CN101787824B - Intelligent anti-theft lock system - Google Patents

Abstract

The invention provides an intelligent anti-theft lock system comprising a processor, a camera module, an iris module, a fingerprint module, an input module, a door switch module, a doorbell module and a communication module. The camera module, the iris module, the fingerprint module, the input module, the door switch module, the doorbell module and the communication module are respectively connected with the processor. The system further comprises a communication terminal connected with the communication module. An intelligent anti-theft lock in the invention transmits unlocking matching information and unlocking person information collected by the camera module to a setting communication terminal in real time through the communication module, so as to learn about state information, visitors and the like, and to determine whether to be normal visiting or thief visiting, thereby improving the safety of doors.

Description

Intelligent anti-theft lock system

Technical field

The present invention relates to a kind of anti-theft lock system, belong to the theftproof lock control field.

Background technology

Present anti-theft lock system mainly is simple monitoring type, such as when there being the guest to visit, just rings doorbell, and owner carries out a management according to customer image or answer that anti-theft lock system transmits, and this anti-theft lock system has brought very large hidden danger.Such as when owner goes out, whether the thief can utilize this system to survey owner and be in, if do not reply, just the thief knows that owner stays out, just may cause property stolen, the unsafe problem of generation property.

Summary of the invention

Goal of the invention:

The present invention has been mainly the unsafe problems that solves existing theftproof lock.In order to address this problem, the invention provides a kind of intelligent anti-theft lock system.

Technical scheme:

The present invention adopts following technical scheme for achieving the above object:

A kind of intelligent anti-theft lock system, comprise indoor set and outdoor location, outdoor location comprises processor and the information acquisition module that is connected with processor, door switch module, doorbell module and communication module, and indoor set comprises the communication terminal that is connected with communication module; Wherein:

Processor be used for to be realized the control to information acquisition module, door switch module, doorbell module and communication module, and to data acquisition and the calculation process of whole system;

Described information acquisition module is used for gathering all kinds of authorization informations of opening the door, comprise for the photographing module that gathers image, be used for gathering iris the iris module, be used for gathering fingerprint fingerprint module, be used for the input module of input door-opening password;

The door switch module is used for receiving the control command of self processor to carry out the switch lock operation;

Communication module is used for the control command that the received communication terminal is sent, and by control command control gate switch module; Also be used for to the communication terminal transmitting state information simultaneously;

The doorbell module is used for realizing auditory tone cues;

Communication terminal is used for by communication module and processor communication, receives the status information of self-locking, and sends control command to processor.

The number of communication terminal is N in the intelligent anti-theft lock system of the present invention, and N is natural number.

The iris module comprises imageing sensor, infrared light LED lamp and white LED lamp in the intelligent anti-theft lock system of the present invention, white LED lamp and infrared light LED lamp alternately be evenly distributed on imageing sensor around.

The method of work of intelligent anti-theft lock system of the present invention, concrete steps comprise:

(1) initial setup procedure:

A, carry out initial key information setting: by the initial iris information of iris module collection; Gather initial fingerprint information by fingerprint module, by input module initial password information is set; Above-mentioned three kinds of key informations are stored to database;

B, communication terminal setting: the transmission of described each information of A step and the terminal of reception are set;

(2) work identification step:

C, when adopting key information to open the door, processor is verified according to the key information of (one) step initial setting up; When the initial key information coupling in key information that processor receives and the database, processor transmission control command control gate switch module is opened door; When the key information that receives when processor and the initial key information in the database do not mate, then enter the D step;

D, the image information that adopts the photographing module collection to visit the guest are sent to processor, and processor is transferred to one or N communication terminal with this image information through communication module, and N is natural number; Communication terminal judges whether to open the door according to image information, and corresponding control command is transferred to processor through communication module, carries out the operation of switch lock by processor according to control command control gate switch module again.

The method that key information is verified in the method for work of intelligent anti-theft lock system of the present invention comprises the identification step of iris module, the identification step of fingerprint module or the identification step of input module.

The identification step of iris module is specially in the method for work of intelligent anti-theft lock system of the present invention:

Gather visit guest's iris and pass to the processor processing, comprise that mainly ripple door location, Iris preprocessing, iris feature extract, relatively judge four steps; Wherein:

Ripple door positioning step arranges the ripple door by the iris image that gathers, and realizes the accurate location to the original image iris region, with the real-time retrieval area-of-interest;

The Iris preprocessing step by to the region of interest area image that retains by various processing to improve iris image quality;

The iris feature extraction step realizes extracting iris image by adopting the multiple dimensioned transcoding, coding transform method based on the Harr small echo;

Relatively the iris of the iris image behind the determining step employing iris feature extraction unit coding and database relatively.

