in first out.
7, according to claim 4 or 5 described embedded head portrait intelligent identification Methods, it is characterized in that: obtaining of image can be by the ccd sensor unit photographs in the step 1), in the shooting process, the ccd sensor signal is done signal shaping, level translation pre-service earlier, convert digital signal to through A/D converter afterwards and be placed on data buffer area, put in order according to image or video format, be transformed into image or video format output by " first in first out " principle again.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100512773A CN100485712C (en) | 2004-08-28 | 2004-08-28 | Embedded head portrait intelligent recognition system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100512773A CN100485712C (en) | 2004-08-28 | 2004-08-28 | Embedded head portrait intelligent recognition system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1655180A CN1655180A (en) | 2005-08-17 |
| CN100485712C true CN100485712C (en) | 2009-05-06 |
Family
ID=34892385
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100512773A Ceased CN100485712C (en) | 2004-08-28 | 2004-08-28 | Embedded head portrait intelligent recognition system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100485712C (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5302581B2 (en) * | 2008-07-02 | 2013-10-02 | 株式会社日立製作所 | Security system and program thereof |
| CN106780768A (en) * | 2016-11-29 | 2017-05-31 | 深圳市凯木金科技有限公司 | A kind of long-range simulation cosmetic system and method for 3D in real time |
-
2004
- 2004-08-28 CN CNB2004100512773A patent/CN100485712C/en not_active Ceased
Also Published As
| Publication number | Publication date |
|---|---|
| CN1655180A (en) | 2005-08-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106157629A (en) | A kind of intelligent car position lock control system and control method | |
| CN105590479B (en) | Parking stall monitoring method and device | |
| CN109409368A (en) | Mine leather belt is vertical to tear detection device and detection method | |
| CN106327642A (en) | Human face and fingerprint access control machine | |
| CN109785471A (en) | The control method of vehicle and its back door, control device | |
| CN108038920A (en) | Parking lot management method, system and the server to parking lot management | |
| CN109035854A (en) | A kind of private parking space management method, system and device | |
| CN107464316A (en) | Intelligent caller machine and its operating system | |
| CN108407759A (en) | Automobile intelligent starting module based on recognition of face and startup method | |
| CN100485712C (en) | Embedded head portrait intelligent recognition system and method | |
| CN108172017A (en) | A kind of intelligent car position ground device and its control method | |
| CN105184236A (en) | Robot-based face identification system | |
| CN109521866A (en) | A kind of low-power consumption face identification method and system | |
| Narang et al. | Robust face recognition method based on SIFT features using Levenberg-Marquardt Backpropagation neural networks | |
| CN201698501U (en) | Schoolyard safety device for sending and picking up students | |
| CN201467311U (en) | Fingerprint identifying digital video camera | |
| CN105809787A (en) | WiFi-based gesture unlocking device and control method thereof | |
| CN201936452U (en) | Automatic license plate recognition system | |
| CN206684751U (en) | People's vehicle body part verification terminal | |
| Patil et al. | Automatic number plate recognition system for vehicle identification | |
| CN109635762A (en) | A kind of city management method, system and device | |
| CN202331500U (en) | Pickup and send-off management device for students in school | |
| CN212724128U (en) | Face recognition processing circuit and access control system | |
| CN210836228U (en) | Bus driver safety attendance system based on face recognition technology | |
| CN208092528U (en) | A kind of building monitoring system based on Internet of Things |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C35 | Partial or whole invalidation of patent or utility model | ||
| IW01 | Full invalidation of patent right |
Decision date of declaring invalidation: 20110921 Decision number of declaring invalidation: 17273 Granted publication date: 20090506 |