CN111914786A

CN111914786A - Finger vein identification method and system

Info

Publication number: CN111914786A
Application number: CN202010802199.5A
Authority: CN
Inventors: 李�杰; 史艺丹; 杨文耀; 刘俊伟; 杨先杰; 聂泽东
Original assignee: Chongqing University of Arts and Sciences
Current assignee: Chongqing University of Arts and Sciences
Priority date: 2020-08-11
Filing date: 2020-08-11
Publication date: 2020-11-10
Anticipated expiration: 2040-08-11
Also published as: CN111914786B

Abstract

The invention discloses a finger vein recognition SOC system and a method based on ARM Cortex-M3, which realize the whole process of the finger vein recognition system, including the processes of finger vein image acquisition, ROI extraction, gray scale and size normalization, feature extraction, connected domain denoising, feature matching and the like, and display the image processing result of each process through an LCD. And hardware acceleration is carried out on the ROI identification through an ROI hardware extraction module, so that the speed of finger vein identification is improved. A large number of tests show that the system can completely realize the function of finger vein recognition, and the recognition accuracy is about 95%. The ROI extraction module is accelerated by adopting the FPGA, so that the total recognition speed can be improved by more than 15%. The system can be used as a scheme of a portable finger vein recognition system for product research and development, can also be used as a process demonstration system for finger vein recognition during experimental teaching, and has higher practical value.

Description

Finger vein identification method and system

Technical Field

The invention relates to the technical field of finger vein recognition, in particular to a finger vein recognition method and a system thereof.

Background

With the development of information technology, the biometric identification technology becomes a key technology in the era of internet artificial intelligence. Nowadays, the research on biological recognition technology is more and more, and the application is more and more extensive. Common biometric techniques are: fingerprints, faces, irises, finger veins, etc. Although the finger vein recognition algorithm has more research results in the last decade, the finger vein recognition products on the market are rare at present. The main reasons are that the vein recognition algorithm is complex, so that hardware is difficult to implement and recognition time is long.

There are three implementations on the market today: the whole recognition algorithm is realized on local hardware, the original image data is transmitted to a PC to be realized, and the original image data is transmitted to a cloud server to be realized. For a portable finger vein recognition system, a scheme of adopting a PC and a cloud server is not convenient. The local hardware implementation scheme is a microprocessor (such as an ARM) and an FPGA. The embedded microprocessor has the characteristics of low cost, low power consumption, simple development and the like, but has no advantage in the running time for an image processing algorithm with a plurality of multiplication operations because the embedded microprocessor executes instructions one by a CPU (central processing unit) during working. For complex algorithms, the FPGA has no advantage in terms of implementation difficulty and power consumption.

Disclosure of Invention

In view of this, the present invention provides a finger vein recognition method and a system thereof, in which the method accelerates the ROI extraction module by using the FPGA, thereby increasing the recognition speed.

In order to achieve the purpose, the invention provides the following technical scheme:

the invention provides a finger vein recognition method, which comprises the following steps:

collecting a finger vein image;

obtaining ROI data of a region of interest of the finger vein image through an ROI hardware extraction module;

reading ROI data from the ROI hardware extraction module and carrying out normalization processing on the gray scale of the ROI data;

extracting the line characteristics of the normalized ROI data;

performing median filtering and connected domain denoising processing on the texture characteristics;

judging whether the registration process is carried out, if so, storing the grain characteristic data into FLASH;

if not, reading the stored grain characteristic data from the FLASH to perform characteristic matching;

and displaying the characteristic matching result through an LCD.

Further, the gray scale of the ROI data is normalized, the texture features of the normalized ROI data are extracted, and results of median filtering and connected domain denoising processing on the texture features are input into an LCD for displaying.

Furthermore, the ROI hardware extraction module adopts an FPGA to realize finger vein image preprocessing, and stores the acquired region of interest of the finger vein image into a dual-port RAM of the FPGA.

Furthermore, the acquisition of the finger vein image is controlled by controlling the corresponding image acquisition equipment through a control instruction stored in a register in a system control module; the control of the LCD display module is realized by controlling the corresponding LCD display module through a control instruction stored by a register in the system control module.

The invention also provides a finger vein recognition system, which comprises a finger vein image acquisition module, an ROI hardware extraction module, a gray scale normalization module, a grain feature extraction module, a filtering and denoising module, a registration judgment module and an LCD display module;

the finger vein image acquisition module is used for acquiring a finger vein image;

the ROI hardware extraction module is used for extracting a region of interest of the finger vein image;

the gray scale normalization module is used for reading the region of interest from the ROI hardware extraction module and normalizing the gray scale of the region of interest;

the texture feature extraction module is used for extracting texture features of the normalized ROI data;

the filtering and denoising module is used for carrying out median filtering and connected domain denoising on the texture characteristics;

the registration judgment module is used for judging whether the registration process is performed, and if the registration process is performed, the texture characteristic data is stored in the FLASH; if not, reading the stored grain characteristic data from the FLASH to perform characteristic matching;

and the LCD display module is used for displaying the intermediate result and the feature matching result in the image acquisition and processing process.