The concrete steps of fingerprint Module recognition are in the method for work of intelligent anti-theft lock system of the present invention:

At first, carry out 3 layers of Harr wavelet transformation for fingerprint image, the wavelet coefficient after the conversion is divided into approximation coefficient LL, horizontal coefficients HL, Vertical factor LH and diagonal coefficient HH;

Then, decomposing each subimage that obtains for the Wavelet image after the conversion proceeds as follows respectively:

I, calculate average gray value for each subgraph:

$\mathrm{Mean}=\frac{1}{M\×N}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}I(i,j);$

Here M * N represents the pixel resolution of subimage, the gray value of the pictorial element of the capable j row of I (i, j) expression i; Wherein M represents total line number of subimage, and N represents total columns of subimage;

II, calculate the gray variance of each subimage:

$\mathrm{Var}=\frac{1}{M\×N}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{[I(i,j)-\mathrm{Mean}]}^2;$

For each subimage, as Var during less than predefined threshold value R, be set and be the background area; Otherwise, as foreground area, keep its gray value; Threshold value R=0.1Mean wherein;

III, carry out binaryzation according to the threshold value of choosing:

$F(i,j)=\left\{\begin{array}{c}1,\mathrm{ifVar}(i,j)\≥R\\ 0,\mathrm{otherwise}\end{array}\right.$

Wherein (i, j) represents the present image pixel coordinate position, i.e. the capable j row of i;

IV, with the image of above-mentioned processing according to the Z-type scanning sequency, output obtains one group of binary sequence, namely obtains fingerprint characteristic; Wherein the Z-type scanning sequency is LL3, HL3, LH3, HH3, HL2, LH2, HH2, HL1, LH1, HH1;

At last, the finger print information that obtains is sent into processor process, compare with the finger print information in the database.

In the method for work of intelligent anti-theft lock system of the present invention:

(1) method of described ripple door location is specially:

At 4 special registers of processor internal structure and 2 counters; 4 special registers are used for preserving the ripple door information of reception, are respectively image row, column initial address csa, rsa and row, column end address cea, rea; Adopt 2 counters to be used for respectively realizing line count to original image; Simultaneously, structure area-of-interest discriminant function f (cx, ry) is as follows:

$f(\mathrm{cx},\mathrm{ry})=\left\{\begin{array}{c}1,\mathrm{ifcsa}\≤\mathrm{cx}\≤\mathrm{cea},\mathrm{rsa}\≤\mathrm{ry}\≤\mathrm{rea}\\ 0,\mathrm{otherwise}\end{array}\right.$

Wherein, cx, ry represent respectively present image row, column address information; If discriminant function is true, represent that then current count value is in the area-of-interest scope.

(2) method of described Iris preprocessing is specially: comprise 1. Iris Location processing, 2. standardization, 3. enhancing processing;

1. Iris Location is processed, and adopts the center of circle and the radius that find the inside and outside circle of iris in a width of cloth eyes image, and iris is cut from out; Comprise coarse positioning and accurate location: coarse positioning is realized the coarse positioning of iris inner circle and cylindrical; Behind coarse positioning, again in Hough change detection iris outer radius and inner and outer boundary center to realize accurate location;

2. standardization comprises: at first, the iris effective coverage that the location is obtained is divided into n concentric circles, and is converted to two dimensionless polar coordinate representation I (r, θ), take pupil center as the origin of coordinates, the I (r, θ) under the polar coordinate system is transformed into the I (x of rectangular coordinate system, y), wherein r represents current radius of a circle, and θ represents the anglec of rotation, (r, θ) be illustrated in that radius is the point at r place on the θ direction, form the endless belt radiation areas take pupil center as the center of circle; Then, it is as a reference point that the concentric circles that each is virtual is divided equally m point, can obtain the rectangular image of n * m; At last, by bilinearity gray-level interpolation algorithm, annular iris image is expanded into the rectangle standardized images of n*m; Total line number of n presentation graphs picture, total columns of n presentation graphs picture, the n*m presentation graphs is as pixel resolution;

3. the dynamic range that is used for adjusting the iris image gray level is processed in described enhancing;

(3) method of described iris feature extraction is specially:

At first, by the Harr wavelet basis, carry out the Harr wavelet transformation for pretreated iris image: image is carried out four layers of wavelet decomposition, these layers minute obtain horizontal coefficients HL1 to HL4, Vertical factor LH1 to LH4, diagonal coefficient HH1 is to HH3 and LL1 to approximation coefficient LL4;

Then, extract the coefficient that those represent the iris pattern: wavelet conversion coefficient obtained above is combined the one-dimensional vector that forms a statement iris pattern by the mode of Z-type scanning, the image that obtains behind the Harr wavelet transformation extracts general profile information and the detailed information that obtains according to frequency band, this one-dimensional vector is characteristic vector, and the binary bit stream of this characteristic vector is the iris feature of extraction;

Wherein the mode of Z-type scanning is LL3, HL3, LH3, HH3, HL2, LH2, HH2, HL1, LH1, HH1;

(4) the described method of relatively judging is specially:

The iris of the iris image behind the iris feature extraction unit coding and database is compared,

$\mathrm{HD}=\underset{i=1}{\overset{L}{\mathrm{\Σ}}}\frac{1}{L}((1-{\mathrm{\δ}}_j^1\left(\mathrm{XOR}\right)B_j^1)(1-{\mathrm{\δ}}_j^2\left(\mathrm{XOR}\right)B_j^2)...(1-{\mathrm{\δ}}_j^n\left(\mathrm{XOR}\right)B_j^2))$

In the following formula, HD represents to extract the iris code metric function distance of iris image feature and database; L representation feature vector length, δ _j ⁿRepresent current subclass characteristic measure function, the iris class coding that namely gathers, B _j ⁿBe the iris-encoding of database storage, total class number of j representative feature metric function, n represents current subclass, XOR represents xor operation; The match is successful greater than 0.6 expression as HD.

