CN106096619A - Based on artificial intelligence technology join spoon method and system - Google Patents
Based on artificial intelligence technology join spoon method and system Download PDFInfo
- Publication number
- CN106096619A CN106096619A CN201610452592.XA CN201610452592A CN106096619A CN 106096619 A CN106096619 A CN 106096619A CN 201610452592 A CN201610452592 A CN 201610452592A CN 106096619 A CN106096619 A CN 106096619A
- Authority
- CN
- China
- Prior art keywords
- key
- image
- picture
- artificial intelligence
- computer vision
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y10/00—Processes of additive manufacturing
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y50/00—Data acquisition or data processing for additive manufacturing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
- G06F21/6209—Protecting access to data via a platform, e.g. using keys or access control rules to a single file or object, e.g. in a secure envelope, encrypted and accessed using a key, or with access control rules appended to the object itself
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Abstract
The present invention relates to join a spoon technical field, be specifically related to a kind of based on artificial intelligence technology join spoon method and system.The invention provides a kind of automatically fitting its key method and system based on artificial intelligence technology, method therein includes: obtains key picture and is stored in high in the clouds；Use computer vision algorithms make that key picture is aligned；The Feature Extraction Technology in computer vision algorithms make is utilized to extract the characteristics of image of the key picture after aligning；Key type is obtained according to characteristics of image, machine learning algorithm and algorithm for pattern recognition；The specifying information of key is obtained according to key type, computer vision algorithms make and image processing algorithm；Specifying information according to key is prepared.Present invention entity Protection Product in the safety-security area join spoon, can solve the problem that key forgetting, lose and cannot open after the problem such as abrasion a difficult problem for lockset, provide a brand-new solution for various emergency case, make to join spoon convenient, simply, safely, quickly simultaneously.
Description
Technical field
The present invention relates to join a spoon technical field, be specifically related to a kind of based on artificial intelligence technology join spoon method and system.
Background technology
In actual life, we can use various key every day, such as car key, antitheft door key, takes out
Drawer key, safety cabinet key, ATM key, cashbox door key, rifle cabinet key, ammunition depot door key etc..Key-cut
People's frequently problems, according to rough Statistics, have every year 100000000 people need prepare key, it addition, have 50,000,000 people lose or
Leave behind key and seek locksmith's help.But, there is again serious potential safety hazard in traditional spoon mode i.e. inconvenience of joining.Run into key
Forget, lose and situation that the problem such as abrasion does not but have key to back up, the company of unblanking can only be sought help from, costly, time-consuming take
Power is dangerous.
At locksmithing, traditional Bitting machine device includes current state-of-the-art numerically controlled machine, needs to buy miscellaneous
Key billet, every kind of key correspondence one key billet, various costs are the highest.For some private keys, it is difficult to preparation, though preparation
Out, due to precision problem, open lockset and have some setbacks, thus damage lockset.
For solving the problems referred to above, the present invention provides a kind of method of new preparation key.
Summary of the invention
In view of the above problems, it is an object of the invention to provide and a kind of based on artificial intelligence technology join spoon method and system,
Utilize artificial intelligence technology automatically to obtain key data, key data is transferred to 3D Bitting machine and enters preparation, solve key and lose
Forget, lose and cannot open after the problem such as abrasion a difficult problem for lockset, provide a brand-new solution party for various emergency case
Case, make simultaneously people to join spoon mode convenient, simply, safely, quickly.
The invention provides and a kind of based on artificial intelligence technology join a spoon method, including:
Obtain key picture and be stored in high in the clouds；
Use computer vision algorithms make that key picture is aligned；
The Feature Extraction Technology in computer vision algorithms make is utilized to extract the characteristics of image of the key picture after aligning；
Key type is obtained according to characteristics of image, machine learning algorithm and algorithm for pattern recognition；
The specifying information of key is obtained according to key type, computer vision algorithms make and image processing algorithm；
Specifying information according to key is prepared.
Further, during obtaining key picture and being stored in high in the clouds, computer applied algorithm is utilized to obtain key
Spoon picture.
Further, during obtaining key picture and being stored in high in the clouds, by the described encrypted process of key picture
After be stored in high in the clouds, wherein, during key image encryption, use encryption technology based on random sequence, utilize pseudorandom
Sequence generator produces the binary sequence of pixel transform, and changes the pixel value in image according to this sequence, thus realizes
Encryption.
Further, during using computer vision algorithms make that described key picture is aligned,
The extractive technique utilizing the interesting target region in computer vision extracts key position in described key image,
And described key position is aligned.
Further, the extractive technique in described interesting target region includes: described key picture is carried out over-segmentation, right
The various combination of overdivided region judges, and finds out most possible key area.
