Palm vein recognition intelligent system
Technical Field
The invention belongs to the field of biological identification, and particularly relates to a palm vein identification intelligent system for intelligent identification.
Background
The veins are the vessels that lead blood back to the heart, originating in the capillaries, ending in the atrium, and the superficial veins are visible subcutaneously. The palm vein, as the name implies, is the internal palm vein. The palm vein recognition is a kind of vein recognition, belongs to biological recognition, and the palm vein recognition system is characterized in that a palm vein distribution diagram of an individual is obtained through a vein recognition instrument, a characteristic value is extracted from the palm vein distribution diagram according to a special comparison algorithm, images of veins of a finger, a palm and a back of the hand are obtained through an infrared CCD camera, a digital image of the veins is stored in a computer system, and the characteristic value is stored. When vein is compared, a vein image is adopted in real time, characteristic values are extracted, advanced filtering, image binarization and refining means are used for extracting characteristics of the digital image, the characteristics are compared with vein characteristic values stored in a host computer, and a complex matching algorithm is used for matching the vein characteristics, so that identity identification is carried out on individuals, and the identities are confirmed.
When the palm vein is used for identity authentication, the image characteristics of the palm vein are acquired, and the image characteristics exist only when the palm is a living body. In this system, the palm of the non-living body is not provided with the vein image features and thus cannot be identified, and thus cannot be counterfeited. When the palm vein is used for identity authentication, the image characteristics of the veins inside the palm are acquired, but not the image characteristics of the palm surface. Therefore, there is no recognition obstacle due to damage, abrasion, dryness, or too wet of the palm surface, etc. The palm vein is used for identity authentication, and when the palm vein image is obtained, the palm does not need to be in contact with equipment, and the palm can be recognized by slightly placing the palm vein image. This approach avoids the problem of insanitation when the hand is touching the device and the safety problems associated with the possible duplication of finger surface features and avoids the psychological discomfort of being considered as an object of examination and also does not become unrecognizable after contamination with dirt. The palm vein mode is almost suitable for all users because the veins are positioned in the palm, and the influence degree of external factors such as air temperature and the like can be ignored. The user acceptance is good. In addition to not requiring direct contact with the scanner surface, this non-invasive scanning process is simple and natural, alleviating the user's resistance that may be present due to concerns about hygiene or bothersome use. Because of the above living body identification, internal features and non-contact 3 aspects of features, the palm vein feature of the user is ensured to be difficult to forge. Therefore, the palm vein recognition system has high safety level and is particularly suitable for places with high safety requirements. The traditional vein recognition algorithm and how to use an expensive DSP processor to process floating point operation and improve the real-time requirement, and shorten the recognition time, the palm vein still has the following technical problems to be solved at present: and (1) belongs to internal physiological characteristics, cannot be worn, is difficult to forge and has high safety. (2) The blood vessel features are usually more obvious, easy to identify and good in anti-interference performance. (3) The non-contact measurement can be realized, the sanitation is good, and the method is easy to accept by users. (4) Is not easily affected by the scars or oil stains on the surface of the hand. The above techniques have the disadvantages that: the acquisition equipment has special requirements, relatively complex design and high manufacturing cost.
Disclosure of Invention
1. The object of the invention is to provide a method for producing a high-quality glass.
The invention provides a palm vein recognition intelligent system for intelligent recognition, which aims to improve the recognition accuracy of palm veins and solve the problem of low accuracy caused by environmental interference or different palm placing angles in the prior art.
2. The technical scheme adopted by the invention is disclosed.
The palm vein identification intelligent system comprises an imaging circuit, a lighting circuit, an identification module and a wireless transmission module; the imaging circuit comprises a CCD camera, an optical lens and an optical filter; the lighting circuit consists of near-infrared LEDs; the identification module comprises an identification unit, an electronic distance measurement unit and a voice unit, the wireless transmission module transmits video information and control signals to a computer and realizes communication with the computer, and the identification module identifies the palm vein according to the following steps:
firstly, training and projecting palm vein images, and obtaining five candidate palm vein images through matching of a classifier;
secondly, finding the best test image from five candidate palm vein images in the classifier:
a. loading all palm vein image training sets, and corresponding to the candidate palm veins;
b. extracting the deformation of each palm vein from the training set by using the fusion palm vein description;
c. extracting deformation of the palm vein test image, which is expressed as Dq;
d. matching the Dq with the five previous candidate palm vein images to find the number of matching key points
And thirdly, determining a verification criterion according to the number of the matching key points, setting the palm vein descriptor with the maximum number of the matching points as a most probable item, and obtaining a target palm vein image through feature matching.
Preferably, the optimal projection matrix obtained by training in the first step is as follows:
is an optimal projection matrix, wherein
And
is defined as follows:
preferably, in the third step, for the matching process, the palm veins are classified by using the euclidean distance based on the nearest neighbor rule, negative samples, that is, the non-training palm vein image and the non-palm vein image, are used to define a threshold for palm vein verification, if the minimum score is lower than the defined threshold, the input data is considered to be a known positive candidate palm vein, and otherwise, the input data is considered to be a negative palm vein image or an unknown palm vein image.
Preferably, in the first step, the image of the palm vein and the image of the palm vein with the query are preprocessed, feature extraction, training and projection are respectively performed, and the trained palm vein stroke projection and the data matrix are performed.