Beneficial effect:

Intelligent anti-theft lock of the present invention, pass through communication module, the unlocking person communication that in real time unblank match information and photographing module thereof is gathered arrives sets communication terminal, thereby can understand status information, the visit people etc. of door, thereby can determine normal visit or thief, improve the safety of door.

Above-mentioned iris module comprises imageing sensor, infrared light LED lamp and white LED lamp, white LED lamp and infrared light LED lamp alternately, and be evenly distributed on imageing sensor around.Arrange in this way, can be so that image brightness distribution be even, and can very well show the textured pattern of iris, such as minutias such as spot, filament, stripeds.

Description of drawings

Fig. 1 is structured flowchart of the present invention.

Fig. 2 is this iris mould block structure layout.

Fig. 3 is wavelet transformation scanning sequency figure.

The specific embodiment

Below in conjunction with accompanying drawing technical scheme of the present invention is elaborated:

Disclosed all features in this manual, or the step in disclosed all methods or the process except mutually exclusive feature and/or step, all can make up by any way.

Disclosed arbitrary feature in this manual (comprising any accessory claim, summary and accompanying drawing) is unless special narration all can be replaced by other equivalences or the alternative features with similar purpose.That is, unless special narration, each feature is an example in a series of equivalences or the similar characteristics.

Intelligent anti-theft lock system as shown in Figure 1 comprises: processor, the photographing module that is connected with processing, iris module, fingerprint module, input module, door switch module, doorbell module and communication module, and the communication terminal that is connected with communication module.

Processor be used for to be realized the control to photographing module, iris module, fingerprint module, input module, door switch module, doorbell module and communication module, and to data acquisition and the calculation process of whole system.Generally select image processing function and the more intense processor of control function, realize such as DSP or ARM.

Photographing module is mainly used in gathering image, and under processor control, Real-time Collection is visited guest's image and according to initial setting up, by communication module, the image transmitting of Real-time Collection arrived one or more communication terminals of setting.

The iris module is used under processor control, gathers visit guest's iris and passes to the processor processing.Because iris is distinguished the difference that mainly is grain details, and under normal illumination condition, be difficult to obtain clearly iris image with common CCD camera.In order to obtain extraordinary collection effect, the present invention has designed a kind of special iris module.

As shown in Figure 2, this iris capturing module comprises imageing sensor 1, infrared light LED 2 lamps and white light LEDs 3 lamps, white LED lamp and infrared light LED lamp are alternately, and be evenly distributed on imageing sensor 1 around, arrange in this way, can be so that image brightness distribution be even, and can very well show the textured pattern of iris, such as minutias such as spot, filament, stripeds.

IMAQ sensor 1 can be selected COMS image sensor chip OV7141, this chip OV7141 operating voltage is 2.5V, resolution ratio is 640 * 480, operating frequency is 27MHz, can export 30 frames (VGA pattern) or 60 frames (QVGA pattern) each second, internal integration analog-to-digital conversion (A/D), automatic gain control (AGC), SCCB bus control ports etc. can directly be exported the 8bit view data.

The iris image that collects is processed in processor, mainly comprises: ripple door location, Iris preprocessing, iris feature extract and relatively judge.

Ripple door positioning unit is realized the accurate location to the original image iris region, with the real-time retrieval area-of-interest by the iris image that gathers is arranged the ripple door.Here method to set up is as follows: at 4 special registers of processor internal structure and 2 counters.Special register (image line, row initial address csa, rsa and row, row end address cea, rea) is used for preserving the ripple door information of reception; Counter is used for respectively realizing the line count to original image.Simultaneously, structure area-of-interest discriminant function f (cx, ry), it is defined as follows:

$f(\mathrm{cx},\mathrm{ry})=\left\{\begin{array}{c}1,\mathrm{ifcsa}\≤\mathrm{cx}\≤\mathrm{cea},\mathrm{rsa}\≤\mathrm{ry}\≤\mathrm{rea}\\ 0,\mathrm{otherwise}\end{array}\right.$

Wherein, cx, ry represent respectively present image row, column address information.

If discriminant function is true, then represent current count value in the area-of-interest scope, then view data is kept, otherwise is given up.

Iris preprocessing by to the region of interest area image that retains by various processing to improve iris image quality, comprise that mainly Iris Location processing, standardization and enhancing process.

Iris Location is processed and to be implemented in the center of circle and the radius that finds the inside and outside circle of iris in the width of cloth eyes image, and iris is cut from out; Here adopt coarse positioning and accurately locate the method that combines.

Coarse positioning is exactly the outer edge of coarse positioning iris. and be coarse positioning iris inner circle and cylindrical, inner circle represents the border of iris and pupil, and cylindrical represents the border of iris and sclera.