Further, the Feature Extraction Technology in utilizing computer vision algorithms make extracts the figure of the key picture after aligning
As, during feature, described characteristics of image includes: the manual characteristics of image set and degree of deep learning algorithm automatically learn and obtain
The characteristics of image taken；Wherein, the described manual characteristics of image set includes histogram of gradients feature and Scale invariant characteristics of image；
The characteristics of image that described degree of deep learning algorithm automatically learns and obtains, utilizes the neural network model of multilamellar automatically learn and extract
The characteristics of image of the key picture after aligning.
Further, key type is being obtained according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition
During, based on described picture characteristics, utilize machine learning algorithm that different key type is trained corresponding grader, and profit
With described grader, pending key picture is identified intelligently, obtain key type.
Further, the specifying information of described key includes the tooth position of key, tooth number, bottom width, opening angle, extreme difference, rail
Mark, location mode.
Further, during preparing according to the specifying information of described key, concrete according to described key
The 3D shape of information architecture key, is transferred to 3D Bitting machine by the information of the 3D shape of described key type and key and joins
System.
Present invention also offers and based on artificial intelligence technology join key system, including:
Key picture obtains and memory element, is used for obtaining key picture and being stored in high in the clouds；
Key picture aligns unit, is used for using computer vision algorithms make to align described key picture；
Image characteristics extraction unit, for utilizing the Feature Extraction Technology in computer vision algorithms make to extract the key after aligning
The characteristics of image of spoon picture；
Key type acquiring unit, for obtaining according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition
Key type；
The specifying information acquiring unit of key, for according to described key type, computer vision algorithms make and image procossing
Algorithm obtains the specifying information of key；
Key-cut unit, for preparing according to the specifying information of described key.
Further, described key picture obtains and memory element includes key picture acquisition module,
Described key picture acquisition module, is used for utilizing computer applied algorithm to obtain key picture.
Further, described key picture obtains and memory element also includes key image ciphering memory module, described key
Spoon image ciphering memory module, for being stored in high in the clouds after the described encrypted process of key picture, wherein, adds at key image
During close, use encryption technology based on random sequence, utilize pseudo random sequence generator to produce the two of pixel transform and enter
Sequence processed, and change the pixel value in image according to this sequence, thus realize encryption.
Further, described key picture align unit use computer vision algorithms make described key picture is carried out right
During just, the extractive technique in the interesting target region in computer vision is utilized to extract key position in described key image
Put, and described key position is aligned.
Further, the extractive technique in described interesting target region includes: described key picture is carried out over-segmentation, right
The various combination of overdivided region judges, and finds out most possible key area.
Further, described image characteristics extraction unit Feature Extraction Technology in utilizing computer vision algorithms make is extracted
During the characteristics of image of the key picture after aligning, described characteristics of image includes: the manual characteristics of image set and the degree of depth
The characteristics of image that learning algorithm automatically learns and obtains；Wherein, the described manual characteristics of image set includes that histogram of gradients is special
Seek peace Scale invariant characteristics of image；The characteristics of image that described degree of deep learning algorithm automatically learns and obtains, utilizes the nerve of multilamellar
Network model automatically learns and extracts the characteristics of image of the key picture after aligning.
Further, described key type acquiring unit is being known according to described characteristics of image, machine learning algorithm and pattern
During other algorithm obtains key type, based on described picture characteristics, utilize machine learning algorithm to different key type
Train corresponding grader, and utilize described grader that pending key picture is identified intelligently, obtain key class
Type.
Further, described in the specifying information acquiring unit of described key, the specifying information of key includes the tooth of key
Position, tooth number, bottom width, opening angle, extreme difference, track, location mode.
Further, described key-cut unit during preparing according to the specifying information of described key, root
Specifying information according to described key builds the 3D shape of key, the information of the 3D shape of described key type and key is transferred to
3D Bitting machine is prepared.
Use technique scheme, including following Advantageous Effects:
1) the mobile applications scanning key of the present invention, key meeting is by " digitized ", and encryption is stored in high in the clouds, when sending out
Raw key forgetting, losing and just the available key picture preserved in advance is prepared time after the problem such as abrasion, solving cannot
Open a difficult problem for lockset；
2) to join spoon method more convenient for the present invention, and user is home-confined can place an order, only need to be by key picture by mobile
Application program passes to closest terminal Bitting machine can prepare key；
3) to join spoon method safer for the present invention, and key image information passes through encryption, and digital copies is stored in high in the clouds,
Do not record the personal information such as user identity, inhabitation address and any other number mating key message with position can be used to
According to, only client knows key corresponding informance, has stopped the probability of key loss cause the user loss；
4) present invention can break away from miscellaneous key billet, reduces locksmith's cost；
5) present invention can replace manually, greatly reduces cost of labor；
6) present invention further increases and joins a spoon precision；
7) present invention make some private keys be configured in order to may；
8) present invention accepts private customized, it is provided that the most interesting user that is designed for selects, including word, letter, symbol,
The 3D models such as photo all can be imprinted on bottom key, the pattern that Customer design is liked well, chooses the material of needs, can join
System.