The system further comprises a two-level AHB bus, wherein the AHB bus comprises a first-level AHB bus and a second-level AHB bus;

the first-level AHB bus is respectively connected with the instruction memory and the data memory through an AHB _ to _ SRAM;

the DDR3 is connected with a first-level AHB bus;

the UART is connected with the first-level AHB bus sequentially through the APBInterconnect and the AHP _ to _ APB;

the second-stage AHB bus is respectively connected with a finger vein image acquisition module, an ROI hardware extraction module, an LCD display module, a FLASH read-write control module and a dual-port Block RAM through a system control module;

and the key, the LED indicator lamp and the buzzer are also hung on the second-level AHB bus through GPIO.

Furthermore, a hardware control LCD display register, a reading Block RAM register, a camera control register, an interesting region extraction control register, a pixel accumulation and lower limit threshold value register, a pixel accumulation and upper limit threshold value register, a writing FLASH register, a reading FLASH register and a FLASH erasing register are arranged in the system control module; the CPU controls the operation of the corresponding equipment by configuring the corresponding register.

Further, the ROI hardware extraction module includes the following states, IDLE state, FIRST _ CUT state, EDGE _ POINT state, CORRECTION state, WIDTH _ DEFINE state, HIGHT _ DEFINE state, and SECOND _ CUT state, respectively:

in the IDLE state, if the camera collects a finger vein image and stores the finger vein image into a Block RAM, entering a FIRST _ CUT state;

the FIRST _ CUT state is used for carrying out background interception on the finger vein image and storing the intercepted image into a Block RAM;

the EDGE _ POINT state is used for reading image data from the Block RAM and calculating boundary POINTs of fingers;

the CORRECTION state is used for calculating a finger deflection angle and a value of each line needing translation according to boundary points, then reading image data from the RAM in sequence, storing the image data into the RAM after translation, and actually correcting the translation of the image is the change of the storage address of each pixel point;

the WIDTH-DEFINE state is used for finding out column numbers needing to be intercepted at the left side and the right side of the finger image according to the boundary points after the translation correction so as to determine the WIDTH of the ROI;

the HIGHT-DEFINE state is used for solving the area of the far-end joint of the finger by adopting a sliding window method so as to determine the height of the ROI area, namely the line numbers intercepted up and down are required to be calculated;

and the SECOND _ CUT state is used for reading and writing the image data into the RAM again according to the column numbers intercepted from the left and the right and the line numbers intercepted from the upper and the lower sides, and finally, the data stored in the RAM by taking 0 as the initial address is the data of the ROI image.

The invention has the beneficial effects that:

the finger vein recognition SOC system and method based on ARM Cortex-M3 provided by the invention realize the whole process of the finger vein recognition system, including the processes of finger vein image acquisition, ROI extraction, gray scale and size normalization, feature extraction, connected domain denoising, feature matching and the like, and display the image processing result of each process through an LCD. And a hardware accelerator is used for identifying the ROI through an ROI hardware extraction module, so that the speed of finger vein identification is improved. A large number of tests show that the system can completely realize the function of finger vein recognition, and the recognition accuracy is about 95%. The ROI extraction module is accelerated by adopting the FPGA, so that the total recognition speed can be improved by more than 15%. The system can be used as a scheme of a portable finger vein recognition system for product research and development, can also be used as a process demonstration system for finger vein recognition during experimental teaching, and has higher practical value.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention may be realized and attained by the means of the instrumentalities and combinations particularly pointed out hereinafter.

Drawings

In order to make the object, technical scheme and beneficial effect of the invention more clear, the invention provides the following drawings for explanation:

fig. 1 shows a finger vein recognition implementation process and software and hardware partitioning.

FIG. 2 is a system software flow diagram.

Fig. 3 is a system block diagram.

FIG. 4 is a block diagram of a Cortex-M3 based SOC system.

Fig. 5 shows Master and Slave access relationships for L1 AhbMtx.

Fig. 6 shows Master and Slave access relationships for L2 AhbMtx.

Fig. 7 is a ROI extraction state transition diagram.

Fig. 8 is an algorithm flow for ROI extraction.

Detailed Description

The present invention is further described with reference to the following drawings and specific examples so that those skilled in the art can better understand the present invention and can practice the present invention, but the examples are not intended to limit the present invention.

As shown in fig. 1, the finger vein recognition method provided in this embodiment is a finger vein recognition SOC method based on ARM Cortex-M3, and includes the following steps:

collecting a finger vein image;

obtaining ROI data of the finger vein image through an ROI hardware extraction module;

extracting the line characteristics of the normalized ROI data;

and displaying the characteristic matching result through an LCD.