The inner circle coarse positioning is the center of circle and the radius of rough estimate pupil. in general, the iris image that collects, its pupil gray scale is less than the iris gray scale, the iris gray scale then is less than the sclera gray scale, and the gray value of pupil region image is minimum. therefore, choose the key that suitable threshold value is the inside and outside circle coarse positioning.According to These characteristics, the concrete grammar of inner circle coarse positioning is as follows: at first, do the grey level histogram of image, according to histogram distribution as can be known, this histogram has " three peaks " characteristic, lesser ring of Merkel in the image is distributed in and forms a crest on the darker gray level, forms second crest on the gray level in the middle of the iris district in the image is distributed in, and the sclerotic zone in the image is distributed in and forms the 3rd crest on the brighter gray level.Then, process as the thresholding that threshold value T carries out image with the peak-to-peak paddy lower of its first and second ripple gray value, just can distinguish lesser ring of Merkel and iris district. last, according to threshold value T image is carried out binaryzation:

$g(x,y)=\left\{\begin{array}{cc}0& f(x,y)<T\\ 255& f(x,y)\&GreaterEqual;T\end{array}\right.$

Because image itself is 2D signal, can resolve into x, y two directions.In the situation of given image resolution ratio (such as 320*240), respectively for image each row and column direction (x, the y direction) binaryzation numerical value addition, find the x of addition summed result minimum, the y direction namely obtains coordinate and is (x0, y0), then (x0, y0) is the center of circle of coarse positioning inner circle.Centered by (x0, y0), optional several angles are done chord length again, choose wherein maximum value as radius of circle in the coarse positioning, the angle of preferably choosing more than three is done chord length, and even angle is distributed on the whole circumference, and the interior radius of circle that obtains of coarse positioning is more accurate in this way.

The coarse positioning cylindrical adopts the method identical with the coarse positioning inner circle: at first, according to histogrammic " three peaks " characteristic, process as the thresholding that threshold value T carries out the coarse positioning cylindrical with the peak-to-peak paddy lower of the second and the 3rd ripple gray value, just iris district and sclerotic zone can be distinguished. then, according to threshold value T image is carried out binaryzation.At last, take in the inner circle center of circle as the cylindrical coarse positioning center of circle, optional several angles are done chord length, choose wherein maximum value as the coarse positioning exradius, the angle of preferably choosing more than three is done chord length, and even angle is distributed on the whole circumference, and the interior radius of circle that obtains of coarse positioning is more accurate in this way.

It should be noted that because the gray value of eyelashes and eyebrow is very close with pupil, have in addition less than the gray value of pupil, there is noise in the lesser ring of Merkel image after therefore cutting apart.For the impact of noise decrease, preferably before coarse positioning, with median filtering algorithm it is carried out denoising.

Behind coarse positioning, again in Hough change detection iris outer radius and inner and outer boundary center to realize accurate location.If all boundary points on the iris image outer edge of extracting are (x _j, y _j), j=1,2 ... n.Wherein n represents the sum of boundary point; Hough is defined as:

$H(x_c,y_c,r)=\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}h(x_j,y_j,x_c,y_c,r)$

Here, $\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}h(x_j,y_j,x_c,y_c,r)=\left\{\begin{array}{c}1,\mathrm{ifg}(x_j,y_j,x_c,y_c,r)=0\\ 0,\mathrm{otherwise}\end{array}\right.$

Wherein, g (x _j, y _j, x _c, y _c, r)=(x _j-x _c) ²+ (y _j-y _c) ²-r ²

For each boundary point (x _j, y _j),, if g (x is arranged _j, y _j, x _c, y _c, r)=0, illustrate by parameter (x _c, y _c, r) definite border circumference has passed through boundary point (x _j, y _j), represent that corresponding to peaked H the boundary point of the round process determined by it is maximum like this, accurate setting circle perimeter then, thus realization is to the accurate location at outer radius in the iris and inner and outer boundary center.

The iris standardization realizes the size of iris region is adjusted to fixing size and corresponding position.Because the iris dimensions of Different Individual is different, even same eyes, iris dimensions also is discrepant under different collection environment.Therefore, in order to obtain better recognition result, should reduce the impact that these deformation bring as far as possible.Not only reduce to a certain extent pupil by standardization and changed the iris deformation that causes, but also simplified subsequent calculations.Concrete processing is as follows:

At first, Iris Location is processed the iris image that obtains polar coordinates I (r, θ) expression, r represents the iris image radius, θ represents angle: namely image is take pupil center as the origin of coordinates, (r, θ) point is illustrated in that radius is the point at r place on the θ angle direction, has formed ring belt area take pupil center as the center of circle (annulus belt area that the inner circle that is namely obtained by positioning unit and cylindrical form).

Then, further the polar form of I (r, θ) is converted to rectangular co-ordinate f (p, q).Wherein,

θ=arctan (q/p).

At last, carry out interpolation to changing rear image, be converted to standard-resolution image.Because at polar coordinates I (r, θ) be converted to rectangular co-ordinate f (p, q) after, resolution ratio after the conversion may be also inconsistent with the matching template resolution ratio of preserving, therefore need to be converted to standard-resolution image (such as 320*240) to image interpolation, namely obtain the standardization image.Wherein interpolation can adopt existing ripe bilinear interpolation or other interpolation methods.

Iris strengthens processes the dynamic range that is used for adjusting the iris image gray level.The dynamic range of iris image gray level is narrow, is unfavorable for being directly used in follow-up texture analysis and coupling, therefore, need to strengthen processing to iris image, and the most frequently used image enchancing method is histogram equalization.Here adopt histogram equalization to improve the picture quality after the standardization. it is take the Histogram Modification Methods of cumulative distribution function converter technique as the basis that histogram equalization is processed.Relevant algorithm of histogram equalization can wait " Digital Image Processing " of writing (p56-p59) etc. referring to what east of publishing house of Xian Electronics Science and Technology University in 2008 is strong.