In order to realize above-mentioned and relevant purpose, one or more aspects of the present invention include will be explained in below and
The feature particularly pointed out in claim.Description below and accompanying drawing are described in detail some illustrative aspects of the present invention.
But, some modes in the various modes of the principle that only can use the present invention of these aspects instruction.Additionally, the present invention
It is intended to include all these aspect and their equivalent.
Accompanying drawing explanation
Fig. 1 is to join a spoon method flow schematic diagram according to the based on artificial intelligence technology of the embodiment of the present invention；
Fig. 2 is to join key system logical structure schematic diagram according to the based on artificial intelligence technology of the embodiment of the present invention；
Fig. 3 is that the key picture according to the embodiment of the present invention obtains and the structural representation of memory element.
Detailed description of the invention
In the following description, for purposes of illustration, in order to provide the comprehensive understanding to one or more embodiments, explain
Many details are stated.It may be evident, however, that these embodiments can also be realized in the case of not having these details.
For aforementioned proposition key forgetting, lose and cannot open after the problem such as abrasion the difficult problem of lockset, the present invention carries
Having gone out a kind of based on artificial intelligence technology to join spoon method and system, user only needs to shoot key figure with intelligent terminal such as mobile phones
Sheet, just can be safely stored in cloud database by key information, and when the user desires, system utilizes artificial intelligence technology foundation
Key snapshot obtains key data, and key data is transferred to 3D Bitting machine, utilizes 3D printing technique copy key.The present invention
Fundamentally changing people and join a spoon mode, the encryption of high in the clouds key data makes only client itself know key corresponding informance,
Simultaneously the joining spoon mode and can rapidly key be prepared of intelligence so that join spoon more safe and efficient.
Below by specific embodiment and combine accompanying drawing the present invention is described in further detail.
Based on artificial intelligence technology joining a spoon method in order to illustrate that the present invention provides, Fig. 1 shows according to the present invention real
Execute the based on artificial intelligence technology of example and join a spoon method flow.
A kind of based on artificial intelligence technology a spoon method is joined as it is shown in figure 1, the invention provides, including: S110: obtain key
Spoon picture is also stored in high in the clouds；
S120: use computer vision algorithms make that key picture is aligned；
S130: utilize the Feature Extraction Technology in computer vision algorithms make to extract the image spy of the key picture after aligning
Levy；
S140: obtain key type according to characteristics of image, machine learning algorithm and algorithm for pattern recognition；
S150: obtain the specifying information of key according to key type, computer vision algorithms make and image processing algorithm；
S160: prepare according to the specifying information of key.
Wherein, in step s 110, computer applied algorithm is utilized to obtain key picture, by acquisition key picture and store
In high in the clouds, wherein, after the described encrypted process of key picture, be stored in high in the clouds, during key image encryption, use based on
The encryption technology of random sequence, utilizes pseudo random sequence generator to produce the binary sequence of pixel transform, and according to this sequence
Row change the pixel value in image, thus realize encryption.
It should be noted that utilize the picture of smart machine (such as: mobile phone app) shooting key.Shooting process makes
Key is positioned at shooting center, and shoots the positive and negative of key respectively.Key picture is encrypted, after encrypting
Key picture be stored in high in the clouds, in case follow-up use.
In the step s 120, the extractive technique utilizing the interesting target region in computer vision extracts described key figure
Key position in Xiang, and described key position is aligned.Specifically, the first step, utilize interesting target district in computer vision
The extractive technique in territory extracts the exact position of key in key picture；Wherein, the extractive technique in interesting target region is mainly divided
It is two steps: (1) carries out over-segmentation to key image；(2) various combination of overdivided region is judged, find out and have most
Possible key area.Second step, finds out key axis according to key area, and utilizes key axis to enter key position
Row accurate contraposition.