The image processing result of each process in this embodiment is displayed through an LCD, and specifically includes normalizing the gray scale size of the ROI data, extracting texture features of the normalized ROI data, and inputting the results of performing median filtering and connected domain denoising on the texture features into the LCD for display;

the ROI hardware extraction module adopts an FPGA to accelerate the ROI extraction module to realize finger vein image preprocessing, the FPGA in the embodiment adopts XC7A75T of an Artix-7 series of Xilinx company, CMSDK is adopted to build an SOC system of Cortex-M3, and the ROI hardware extraction module is used for acquiring an interested area of a finger vein image and storing the interested area into a dual-port RAM of the FPGA;

the FLASH provided by the embodiment is MT25QL128, and the read-write control module of the FLASH is more conveniently implemented by hardware and saves more time, so that the read-write control of the FLASH is implemented by a hardware logic circuit.

As shown in fig. 2, in this embodiment, in order to visually display the result of image processing in each stage of the finger vein recognition process, the LCD needs to display the processed image in real time, so that the LCD display is controlled by hardware in the image acquisition and ROI extraction stages.

The collection of finger vein image adopts the camera to catch the image and show in real time in this embodiment, and this camera is on lug connection to FPGA's IO mouth, has practiced thrift image acquisition time, and the control of camera adopts the hardware to realize, detects when the camera module that the finger stretches into collection system's draw-in groove after, then saves the image of gathering to FPGA's dual port RAM.

As shown in fig. 3, the finger vein recognition system provided in this embodiment is based on the SOC design of Xilinx Artix-7, and includes a peripheral circuit, where the peripheral circuit includes a near-infrared LED lighting circuit, a power conversion circuit, a key circuit, an LED indicator circuit, and a buzzer circuit, which are respectively connected to an FPGA.

As shown in fig. 4, in this embodiment, a Cortex-M3 kernel of design start Eval is adopted for development, and the development is performed on the basis of a CMSDK bus and peripheral devices provided by ARM company, where the bus includes two levels of AHB buses, that is, a first-level AHB bus and a second-level AHB bus. The first-level AHB bus is connected with the core processor;

in the figure, I-code represents an instruction bus, which is mainly used for fetching instructions from a memory; d-code represents a data bus, which is mainly used for fetching data from a memory; the System represents a System bus, which is mainly used to access System peripheral modules.

The first-level AHB bus is connected with the second-level AHB bus through an AHB bridge; the second-stage AHB bus is respectively connected with a finger vein image acquisition module, an ROI hardware extraction module, an LCD display module, a FLASH read-write control module and a dual-port Block RAM through a system control module; and the second-level AHB bus is connected with the key, the LED indicator lamp and the buzzer through GPIO.

The Cortex-M3 kernel provided in this embodiment is provided with an instruction memory and a data memory, which are respectively hung on the first-level AHB bus through an AHB _ to _ SRAM; DDR3 is also hung on the first-stage AHB bus; the UART is connected with the first-stage AHB bus sequentially through the APB bus and an AHB-to-APB bridge APBInterconnect;

the instruction memory is used for storing instructions run by the Cortex-M3 kernel; the instruction data memory represents a program memory for storing data run by the Cortex-M3 kernel; DDR3 represents memory for image capture caching; the UART represents a universal asynchronous receiver/transmitter, i.e., a serial port, for transmitting the feature data and the matching result to an upper computer (e.g., a computer).

The finger vein image acquisition module, the LCD display control module, the FLASH read-write control module, the dual-port Block RAM and the ROI hardware extraction module are all hung on the second-level AHB bus through the system control module, and the keys, the LED indicator lamps and the buzzer are also hung on the second-level AHB bus through GPIOs.

The GPIO represents a general purpose input/output port, and has 16 ports, which can be configured to an input mode or an output mode through a register.

As shown in fig. 5, the first-level AHB bus is provided with 5 Slave interfaces (S0-S4) and 5 Master interfaces (M0-M4), the S0 interface is connected to the I-code bus of the Cortex-M3 core, the S1 interface is connected to the D-code bus of the Cortex-M3 core, the S3 interface is connected to the system bus of the Cortex-M3 core, and S3 and S4 are reserved for expanding DMA modules and coprocessors for future systems. The M0 and M1 interfaces are respectively connected with an instruction memory and an instruction data memory, the M2 is connected with an AHB-to-APB bridge, the M3 is connected with a DDR, and the M4 is connected with the AHB-to-AHB bridge to expand a second-level AHB bus. The access relationship between each Slave and the Master is shown in fig. 4.

As shown in fig. 6, the second-level AHB bus is provided with 1 Slave interface (S0) and 2 Master interfaces (M0, M1). The S0 interface is connected with the AHB-to-AHB bridge, the M0 interface is connected with the system control module, the M1 interface is connected with GPIO,

in this embodiment, when the Cortex-M3 core needs to access the modules on each bus, it is necessary to allocate address spaces to each module, and the allocated address spaces are shown in table 1.