Preferably, in the imaging circuit, the optical lens is a large-view-field optical lens, the optical filter is an infrared-transmitting optical filter with a cut-off wavelength of 720nm, the optical lens is mounted on the near-infrared camera, and the infrared-transmitting optical filter is embedded in the front panel of the shell and positioned in front of the optical lens.
Preferably, the recognition module measures the palm distance information of the user through the electronic distance measuring unit, the control circuit sends a corresponding instruction to the voice unit, and the voice unit gives a prompt of the moving direction of the user.
Preferably, the electronic distance measuring unit in the identification module is an ultrasonic or infrared distance measuring module, and measures the palm distance information of the user in real time; the voice unit adopts a special voice processing chip and is provided with a high-sensitivity loudspeaker, and voice to be broadcasted is burnt into FLASH of the voice processing chip in advance.
Preferably, the single chip microcomputer in the identification module is used as a main control chip of the system, and the single chip microcomputer reads the distance information of the electronic distance measuring unit in real time and judges whether the palm of the user is in a clear imaging range or not; if the imaging range exceeds the clear imaging range, the singlechip sends a corresponding instruction to the voice unit to control the voice unit to broadcast corresponding voice to prompt the user to move; the illumination light source can be adjusted according to the requirement of the actual acquisition environment; the computer sends a corresponding dimming instruction to the single chip microcomputer, and the single chip microcomputer performs PWM (pulse width modulation) or IO (input/output) port level change after receiving the instruction to control the brightness and the quantity of the near-infrared LEDs so as to adjust the illumination brightness.
3. The invention has the beneficial effects.
By means of the training and feature matching method, the response speed can be kept within 8s, interference of environmental factors on palm vein recognition can be accurately eliminated, and accuracy and efficiency of palm vein recognition are improved.
Drawings
Fig. 1 is a flow chart of palm vein recognition according to the present invention.
Detailed Description
In order to make the technical spirit and advantages of the present invention more clearly understandable to examiners of the patent office and particularly to the public, the applicant shall describe in detail below by way of examples, but the description of the examples is not a limitation of the present invention, and any equivalent changes made according to the present inventive concept, which are merely formal and insubstantial, shall be considered to be within the scope of the present invention.
Examples
The palm vein identification intelligent system comprises an imaging circuit, a lighting circuit, an identification module and a wireless transmission module; the imaging circuit comprises a CCD camera, an optical lens and an optical filter; in the imaging circuit, the optical lens is a large-view-field optical lens, the optical filter adopts an infrared-transmitting optical filter with the cut-off wavelength of 720nm, the optical lens is arranged on the near-infrared camera, and the infrared-transmitting optical filter is embedded on the front panel of the shell and positioned in front of the optical lens. The lighting circuit consists of near-infrared LEDs; the recognition module comprises an electronic distance measuring unit and a voice unit, the wireless transmission module transmits video information and control signals to a computer and realizes communication with the computer, the recognition module measures palm distance information of a user through the electronic distance measuring unit, the control circuit sends a corresponding instruction to the voice unit, and a prompt of the moving direction of the user is given through the voice unit. The electronic distance measuring unit in the identification module is an ultrasonic or infrared distance measuring module and is used for measuring the palm distance information of the user in real time; the voice unit adopts a special voice processing chip and is provided with a high-sensitivity loudspeaker, and voice to be broadcasted is burnt into FLASH of the voice processing chip in advance. The single chip microcomputer in the identification module is used as a main control chip of the system, and the single chip microcomputer reads the distance information of the electronic distance measuring unit in real time and judges whether the palm of the user is in a clear imaging range or not; if the imaging range exceeds the clear imaging range, the singlechip sends a corresponding instruction to the voice unit to control the voice unit to broadcast corresponding voice to prompt the user to move; the illumination light source can be adjusted according to the requirement of the actual acquisition environment; the computer sends a corresponding dimming instruction to the single chip microcomputer, and the single chip microcomputer performs PWM (pulse width modulation) or IO (input/output) port level change after receiving the instruction to control the brightness and the quantity of the near-infrared LEDs so as to adjust the illumination brightness.
The identification module identifies the palm vein according to the following steps:
firstly, preprocessing an image of a palm vein and a palm vein image with query, respectively extracting features, training and projecting, and projecting a trained palm vein stroke and a data matrix. Training and projecting the palm vein image, obtaining five candidate palm vein images through matching of the classifier, wherein an optimal projection matrix obtained through training is as follows:
is an optimal projection matrix, wherein
And
is defined as follows:
secondly, finding the best test image from five candidate palm vein images in the classifier:
a. loading all palm vein image training sets, and corresponding to the candidate palm veins;
b. extracting the deformation of each palm vein from the training set by using the fusion palm vein description;
c. extracting deformation of the palm vein test image, which is expressed as Dq;
d. matching the Dq with the five previous candidate palm vein images to find the number of matching key points
And thirdly, determining a verification criterion according to the number of the matching key points, setting the palm vein descriptor with the maximum number of the matching points as a most probable item, and obtaining a target palm vein image through feature matching. For the matching process, the palm veins are classified by using the Euclidean distance based on the nearest neighbor rule, negative samples, namely a non-training palm vein image and a non-palm vein image are used for defining a threshold value for palm vein verification, if the minimum score is lower than the defined threshold value, input data is considered to be a known positive candidate palm vein, and otherwise, input data is considered to be a negative palm vein image or an unknown palm vein image.