In step s 130, the extraction of characteristics of image；Specifically, the first step, utilize computer vision algorithms make to extract and align
After the multiple characteristics of image of key picture, automatically learn also including the manual characteristics of image set and degree of deep learning algorithm
The characteristics of image obtained；The manual characteristics of image set includes histogram of gradients feature and Scale invariant characteristics of image.The degree of depth
Practise the characteristics of image that algorithm automatically learns and obtains, utilize the neural network model of multilamellar automatically learn and extract the key after aligning
The characteristics of image of spoon picture.Second step, merges the multiple characteristics of image of above-mentioned key picture, obtains key picture
Whole feature representation.Wherein, degree of depth convolutional neural networks mainly includes convolution and pondization operation composition, and by training end to end
Procedural learning image feature representation.
Specifically, in an embodiment of the present invention, the method obtaining histogram of gradients feature comprises the steps:
The first step: by coloured image gray processing, is i.e. converted into gray level image；
Second step: calculate gradient magnitude and the direction of each pixel；Formula is as follows:
θ (x, y)=arctan ((I (x+1, y)-I (x-1, y))/(I (x, y+1)-I (x, y-1)))
Wherein, (x, y), (x, y), (x y) is respectively (x, y) the pixel value size at place, the size and Orientation of gradient to θ to M to I；
3rd step: divide an image into fritter, such as the fritter of 6 × 6；
4th step: add up the histogram of gradients of each fritter, the i.e. number of different directions gradient, then form each fritter
Feature Descriptor；
5th step: every several fritters form a region (such as 3 × 3 fritter/regions) is all little in a region
The Feature Descriptor of block is together in series and just obtains the histogram of gradients feature in this region.
Wherein, in an embodiment of the present invention, Scale invariant characteristics of image mainly includes direction and the life determining characteristic point
Becoming two steps of Feature Descriptor, the method obtaining Scale invariant characteristics of image comprises the steps:
The first step: according to the principal direction of the gradient direction distribution property calculation characteristic point of characteristic point neighborhood territory pixel:
θ (x, y)=arctan ((I (x+1, y)-I (x-1, y))/(I (x, y+1)-I (x, y-1)))
Wherein, (x, y), (x, y), (x y) is respectively (x, y) the pixel value size at place, the size and Orientation of gradient to θ to M to I.
When Practical Calculation, sample in the neighborhood window centered by key point, and with statistics with histogram neighborhood territory pixel
Gradient direction.The scope of histogram of gradients is 0～360 degree, the most every 45 degree of posts, altogether 8 posts, or every 10 degree one
Post, altogether 36 posts.Histogrammic peak value then represents the principal direction of neighborhood gradient at this key point, i.e. as this key point
Direction.
Second step: choose the region of 4 × 4 in the characteristic point field with R as radius, and add up each 4 × 4 pieces
Histograms of oriented gradients.For removing the illumination impact on Feature Descriptor, histogram of gradients is normalized.By each
The series connection of rectangular histogram groove data i.e. obtains scale invariant feature and describes son.
In an embodiment of the present invention, when obtaining characteristics of image, degree of depth convolutional neural networks is mainly by input layer, convolution
Layer, sample level, output layer form.Degree of depth convolutional neural networks obtains characteristics of image method and comprises the steps:
The first step: the image of required process is transferred to input layer as the input of convolutional neural networks, and input layer
Data are transferred to Internet below with a matrix type.
Second step: the input of convolutional layer is the output of last network, and it can be original image data, it is also possible to be it
The characteristic pattern of his Internet.In convolutional layer, the characteristic pattern of each passage has a convolution kernel responded, the volume of each passage
Long-pending core size is identical.Doing convolution operation on picture with this convolution kernel and obtain the output of convolutional layer, its formula is as follows:
Wherein, the size of convolution kernel is m × m, zijFor at (i, j) the convolution response value at place, wabFor convolution kernel (a, b)
Value, y(i+a)(j+b)The value gone out at (i+a) (j+b) for convolutional layer input feature vector figure.
3rd step: sample level: sample level is a sampling processing to the output of last layer network, and sample mode here is
The territory, neighbor cell of last layer characteristic pattern is carried out aggregate statistics.The most commonly used has average polymerization and maximizes polymerization, averagely
Polymerization is using the meansigma methods in adjacent area as output, and maximizes polymerization using peak response as output.Such as at characteristic pattern
The region of 2 × 2 in value be respectively 1,2, Isosorbide-5-Nitrae, then the output valve of average polymerization is 2, and maximizes the output of polymerization
Value is 4.In degree of depth convolutional network, the output after the convolution and sample level of n-layer is as the final feature of image.Wherein
Number of plies n of network designs according to the complexity of problem, and in general, problem is the most complicated, and network number of plies n is the biggest.
In step S140, based on described picture characteristics, utilize machine learning algorithm to different key type training phases
The grader answered, and utilize grader that pending key picture is identified intelligently, obtain key type.