Table 1 mapping of module addresses in a system

The system control module provided by the embodiment is an important control center shaft for connecting software and hardware, and converts a control instruction into a hardware control signal, so that the camera module, the ROI accelerator module, the LCD display module and the FLASH read-write module are controlled.

The system control module provided by this embodiment has 9 registers, and software can control the operation of corresponding hardware by configuring these 9 registers. These 9 registers are shown in table 2.

TABLE 2 register List for System control Module

The embodiment also provides a register setting method for controlling the finger vein image recognition chip, wherein the register setting comprises a hardware control LCD display register, a reading Block RAM register, a camera control register, an interested region extraction control register, a pixel accumulation and lower limit threshold value register, a pixel accumulation and upper limit threshold value register, a writing FLASH register, a reading FLASH register and a FLASH erasing register; the specific settings are as follows:

the hardware control LCD display register (LCD _ HW _ CTRL _ DISP) is configured with an LCD hardware control bit; when the LCD hardware control bit is configured to be '1', starting a hardware logic circuit to control the signal of the LCD and drive the LCD to display; when the LCD hardware control bit is configured to be '0', the LCD is controlled by software to display, and the detailed characteristics are as follows:

the address is 0x50000000, and software reads and writes the address through Cortex-M3 kernel to access the register.

The reset value is 0x 00000000, i.e. the default is displayed by software controlling the LCD.

Bits [31:1] are reserved bits.

Bit [0] is the LCD hardware control bit, named LCD _ hw _ ctrl, which is readable and writable, and is displayed by the hardware control LCD when the bit is configured as "1", and is displayed by the software control LCD when the bit is configured as "0".

The use scene and the configuration method of the register are as follows: when the system is started and in a registration and identification mode, firstly, 0x1 is written into a 0x50000000 address, namely the LCD _ hw _ ctrl bit of the register is configured to be '1', a hardware logic circuit of the FPGA is started to control signals of LCD _ CS, LCD _ RS, LCD _ WR, LCD _ RD and the like of the LCD and drive the LCD to display, and therefore, an image acquired by the camera is displayed on the LCD screen in real time. When the software detects that the ROI extraction is completed, 0x0 is written into an address of 0x50000000, namely the LCD _ hw _ ctrl bit of the register is configured to be 0, and the control right of the LCD display is given to the software, so that signals such as LCD _ CS, LCD _ RS, LCD _ WR and LCD _ RD are controlled in a software mode, and the LCD display is driven. And the ROI extraction result, normalization, feature extraction, connected domain denoising, median filtering and display of feature matching results are controlled by software.

The reading Block RAM register is configured with a Block RAM reading control bit and a Block RAM address bit;

when the reading control bit of the Block RAM is '1', reading back the data in the corresponding address of the Block RAM through software, and when the bit is configured to be '0', not reading the Block RAM;

when the address bit of the Block RAM is 1, reading the data in the address specified by the bit;

the READ Block RAM register (READ _ BRAM) is used for reading data from the Block RAM inside the FPGA, and the use of the READ Block RAM register is divided into two steps of reading and writing. When writing this register, its detailed features are:

the address is 0x 50000004, which the software will access to this register by writing to it through the Cortex-M3 kernel.

The reset value is 0x 00000000, namely the Block RAM inside the FPGA is not read by default.

Bits [31:21] are reserved bits.

Bit [20] is the Block RAM read control bit, named bram _ rd, which is writable only. When the bit is configured to be 1, the software reads back the data in the corresponding address of the Block RAM, and when the bit is configured to be 0, the Block RAM is not read.

Bits [19:17] are reserved bits.

Bits [16:0] are the Block RAM address bits, named bram _ addr, which are only writable. When the bram rd bit is 1, the data in the address specified by the bit will be read. Since the image pixel collected by the camera is 320 × 240, the maximum value of the address represented by the bit is 320 × 240-1, i.e., 0 × 12 BFF.

When reading the register, the detailed characteristics are

The address is 0x 50000004, which the software will access the register by reading it through the Cortex-M3 kernel.

The reset value is 0x 00000000, i.e. the data read back is 0 by default.

Bits [31:8] are reserved bits.

Bits [7:0] are read data bits, named bram _ data, which are only readable. The data of the address corresponding to the Block RAM read will be stored in this bit.

The use scene and the configuration method of the register are as follows: when an ROI extraction hardware accelerator module in the FPGA finishes ROI data extraction, the ROI data is stored in a Block RAM in the FPGA, software needs to read back the ROI data for subsequent algorithm processing, and at the moment, the data in the Block RAM needs to be read by configuring a register. For example, it is necessary to read the data in address 0x1234, then it is first necessary to write to the register 0x00101234 and then read the register whose data in bits 7:0 is the data in Block RAM address 0x 1234.

