WO2018049858A1 - 一种指静脉识别装置的校准方法 - Google Patents

一种指静脉识别装置的校准方法 Download PDF

Info

Publication number
WO2018049858A1
WO2018049858A1 PCT/CN2017/087841 CN2017087841W WO2018049858A1 WO 2018049858 A1 WO2018049858 A1 WO 2018049858A1 CN 2017087841 W CN2017087841 W CN 2017087841W WO 2018049858 A1 WO2018049858 A1 WO 2018049858A1
Authority
WO
WIPO (PCT)
Prior art keywords
finger vein
pixel
section
cross
images
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2017/087841
Other languages
English (en)
French (fr)
Inventor
梁添才
刘建平
金晓峰
黎明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
GRG Banking Equipment Co Ltd
GRG Banking IT Co Ltd
Original Assignee
GRG Banking Equipment Co Ltd
GRG Banking IT Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by GRG Banking Equipment Co Ltd, GRG Banking IT Co Ltd filed Critical GRG Banking Equipment Co Ltd
Publication of WO2018049858A1 publication Critical patent/WO2018049858A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/22Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/26Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/14Vascular patterns

Definitions

  • the present invention relates to the field of image acquisition, and in particular to a calibration method for a finger vein recognition device.
  • Vein recognition technology achieves the purpose of identification by in vivo recognition of vein images in the fingers or palms. It has high anti-counterfeiting, in vivo detection, high accuracy, adaptability and ease of use.
  • the finger vein recognition device involves a series of contents such as a light source, an optical lens, a photosensitive chip, etc.
  • the application area of the identification device covers various regions of China, and the external climate and temperature affect the luminous intensity of the near-infrared LED lamp of the device, and the optical lens transmittance At the same time, the external light directly interferes with the light received by the photosensitive chip, thereby affecting the acquired image, and finally results in a decrease in the recognition success rate of the finger vein recognition device.
  • the invention provides a calibration method for a finger vein recognition device, which can effectively reduce the influence of factors such as climate, temperature and illumination in different regions on the recognition success rate of the finger vein recognition device.
  • an embodiment of the present invention provides a calibration method for a finger vein recognition device, including the steps of:
  • ⁇ s is a preset standard value of the finger vein recognition device
  • ( ⁇ , ⁇ ) is a Gaussian distribution parameter obtained by the above step
  • sign is a symbol function
  • is the infrared image.
  • the recognition algorithm of the finger vein recognition device provided by the invention directly recognizes on the collected image, and the consistency of the collected image directly affects the success rate of the device identification, and the above steps improve the collection of different devices in different environments.
  • the consistency of the image can effectively reduce the influence of image acquisition consistency on the recognition algorithm of the finger vein recognition device and improve the recognition success rate.
  • the Gaussian distribution parameter ( ⁇ , ⁇ ) of the gray value of the cross-section pixel of the collected N*L finger vein images is calculated by the following steps;
  • S21 Perform binarization on each of the finger vein images to divide a finger vein region, and obtain a finger vein line based on the finger vein region;
  • the gray level of the cross-section of each vein basically conforms to the Gaussian distribution, and the above steps are used to enrich the statistical data of each cross-section of each finger vein line, so that the statistical result is more accurate.
  • the invention also proposes a calibration method for another finger vein recognition device, comprising the steps of:
  • ⁇ s is a preset standard value of the finger vein recognition device
  • ( ⁇ r , ⁇ r ) is a Gaussian distribution parameter obtained by the above step
  • is a gamma of the infrared image sensor Horse mapping coefficient
  • mapping coefficient ⁇ calculated the gamma adjusted values of the input pixel f and pixel output value f. 1 mapping relationship in the current state of the infrared image sensor is acquired, then the vein authentication device adjusted by said means The pixel output value f 1 is identified to obtain the recognition result.
  • the following steps and formulas may be used to calculate the Gaussian distribution parameter of the gray value of the cross-section pixel of each of the finger vein images collected:
  • M t is the number of pixels corresponding to the t-th cross-section, T ⁇ 3;
  • the following formula can be used to calculate a Gaussian distribution parameter ( ⁇ r , ⁇ r ) of the gray value of the cross-section pixel of all selected finger vein images that meet the requirements of the finger vein image width:
  • ⁇ r ( ⁇ 1 + ⁇ 2 + ⁇ 3 +...+ ⁇ p )/P
  • ⁇ 1 , ⁇ 2 , ⁇ 3 , ..., ⁇ p are respectively pixel gray values of the gray values of the cross-section pixels of each finger vein image selected Degree average.
  • the following formula can also be used to calculate the Gaussian distribution parameter ( ⁇ r , ⁇ r ) of the gray value of the cross-section pixel of all the finger vein images selected:
  • the NIBLACK image binarization method is used to binarize each of the finger vein images to extract a finger vein region; after the extracted finger vein region is segmented, skeleton extraction is performed. Finger vein line.
  • the calibration method of the finger vein recognition device of the present invention obtains the adjustment parameters of the infrared image sensor corresponding to the finger vein recognition device according to the gray scale change of the vein region of the finger vein acquisition image, which can effectively reduce the difference The effects of regional climate, temperature and light.
  • FIG. 1 is a schematic flow chart of a calibration method of a finger vein recognition device according to a first embodiment of the present invention.
  • FIG. 2 is a schematic diagram of a specific process of step S102 in FIG. 1.
  • Fig. 3 is a schematic view showing the vein extraction in the first embodiment of the present invention.
  • Fig. 4 is a schematic cross-sectional view showing a vein in the first embodiment of the present invention.
  • FIG. 5 is a schematic flow chart of a calibration method of a finger vein recognition device according to Embodiment 2 of the present invention.
  • FIG. 6 is a schematic diagram of a specific process of step S202 in FIG. 5.
  • FIG. 7 is a schematic flow chart of a calibration method of a finger vein recognition device according to a third embodiment of the present invention.
  • FIG. 8 is a schematic diagram of a specific process of step S302 in FIG. 7.
  • FIG. 9 is a schematic diagram of a specific process of step S304 in FIG. 7.