Specifically, in an embodiment of the present invention, the step of training grader and Intelligent Recognition algorithm is as follows:
The first step: collect key image data, and mark key type, it is assumed that collect N number of key picture sample: (x1,
y1),(x2,y2),...(xN,yN), wherein, x, y represent the mark of key sample and key type respectively.
Second step: design key recognition classifier, formula is as follows:
G (x)=wx+b
Wherein, w, b are respectively parameter and the bias term of grader, and x is the characteristic vector of key picture, and g (x) is grader
Prediction to key type.Utilize the method minimizing empirical loss that the w in above formula, b are trained, thus obtain optimum
Key type grader.
3rd step: during key identification, utilizes the good grader of training to treat test key picture and is predicted,
The type identifying key of output g (x) intelligence according to grader.
In step S150, the key picture identifying key type is extracted the specifying information of key, wherein key
Specifying information includes the tooth position of key, tooth number, bottom width, opening angle, extreme difference, track, location mode etc..
In step S160, build the 3D shape of key according to the specifying information of key, by key type and the 3D of key
The information of shape is transferred to 3D Bitting machine and prepares.It is to say, specifying information based on key builds the 3D shape of key,
The type of key and the 3D information of key are transferred to terminal Bitting machine, utilize 3D Bitting machine to be printed by key.Wherein,
It should be noted that in one specific embodiment of the present invention, the material of configuration key can be metal or plastics, according to
The key being actually needed the required material of customization of user.
Corresponding with said method, the present invention also provides for a kind of based on artificial intelligence technology joining key system, and Fig. 2 shows
According to embodiments of the present invention based on artificial intelligence technology joins key system logical structure.
As in figure 2 it is shown, a kind of based on artificial intelligence technology the key system 200 of joining that the present invention provides includes that key picture obtains
Take with memory element 210, key picture align unit 220, image characteristics extraction unit 230, key type acquiring unit 240,
The specifying information acquiring unit 250 of key and key-cut unit 260.
Wherein, key picture obtains and memory element 210, is used for obtaining key picture and being stored in high in the clouds；Key picture
Align unit 220, be used for using computer vision algorithms make that described key picture is aligned；Image characteristics extraction unit 230,
For utilizing computer vision algorithms make to extract the characteristics of image of the key picture after aligning；Key type acquiring unit 240, is used for
Key type is obtained according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition；The specifying information of key obtains single
Unit 250, for obtaining the specifying information of key according to described key type, computer vision algorithms make and image processing algorithm；Key
Spoon preparation unit 260, for preparing according to the specifying information of described key.
Wherein, Fig. 3 shows that key picture according to embodiments of the present invention obtains and the structural representation of memory element, as
Shown in Fig. 3, key picture obtains and memory element 210 includes key picture acquisition module 211 and key image ciphering storage mould
Block 212, key picture acquisition module 211, it is used for utilizing computer applied algorithm to obtain key picture.Key picture obtains and deposits
Storage unit also includes, key image ciphering memory module 212, for being stored in cloud after the described encrypted process of key picture
End, wherein, during key image encryption, uses encryption technology based on random sequence, utilizes pseudo random sequence generator
Produce the binary sequence of pixel transform, and change the pixel value in image according to this sequence, thus realize encryption.
Wherein, key picture aligns what described key picture was aligned by unit 220 in employing computer vision algorithms make
During, utilize the extractive technique in the interesting target region in computer vision to extract key position in described key image,
And described key position is aligned.
Wherein, the extractive technique in interesting target region includes: described key picture is carried out over-segmentation, to over-segmentation district
The various combination in territory judges, and finds out most possible key area.
Wherein, the image characteristics extraction unit 230 Feature Extraction Technology in utilizing computer vision algorithms make is extracted and is aligned
After key picture characteristics of image during, described characteristics of image includes: the manual characteristics of image set and degree of depth study
The characteristics of image that algorithm automatically learns and obtains；Wherein, the described manual characteristics of image set include histogram of gradients feature and
Scale invariant characteristics of image；The characteristics of image that described degree of deep learning algorithm automatically learns and obtains, utilizes the neutral net of multilamellar
Model automatically learns and extracts the characteristics of image of the key picture after aligning.
Wherein, key type acquiring unit 240 is according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition
During obtaining key type, based on described picture characteristics, utilize machine learning algorithm to different key type training phases
The grader answered, and utilize described grader that pending key picture is identified intelligently, obtain key type.
Wherein, described in the specifying information acquiring unit 250 of key, the specifying information of key includes the tooth position of key, tooth
Number, bottom width, opening angle, extreme difference, track, location mode etc..