Iris feature extracts the multiple dimensioned transcoding, coding transform method that adopts based on the Harr small echo and realizes.At first, by the Harr wavelet basis, carry out the Harr wavelet transformation for pretreated iris image: image is carried out four layers of wavelet decomposition, and these layers are divided into: HL1 is to HL4 (horizontal coefficients), LH1 to LH4 (Vertical factor), HH1 to HH3 (diagonal coefficient) and LL1 to LL4 (approximation coefficient); Then, extract the coefficient that those represent the iris pattern: wavelet conversion coefficient obtained above is combined the one-dimensional vector that forms a statement iris pattern by the mode of Z scanning (LL3, HL3, LH3, HH3, HL2, LH2, HH2, HL1, LH1, HH1), and the image that has namely obtained behind the Harr wavelet transformation extracts general profile information and the detailed information that obtains according to frequency band.This one-dimensional vector is characteristic vector, and the binary bit stream of this characteristic vector is the iris feature of extraction.

Judge that relatively the iris of realizing the iris image behind the iris feature extraction unit coding and database compares.During identity verification, the iris code that need to again extract iris image feature and database compares, and distinguish true from false.The Hamming distance that present conventional method is utilized the iris code is from the similarities and differences that can compare two iris codes, and the Hamming distance of any two iris intersymbols is from being defined as:

$\mathrm{HD}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{A}_i\left(\mathrm{XOR}\right)B_i$

Wherein, A represents different iris codes with B, and i represents current iris-encoding figure place, and n iris code total bit is if XOR represents xor operation. two condition codes are identical, then HD=0; If two condition codes are fully different, then HD=1. so pattern match degree represent with the number between 0 and 1, the Hamming distance of " forger " is distributed between 0.35 to 0.5 from concentrating, the Hamming distance of " credible person " is distributed between 0 to 0.25 from then concentrating. and final, neighbouring as door interlock switch by selected threshold 0.3.

As further improvement, propose a kind of based on classification piecewise nonlinear method of discrimination here. establishing δ is the unknown characteristics vector, and it is divided into some subclasses as a class.Even

To non-line of each subclass definitions

The property metric function:

${\mathrm{\&delta;}}_j^n=\left\{\begin{array}{c}H,{\mathrm{\&delta;}}_j^n>{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="HSA00000017015900021.png"\; state="\{\{state\}\}">T</figure-callout>}}_1\\ 0,{\mathrm{\&delta;}}_j^n\&le;{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="HSA00000017015900021.png"\; state="\{\{state\}\}">T</figure-callout>}}_1\end{array}\right.$

Total class number of j representative feature metric function in the formula, n represents current subclass; H, T ₁Be nonlinear metric parameter .T ₁Represent the reference threshold that current subclass is chosen, generally get current subclass average value of a function, H represents the boundary value that subclass is set, and gets H=1.2T here ₁. according to above-mentioned metric function, it is as follows to define improved distance:

$\mathrm{HD}=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}\frac{1}{L}((1-{\mathrm{\&delta;}}_j^1\left(\mathrm{XOR}\right)B_j^1)(1-{\mathrm{\&delta;}}_j^2\left(\mathrm{XOR}\right)B_j^2)...(1-{\mathrm{\&delta;}}_j^n\left(\mathrm{XOR}\right)B_j^2))$

Wherein, L representation feature vector length, δ _j ⁿRepresent current subclass characteristic measure function, the iris class coding that namely gathers, B _j ⁿBe the iris-encoding of database storage, XOR represents xor operation.Total class number of j representative feature metric function, n represents current subclass.

Within 0 and 1 scope, high near 1 expression matching degree according to the output valve of HD.Arranging in the situation of fixed threshold, such as 0.6, the automatic opening door lock.

Processor is processed according to comparative result after relatively judging iris image.When the iris coupling, greater than 0.6, then the control gate switch module is opened lock such as matching degree; If iris does not mate, then processor is transferred to the signal that it fails to match the communication terminal of setting together with the image of photographing module collection.

Fingerprint module mainly comprises fingerprint collecting unit, coding unit and ciphering unit thereof.Fingerprint module is sent into processor to the finger print information after collection, coding and the encryption and is processed under processor control.

Wherein the fingerprint collecting unit gathers fingerprint image, selects the pressure-sensitive fingerprint sensor here, and the variation by electric capacity in the touch process gathers finger print information.

The finger-print codes unit finds the bifurcated of lines, the coordinate position that stops or loop and locate by binary conversion treatment and coding to fingerprint image, extracts with good realization fingerprint characteristic.For improving parallel processing speed and processing capability in real time, taked a kind of parallel binary sequential coding scheme here:

At first, carry out 3 layers of Harr wavelet transformation for fingerprint image. then the wavelet coefficient after the conversion is divided into approximation coefficient LL, horizontal coefficients HL, Vertical factor LH and diagonal coefficient HH, is illustrated in fig. 3 shown below.