The CAMERA control register (CAMERA _ CTRL) is configured with an image pixel accumulation sum bit, an image acquisition completion flag bit and a CAMERA control bit;

the image pixel sum bits are used to calculate an image pixel sum for each frame of the acquisition,

the image acquisition completion zone bit is used for storing image information into a Block RAM and enabling the position to be 1 when the camera detects that a finger extends into a clamping groove of the image acquisition device;

when the configuration bit of the camera control bit is '1', starting a camera to collect images, and when the configuration bit is '0', closing the camera;

it is characterized in that

The address is 0x 50000008, and software reads and writes the address through Cortex-M3 kernel to access the register.

The reset value is 0x 00000000, i.e. the camera is not turned on by default.

Bits [31:29] are reserved bits.

Bits [28:4] are the image pixel accumulated sum bits, named pix _ color _ sum. When the camera collects images, the value of the bit is updated after each frame of image, and the value is used for storing the pixel accumulation sum of each frame of image.

Bits [3:2] are reserved bits.

And the bit [1] is an image acquisition completion flag bit, namely a camera _ finish _ flag, when the camera detects that a finger extends into a card slot of the image acquisition device, the camera stores the image information into a Block RAM and stores the position to be '1'.

And the bit [0] is a camera control bit, the name of the camera _ ctrl is taken, when the bit is configured to be '1', the camera is started to collect images, and when the bit is configured to be '0', the camera is closed.

The use scene and the configuration method of the register are as follows: when a registration or identification mode is started, firstly, configuring the camera _ ctrl bit of the register as '1', starting a camera, when the camera _ finish _ flag bit of the register is read as '1', indicating that a finger image is acquired, and writing the camera _ ctrl bit and the camera _ finish _ flag bit of the register as 0 so as to close the camera and clear the flag bit. The value of pix _ color _ sum of the register can be read when the camera is started so as to obtain the pixel accumulated value of the image when the finger extends into the card slot and does not extend into the card slot, and therefore the pixel accumulated value and the threshold value of whether the finger extends into the card slot or not are detected through adjustment according to the value.

The region of interest extraction control register (ROI _ CTRL) is configured with a region of interest width bit, a region of interest height bit, a region of interest extraction completion flag bit and a region of interest extraction control bit;

the width of the region of interest is set; for storing the acquired width data of the region of interest;

the region of interest height position; the height data is used for storing the acquired region of interest;

the region of interest extraction completion flag bit; when the extraction of the region of interest is completed, the configuration bit is '1';

extracting a control bit from the region of interest; when the configuration bit is '1', starting the ROI hardware accelerator module to extract the region of interest, and when the configuration bit is '0', closing the ROI module;

it is characterized in that

The address is 0x 5000000C, and software reads and writes the address through Cortex-M3 kernel to access the register.

The reset value is 0x 00000000, i.e. the ROI extraction function is not turned on by default.

Bits [31:24] are region of interest width bits, ROI _ width. After ROI extraction, the resulting width data of the region of interest will be stored into this bit.

Bits [23:16] are the region of interest height bits, ROI _ height. After ROI extraction, the resulting height data of the region of interest will be stored into this bit.

Bits [15:2] are reserved bits.

And bit [1] is a region-of-interest extraction completion flag bit, namely ROI _ finish _ flag, and when the operation of an ROI hardware accelerator module in the FPGA is finished, namely the region-of-interest extraction is completed, the position is '1'.

Bit [0] is the region of interest extraction control bit, named ROI _ ctrl, when the bit is configured to be '1', starting the ROI hardware accelerator module to extract the region of interest, and when the bit is '0', closing the ROI module.

The use scene and the configuration method of the register are as follows: when the CAMERA _ finish _ flag bit of a CAMERA control register (CAMERA _ CTRL) is read to be "1", it is indicated that a finger image is acquired, at this time, the ROI _ CTRL bit of the register is configured to be "1", an ROI hardware accelerator module is started to start ROI extraction, when the ROI _ finish _ flag bit of the register is read to be "1", it is indicated that the region of interest has been extracted, the ROI _ CTRL bit and the ROI _ finish _ flag bit of the register are written to be 0, so as to close the ROI module and clear flag bit, at this time, the ROI _ width and the ROI _ height bit of the register can be read to obtain the width and height of the extracted region of interest, so as to be used in subsequent algorithm processing.

The pixel SUM and lower threshold register (PIX _ SUM _ MIN) is configured with an image pixel SUM and lower threshold bit;

the image pixel accumulation and lower threshold bit; the pixel accumulation lower limit threshold is used for storing a frame of the collected image;

it is characterized in that

The address is 0x 50000010, which the software will access the present registers by reading and writing it through the Cortex-M3 kernel.