Wherein, key-cut unit 260 is during preparing according to the specifying information of described key, according to described
The specifying information of key builds the 3D shape of key, the information of the 3D shape of described key type and key is transferred to 3D and joins spoon
Machine is prepared.
By above-mentioned embodiment it can be seen that the based on artificial intelligence technology of present invention offer is joined spoon method and be
System, utilizes artificial intelligence technology to obtain key data, and key data is transferred to 3D Bitting machine and prepares the most at last, energy of the present invention
Enough solve key forgetting, lose and cannot open after the problem such as abrasion a difficult problem for lockset, provide one for various emergency case
Brand-new solution, makes key-cut convenient, simply, safely, quickly simultaneously.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for the skill of this area
For art personnel, the present invention can have various modifications and variations.All within the spirit and principles in the present invention, that is made any repaiies
Change, equivalent, improvement etc., should be included within the scope of the present invention.
Claims (10)
1. based on artificial intelligence technology join a spoon method, including:
Obtain key picture and be stored in high in the clouds: during obtaining key picture and being stored in high in the clouds, utilizing computer to answer
Obtain key picture by program, wherein, during key image encryption, use encryption technology based on random sequence, utilize
Pseudo random sequence generator produces the binary sequence of pixel transform, and changes the pixel value in image according to this sequence, from
And realize encryption；
Use computer vision algorithms make that described key picture is aligned: to utilize the interesting target region in computer vision
Extractive technique extract key position in described key image, and described key position is aligned；
The Feature Extraction Technology in computer vision algorithms make is utilized to extract the characteristics of image of the key picture after aligning；
Key type is obtained according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition；
The specifying information of key is obtained according to described key type, computer vision algorithms make and image processing algorithm；
Specifying information according to described key is prepared.
The most according to claim 1 based on artificial intelligence technology join a spoon method, it is characterised in that described interesting target
The extractive technique in region includes: described key picture is carried out over-segmentation, judges the various combination of overdivided region, and
Find out most possible key area.
The most according to claim 1 based on artificial intelligence technology join a spoon method, it is characterised in that regard utilizing computer
During Feature Extraction Technology in feel algorithm extracts the characteristics of image of the key picture after aligning,
Described characteristics of image includes: the image spy that the manual characteristics of image set and degree of deep learning algorithm automatically learn and obtain
Levy；Wherein,
The described manual characteristics of image set includes histogram of gradients feature and Scale invariant characteristics of image；The study of the described degree of depth is calculated
The characteristics of image that method automatically learns and obtains, utilizes the neural network model of multilamellar automatically to learn and extracts the key figure after aligning
The characteristics of image of sheet.
The most according to claim 1 based on artificial intelligence technology join a spoon method, it is characterised in that according to described image
During feature, machine learning algorithm and algorithm for pattern recognition obtain key type,
Based on described picture characteristics, utilize machine learning algorithm that different key type is trained corresponding grader, and utilize
Pending key picture is identified by described grader intelligently, obtains key type.
The most according to claim 1 based on artificial intelligence technology join a spoon method, it is characterised in that described key concrete
Information includes the tooth position of key, tooth number, bottom width, opening angle, extreme difference, track, location mode.
The most according to claim 1 based on artificial intelligence technology join a spoon method, it is characterised in that according to described key
Specifying information prepare during,
Specifying information according to described key builds the 3D shape of key, by the information of the 3D shape of described key type and key
It is transferred to 3D Bitting machine prepare.
7. based on artificial intelligence technology join a key system, including:
Key picture obtains and memory element, is used for obtaining key picture and being stored in high in the clouds, and described key picture obtains and deposits
Storage unit includes key picture acquisition module,
Described key picture acquisition module, be used for utilizing computer applied algorithm obtain key picture described in key picture obtain and
Memory element also includes key image ciphering memory module,
Described key image ciphering memory module, for being stored in high in the clouds after the described encrypted process of key picture, wherein,
During key image encryption, use encryption technology based on random sequence, utilize pseudo random sequence generator to produce pixel
The binary sequence of conversion, and change the pixel value in image according to this sequence, thus realize encryption；
Key picture aligns unit, is used for using computer vision algorithms make to align described key picture, described key figure
Sheet aligns unit and is using during described key picture aligns by computer vision algorithms make,
The extractive technique utilizing the interesting target region in computer vision extracts key position in described key image, and will
Described key position aligns；
Image characteristics extraction unit, for utilizing the Feature Extraction Technology in computer vision algorithms make to extract the key figure after aligning
The characteristics of image of sheet；
Key type acquiring unit, for obtaining key according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition
Type；
The specifying information acquiring unit of key, for according to described key type, computer vision algorithms make and image processing algorithm
Obtain the specifying information of key；
Key-cut unit, for preparing according to the specifying information of described key.