Then, decomposing each subimage that obtains for the Wavelet image after the conversion (is LL _nLH _nHL _nAnd HH _n) carry out respectively following 4 step operations:

1. calculate average gray value for each subgraph:

$\mathrm{Mean}=\frac{1}{M\&times;N}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}I(i,j)$

Here M * N represents the pixel resolution of subimage, the gray value of the pictorial element of the capable j row of I (i, j) expression i.

2. calculate the gray variance of each subimage:

$\mathrm{Var}=\frac{1}{M\&times;N}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{[I(i,j)-\mathrm{Mean}]}^2$

3. for each subimage, as Var during less than predefined threshold value R, be set and be the background area; Otherwise, as foreground area, keep its gray value, to make subsequent treatment.Here, threshold value R=0.1Mean.

4. carry out binaryzation according to the threshold value of choosing:

$F(i,j)=\left\{\begin{array}{c}1,\mathrm{ifVar}(i,j\&GreaterEqual;R)\\ 0,\mathrm{otherwise}\end{array}\right.$

Owing to image has been carried out 3 layers of wavelet decomposition here, for improving real-time processing speed, can process simultaneously 3 layers of wavelet decomposition, according to the image of above-mentioned processing according to Z-type scanning sequency (LL3, HL3, LH3, HH3, HL2, LH2, HH2, HL1, LH1, HH1), output obtains one group of binary sequence, namely consists of the fingerprint characteristic sequence.

At last, the fingerprint characteristic sequence is inputed to the matching template that prestores and carry out fingerprint matching.At present, fingerprint matching algorithm is ripe, and matching process please refer to Journal of UEST of China in April, 2004,1.1.2 joint characteristic vector coupling (p154-157) in the 2nd phase of 33 volumes " dot pattern fingerprint matching algorithm research and implementation ".

Ciphering unit is mainly by encrypted circuit, prevents the disabled user without the fingerprint characteristic extraction unit and directly a successful simulate signal of self-control comparison delivered on the output port illegally to open door lock; Because in the side circuit, the signal of telecommunication of expression fingerprint recognition success is that simulation produces easily, directly a self-control simulate signal is delivered on the successful port of comparison if the disabled user is arranged without finger-print recognising instrument, system is exactly unsafe so.Need to take encryption measures, encrypted circuit often adopts feedback or pipeline organization at present, proposes a kind of parallel step-by-step operation here for this reason, and concrete mode is as follows:

The first step: to the controlled binary sequence P that obtains behind the 2-d wavelet transcoding, coding transform _{N, z} ^kThe data flow of (n, z, k) (k represents the active user, and n represents to decompose the number of plies, and z represents code word size) deposits buffer unit in.

Second step: the controlled binary sequence P that the first step is obtained _{N, z} ^k(n, z, k) carries out the two-value inverted order and arranges, by the real-time keeping records of storehouse mode.

The 3rd step: for above-mentioned output binary sequence P _{N, z} ^k(n, z, k) carries out the step-by-step xor operation, obtains a sequence U _{N, z} ^k(u, z, k), namely

U _{N, z} ^k(n, z, k) be Buffer output after the even-odd check step-by-step operation.

This AES needs computational resource few, simple efficient, is easy to hardware and realizes.

Identification output mainly compares identification by the Output rusults with the coding encrypting unit with the fingerprint template that prestores, and according to matching result output, comes opening door lock.For coupling, can adopt existing image matching algorithm, also can adopt other existing fingerprint identification methods.

Fingerprint, iris etc. are present the most safe and reliable, noninvasive effective identity validation technologies based on biometrics identification technology.Use it for intelligent anti-theft, can greatly add the strong authentication accuracy, for identity verification accurately and quickly provides reliable method.

Input module is mainly keyboard, button or touch-screen etc., is used for receiving request call, input password etc.Such as being provided with password input mode, as opening the door key, when the input secret was correct, door was just opened automatically by the input password; In case three input errors, processor are transferred to the information of secret input error the communication terminal of appointment by communication module together with the image of photographing module collection.

Doorbell is mainly realized the auditory tone cues when the guest is arranged.

The door switch module, switch lock under processor control.Unblank such as the extraction realization intelligentized design by unlocking person biological information fingerprint, iris.

Communication module is used for the control command that the received communication terminal is sent under processor control, and operates by control command; Also be used for to the communication terminal transmission information simultaneously, these information comprise the image that photographing module gathers, the information that door is opened, password error message, fingerprint recognition error message, iris identification information etc.

The doorbell module mainly realizes the auditory tone cues when the guest is arranged.

Communication terminal mainly by communication module and processor communication, receives the information of self-locking, thus the switching information etc. of release; Send control command etc. to anti-theft lock system.

Intelligent anti-theft of the present invention comprises the pattern of setting and mode of operation:

Pattern is being set, is mainly finishing:

First: initial key information setting, key information comprises fingerprint, iris and/or password etc., mainly by the initial iris information of iris module collection as key; By the initial fingerprint information of fingerprint module collection as key, input password etc. by input module.Each key information is drawn together fingerprint, iris and/or password combination, but for safety, preferably not separately password as key.

The second, communication terminal setting, such as being mobile phone, pda, computer etc., the information of theftproof lock is accessed and received to main configuration information transmission and the terminal that receives to prevent the disabled user.Communication terminal at least one, be preferably two or more, to prevent that after a lost terminal another terminal also can be understood the relevant information of theftproof lock in real time.