The reset value is 0x 00400000, i.e. by default no finger is inserted into the card slot of the image capturing device when the sum of the pixel sums of each frame of image is less than 0x 400000.

Bits [31:25] are reserved bits.

Bits [24:0] are the image pixel accumulation and lower threshold bits, named pix _ sum _ min, which are readable and writable. When the sum of pixel accumulation of one frame of the image captured by the camera is smaller than the threshold value, the finger does not extend into the card slot, otherwise, if the sum of pixel accumulation is larger than the threshold value and is smaller than or equal to the value of pix _ sum _ max bit in the register of pixel accumulation and upper limit threshold value, the finger extends into the card slot of the image acquisition device.

The use scene and the configuration method of the register are as follows: if the detection that the finger extends into the card slot is inaccurate when the system is debugged, when the CAMERA is started, pixel accumulated sum values when the finger extends into the card slot and does not extend into the card slot are obtained by reading a pix _ color _ sum bit in a CAMERA control register (CAMERA _ CTRL), and the pix _ sum _ min bit of the register is configured according to the pixel accumulated sum value and the lower limit threshold value to set the image pixel accumulated sum.

The pixel SUM and ceiling threshold register (PIX _ SUM _ MAX) is configured with image pixel SUM and ceiling threshold bits;

the image pixel accumulation and upper threshold bit; the pixel accumulation upper limit threshold is used for storing one frame of the collected image;

it is characterized in that

The address is 0x 50000014, which the software will access the present registers by reading and writing it through the Cortex-M3 kernel.

The reset value is 0x 00B 00000, that is, it is default that no finger is inserted into the card slot of the image capturing device when the sum of the pixel sums of each frame of image is greater than 0xB 00000.

Bits [31:25] are reserved bits.

Bits [24:0] are the image pixel accumulation and upper threshold bits, named pix _ sum _ max, which are readable and writable. When the sum of pixel accumulation of one frame of the image captured by the camera is larger than the threshold value, the finger does not extend into the card slot, otherwise, if the sum of pixel accumulation is smaller than the threshold value and is larger than or equal to the value of pix _ sum _ min bit in the register of the pixel accumulation sum lower limit threshold value, the finger extends into the card slot of the image acquisition device.

The use scene and the configuration method of the register are as follows: if the detection that the finger extends into the card slot is inaccurate when the system is debugged, when the CAMERA is started, pixel accumulated sum values when the finger extends into the card slot and does not extend into the card slot are obtained by reading pix _ color _ sum bits in a CAMERA control register (CAMERA _ CTRL), and pix _ sum _ max bits of the register are configured according to the pixel accumulated sum values to set an image pixel accumulated sum upper limit threshold value.

The WRITE FLASH register (FLASH _ WRITE),

configuring FLASH write data bits and FLASH write address bits;

the FLASH data writing bit is configured for writing FLASH data;

the FLASH write address bit is used for configuring an address for writing FLASH;

it is characterized in that

The address is 0x 50000018, which the software will access the present registers by reading and writing it through the Cortex-M3 kernel.

The reset value is 0x 00800000, i.e. the default FLASH memory stores the starting address of the vein print feature as 0x 00800000.

Bits [31:24] are the FLASH write data bits, naming FLASH _ wdata, i.e., the data that needs to be written into FLASH. The bit is readable and writable.

Bits [23:0] are the FLASH write address bits, naming FLASH _ waddr, i.e., the address to be written to FLASH. The bit is readable and writable.

The use scene and the configuration method of the register are as follows: if the current mode is the registration mode, after the connected domain denoising and median filtering of the vein lines are carried out, the feature data needs to be stored into the FLASH, and at the moment, the software stores the feature data into the corresponding address of the FLASH by configuring the register. For example, if the 3 rd pixel value of the feature data is 0xFF and the address to be stored is 0x800003, the register is configured to be 0xFF 800003. In the embodiment, the FLASH memory is selected from MT25QL128, the storage space is 16M bytes, the address range is 0x 000000-0 xFFFFFF, and the storage space of 8MB is reserved from 0x800000 to the end, wherein the 8MB is specially used for storing the vein line characteristics. One vein print is characterized by 96 x 64 bytes, i.e., 6144 bytes. In this embodiment, the initial addresses of the vein texture feature images stored in the image memory are aligned to 8KB, so that 1024 vein texture feature images can be stored in total. The first feature store header address is 0x800000, the second feature store header address is 0x802000, the third feature store header address is 0x804000, and so on.

The FLASH reading register (FLASH _ READ is configured with a FLASH reading address bit which is used for configuring the address of the FLASH to be READ;

it is characterized in that

The address is 0x 5000001C, and software reads and writes this address through the Cortex-M3 kernel will access this register.

The reset value is 0x 00800000, i.e. the starting address of the default read FLASH is 0x 00800000.

Bits [31:25] are reserved bits.