The most according to claim 7 based on artificial intelligence technology join key system, it is characterised in that described interesting target
The extractive technique in region includes: described key picture is carried out over-segmentation, judges the various combination of overdivided region, and
Find out most possible key area.
The most according to claim 7 based on artificial intelligence technology join key system, it is characterised in that described characteristics of image carries
Take the mistake that unit Feature Extraction Technology in utilizing computer vision algorithms make extracts the characteristics of image of the key picture after aligning
Cheng Zhong,
Described characteristics of image includes: the image spy that the manual characteristics of image set and degree of deep learning algorithm automatically learn and obtain
Levy；Wherein,
The described manual characteristics of image set includes histogram of gradients feature and Scale invariant characteristics of image；The study of the described degree of depth is calculated
The characteristics of image that method automatically learns and obtains, utilizes the neural network model of multilamellar automatically to learn and extracts the key figure after aligning
The characteristics of image of sheet.
The most according to claim 7 based on artificial intelligence technology join key system, it is characterised in that described key type
Acquiring unit is obtaining during key type according to described characteristics of image, machine learning algorithm and algorithm for pattern recognition,
Based on described picture characteristics, utilize machine learning algorithm that different key type is trained corresponding grader, and utilize
Pending key picture is identified by described grader intelligently, obtains key type；
The specifying information of key described in the specifying information acquiring unit of described key includes the tooth position of key, tooth number, bottom width, opens
Bicker degree, extreme difference, track, location mode；
Described key-cut unit during preparing according to the specifying information of described key,
Specifying information according to described key builds the 3D shape of key, by the information of the 3D shape of described key type and key
It is transferred to 3D Bitting machine prepare.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610452592.XA CN106096619A (en) | 2016-06-21 | 2016-06-21 | Based on artificial intelligence technology join spoon method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610452592.XA CN106096619A (en) | 2016-06-21 | 2016-06-21 | Based on artificial intelligence technology join spoon method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106096619A true CN106096619A (en) | 2016-11-09 |
Family
ID=57238111
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610452592.XA Pending CN106096619A (en) | 2016-06-21 | 2016-06-21 | Based on artificial intelligence technology join spoon method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106096619A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107931606A (en) * | 2017-11-07 | 2018-04-20 | 昆明理工大学 | A kind of bitting device and method |
| CN108053024A (en) * | 2017-11-17 | 2018-05-18 | 江西理工大学 | Key tooth flower recognition methods |
| CN108171856A (en) * | 2018-01-22 | 2018-06-15 | 珠海优特电力科技股份有限公司 | Lock control system and its control method and the method for replicating unlocker device |
| CN109049716A (en) * | 2018-10-29 | 2018-12-21 | 北京航空航天大学 | Generation method, device, electronic equipment and the storage medium of 3 D-printing illustraton of model |
| CN109794728A (en) * | 2019-03-25 | 2019-05-24 | 湖南森普罗科技有限公司 | A kind of system and method configuring special-section key |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102196320A (en) * | 2011-04-15 | 2011-09-21 | 江苏省现代企业信息化应用支撑软件工程技术研发中心 | Image encrypting and decrypting system |
| CN103218831A (en) * | 2013-04-21 | 2013-07-24 | 北京航空航天大学 | Video moving target classification and identification method based on outline constraint |
| CN103942863A (en) * | 2014-03-15 | 2014-07-23 | 杨密凯 | Allocation method of key chain of intelligent lock |
| US9082003B1 (en) * | 2011-11-22 | 2015-07-14 | The United States Of America As Represented By The Secretary Of The Navy | System and method for adaptive face recognition |
| CN104866855A (en) * | 2015-05-07 | 2015-08-26 | 华为技术有限公司 | Image feature extraction method and apparatus |
| CN104933420A (en) * | 2015-07-03 | 2015-09-23 | 北京旷视科技有限公司 | A scene image identification method and a scene image identification device |
| CN105426900A (en) * | 2014-09-18 | 2016-03-23 | 董醒华 | Shadow analysis method of key photograph |
-
2016
- 2016-06-21 CN CN201610452592.XA patent/CN106096619A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102196320A (en) * | 2011-04-15 | 2011-09-21 | 江苏省现代企业信息化应用支撑软件工程技术研发中心 | Image encrypting and decrypting system |