  • the invention provides a calibration method for a finger vein recognition device for performing corresponding calibration of a finger vein recognition device according to a current environment before using a finger vein recognition device.
  • the external environment including external climate, temperature and illumination
  • the present invention provides a calibration method for a finger vein recognition device, which is to improve the consistency of image collection under different environments in different finger vein recognition devices, thereby reducing the influence of the finger vein recognition device on environmental interference, thereby improving The recognition accuracy of the finger vein recognition device.
  • the calibration method of the finger vein recognition device of the present invention will be specifically described below by way of various embodiments.
  • the calibration method for a finger vein recognition device according to the first embodiment of the present invention includes steps S101-S104:
  • This step is used to acquire a finger vein image.
  • the infrared image sensor using the finger vein recognition device irradiates the fingers of the N collectors with the finger vein infrared light source of the preset intensity, and collects the L persons of the N state under the current state (in the same regional environment). Refers to the vein image to obtain N*L finger vein images, where N ⁇ 2, L ⁇ 1.
  • S102 Calculate a Gaussian distribution parameter ( ⁇ , ⁇ ) of the cross-sectional pixel gray value of the collected N*L finger vein images; wherein, the cross-sectional pixel gray value of each of the finger vein images conforms to a Gaussian distribution.
  • This step is used to count the gray values of the collected N*L finger vein images.
  • the step may be implemented by the following steps, including steps S1021-S1023, where:
  • S1021 Perform binarization on each of the finger vein images to divide a finger vein region, and obtain a finger vein line based on the finger vein region;
  • FIG. 3 is a schematic diagram of vein extraction in the first embodiment of the present invention, and the vein line 32 is extracted from the vein image 31.
  • each of the finger vein images is binarized by using a NIBLACK image binarization method. Thereby extracting the finger vein area. After the extracted finger vein region is divided, skeleton extraction is performed to obtain a finger vein line.
  • the gray value of the i-th pixel of the b-th cross-section, M b is the number of pixels corresponding to the b-th cross-section, T ⁇ 3; wherein the gray value of the cross-section pixel of each of the finger vein images conforms to Gaussian distributed;
  • FIG. 4 is a schematic cross-sectional view of a vein in the first embodiment of the present invention. Specifically, the extracted vein lines 40 are equally spaced T sections, and a vein section 41, a vein section 42, and a vein section 43 are obtained.
  • S103 Pass the formula: Calculating a gamma mapping coefficient of the infrared image sensor, wherein ( ⁇ s , ⁇ s ), ⁇ s is a preset standard value of the finger vein recognition device, and ( ⁇ , ⁇ ) is a Gaussian distribution parameter obtained by the above step , ⁇ is the gamma mapping coefficient of the infrared image sensor.
  • the gamma mapping coefficient GAMMA is used to correct the infrared image sensor in order to make the output images of the different devices, in different temperatures and regions, when the same finger is collected, as much as possible, thereby ensuring the device. Identify the effect.
  • the preset adjustment formula can be adopted.
  • the image pixels in the current state acquired by the finger vein recognition device through the infrared image sensor are calibrated, and the pixel-calibrated image is subjected to finger vein recognition to obtain a recognition result.
  • the algorithm process used for the identification may be in a manner well known to those skilled in the art, and the description is omitted here.
  • the infrared image sensor using the finger vein recognition device collects at least one finger vein image of each person and each finger, and takes multiple sections for each finger vein line, enriching the statistical data and making the statistical result more Accurately, at the same time, the correction of the infrared image sensor by the gamma mapping coefficient in the first embodiment enables different finger vein recognition devices to output the same finger as uniformly as possible when the same finger is collected under different temperatures and regions. The recognition success rate of the finger vein recognition device is improved.
  • the calibration method of the finger vein recognition device provided by the second embodiment of the present invention includes steps S201-S206:
  • S201 using an infrared image sensor of the finger vein recognition device to collect L finger images of N individuals in the current state, wherein N ⁇ 2, L ⁇ 1;
  • This step is used to acquire a finger vein image.
  • the infrared image sensor using the finger vein recognition device irradiates the fingers of the N collectors with the finger vein infrared light source of the preset intensity, and collects the L persons of the N state under the current state (in the same regional environment). Refers to the vein image to obtain N*L finger vein images, where N ⁇ 2, L ⁇ 1.
  • S202 Calculate a Gaussian distribution parameter of the gray value of the cross-section pixel of each of the finger vein images collected respectively; wherein the gray value of the cross-section pixel of each of the finger vein images conforms to a Gaussian distribution, and the Gaussian distribution parameter includes Pixel gray mean and standard deviation;
  • This step is used to count the gray values of the collected N*L finger vein images.
  • the step may be implemented by the following steps, including steps S2021-S2023, where:
  • S2021 Perform binarization on each of the finger vein images to divide a finger vein region, and obtain a finger vein line based on the finger vein region;
  • each of the finger vein images is binarized by using a NIBLACK image binarization method to extract a finger vein region. After the extracted finger vein region is divided, skeleton extraction is performed to obtain a finger vein line.
  • S2022 taking equal intervals of T sections for each of the finger vein lines, taking the gray value of all the pixels of each section, and recording among them
  • the gray value of the i-th pixel of the t-th cross-section, M t is the number of pixels corresponding to the t-th cross-section, T ⁇ 3; wherein the gray value of the cross-section pixel of each of the finger vein images conforms to Gauss distributed.
  • S203 Calculate a width of each of the finger vein images according to a Gaussian distribution parameter of a cross-sectional pixel gray value of each of the finger vein images, and then calculate an average value h w of widths of all the finger vein images, and select the All finger vein images having a width between h w (1 ⁇ A%) in N*L finger vein images; wherein 0 ⁇ A ⁇ 50, the width of each of the finger vein images refers to each of the fingers The number of pixels of the cross-section pixel gray value on the vein image is smaller than the corresponding pixel gray mean value;