In mode of operation, theftproof lock is in running order, when useful key information opens the door, such as being password, fingerprint or iris etc., theftproof lock verifies according to the key information of initial setting up, when used key information and initial key information coupling, opens asking with regard to the control gate switch module; When used key information and initial key information do not mate, just the information of key matching error information and photographing module collection is transferred to communication terminal together, it is normal guest or thief that communication terminal just can be understood the people who unblanks.

In addition, when there is a ring at the door when request, processor also is transferred to setting terminal to the information that people's door opening request is arranged together with the information of camera collection.In this way, in time whether the family no one if there is the guest to visit, also can understand by communication terminal visit people's situation, thereby determine to allow the visit people enter house.If the guest then can send the order of opening the door to processor by communication terminal, thereby door is opened.If the thief then can also report to the police.

Intelligent anti-theft lock of the present invention, pass through communication module, the unlocking person communication that in real time unblank match information and photographing module thereof is gathered arrives sets communication terminal, thereby can understand status information, the visit people etc. of door, thereby can determine normal visit or thief, improve the safety of door.

As seen, this device has intelligent and visual advantage concurrently, and real-time is good, is the various antitheft occasion that is fit to the household office.

Claims (8)

1. intelligent anti-theft lock system, it is characterized in that: comprise indoor set and outdoor location, outdoor location comprises processor and the information acquisition module that is connected with processor, door switch module, doorbell module and communication module, and indoor set comprises the communication terminal that is connected with communication module; Wherein:

Processor be used for to be realized the control to information acquisition module, door switch module, doorbell module and communication module, and to data acquisition and the calculation process of whole system;

Described information acquisition module is used for gathering all kinds of authorization informations of opening the door, comprise for the photographing module that gathers image, be used for gathering iris the iris module, be used for gathering fingerprint fingerprint module, be used for the input module of input door-opening password;

The door switch module is used for receiving the control command of self processor to carry out the switch lock operation;

Communication module is used for the control command that the received communication terminal is sent, and by control command control gate switch module; Also be used for to the communication terminal transmitting state information simultaneously;

The doorbell module is used for realizing auditory tone cues;

Communication terminal is used for by communication module and processor communication, receives the status information of self-locking, and sends control command to processor.

2. intelligent anti-theft lock system according to claim 1 is characterized in that: the number of described communication terminal is N, and N is natural number.

3. intelligent anti-theft lock system according to claim 1, it is characterized in that: described iris module comprises imageing sensor, infrared light LED lamp and white LED lamp, white LED lamp and infrared light LED lamp alternately be evenly distributed on imageing sensor around.

4. method of work based on the arbitrary described intelligent anti-theft lock system of claims 1 to 3, it is characterized in that: concrete steps comprise:

(1) initial setup procedure:

A, carry out initial key information setting: by the initial iris information of iris module collection; Gather initial fingerprint information by fingerprint module, by input module initial password information is set; Above-mentioned three kinds of key informations are stored to database;

B, communication terminal setting: the transmission of the described iris information of A step, finger print information, encrypted message and the terminal of reception are set;

(2) work identification step:

C, when adopting key information to open the door, processor is verified according to the key information of (one) step initial setting up; When the initial key information coupling in key information that processor receives and the database, processor transmission control command control gate switch module is opened door; When the key information that receives when processor and the initial key information in the database do not mate, then enter the D step;

D, the image information that adopts the photographing module collection to visit the guest are sent to processor, and processor is transferred to one or N communication terminal with this image information through communication module, and N is natural number; Communication terminal judges whether to open the door according to image information, and corresponding control command is transferred to processor through communication module, carries out the operation of switch lock by processor according to control command control gate switch module again.

5. the method for work of intelligent anti-theft lock system according to claim 4, it is characterized in that: the method that described key information is verified comprises the identification step of iris module, the identification step of fingerprint module or the identification step of input module.

6. the method for work of intelligent anti-theft lock system according to claim 5, it is characterized in that: the identification step of iris module is specially:

Gather visit guest's iris and pass to the processor processing, comprise that mainly ripple door location, Iris preprocessing, iris feature extract, relatively judge four steps; Wherein:

Ripple door positioning step arranges the ripple door by the iris image that gathers, and realizes the accurate location to the original image iris region, with the real-time retrieval area-of-interest;

The Iris preprocessing step by to the region of interest area image that retains by various processing to improve iris image quality;

The iris feature extraction step realizes extracting iris image by adopting the multiple dimensioned transcoding, coding transform method based on the Harr small echo;

Relatively the iris of the iris image behind the determining step employing iris feature extraction unit coding and database relatively.