Bits [23:0] are FLASH read address bits, namely the FLASH _ raddr bit, which is the address of the FLASH to be read. The bit is write only.

When reading the register, the detailed characteristics are

The reset value is 0x 00000000, i.e. the data of FLASH read by default is 0x 0.

Bits [31:8] are reserved bits.

Bits [7:0] are FLASH read address bits, naming FLASH _ rdata bits, i.e. the data of the address corresponding to the FLASH to be read. This bit is read-only.

The use scene and the configuration method of the register are as follows: if the current mode is the identification mode, after the connected domain denoising and the median filtering of the vein lines are carried out, the feature data stored in the FLASH before are required to be read out for feature matching, and at the moment, the software is used for reading out the feature data of the corresponding address of the FLASH by configuring the register. For example, if the data in FLASH address 0x801234 needs to be read, the register needs to be written to 0x00801234 first, and then the register is read, and the data in bits 7:0 is the data in FLASH address 0x 801234.

The FLASH reading eraser (FLASH _ CLR) is configured with an erasing selection bit and a starting address bit of FLASH erasing;

when the erasing selection bit is configured to be '1', all vein texture characteristics in the FLASH are erased, and when the erasing selection bit is configured to be '0', only one texture characteristic image at the beginning of the FLASH _ caddr address is erased;

it is characterized in that

The address is 0x 50000020, which the software will access the present registers by reading and writing it through the Cortex-M3 kernel.

The reset value is 0x 00800000, i.e. the default FLASH erase address is 0x 00800000, and only the texture image starting at that address is erased.

Bits [31:25] are reserved bits.

Bit [24] is an erase select bit, named CA, which erases all vein texture features in FLASH when the bit is configured as "1" and erases only one texture feature image from the FLASH _ caddr address when the bit is configured as "0". The bit is write only.

Bits [23:0] are the starting address of FLASH erase, naming the FLASH _ caddr bit, i.e. the first address of FLASH that needs to be erased. The bit is write only.

The use scene and the configuration method of the register are as follows: if the vein texture features stored in the FLASH are to be deleted currently, the register needs to be configured. If the CA bit configuring this register is "1", then whatever the value of the FLASH _ caddr bit, all the feature data in FLASH after address 0x 00800000 will be erased. If only a certain feature image is to be erased, the CA bit of the register is configured to be "0", and the initial address of the erasure is specified by flash _ caddr. The 8KB of data starting with the flash _ caddr address will be erased. The configuration of the flash _ caddr bit at this point requires 8KB alignment.

After the CAMERA controller provided in this embodiment opens the CAMERA module by configuring the CAMERA control register (CAMERA _ CTRL), the system control module generates a CAMERA _ start signal to start the CAMERA to acquire an image in real time, and stores the acquired image data into the Block RAM, and the LCD controller module reads data from the Block RAM in real time to display the data.

The camera controller in this embodiment is provided with a finger position detection module for detecting whether a finger extends into the card slot of the acquisition device, and the finger position detection module detects according to the following steps:

acquiring a vein image when a finger is present;

calculating the pixel sum of the vein image;

judging whether the sum of the pixels of the vein image is within a preset threshold range, if so, extending the finger into the card slot;

if not, the finger does not extend into the clamping groove;

the cycle is repeated by moving the finger to the position of the slot.

The default range of the preset threshold provided by the embodiment is as follows: the lower threshold is 0x400000 and the upper threshold is 0xB 00000.

When the pixel accumulated value of the image is detected to be in the interval, the finger stretches into the card slot, then the image after 100 frames of delay is collected to be used as an original image of the finger vein, and a camera _ finish signal is generated to be used as a mark for collection ending. The effect of delaying 100 frames is to stabilize the image, the accumulated SUM of image pixels can be read out through the PIX _ color _ SUM bit of the register CAMERA _ CTRL, and the lower threshold and the upper threshold of the detection finger can be configured through the registers PIX _ SUM _ MIN and PIX _ SUM _ MAX.

As shown in fig. 7 and 8, the ROI hardware extraction module writes the ROI hardware accelerator using Verilog hardware description language, and when the camera acquires the finger vein image, the ROI hardware extraction module starts to work,

the ROI algorithm in this embodiment is developed by using a hardware description language Verilog for FPGA, a main program is designed by using a Finite-state machine (FSM), and the ROI extraction process is completed by 7 states, which are IDLE states: namely an idle state; FIRST _ CUT state: namely a primary interception state; EDGE _ POINT status: namely the detection state of the finger boundary point; CORRECTION state: i.e. the rotation correction state; WIDTH _ DEFINE state: i.e. width-defining state; high _ default state: namely a height-defining state; SECOND _ CUT state: i.e. the secondary intercept state.