  • This step is used to count the gray value of the selected finger vein image.
  • the Gaussian distribution parameter ( ⁇ r , ⁇ r ) of the gray value of the cross-section pixel of all the finger vein images selected is calculated using the following formula:
  • ⁇ r ( ⁇ 1 + ⁇ 2 + ⁇ 3 +...+ ⁇ p )/P
  • ⁇ 1 , ⁇ 2 , ⁇ 3 , ..., ⁇ p are respectively pixel gray values of the gray values of the cross-section pixels of each finger vein image selected Degree average.
  • S206 Pass the formula according to the calculated gamma mapping coefficient ⁇ Adjusting the mapping relationship between the pixel input value f and the pixel output value f 1 in the current state acquired by the infrared image sensor, and then identifying the adjusted pixel output value f 1 by the finger vein recognition device to obtain the recognition result.
  • the preset adjustment formula can be adopted.
  • the image pixels in the current state acquired by the finger vein recognition device through the infrared image sensor are calibrated, and the pixel-calibrated image is subjected to finger vein recognition to obtain a recognition result.
  • the finger vein image is selected when calculating the Gaussian distribution parameter of the gray value of the cross-section pixel of the finger vein image, and it is preferable to select all the finger vein images collected.
  • the experimental result is more stable.
  • the width of the finger vein image is 0.9 to 1.1 times the average width of the finger vein image, and the gamma mapping coefficient is obtained according to the Gaussian distribution parameter of the gray value of the cross-section pixel of the selected finger vein image, and the gamma is used.
  • the recognition success rate of the finger vein identification device after the mapping coefficient calibration is higher.
  • FIG. 7 is a schematic flowchart of a method for calibrating a finger vein recognition device according to Embodiment 3 of the present invention.
  • the calibration method for a finger vein recognition device according to Embodiment 3 of the present invention includes steps S301-S306:
  • the infrared image sensor of the finger vein recognition device is used to collect the image of each finger vein of the N person in the current state, wherein N ⁇ 2, L ⁇ 1;
  • This step is used to acquire a finger vein image.
  • the infrared image sensor using the finger vein recognition device irradiates the fingers of the N collectors with the finger vein infrared light source of the preset intensity, and collects the L persons of the N state under the current state (in the same regional environment). Refers to the vein image to obtain N*L finger vein images, where N ⁇ 2, L ⁇ 1.
  • S302 Calculate a Gaussian distribution parameter of the gray value of the cross-section pixel of each of the finger vein images collected, wherein the gray value of the cross-section pixel of each of the finger vein images conforms to a Gaussian distribution, and the Gaussian distribution parameter includes Pixel gray mean and standard deviation;
  • This step is used to count the gray values of the collected N*L finger vein images.
  • the step may be implemented by the following steps, including steps S3021-S3023, where:
  • S3021 Perform binarization on each of the finger vein images to divide a finger vein region, and obtain a finger vein line based on the finger vein region;
  • each of the finger vein images is binarized by using a NIBLACK image binarization method to extract a finger vein region. After the extracted finger vein region is divided, skeleton extraction is performed to obtain a finger vein line.
  • S3022 Taking equal intervals of T sections for each of the finger vein lines, taking the gray value of all the pixels of each section, and recording among them
  • the gray value of the i-th pixel of the t-th cross-section, M t is the number of pixels corresponding to the t-th cross-section, T ⁇ 3, wherein the gray value of the cross-section pixel of each of the finger vein images conforms to Gaussian distributed.
  • S303 Calculate a width of each of the finger vein images according to a Gaussian distribution parameter of a gray value of a cross-sectional pixel of each of the finger vein images, and then calculate an average value h w of widths of all the finger vein images, and select the All finger vein images having a width between h w (1 ⁇ A%) in N*L finger vein images; wherein 0 ⁇ A ⁇ 50, the width of each of the finger vein images refers to each of the fingers
  • the grayscale value of the cross-section pixel on the vein image is smaller than the number of pixels of the corresponding grayscale mean of the pixel.
  • This step is used to count the gray value of the selected finger vein image.
  • the step can be as follows:
  • the implementation includes steps S3041-S3042, wherein:
  • S305 Pass the formula: Calculating a gamma mapping coefficient of the infrared image sensor, wherein ( ⁇ s , ⁇ s ), ⁇ s is a preset standard value of the finger vein recognition device, and ( ⁇ r , ⁇ r ) is calculated by the above step S304
  • the cross-sectional pixel gray value Gaussian distribution parameter of the selected finger vein image, ⁇ is the gamma mapping coefficient of the infrared image sensor.
  • the preset adjustment formula can be adopted.
  • the image pixels in the current state acquired by the finger vein recognition device through the infrared image sensor are calibrated, and the pixel-calibrated image is subjected to finger vein recognition to obtain a recognition result.
  • the finger vein image is selected when calculating the Gaussian distribution parameter of the gray value of the cross-section pixel of the finger vein image, and it is preferable to select all the finger vein images collected.
  • the experimental result is more stable.
  • the width of the finger vein image is 0.9 to 1.1 times the average width of the finger vein image, and the gamma mapping coefficient is obtained according to the Gaussian distribution parameter of the gray value of the cross-section pixel of the selected finger vein image, and the gamma is used.
  • the recognition success rate of the finger vein identification device after the mapping coefficient calibration is higher.
  • the difference from the second embodiment is that the method for calculating the Gaussian distribution parameter of the cross-sectional pixel gray value of the selected finger vein image is different.
  • the second embodiment is to first find the interface pixel gray scale of each selected finger vein image.
  • the Gaussian distribution parameter is calculated, and the Gaussian distribution parameters of all the finger vein images are calculated.
  • the Gaussian distribution parameters of all the gray values of the finger vein segments are directly obtained.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
  • Image Input (AREA)