7. the method for work of intelligent anti-theft lock system according to claim 5 is characterized in that: the concrete steps of described fingerprint module identification are:

At first, carry out 3 layers of Harr wavelet transformation for fingerprint image, the wavelet coefficient after the conversion is divided into approximation coefficient LL, horizontal coefficients HL, Vertical factor LH and diagonal coefficient HH;

Then, decomposing each subimage that obtains for the Wavelet image after the conversion proceeds as follows respectively:

I, calculate average gray value for each subgraph:

$\mathrm{Mean}=\frac{1}{M\&times;N}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}I(i,j);$

Here M * N represents the pixel resolution of subimage, the gray value of the pictorial element of the capable j row of I (i, j) expression i; Wherein M represents total line number of subimage, and N represents total columns of subimage;

II, calculate the gray variance of each subimage:

$\mathrm{Var}=\frac{1}{M\&times;N}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{[I(i,j)-\mathrm{Mean}]}^2;$

For each subimage, as Var during less than predefined threshold value R, be set and be the background area; Otherwise, as foreground area, keep its gray value; Threshold value R=0.1Mean wherein;

III, carry out binaryzation according to the threshold value of choosing:

$F(i,j)=\left\{\begin{array}{c}1,\mathrm{ifVar}(i,j)\&GreaterEqual;R\\ 0,\mathrm{otherwise}\end{array}\right.$

Wherein (i, j) represents the present image pixel coordinate position, i.e. the capable j row of i;

IV, the image that will cross through above-mentioned I, II, III step process are according to the Z-type scanning sequency, and output obtains one group of binary sequence, namely obtains fingerprint characteristic; Wherein the Z-type scanning sequency is LL3, HL3, LH3, HH3, HL2, LH2, HH2, HL1, LH1, HH1;

At last, the finger print information that obtains is sent into processor process, compare with the finger print information in the database.

8. the method for work of intelligent anti-theft lock system according to claim 6 is characterized in that:

(1) method of described ripple door location is specially:

At 4 special registers of processor internal structure and 2 counters; 4 special registers are used for preserving the ripple door information of reception, are respectively image row, column initial address csa, rsa and row, column end address cea, rea; Adopt 2 counters to be used for respectively realizing line count to original image; Simultaneously, structure area-of-interest discriminant function f (cx, ry) is as follows:

$f(\mathrm{cx},\mathrm{ry})=\left\{\begin{array}{c}1,\mathrm{ifcsa}\&le;\mathrm{cx}\&le;\mathrm{cea},\mathrm{rsa}\&le;\mathrm{ry}\&le;\mathrm{rea}\\ 0,\mathrm{otherwise}\end{array}\right.$

Wherein, cx, ry represent respectively present image row, column address information; If discriminant function is true, represent that then current count value is in the area-of-interest scope;

(2) method of described Iris preprocessing is specially: comprise 1. Iris Location processing, 2. standardization, 3. enhancing processing;

1. Iris Location is processed, and adopts the center of circle and the radius that find the inside and outside circle of iris in a width of cloth eyes image, and iris is cut from out; Comprise coarse positioning and accurate location: coarse positioning is realized the coarse positioning of iris inner circle and cylindrical; Behind coarse positioning, again in Hough change detection iris outer radius and inner and outer boundary center to realize accurate location;

2. standardization comprises: at first, the iris effective coverage that the location is obtained is divided into n concentric circles, and is converted to two dimensionless polar coordinate representation I (r, θ), take pupil center as the origin of coordinates, the I (r, θ) under the polar coordinate system is transformed into the I (x of rectangular coordinate system, y), wherein r represents current radius of a circle, and θ represents the anglec of rotation, (r, θ) be illustrated in that radius is the point at r place on the θ direction, form the endless belt radiation areas take pupil center as the center of circle; Then, it is as a reference point that the concentric circles that each is virtual is divided equally m point, can obtain the rectangular image of n * m; At last, by bilinearity gray-level interpolation algorithm, annular iris image is expanded into the rectangle standardized images of n*m; Total line number of n presentation graphs picture, total columns of m presentation graphs picture, the n*m presentation graphs is as pixel resolution;

3. the dynamic range that is used for adjusting the iris image gray level is processed in described enhancing;

(3) method of described iris feature extraction is specially:

At first, by the Harr wavelet basis, carry out the Harr wavelet transformation for pretreated iris image: image is carried out four layers of wavelet decomposition, these layers minute obtain horizontal coefficients HL1 to HL4, Vertical factor LH1 to LH4, diagonal coefficient HH1 is to HH3 and approximation coefficient LL1 to LL4;

Then, extract the coefficient that those represent the iris pattern: wavelet conversion coefficient obtained above is combined the one-dimensional vector that forms a statement iris pattern by the mode of Z-type scanning, the image that obtains behind the Harr wavelet transformation extracts general profile information and the detailed information that obtains according to frequency band, this one-dimensional vector is characteristic vector, and the binary bit stream of this characteristic vector is the iris feature of extraction;

Wherein the mode of Z-type scanning is LL3, HL3, LH3, HH3, HL2, LH2, HH2, HL1, LH1, HH1;

(4) the described method of relatively judging is specially:

The iris of the iris image behind the iris feature extraction unit coding and database is compared,

$\mathrm{HD}=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}\frac{1}{L}((1-{\mathrm{\&delta;}}_j^1\left(\mathrm{XOR}\right)B_j^1)(1-{\mathrm{\&delta;}}_j^2\left(\mathrm{XOR}\right)B_j^2)...(1-{\mathrm{\&delta;}}_j^n\left(\mathrm{XOR}\right)B_j^n))$

In the following formula, HD represents to extract the iris code metric function distance of iris image feature and database; L representation feature vector length,

Represent current subclass characteristic measure function, the iris class coding that namely gathers, Be the iris-encoding of database storage, total class number of j representative feature metric function, n represents current subclass, XOR represents xor operation; The match is successful greater than 0.6 expression as HD.

Images