| US9082003B1 (en) * | 2011-11-22 | 2015-07-14 | The United States Of America As Represented By The Secretary Of The Navy | System and method for adaptive face recognition |
| CN103218831A (en) * | 2013-04-21 | 2013-07-24 | 北京航空航天大学 | Video moving target classification and identification method based on outline constraint |
| CN103942863A (en) * | 2014-03-15 | 2014-07-23 | 杨密凯 | Allocation method of key chain of intelligent lock |
| CN105426900A (en) * | 2014-09-18 | 2016-03-23 | 董醒华 | Shadow analysis method of key photograph |
| CN104866855A (en) * | 2015-05-07 | 2015-08-26 | 华为技术有限公司 | Image feature extraction method and apparatus |
| CN104933420A (en) * | 2015-07-03 | 2015-09-23 | 北京旷视科技有限公司 | A scene image identification method and a scene image identification device |
Non-Patent Citations (3)
| Title |
|---|
| 欧珠著: "《藏语模式识别技术及工程实践》", 31 March 2015, 西南交通大学出版社 * |
| 汪洪桥: "《模式分析的多核方法及其应用》", 31 March 2014, 国防工业出版社 * |
| 葛致磊等编著: "《导弹导引系统原理》", 31 March 2016, 国防工业出版社 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107931606A (en) * | 2017-11-07 | 2018-04-20 | 昆明理工大学 | A kind of bitting device and method |
| CN108053024A (en) * | 2017-11-17 | 2018-05-18 | 江西理工大学 | Key tooth flower recognition methods |
| CN108171856A (en) * | 2018-01-22 | 2018-06-15 | 珠海优特电力科技股份有限公司 | Lock control system and its control method and the method for replicating unlocker device |
| CN109049716A (en) * | 2018-10-29 | 2018-12-21 | 北京航空航天大学 | Generation method, device, electronic equipment and the storage medium of 3 D-printing illustraton of model |
| CN109794728A (en) * | 2019-03-25 | 2019-05-24 | 湖南森普罗科技有限公司 | A kind of system and method configuring special-section key |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106096619A (en) | Based on artificial intelligence technology join spoon method and system | |
| CN105719188B (en) | The anti-method cheated of settlement of insurance claim and server are realized based on plurality of pictures uniformity | |
| CN108446700B (en) | License plate attack generation method based on anti-attack | |
| CN100414558C (en) | Automatic fingerprint distinguishing system and method based on template learning | |
| CN107123123A (en) | Image segmentation quality evaluating method based on convolutional neural networks | |
| CN101894260A (en) | Method for identifying forgery seal based on feature line randomly generated by matching feature points | |
| CN107704877A (en) | A kind of image privacy cognitive method based on deep learning | |
| CN106326886A (en) | Finger-vein image quality evaluation method and system based on convolutional neural network | |
| CN105631272B (en) | A kind of identity identifying method of multiple security | |
| CN109657716A (en) | A kind of vehicle appearance damnification recognition method based on deep learning | |
| CN107330396A (en) | A kind of pedestrian's recognition methods again based on many attributes and many strategy fusion study | |
| CN108229444A (en) | A kind of pedestrian's recognition methods again based on whole and local depth characteristic fusion | |
| CN107563328A (en) | A kind of face identification method and system based under complex environment | |
| CN112801054B (en) | Face recognition model processing method, face recognition method and device | |
| CN108596211A (en) | It is a kind of that pedestrian's recognition methods again is blocked based on focusing study and depth e-learning | |
| CN111832615A (en) | Sample expansion method and system based on foreground and background feature fusion | |
| CN103955926A (en) | Method for remote sensing image change detection based on Semi-NMF | |
| CN104636764B (en) | A kind of image latent writing analysis method and its device | |
| CN104700094A (en) | Face recognition method and system for intelligent robot | |
| CN1760887A (en) | The robust features of iris image extracts and recognition methods | |
| Zhang et al. | Deep multiple instance learning for landslide mapping | |
| CN104102909A (en) | Vehicle characteristic positioning and matching method based on multiple-visual information | |
| Liu et al. | Hierarchical multimodal fusion for ground-based cloud classification in weather station networks | |
| CN102163285A (en) | Cross-domain video semantic concept detection method based on active learning | |
| CN103268498A (en) | Method for area-of-interest blurred image semantic comprehension |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20170930 Address after: Qian Zhang Cun town Shandong 271401 Center Street East sparse city in Tai'an province Ningyang County 45 Applicant after: Zhao Yi Address before: 266000 room 306, building A, science and Technology Park, Qiingdao University, 102-13 Ningxia Road, Laoshan District, Shandong, Qingdao Applicant before: Qingdao translation lock Tong Safe Technology Ltd |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161109 |
|
| RJ01 | Rejection of invention patent application after publication |