Abstract

一种指静脉识别装置的校准方法,包括以下步骤:利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1 (S101);计算采集到的所述N*L个指静脉图像的截面像素灰度值的高斯分布参数;其中,每一所述指静脉图像的截面像素灰度值符合高斯分布 (S102);通过公式计算所述红外图像传感器的伽马映射系数,其中,(μ s,σ s)、γ s为所述指静脉识别装置预设的标准值,(μ,σ)为上述步骤得到的高斯分布参数,γ为所述红外图像传感器的伽马映射系数 (S103);根据计算得到的所述伽马映射系数γ,调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f 1的映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f 1进行识别以获取识别结果 (S104)。上述校准方法可有效降低不同地域气候、温度和光照等因素对指静脉识别装置的识别成功率的影响。

Description

一种指静脉识别装置的校准方法 技术领域
本发明涉及图像采集领域,尤其涉及一种指静脉识别装置的校准方法。
背景技术
随着时代的发展,个人信息安全越来越重要。选择合理的认证技术是保证信息安全的必要因素。传统的认证技术是基于个人密码,而密码被破解的概率越来越高。生物认证将成为今后几年信息产业的重要变革,越来越多的个人、消费者、公司乃至政府机构都承认,现有的基于智能卡、身份证号和密码的身份识别系统是远远不够的,生物特征识别技术将在未来提供安全认证方面占据重要的地位。
静脉识别技术是通过对手指或手掌中静脉图像进行活体识别来达到认证目的,具有高度防伪、活体检测、高度准确、适应性强和简便易用的特性。
指静脉识别装置涉及光源、光学镜头、光敏芯片等一系列内容,该识别装置的应用区域覆盖我国的各个地域,外界的气候和温度会影响设备近红外LED灯的发光强度,光学镜头透光率,同时外部光照会直接干扰光敏芯片接收到的光线,进而影响到采集到的图像,最终导致指静脉识别装置的识别成功率的降低。为了有效的提高指静脉识别装置的识别成功率,需要在使用前对指静脉识别装置进行相应校准。
发明内容
本发明提出了一种指静脉识别装置的校准方法,能够有效的降低不同地域气候、温度和光照等因素对指静脉识别装置的识别成功率的影响。
为实现上述目的,本发明实施例提出了一种指静脉识别装置的校准方法,包括步骤:
S1、利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
S2、计算采集到的所述N*L个指静脉图像的截面像素灰度值的高斯分布参数(μ,σ);其中,μ为均值,σ为标准差,每一所述指静脉图像的截面像素灰度值符合高斯分布;
S3、通过以下公式计算所述红外图像传感器的伽马映射系数:
Figure PCTCN2017087841-appb-000001
其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μ,σ)为上述步骤得到的高斯分布参数,sign为符号函数,γ为所述红外图像传感器的伽马映射系数;
S4、根据计算得到的所述伽马映射系数γ,调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
本发明提供的指静脉识别装置的识别算法是直接在采集到的图像上进行识别,采集到的图像的一致性会直接影响到装置识别成功率,通过上述步骤提高了不同设备在不同环境下采集图像的一致性,能有效的减小图像采集一致性对指静脉识别装置的识别算法的影响,提高识别成功率。
作为上述方案的改进,通过以下步骤计算采集到的所述N*L个指静脉图像的截面像素灰度值的高斯分布参数(μ,σ);
S21、对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
S22、对每一所述指静脉线条取等间距的T个截面,共得到B=N*L*T个截面;取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000002
其中
Figure PCTCN2017087841-appb-000003
为第b个截面的第i个像素的灰度值,Mb为第b个截面所对应的像素个数,T≥3;
S23、通过以下公式计算得到所述B个截面像素灰度值的高斯分布参数(μ,σ):
Figure PCTCN2017087841-appb-000004
Figure PCTCN2017087841-appb-000005
作为上述方案的改进,通过公式
Figure PCTCN2017087841-appb-000006
调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系。
根据实验数据统计可知静脉图像中,每根静脉的截面像素灰度基本符合高斯分布,通过上述步骤对每根指静脉线条取多个截面丰富了统计数据,使统计结果更精确。
本发明还提出了另一种指静脉识别装置的校准方法,包括步骤:
S1、利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
S2、分别计算采集到的每一所述指静脉图像的截面像素灰度值的高斯分布参数;其中,每一所述指静脉图像的截面像素灰度值符合高斯分布,所述高斯分布参数包括像素灰度均值和标准差;
S3、根据每一所述指静脉图像的截面像素灰度值的高斯分布参数计算得到每一所述指静脉图像的 宽度,然后计算所有指静脉图像的宽度的平均值hw,并选取所述N*L个指静脉图像中宽度在hw(1±A%)之间的所有指静脉图像;其中,0<A≤50,每一所述指静脉图像的宽度是指每一所述指静脉图像上的截面像素灰度值小于其对应的像素灰度均值的像素个数;
S4、计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr);
S5、通过以下公式计算所述红外图像传感器的伽马映射系数:
Figure PCTCN2017087841-appb-000007
其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μr,σr)为上述步骤得到的高斯分布参数,γ为所述红外图像传感器的伽马映射系数;
S6、根据计算得到的所述伽马映射系数γ,调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
作为上述方案的改进,可采用以下步骤和公式来计算采集到的每一所述指静脉图像的截面像素灰度值的高斯分布参数:
对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
对每一所述指静脉线条取等间距的T个截面,取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000008
其中
Figure PCTCN2017087841-appb-000009
为第t个截面的第i个像素的灰度值,Mt为第t个截面所对应的像素个数,T≥3;
Figure PCTCN2017087841-appb-000010
Figure PCTCN2017087841-appb-000011
作为上述方案的改进,可采用以下公式来计算符合指静脉图像宽度要求的所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr):
μr=(μ123+...+μp)/P
Figure PCTCN2017087841-appb-000012
其中,P为所选取的所有指静脉图像的总个数,μ1,μ2,μ3,...,μp分别为所选取的每一指静脉图像的截面像素灰度值的像素灰度均值。
作为上述方案的改进,还可采用以下公式来计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr):
Figure PCTCN2017087841-appb-000013
Figure PCTCN2017087841-appb-000014
对每一所述指静脉线条取等间距的T2个截面,共得到B2=P*T2个截面;取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000015
其中
Figure PCTCN2017087841-appb-000016
为第b2个截面的第i个像素的灰度值,Mb2为第b2个截面所对应的像素个数,P为所选取的所有指静脉图像的总个数,T2≥3。
作为上述方案的改进,采用NIBLACK图像二值化方法对每一所述指静脉图像进行二值化处理,从而提取指静脉区域;对提取的所述指静脉区域进行分割后,进行骨架提取,得到指静脉线条。
作为上述方案的改进,A=10。
作为上述方案的改进,通过公式
Figure PCTCN2017087841-appb-000017
调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系。
综上所述,本发明所述指静脉识别装置的校准方法,根据指静脉采集图像的静脉区域的灰度变化,得到该指静脉识别装置对应的红外图像传感器的调整参数,可有效的降低不同地域气候、温度和光照等因素的影响。
附图说明
图1是本发明实施例一中一种指静脉识别装置的校准方法的流程示意图。
图2是图1中步骤S102的具体流程示意图。
图3是本发明实施例一中静脉提取示意图。
图4是本发明实施例一中静脉截面示意图。
图5是本发明实施例二中一种指静脉识别装置的校准方法的流程示意图。
图6是图5中步骤S202的具体流程示意图。
图7是本发明实施例三中一种指静脉识别装置的校准方法的流程示意图。
图8是图7中步骤S302的具体流程示意图。
图9是图7中步骤S304的具体流程示意图。
具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
本发明提供一种指静脉识别装置的校准方法,用于在使用指静脉识别装置前根据当前环境对指静脉识别装置进行相应校准。如前所述,外界环境(包括外界的气候、温度和光照等因素)会影响到红外图像传感器所采集的图像,进而影响到指静脉识别装置的识别准确率。因此,本发明提供一种指静脉识别装置的校准方法,是为了提高不同的指静脉识别装置,在不同的环境下图像采集的一致性,从而减少指静脉识别装置受环境干扰的影响,从而提高指静脉识别装置的识别准确率。下面通过多个实施例对本发明的指静脉识别装置的校准方法进行具体描述。
参见图1,是本发明实施例一提供的一种指静脉识别装置的校准方法的流程示意图,本发明实施例一提供的一种指静脉识别装置的校准方法包括步骤S101-S104:
S101、利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
该步骤用于采集指静脉图像。具体的,在该步骤中,利用指静脉识别装置的红外图像传感器采用预设强度的指静脉红外光源照射N个采集者的手指,采集当前状态(在同一区域环境)下的N个人各L幅指静脉图像,从而得到N*L个指静脉图像,其中N≥2,L≥1。
S102:计算采集到的所述N*L个指静脉图像的截面像素灰度值的高斯分布参数(μ,σ);其中,每一所述指静脉图像的截面像素灰度值符合高斯分布。
该步骤用于对采集到的N*L个指静脉图像的灰度值进行统计。具体的,参考图2,该步骤可以通过以下步骤实现,包括步骤S1021-S1023,其中:
S1021:对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
参见图3,是本发明实施例一中静脉提取示意图,从静脉图像31中提取到静脉线条32,具体的,采用NIBLACK图像二值化方法对每一所述指静脉图像进行二值化处理,从而提取指静脉区域。对提取的所述指静脉区域进行分割后,进行骨架提取,从而得到指静脉线条。
S1022:对每一所述指静脉线条取等间距的T个截面,共得到B=N*L*T个截面;取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000018
其中
Figure PCTCN2017087841-appb-000019
为第b个截面的第i个像素的灰度值,Mb为第b 个截面所对应的像素个数,T≥3;其中,每一所述指静脉图像的截面像素灰度值符合高斯分布;
参见图4,是本发明实施例一中静脉截面示意图,具体的,对提取到的静脉线条40取等间距的T个截面,得到静脉截面41、静脉截面42、静脉截面43。
S1023:通过以下公式计算得到所述B个截面像素灰度值的高斯分布参数(μ,σ):
Figure PCTCN2017087841-appb-000020
Figure PCTCN2017087841-appb-000021
S103:通过公式:
Figure PCTCN2017087841-appb-000022
计算所述红外图像传感器的伽马映射系数,其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μ,σ)为上述步骤得到的高斯分布参数,γ为所述红外图像传感器的伽马映射系数。
本实施例通过伽马映射系数GAMMA对红外图像传感器进行校正的的作用是为了使得不同设备、在不同温度、区域等情况下,在采集同一个手指时,输出的图像尽可能一致,从而保证设备识别效果。
根据计算得到的所述伽马映射系数γ,通过公式
Figure PCTCN2017087841-appb-000023
调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
在该步骤中,根据步骤S103计算得到的伽马映射系数γ,即可通过预设的调整公式
Figure PCTCN2017087841-appb-000024
对指静脉识别装置通过红外图像传感器获取的当前状态下的图像像素进行校准,并将经过像素校准后的图像才进行指静脉识别,以得到识别结果。其中,识别所采用的算法过程可采用本领域技术人员公知的方式,在此省略描述。
如上所述实施例一,使用指静脉识别装置的红外图像传感器采集多人、每人至少一幅指静脉图像,并对每根指静脉线条取多个截面,丰富了统计数据,使统计结果更精确,同时,通过实施例一中伽马映射系数对红外图像传感器的校正,使得不同指静脉识别装置、在不同温度、区域等情况下,在采集同一个手指时,输出的图像尽可能一致,提高了指静脉识别装置的识别成功率。
参见图5,是本发明实施例二提供的一种指静脉识别装置的校准方法的流程示意图,本发明实施例二提供的一种指静脉识别装置的校准方法包括步骤S201-S206:
S201:利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
该步骤用于采集指静脉图像。具体的,在该步骤中,利用指静脉识别装置的红外图像传感器采用预设强度的指静脉红外光源照射N个采集者的手指,采集当前状态(在同一区域环境)下的N个人各L幅指静脉图像,从而得到N*L个指静脉图像,其中N≥2,L≥1。
S202:分别计算采集到的每一所述指静脉图像的截面像素灰度值的高斯分布参数;其中,每一所述指静脉图像的截面像素灰度值符合高斯分布,所述高斯分布参数包括像素灰度均值和标准差;
该步骤用于对采集到的N*L个指静脉图像的灰度值进行统计。具体的,参见图4,该步骤可以通过以下步骤实现,包括步骤S2021-S2023,其中:
S2021:对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
具体的,采用NIBLACK图像二值化方法对每一所述指静脉图像进行二值化处理,从而提取指静脉区域。对提取的所述指静脉区域进行分割后,进行骨架提取,从而得到指静脉线条。
S2022:对每一所述指静脉线条取等间距的T个截面,取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000025
其中
Figure PCTCN2017087841-appb-000026
为第t个截面的第i个像素的灰度值,Mt为第t个截面所对应的像素个数,T≥3;其中,每一所述指静脉图像的截面像素灰度值符合高斯分布。
S2023:通过以下公式计算得到每一所述指静脉图像的截面像素灰度值的高斯分布参数(μ,σ):
Figure PCTCN2017087841-appb-000027
Figure PCTCN2017087841-appb-000028
S203:根据每一所述指静脉图像的截面像素灰度值的高斯分布参数计算得到每一所述指静脉图像的宽度,然后计算所有指静脉图像的宽度的平均值hw,并选取所述N*L个指静脉图像中宽度在hw(1±A%)之间的所有指静脉图像;其中,0<A≤50,每一所述指静脉图像的宽度是指每一所述指静脉图像上的截面像素灰度值小于其对应的像素灰度均值的像素个数;
该步骤用于对采集到的指静脉图像进行进一步的选取,选取宽度符合要求的指静脉图像。具体的, 在该步骤中,首先计算出每一个指静脉图像的宽度,然后求出宽度平均值hw,本实施例中,A=10,即选取宽度为平均宽度的0.9倍至1.1倍的指静脉图像。
S204:计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr);
该步骤用于对选取的指静脉图像的灰度值进行统计。具体的,在该步骤中使用以下公式计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr):
μr=(μ123+...+μp)/P
Figure PCTCN2017087841-appb-000029
其中,P为所选取的所有指静脉图像的总个数,μ1,μ2,μ3,...,μp分别为所选取的每一指静脉图像的截面像素灰度值的像素灰度均值。
S205:通过公式:
Figure PCTCN2017087841-appb-000030
计算所述红外图像传感器的伽马映射系数,其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μr,σr)为上述步骤S204计算得到的选取的指静脉图像截面像素灰度值的高斯分布参数,γ为所述红外图像传感器的伽马映射系数。
S206:根据计算得到的所述伽马映射系数γ,通过公式
Figure PCTCN2017087841-appb-000031
调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
在该步骤中,根据步骤S205计算得到的伽马映射系数γ,即可通过预设的调整公式
Figure PCTCN2017087841-appb-000032
对指静脉识别装置通过红外图像传感器获取的当前状态下的图像像素进行校准,并将经过像素校准后的图像才进行指静脉识别,以得到识别结果。
如上所述实施例二,不同于实施例一的是在计算指静脉图像的截面像素灰度值的高斯分布参数时对指静脉图像进行了选取,在采集到的所有指静脉图像中优选了使实验结果更稳定的指静脉图像的宽度为平均宽度的0.9倍至1.1倍的指静脉图像,根据选取的指静脉图像的截面像素灰度值的高斯分布参数得到伽马映射系数,使用该伽马映射系数校准后的指静脉识别装置的识别成功率更高。
参见图7,是本发明实施例三提供的一种指静脉识别装置的校准方法的流程示意图,本发明实施例三提供的一种指静脉识别装置的校准方法包括步骤S301-S306:
S301:利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
该步骤用于采集指静脉图像。具体的,在该步骤中,利用指静脉识别装置的红外图像传感器采用预设强度的指静脉红外光源照射N个采集者的手指,采集当前状态(在同一区域环境)下的N个人各L幅指静脉图像,从而得到N*L个指静脉图像,其中N≥2,L≥1。
S302:分别计算采集到的每一所述指静脉图像的截面像素灰度值的高斯分布参数,其中,每一所述指静脉图像的截面像素灰度值符合高斯分布,所述高斯分布参数包括像素灰度均值和标准差;
该步骤用于对采集到的N*L个指静脉图像的灰度值进行统计。具体的,参见图6,该步骤可以通过以下步骤实现,包括步骤S3021-S3023,其中:
S3021:对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
具体的,采用NIBLACK图像二值化方法对每一所述指静脉图像进行二值化处理,从而提取指静脉区域。对提取的所述指静脉区域进行分割后,进行骨架提取,从而得到指静脉线条。
S3022:对每一所述指静脉线条取等间距的T个截面,取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000033
其中
Figure PCTCN2017087841-appb-000034
为第t个截面的第i个像素的灰度值,Mt为第t个截面所对应的像素个数,T≥3,其中,每一所述指静脉图像的截面像素灰度值符合高斯分布。
S3023:通过以下公式计算得到每一所述指静脉图像的截面像素灰度值的高斯分布参数(μ,σ):
Figure PCTCN2017087841-appb-000035
Figure PCTCN2017087841-appb-000036
S303:根据每一所述指静脉图像的截面像素灰度值的高斯分布参数计算得到每一所述指静脉图像的宽度,然后计算所有指静脉图像的宽度的平均值hw,并选取所述N*L个指静脉图像中宽度在hw(1±A%)之间的所有指静脉图像;其中,0<A≤50,每一所述指静脉图像的宽度是指每一所述指静脉图像上的截面像素灰度值小于其对应的像素灰度均值的像素个数。
该步骤用于对采集到的指静脉图像进行进一步的选取,选取宽度符合要求的指静脉图像。具体的,在该步骤中,首先计算出每一个指静脉图像的宽度,然后求出宽度平均值hw,本实施例中,A=10,即选取宽度为平均宽度的0.9倍至1.1倍的指静脉图像。
S304:计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr);
该步骤用于对选取的指静脉图像的灰度值进行统计。具体的,参见图7,该步骤可以通过以下步 骤实现,包括步骤S3041-S3042,其中:
S3041:对每一所述指静脉线条取等间距的T2个截面,共得到B2=P*T2个截面;取每个截面所有像素的灰度值,记为
Figure PCTCN2017087841-appb-000037
其中
Figure PCTCN2017087841-appb-000038
为第b2个截面的第i个像素的灰度值,Mb2为第b2个截面所对应的像素个数,P为所选取的所有指静脉图像的总个数,T2≥3。
S3042:通过以下公式计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr)
Figure PCTCN2017087841-appb-000039
Figure PCTCN2017087841-appb-000040
S305:通过公式:
Figure PCTCN2017087841-appb-000041
计算所述红外图像传感器的伽马映射系数,其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μr,σr)为上述步骤S304计算得到的选取的指静脉图像的截面像素灰度值高斯分布参数,γ为所述红外图像传感器的伽马映射系数。
S306:根据计算得到的所述伽马映射系数γ,通过公式
Figure PCTCN2017087841-appb-000042
调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
在该步骤中,根据步骤S305计算得到的伽马映射系数γ,即可通过预设的调整公式
Figure PCTCN2017087841-appb-000043
对指静脉识别装置通过红外图像传感器获取的当前状态下的图像像素进行校准,并将经过像素校准后的图像才进行指静脉识别,以得到识别结果。
如上所述实施例三,不同于实施例一的是在计算指静脉图像的截面像素灰度值的高斯分布参数时对指静脉图像进行了选取,在采集到的所有指静脉图像中优选了使实验结果更稳定的指静脉图像的宽度为平均宽度的0.9倍至1.1倍的指静脉图像,根据选取的指静脉图像的截面像素灰度值的高斯分布参数得到伽马映射系数,使用该伽马映射系数校准后的指静脉识别装置的识别成功率更高。与实施例二的区别在于计算所选取的指静脉图像的截面像素灰度值的高斯分布参数时方法有所不同,实施例二是先求出每一幅选取的指静脉图像的界面像素灰度值高斯分布参数,再计算所有指静脉图像的高斯分布参数,而实施例三则是直接求出所选取的所有指静脉截面像素灰度值的高斯分布参数。
以上所述是本发明的优选实施方式,应当指出,对于本技术领域的普通技术人员来说,在不脱离 本发明原理的前提下,还可以做出若干改进和润饰,这些改进和润饰也视为本发明的保护范围。

Claims (10)

  1. 一种指静脉识别装置的校准方法,其特征在于,包括步骤:
    S1、利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
    S2、计算采集到的所述N*L个指静脉图像的截面像素灰度值的高斯分布参数(μ,σ);其中,μ为均值,σ为标准差,每一所述指静脉图像的截面像素灰度值符合高斯分布;
    S3、通过以下公式计算所述红外图像传感器的伽马映射系数:
    Figure PCTCN2017087841-appb-100001
    其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μ,σ)为上述步骤得到的高斯分布参数,γ为所述红外图像传感器的伽马映射系数;
    S4、根据计算得到的所述伽马映射系数γ,调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
  2. 如权利要求1所述的指静脉识别装置的校准方法,其特征在于,所述步骤S2具体包括步骤:
    S21、对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
    S22、对每一所述指静脉线条取等间距的T个截面,共得到B=N*L*T个截面;取每个截面所有像素的灰度值,记为
    Figure PCTCN2017087841-appb-100002
    其中
    Figure PCTCN2017087841-appb-100003
    为第b个截面的第i个像素的灰度值,Mb为第b个截面所对应的像素个数,T≥3;
    S23、通过以下公式计算得到所述B个截面像素灰度值的高斯分布参数(μ,σ):
    Figure PCTCN2017087841-appb-100004
    Figure PCTCN2017087841-appb-100005
  3. 如权利要求1或2所述的指静脉识别装置的校准方法,其特征在于,在步骤S4中,通过公式
    Figure PCTCN2017087841-appb-100006
    调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系。
  4. 一种指静脉识别装置的校准方法,其特征在于,包括步骤:
    S1、利用指静脉识别装置的红外图像传感器采集当前状态下的N个人各L幅指静脉图像,其中N≥2,L≥1;
    S2、分别计算采集到的每一所述指静脉图像的截面像素灰度值的高斯分布参数;其中,每一所述指静脉图像的截面像素灰度值符合高斯分布,所述高斯分布参数包括像素灰度均值和标准差;
    S3、根据每一所述指静脉图像的截面像素灰度值的高斯分布参数计算得到每一所述指静脉图像的宽度,然后计算所有指静脉图像的宽度的平均值hw,并选取所述N*L个指静脉图像中宽度在hw(1±A%)之间的所有指静脉图像;其中,0<A≤50,每一所述指静脉图像的宽度是指每一所述指静脉图像上的截面像素灰度值小于其对应的像素灰度均值的像素个数;
    S4、计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr);
    S5、通过以下公式计算所述红外图像传感器的伽马映射系数:
    Figure PCTCN2017087841-appb-100007
    其中,(μs,σs)、γs为所述指静脉识别装置预设的标准值,(μr,σr)为上述步骤得到的高斯分布参数,γ为所述红外图像传感器的伽马映射系数;
    S6、根据计算得到的所述伽马映射系数γ,调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系,然后通过所述指静脉识别装置对调整后的像素输出值f1进行识别以获取识别结果。
  5. 如权利要求4所述的指静脉识别装置的校准方法,其特征在于,所述步骤S2具体包括步骤:
    S21、对每一所述指静脉图像进行二值化以划分出指静脉区域,基于所述指静脉区域得到指静脉线条;
    S22、对每一所述指静脉线条取等间距的T个截面,取每个截面所有像素的灰度值,记为
    Figure PCTCN2017087841-appb-100008
    其中
    Figure PCTCN2017087841-appb-100009
    为第t个截面的第i个像素的灰度值,Mt为第t个截面所对应的像素个数,T≥3;
    S23、通过以下公式计算得到每一所述指静脉图像的截面像素灰度值的高斯分布参数(μ,σ):
    Figure PCTCN2017087841-appb-100010
    Figure PCTCN2017087841-appb-100011
  6. 如权利要求5所述的指静脉识别装置的校准方法,其特征在于,在步骤S4中,通过以下公式计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr):
    μr=(μ123+...+μp)/P
    Figure PCTCN2017087841-appb-100012
    其中,P为所选取的所有指静脉图像的总个数,μ1,μ2,μ3,...,μp分别为所选取的每一指静脉图像的截面像素灰度值的像素灰度均值。
  7. 如权利要求5所述的指静脉识别装置的校准方法,其特征在于,在步骤S4中,通过以下步骤计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr):
    S41、对每一所述指静脉线条取等间距的T2个截面,共得到B2=P*T2个截面;取每个截面所有像素的灰度值,记为
    Figure PCTCN2017087841-appb-100013
    其中
    Figure PCTCN2017087841-appb-100014
    为第b2个截面的第i个像素的灰度值,Mb2为第b2 个截面所对应的像素个数,P为所选取的所有指静脉图像的总个数,T2≥3;
    S42、通过以下公式计算所选取的所有指静脉图像的截面像素灰度值的高斯分布参数(μr,σr):
    Figure PCTCN2017087841-appb-100015
    Figure PCTCN2017087841-appb-100016
  8. 如权利要求5所述的指静脉识别装置的校准方法,其特征在于,所述步骤S21具体包括:
    S211、采用NIBLACK图像二值化方法对每一所述指静脉图像进行二值化处理,从而提取指静脉区域;
    S212、对提取的所述指静脉区域进行分割后,进行骨架提取,得到指静脉线条。
  9. 如权利要求4或5所述的指静脉识别装置的校准方法,其特征在于,在步骤S3中,A=10。
  10. 如权利要求4或5所述的指静脉识别装置的校准方法,其特征在于,在步骤S6中,通过公式
    Figure PCTCN2017087841-appb-100017
    调整所述红外图像传感器获取的当前状态下的像素输入值f和像素输出值f1映射关系。
PCT/CN2017/087841 2016-09-13 2017-06-10 一种指静脉识别装置的校准方法 Ceased WO2018049858A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610820854.3 2016-09-13
CN201610820854.3A CN106446813B (zh) 2016-09-13 2016-09-13 一种指静脉识别装置的校准方法

Publications (1)

Publication Number Publication Date
WO2018049858A1 true WO2018049858A1 (zh) 2018-03-22

Family

ID=58168857

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/087841 Ceased WO2018049858A1 (zh) 2016-09-13 2017-06-10 一种指静脉识别装置的校准方法

Country Status (2)

Country Link
CN (1) CN106446813B (zh)
WO (1) WO2018049858A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110826513A (zh) * 2019-11-13 2020-02-21 圣点世纪科技股份有限公司 一种指静脉设备一致性的校准算法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106446813B (zh) * 2016-09-13 2019-10-11 广州广电运通金融电子股份有限公司 一种指静脉识别装置的校准方法
CN110188675B (zh) * 2019-05-29 2021-04-13 Oppo广东移动通信有限公司 静脉采集方法及相关产品
CN117631717B (zh) * 2024-01-10 2024-06-28 深圳豪达尔机械有限公司 变频冷干机传感器控制调节系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6763127B1 (en) * 2000-10-06 2004-07-13 Ic Media Corporation Apparatus and method for fingerprint recognition system
CN101331515A (zh) * 2005-12-21 2008-12-24 日本电气株式会社 色调校正方法、色调校正装置、色调校正程序以及图像设备
CN101650439A (zh) * 2009-08-28 2010-02-17 西安电子科技大学 基于差异边缘和联合概率一致性的遥感图像变化检测方法
JP2011091595A (ja) * 2009-10-22 2011-05-06 Kyocera Mita Corp 画像処理装置、画像処理方法及び画像形成装置
CN104794689A (zh) * 2015-03-12 2015-07-22 哈尔滨工程大学 一种增强声呐图像对比度的预处理方法
CN106446813A (zh) * 2016-09-13 2017-02-22 广州广电运通金融电子股份有限公司 一种指静脉识别装置的校准方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4341597B2 (ja) * 2005-08-25 2009-10-07 セイコーエプソン株式会社 ガンマカーブ調整装置及び調整ポイント設定方法
CN101982826B (zh) * 2010-11-10 2013-03-06 中国船舶重工集团公司第七一○研究所 一种光源亮度自动调整的手指静脉采集识别方法
CN103996209B (zh) * 2014-05-21 2017-01-11 北京航空航天大学 一种基于显著性区域检测的红外舰船目标分割方法
CN104688184B (zh) * 2014-12-05 2017-08-04 南京航空航天大学 可见光皮肤图像的静脉显像方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6763127B1 (en) * 2000-10-06 2004-07-13 Ic Media Corporation Apparatus and method for fingerprint recognition system
CN101331515A (zh) * 2005-12-21 2008-12-24 日本电气株式会社 色调校正方法、色调校正装置、色调校正程序以及图像设备
CN101650439A (zh) * 2009-08-28 2010-02-17 西安电子科技大学 基于差异边缘和联合概率一致性的遥感图像变化检测方法
JP2011091595A (ja) * 2009-10-22 2011-05-06 Kyocera Mita Corp 画像処理装置、画像処理方法及び画像形成装置
CN104794689A (zh) * 2015-03-12 2015-07-22 哈尔滨工程大学 一种增强声呐图像对比度的预处理方法
CN106446813A (zh) * 2016-09-13 2017-02-22 广州广电运通金融电子股份有限公司 一种指静脉识别装置的校准方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110826513A (zh) * 2019-11-13 2020-02-21 圣点世纪科技股份有限公司 一种指静脉设备一致性的校准算法
CN110826513B (zh) * 2019-11-13 2022-04-19 圣点世纪科技股份有限公司 一种指静脉设备一致性的校准方法

Also Published As

Publication number Publication date
CN106446813A (zh) 2017-02-22
CN106446813B (zh) 2019-10-11

Similar Documents

Publication Publication Date Title
Miao et al. A semi-automatic method for road centerline extraction from VHR images
CN111626123B (zh) 视频数据处理方法、装置、计算机设备及存储介质
CN112991345B (zh) 图像真伪检测方法、装置、计算机设备和存储介质
Shen et al. An automatic diagnosis method of facial acne vulgaris based on convolutional neural network
US20120099792A1 (en) Adaptive optical character recognition on a document with distorted characters
WO2019109526A1 (zh) 人脸图像的年龄识别方法、装置及存储介质
CN112507922A (zh) 人脸活体检测方法、装置、电子设备及存储介质
CN111639629B (zh) 一种基于图像处理的猪只体重测量方法、装置及存储介质
WO2017031716A1 (zh) 自然场景图像中手写体数学公式结构分析与识别方法
WO2018049858A1 (zh) 一种指静脉识别装置的校准方法
CN111260645A (zh) 基于分块分类深度学习的篡改图像检测方法及系统
CN114677333B (zh) 一种基于直方图的图像对比度增强检测方法
CN110647889A (zh) 医学图像识别方法、医学图像识别装置、终端设备及介质
Sharan et al. Classification of Medicinal Leaf by Using Canny Edge Detection and SVM Classifier
WO2018107574A1 (zh) 一种对印防伪特征的检测方法及装置
Rosebrock et al. Quantitative analysis of TDLUs using adaptive morphological shape techniques
CN108416357A (zh) 一种彩绘类文物隐含信息的提取方法
CN108399335A (zh) 一种基于局部熵的恶意代码可视化分析方法
CN117132484A (zh) 图像矫正及表计图像识别方法、装置、设备及介质
Liu et al. Latent fingerprint segmentation based on ridge density and orientation consistency
CN103617423B (zh) 基于色彩参数的图像拆分识别方法
CN114882360A (zh) 一种基于PCANet的高光谱图像分类方法及系统
Schumacher Fingerprint recognition for children
CN112329791A (zh) 一种高光谱影像水域自动提取方法
CN115759964A (zh) 一种投标文件通用资格验证的方法、系统及设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17850063

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 29/07/2019)

122 Ep: pct application non-entry in european phase

Ref document number: 17850063

Country of ref document: EP

Kind code of ref document: A1