As shown in fig. 7, fig. 7 is a transition diagram of an ROI extraction state, where the IDLE state is always in a state when a finger does not extend into a card slot, and enters a FIRST _ CUT state after the finger is detected to extend into the card slot and an image is stored into a Block RAM;

the FIRST _ CUT state aims at cutting off a background area and reducing the size of an image, the image enters an EDGE _ POINT state after being CUT off, and the image cutting is finished by reading and writing a Block RAM;

reading image data from the RAM in sequence in an EDGE _ POINT state, and calculating 4 boundary POINTs of the finger;

under a CORRECTION state, firstly calculating a finger deflection angle and a value needing translation of each line according to boundary points obtained in the previous state, then reading image data from an RAM in sequence, and storing the image data into the RAM after translation, wherein the translation CORRECTION of the image is actually the change of storage addresses of all pixel points;

under the WIDTH _ DEFINE state, according to the boundary point after the translation correction, column numbers needing to be intercepted on the left side and the right side of the finger image are found out so as to determine the WIDTH of the ROI;

in a HIGHT-DEFINE state, a sliding window method is adopted to obtain the region of the far-end joint of the finger, so that the height of the ROI region is determined, namely the line numbers intercepted up and down are required to be calculated;

in the state of SECOND _ CUT, reading and writing the image data into the RAM again according to the column numbers intercepted from the left and the right and the line numbers intercepted from the upper and the lower sides, and finally, the data stored in the RAM by taking 0 as the initial address is the data of the ROI image;

the extraction of ROI is completed by the above 7 states, the extracted image data is stored in the Block RAM, and then the ROI _ finish signal is generated to notify software that data can be read from the RAM and subsequent arithmetic processing is performed.

The ROI module is designed to have the advantage of a hardware accelerator in speed, and a software scheme is adopted to realize the same ROI algorithm. The clock frequency of both hardware and software was 36MHz, and the run time of each step and the total run time pair ratio are shown in table 3. It can be seen that the total running time of the ROI method implemented by FPGA is 3.822ms, and 78.547ms is required by software. At the same clock frequency, hardware is more than 20 times faster than software.

TABLE 3ROI Module hardware and software implementation run time comparison (ms)

The algorithm behind the ROI of the system, such as normalization, direction segmentation, connected domain denoising and the like, is realized in a software mode, and the running time of the subsequent algorithm is about 480ms in total through MDK running test, so that the result can be obtained, the ROI module is accelerated by hardware, and the total recognition speed can be improved by more than 15%.

The above-mentioned embodiments are merely preferred embodiments for fully illustrating the present invention, and the scope of the present invention is not limited thereto. The equivalent substitution or change made by the technical personnel in the technical field on the basis of the invention is all within the protection scope of the invention. The protection scope of the invention is subject to the claims.

Claims

1. The finger vein recognition method is characterized by comprising the following steps: the method comprises the following steps:

collecting a finger vein image;

extracting the line characteristics of the normalized ROI data;

and displaying the characteristic matching result through an LCD.

2. The method of claim 1, wherein: and normalizing the gray scale of the ROI data, extracting the texture characteristics of the normalized ROI data, and inputting the results of median filtering and connected domain denoising on the texture characteristics into an LCD for displaying.

3. The method of claim 1, wherein: the ROI hardware extraction module adopts an FPGA to realize finger vein image preprocessing, and stores the acquired region of interest of the finger vein image into a dual-port RAM of the FPGA.

4. The method of claim 1, wherein: the acquisition of the finger vein image is controlled by controlling corresponding image acquisition equipment through a control instruction stored in a register in a system control module; the control of the LCD display module is realized by controlling the corresponding LCD display module through a control instruction stored by a register in the system control module.

5. Finger vein identification system, its characterized in that: the system comprises a finger vein image acquisition module, an ROI hardware extraction module, a gray scale size normalization module, a grain feature extraction module, a filtering and denoising module, a registration judgment module and an LCD display module;

6. The system of claim 5, wherein: the system also comprises a two-level AHB bus, wherein the AHB bus comprises a first-level AHB bus and a second-level AHB bus;

the DDR3 is connected with a first-level AHB bus;

7. The system of claim 6, wherein: the system further comprises a core processor, and the first-level AHB bus is connected with the core processor.

8. The system of claim 6, wherein: the core processor adopts an ARM Cortex-M3 core processor.

9. The system of claim 5, wherein: the system control module is internally provided with a hardware control LCD display register, a Block RAM reading register, a camera control register, an interesting region extraction control register, a pixel accumulation and lower limit threshold value register, a pixel accumulation and upper limit threshold value register, a FLASH writing register, a FLASH reading register and a FLASH erasing register; and each register is provided with a control instruction for controlling corresponding equipment.

10. The system of claim 5, wherein: the ROI hardware extraction module includes the following states, IDLE state, FIRST _ CUT state, EDGE _ POINT state, CORRECTION state, WIDTH _ DEFINE state, HIGHT _ DEFINE state, and SECOND _ CUT state, respectively: