WO2012083860A1 - 一种人脸识别装置和方法 - Google Patents

一种人脸识别装置和方法 Download PDF

Info

Publication number
WO2012083860A1
WO2012083860A1 PCT/CN2011/084408 CN2011084408W WO2012083860A1 WO 2012083860 A1 WO2012083860 A1 WO 2012083860A1 CN 2011084408 W CN2011084408 W CN 2011084408W WO 2012083860 A1 WO2012083860 A1 WO 2012083860A1
Authority
WO
WIPO (PCT)
Prior art keywords
face
image
camera
face recognition
image information
Prior art date
Application number
PCT/CN2011/084408
Other languages
English (en)
French (fr)
Inventor
曾建军
信强
宋柏君
Original Assignee
汉王科技股份有限公司
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 汉王科技股份有限公司 filed Critical 汉王科技股份有限公司
Publication of WO2012083860A1 publication Critical patent/WO2012083860A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR 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
    • G06V40/16Human faces, e.g. facial parts, sketches or expressions
    • G06V40/161Detection; Localisation; Normalisation
    • G06V40/166Detection; Localisation; Normalisation using acquisition arrangements

Definitions

  • the present disclosure relates to the field of biometric identification, and more particularly to a face recognition apparatus and method for preventing light interference. Background technique
  • Face recognition is a biometric recognition technology based on techniques such as computer, image processing and pattern recognition.
  • face recognition technology has received more and more attention as face recognition has been widely used by commercial and law enforcement agencies in areas such as form identification, credit card identification, security systems, and on-site monitoring.
  • the existing face recognition technology is generally based on a face image, that is, a face image is captured by a camera, and then face detection, human eye positioning, and feature extraction are performed. The obtained image is then compared with the template image in the template library to make a recognition determination.
  • a face image that is, a face image is captured by a camera
  • face detection, human eye positioning, and feature extraction are performed.
  • the obtained image is then compared with the template image in the template library to make a recognition determination.
  • ZL200820109477.3 discloses "a face recognition device", comprising a face recognition processing unit, a display unit, a first camera connected to the face recognition processing unit, and a second camera connected to the display unit And selecting, by the first camera and the second camera, the candidate image whose similarity degree is in accordance with the rule as the final recognition result.
  • Chinese patent ZL200820146792.3 discloses "a face recognition attendance machine” comprising a base and a panel, and a main PCB board on the base, the main PCB board being provided with a high-pixel CMOS infrared camera. The high-pixel CMOS infrared camera mode can achieve fast face recognition and low false positive rate.
  • the face recognition technology described in the above scheme can increase the accuracy of face recognition and reduce the false positive rate, but does not design from the perspective of resisting natural light interference. Therefore, in the case of strong solar rays, It will cause great interference to the captured image, and the method of increasing the resolution of the camera and the number of cameras can not achieve the purpose of resisting strong light interference. Therefore, how to accurately collect effective image information and reduce the influence of light interference on face recognition has become an urgent problem to be solved. Summary of the invention
  • the present disclosure provides a face recognition device and method for accurately collecting effective image information and reducing The effect of light interference on face recognition.
  • the face recognition device provided by the present disclosure includes:
  • a modulated light source for emitting light of different wavelength bands
  • At least one camera for collecting corresponding face image information for the different bands; and a processing unit configured to perform image processing on the face image information collected by the at least one camera, and obtain a face recognition result.
  • the face recognition device may further include at least one filter device disposed at the front end of the at least one camera for filtering out light outside the different wavelength bands, wherein: the at least one camera is collected by a person The light reflected from the modulated light source reflected by the face.
  • processing unit may be further configured to control the modulated light source to emit the light of the different wavelength bands in time series during an image acquisition period.
  • processing unit may be further configured to perform image fusion on the corresponding face image information collected by the at least one camera for the different bands for each face in the input stage of the face information.
  • a reference image is obtained and the reference image is stored.
  • the processing unit performing image processing on the face image information collected by the at least one camera and obtaining the face recognition result may include: the processing unit collecting the at least one camera for the different bands in the field The obtained face image information is image-fused, and the obtained fused image is compared with all the reference images stored in advance, and a reference image closest to the fused image is used as a face recognition result; and/or Comparing the face images corresponding to the face image information collected by the at least one camera to the different bands in real time with all the reference images stored in advance, and selecting the closest pair of face images and the reference images The reference image is used as the result of face recognition.
  • the modulation light source may be a modulation laser.
  • the at least one camera may be a CCD infrared camera or a CMOS infrared camera.
  • the front end of the modulating light source may be provided with light adjusting means for converting light of different wavelength bands emitted by the modulating light source into uniform light.
  • the present disclosure also provides a face recognition method, including the following steps:
  • Image processing is performed on the collected face image information, and a face recognition result is obtained. Further, the method may further include: performing image fusion on the face image information collected by the at least one camera for the different bands for each face, in the input stage of the face information, to obtain the reference image and The reference image is stored.
  • performing image processing on the collected face image information and obtaining the face recognition result may include: performing image fusion on the face image information collected by the at least one camera for different bands in the field, and obtaining The fused image is compared with all the pre-stored reference images, and a reference image closest to the fused image is used as a face recognition result; and/or, the face image information collected by the camera for different bands is collected.
  • the corresponding face images are respectively compared with all the reference images stored in advance, and the one of the closest pair of face images and the reference image is used as the face recognition result.
  • the modulating light source may be infrared light in a wavelength band in which the solar spectroscopy is irradiance and the spectral photon irradiance is low.
  • FIG. 1 is a schematic structural diagram of a face recognition device according to an embodiment of the present disclosure
  • FIG. 2 is a schematic structural diagram of another face recognition device according to an embodiment of the present disclosure.
  • FIG. 3 is a flow chart of a face recognition method in accordance with an embodiment of the present disclosure. detailed description
  • FIG. 1 is a schematic structural diagram of a face recognition device according to an embodiment of the present disclosure.
  • the face recognition device includes: a modulation light source 1, a camera 5, a filter device 4, and a processing unit 6.
  • the modulating light source 1 is used to emit light of different wavelength bands to illuminate the human face 3 .
  • the camera 5 is for collecting corresponding image information of the face 3 for the different bands.
  • the filter device 4 is placed at the front end of the camera 5 for filtering out light rays outside the different bands, so that the camera 5 collects only light in the different bands.
  • the processing unit 6 is electrically connected to the camera 5 and the modulation light source 1, respectively, for performing image processing on the image information of the face collected by the camera 5, and obtaining a face recognition result.
  • the processing unit 6 is further configured to control the modulated light source 1 to emit the light of the different wavelength bands in a time during an image acquisition period, so that the camera head acquires corresponding face image information for the different wavelength bands.
  • the face recognition device of the present embodiment by collecting face image information of different bands, Reduce the impact of light interference on face recognition.
  • the processing unit 6 is further configured to perform image fusion on the face image information collected by the camera 5 in the different wavelength bands for each face to obtain a reference image and store The reference image.
  • a large number of reference images corresponding to different faces are stored in the template image library for subsequent face recognition.
  • the processing unit 6 performs image processing on the face image information collected by the camera 5, and obtains a face recognition result, including: the processing unit 6 sets the camera 5 in the different wavelength band.
  • the image information of the face image collected in the field is image-fused, and the obtained fused image is compared with all the reference images stored in advance, and a reference image closest to the fused image is used as a face recognition result; and/or And comparing the face image corresponding to the face image information collected by the camera 5 in the different wavelength bands to all the reference images stored in advance, and the closest pair of face images and the reference image That reference image is used as the result of face recognition.
  • FIG. 2 is a schematic structural diagram of a face recognition device according to an embodiment of the present disclosure.
  • the face recognition device of the present embodiment is different from the face recognition device shown in Fig. 1 in that the number of cameras is increased, and two cameras are provided in this embodiment.
  • the present disclosure does not limit the number of cameras.
  • the face recognition device includes a first camera 5A and a second camera 5B disposed at different positions, and collects corresponding face image information for different bands.
  • Two filter units 4A and 4B which are the same as or similar to those in the embodiment described in Fig. 1 are provided at the front ends of the two cameras, respectively.
  • the first camera 5A and the second camera 5B may be spaced apart in the horizontal direction by a certain distance, or may be spaced apart by a certain distance in the vertical direction, or may be placed according to factors such as placement conditions, aesthetics, concealability, and the like. In other locations that facilitate image acquisition.
  • the first camera 5A and the second camera 5B can collect corresponding face image information in different bands, and transmit the collected image information to the processing module 6 for image processing. In this way, the ability of the face recognition device to adapt to light, posture, and expression can be enhanced.
  • infrared light in a wavelength band in which the solar spectrum irradiance and the spectral photon irradiance are low is taken as Modulate the light source.
  • the solar spectroscopy irradiance at a wavelength of about 825 nm is 785.0 (W*m - 2 * ⁇ m" 1 ), and the spectral photon irradiance is 3.255E+017 (cm- 2 *s -1 * ⁇ m" 1 ) ;
  • the spectral irradiance of the solar light with a wavelength of about 900 nm is 745 (W*m" 2 * ⁇ m" 1 ), and the spectral photon irradiance is S ⁇ lOE+OnCcm ⁇ s ⁇ y! ⁇ 1 ).
  • infrared light having a center wavelength of 825 nm or 900 nm can be selected as the modulation light source.
  • the modulated light source 1 can be a modulated laser.
  • the modulated laser has the advantages of good monochromaticity, high brightness, directivity and coherence, and has better resistance to light interference than ordinary light-emitting diodes.
  • the camera 5, the first camera 5A, and the second camera 5B may be a CCD infrared camera or a CMOS infrared camera.
  • the filter device may be a filter film or a filter such as a band pass filter or a cut filter.
  • the filter device may comprise two passbands or more passbands corresponding to a central wavelength of the modulating light source 1, such as 840 nm, 880 nm, and a passband bandwidth of up to 8 nm.
  • a light adjusting device 2 may be disposed at the front end of the modulating light source 1, as shown in Figs. 1 and 2, for converting the light emitted by the modulating light source 1 into uniform light.
  • the light adjusting device 2 may be frosted glass, and scatters the light emitted by the modulating light source 1 to form uniform light to illuminate the human face 3.
  • the present disclosure also provides a face recognition method, as shown in FIG. 3, including:
  • the processing unit controls the modulated light source to emit light of different wavelength bands, and the at least one camera collects corresponding face image information for the different wavelength bands, and the processing unit pairs
  • the collected face image information is subjected to image fusion to obtain a reference image and store the reference image.
  • the image processing of the face image information collected by the at least one camera for the different bands is performed, and a face recognition result is obtained.
  • the method includes: performing image fusion on the face image information collected by the at least one camera for the different bands, and comparing the obtained fused image with all the reference images stored in advance, and is closest to the fused image. a reference image as a face recognition result; and/or comparing the face image corresponding to the face image information collected by the at least one camera to the different bands in real time with all the reference images stored in advance , the closest pair of face images and the reference image That reference image is used as the result of face recognition.
  • the present disclosure can enhance the ability of the face recognition device to adapt to light, face gestures and expressions by acquiring facial image information under different wavelengths of light, and performing image fusion, accurately collecting effective image information, and reducing light interference on the face. The impact of recognition.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)
  • Collating Specific Patterns (AREA)
  • Studio Devices (AREA)

Description

一种人脸识别装置和方法 技术领域
本公开涉及生物特征识别领域, 尤其是指一种防光干扰的人脸识别装置和方 法。 背景技术
人脸识别是基于计算机、 图像处理及模式识别等技术的一种生物特征识别技 术。 在过去的十几年里, 随着人脸识别被商业和执法部门广泛应用于例如形式鉴 定, 信用卡识别, 安全系统, 现场监控等领域, 人脸识别技术得到了越来越多的 关注。
现有的人脸识别技术一般是基于人脸图像的, 即通过摄像头采集人脸图像, 然后进行人脸检测、 人眼定位和特征提取。 然后将所获得的图像与模板库中的模 板图像进行比对, 做出识别判别。 但是, 在可见光环境下, 尤其是当太阳光较强 的情况下, 由于环境光光谱范围很宽, 会给图像的采集造成较大干扰, 产生误识 别或者不识别的问题。中国专利 ZL200820109477.3公开了 "一种人脸识别装置", 包括人脸识别处理单元, 显示单元, 与所述人脸识别处理单元连接的第一摄像头, 与所述显示单元连接的第二摄像头; 通过第一摄像头与第二摄像头采集的人脸图 像选取相似度高度符合规则的候选图像作为最终识别结果。 中国专利 ZL200820146792.3公开了 "一种人脸识别考勤机", 包括底座和面板, 在底座上 有主 PCB板, 所述主 PCB板设有一高像素 CMOS红外摄像方式的摄像头。 高像 素 CMOS红外摄像方式可实现人脸识别速度快和误判率小的特点。
以上方案所述的人脸识别技术, 虽然可以增加人脸识别的精度、 以及减小误 判率, 但是并未从抗自然光干扰的角度去进行设计, 因而在太阳光线较强的情况 下, 仍然会对采集的图像造成较大的干扰, 而只单纯的采用增加摄像头分辨率与 摄像头的个数的方法, 均不能达到抗强光干扰的目的。 因而, 如何能够准确采集 有效的图像信息, 降低光干扰对人脸识别的影响成为一个亟待解决的问题。 发明内容
本公开提供了一种人脸识别装置和方法, 以准确采集有效的图像信息, 降低 光干扰对人脸识别的影响。
本公开提供的人脸识别装置包括:
调制光源, 用于发射不同波段的光线;
至少一个摄像头, 用于针对所述不同波段, 采集相应的人脸图像信息; 以及 处理单元, 用于对所述至少一个摄像头采集的人脸图像信息进行图像处理, 并得到人脸识别结果。
进一步地, 该人脸识别装置还可以包括至少一个滤光装置, 分别置于所述至 少一个摄像头前端, 用于滤除所述不同波段之外的光线, 其中: 所述至少一个摄 像头采集被人脸反射的从调制光源发射的光线。
进一步地, 所述处理单元, 还可以用于在一个图像采集周期内, 控制所述调 制光源按时序发射所述不同波段的光线。
进一步地, 所述处理单元, 还可以用于在人脸信息的录入阶段, 针对每一张 人脸, 对所述至少一个摄像头针对所述不同波段采集的相应的人脸图像信息进行 图像融合, 以得到基准图像并存储所述基准图像。
进一步地, 所述处理单元对所述至少一个摄像头采集的人脸图像信息进行图 像处理, 并得到人脸识别结果, 可以包括: 所述处理单元将所述至少一个摄像头 针对所述不同波段现场采集到的人脸图像信息进行图像融合, 并将得到的融合图 像与预先存储的所有基准图像进行比对, 将与所述融合图像最接近的一幅基准图 像作为人脸识别结果; 和 /或, 将所述至少一个摄像头针对所述不同波段现场采集 到的人脸图像信息所对应的人脸图像分别与预先存储的所有基准图像进行比对, 将最接近的一对人脸图像和基准图像中的那一幅基准图像作为人脸识别结果。
进一步地, 所述调制光源可以为调制激光器。
进一步地, 所述至少一个摄像头可以为 CCD红外摄像头或 CMOS红外摄像 头。
进一步地, 所述调制光源前端可以设置光线调节装置, 将所述调制光源发出 的不同波段的光线变为均匀光线。
本公开还提供了一种人脸识别方法, 包括以下步骤:
针对调制光源发射的光线的不同波段, 通过至少一个摄像头采集相应的人脸 图像信息; 以及
对采集的人脸图像信息进行图像处理, 并得到人脸识别结果。 进一步地, 该方法还可以包括: 在人脸信息的录入阶段, 针对每一张人脸, 对所述至少一个摄像头针对所述不同波段采集的人脸图像信息进行图像融合, 以 得到基准图像并存储所述基准图像。
进一步地, 所述对采集到的人脸图像信息进行图像处理, 并得到人脸识别结 果, 可以包括: 将所述至少一个摄像头针对不同波段现场采集到的人脸图像信息 进行图像融合, 将得到的融合图像与预先存储的所有基准图像进行比对, 将与所 述融合图像最接近的一幅基准图像作为人脸识别结果; 和 /或, 将摄像头针对不同 波段现场采集到的人脸图像信息所对应的人脸图像分别与预先存储的所有基准图 像进行比对, 将最接近的一对人脸图像和基准图像中的那一幅基准图像作为人脸 识别结果。
所述调制光源可以为处于太阳光光谱辐照度及光谱光子辐照度较低的波段的 红外光。 附图说明
图 1是根据本公开实施例中一种人脸识别装置的结构示意图;
图 2是根据本公开实施例中另一种人脸识别装置的结构示意图;
图 3是根据本公开的实施例中人脸识别方法的流程图。 具体实施方式
下面结合附图详细说明本公开的具体实施方式。
图 1是根据本公开的一个实施方式的人脸识别装置的结构示意图。
如图 1所示, 人脸识别装置包括: 调制光源 1、 摄像头 5、 滤光装置 4和处理 单元 6。 调制光源 1用于发射不同波段的光线, 以照射人脸 3。 摄像头 5用于针对 所述不同波段, 采集人脸 3的相应的图像信息。 滤光装置 4置于摄像头 5的前端, 用于滤除所述不同波段之外的光线, 使摄像头 5仅采集所述不同波段内的光线。 处理单元 6分别与摄像头 5及调制光源 1进行电气连接, 用于对摄像头 5采集的 人脸的图像信息进行图像处理, 并得到人脸识别结果。 处理单元 6还可用于控制 调制光源 1在一个图像采集周期内, 按时序发射所述不同波段的光线, 从而使摄 像头针对所述不同波段, 采集相应的人脸图像信息。
在本实施方式的人脸识别装置中, 通过采集不同波段的人脸图像信息, 可以 降低光干扰对人脸识别的影响。
在人脸识别装置进行人脸识别之前, 需要预先进行人脸信息的录入工作, 以 便于后续的人脸识别。 可选地, 在人脸信息的录入阶段, 处理单元 6还用于针对 每一张人脸, 对摄像头 5在所述不同波段下采集的人脸图像信息进行图像融合, 以获得基准图像并存储所述基准图像。 这样, 在模板图像库中存储了大量对应于 不同人脸的基准图像, 以便后续的人脸识别。
可选地, 在人脸识别阶段, 所述处理单元 6对摄像头 5采集的人脸图像信息 进行图像处理, 并得到人脸识别结果, 包括: 所述处理单元 6将摄像头 5在所述 不同波段下现场采集到的人脸图像信息进行图像融合, 并将得到的融合图像与预 先存储的所有基准图像进行比对, 将与融合图像最接近的一幅基准图像作为人脸 识别结果; 和 /或, 将摄像头 5在所述不同波段下现场采集到的人脸图像信息所对 应的人脸图像分别与预先存储的所有基准图像进行比对, 将最接近的一对人脸图 像和基准图像中的那一幅基准图像作为人脸识别结果。
图 2是根据本公开的一个实施方式的人脸识别装置的结构示意图。
如图 2所示, 本实施方式的人脸识别装置与图 1所示的人脸识别装置的不同 之处在于, 增加了摄像头的数量, 本实施例中设置了两个摄像头。 本公开对摄像 头数量不加以限制。 如图 2所示, 所述人脸识别装置包括设置在不同位置的第一 摄像头 5A和第二摄像头 5B, 针对不同波段采集相应的人脸图像信息。 分别在两 个摄像头前端设置与图 1所述的实施例中相同或类似的两个滤光装置 4A和 4B。
本实施例中, 第一摄像头 5A与第二摄像头 5B可以在水平方向上间隔一定距 离, 或者在垂直方向上间隔一定距离, 也可以根据放置条件、 美观、 隐蔽性等因 素, 将两个摄像头放置在便于图像采集的其他位置。 第一摄像头 5A 与第二摄像 头 5B可以在不同波段下采集相应的人脸图像信息,并将采集到的图像信息传输到 处理模块 6进行图像处理。 以此方式, 能够增强人脸识别装置对光线、 姿态和表 情的适应能力。
在图 1和图 2所示的人脸识别装置中, 为有效降低太阳光干扰对人脸识别的 影响, 将处于太阳光光谱辐照度及光谱光子辐照度较低的波段的红外光作为调制 光源。 根据标准太阳光光谱辐照度分布表, 波长为 825nm左右的太阳光光谱辐照 度为 785.0(W*m-2* μ m"1), 光谱光子辐照度为 3.255E+017(cm-2*s-1 * μ m"1) ; 波长 为 900nm 左右的太阳光光谱辐照度为 745 ( W*m"2* μ m"1) , 光谱光子辐照度为 S^lOE+OnCcm^s^ y!^1)。 以上两个波长相对于其他波长的太阳光光谱辐照度 及光谱光子辐照度较低, 因而, 可选用中心波长为 825nm或 900nm的红外光作为 调制光源。
可选地, 调制光源 1可以为调制激光器。 调制激光器具有单色性好、 亮度高、 方向性及相干性强等优点, 相比普通发光二极管具有更好的抗光干扰能力。
可选地,摄像头 5、第一摄像头 5A、第二摄像头 5B可以为 CCD红外摄像头、 或 CMOS红外摄像头。
可选地, 滤光装置可以为滤光膜或滤光片, 如: 带通滤光膜或截止滤光膜。 所述滤光装置可以包含与调制光源 1 的中心波长, 如 840nm、 880nm对应的两个 通带或更多个通带, 通带带宽可以达到 8nm。
可选地, 在调制光源 1前端可以设置光线调节装置 2, 如图 1和图 2所示, 用于将所述调制光源 1发出的光线变为均匀光线。 例如: 光线调节装置 2可以为 毛玻璃, 将调制光源 1发射的光线进行散射, 从而形成均匀光线照射在人脸 3上。
本公开还提供了一种人脸识别方法, 如图 3所示, 包括:
- 在人脸信息的录入阶段, 针对每一张人脸, 对至少一个摄像头针对光线的 不同波段采集的人脸图像信息进行图像融合, 以得到基准图像并存储所述基准图 像 (301 ) ; 和
- 在人脸识别阶段, 对所述至少一个摄像头针对所述不同波段采集到的人脸 图像信息进行图像处理, 并得到人脸识别结果 (302) 。
在人脸信息的录入阶段中, 针对每一张人脸, 处理单元控制调制光源发出不 同波段的光线, 并由所述至少一个摄像头针对所述不同波段采集相应的人脸图像 信息, 处理单元对采集到的人脸图像信息进行图像融合, 以得到基准图像并存储 所述基准图像。
在人脸识别阶段中, 在人脸识别阶段的一个图像采集周期内, 所述对所述至 少一个摄像头针对所述不同波段采集到的人脸图像信息进行图像处理, 并得到人 脸识别结果, 包括: 将所述至少一个摄像头针对所述不同波段现场采集到的人脸 图像信息进行图像融合,将得到的融合图像与预先存储的所有基准图像进行比对, 将与所述融合图像最接近的一幅基准图像作为人脸识别结果; 和 /或, 将所述至少 一个摄像头针对所述不同波段现场采集到的人脸图像信息所对应的人脸图像分别 与预先存储的所有基准图像进行比对, 将最接近的一对人脸图像和基准图像中的 那一幅基准图像作为人脸识别结果。
本公开通过在不同波段光线下采集人脸图像信息, 并进行图像融合, 能够增 强人脸识别装置对光线、人脸姿态和表情的适应能力, 准确采集有效的图像信息, 降低光干扰对人脸识别的影响。
在上述实施方式中, 为便于理解, 仅说明了本公开的一些例子。 这些实施方 式不是对本公开的限制, 通过适当修改这些实施方式, 本公开的原理和思想也可 以用等同或类似的实施方式加以实施。

Claims

权 利 要 求
1、 一种人脸识别装置, 包括:
调制光源, 用于发射不同波段的光线;
至少一个摄像头, 用于针对所述不同波段, 采集相应的人脸图像信息; 以及 处理单元, 用于对所述至少一个摄像头采集的人脸图像信息进行图像处理, 并得到人脸识别结果。
2、 根据权利要求 1所述的人脸识别装置, 还包括至少一个滤光装置, 分别置 于所述至少一个摄像头前端, 用于滤除所述不同波段之外的光线, 其中:
所述至少一个摄像头采集被人脸反射的从调制光源发射的光线。
3、 根据权利要求 1或 2所述的人脸识别装置, 其中, 所述处理单元还用于在 一个图像采集周期内, 控制所述调制光源按时序发射所述不同波段的光线。
4、 根据权利要求 3所述的人脸识别装置, 其中, 所述处理单元还用于在人脸 信息的录入阶段, 针对每一张人脸, 对所述至少一个摄像头针对所述不同波段采 集的相应人脸图像信息进行图像融合, 以得到基准图像并存储所述基准图像。
5、 根据权利要求 4所述的人脸识别装置, 其中, 所述处理单元对所述至少一 个摄像头采集的人脸图像信息进行图像处理, 并得到人脸识别结果, 包括: 所述 处理单元将所述至少一个摄像头针对所述不同波段现场采集到的人脸图像信息进 行图像融合, 并将得到的融合图像与预先存储的所有基准图像进行比对, 将与所 述融合图像最接近的一幅基准图像作为人脸识别结果; 和 /或, 将所述至少一个摄 像头针对所述不同波段现场采集到的人脸图像信息所对应的人脸图像分别与预先 存储的所有基准图像进行比对, 将最接近的一对人脸图像和基准图像中的那一幅 基准图像作为人脸识别结果。
6、根据权利要求 1所述的人脸识别装置,其中,所述调制光源为调制激光器。
7、根据权利要求 1所述的人脸识别装置, 其中, 所述至少一个摄像头为 CCD 红外摄像头或 CMOS红外摄像头。
8、 根据权利要求 1所述的人脸识别装置, 其中, 所述调制光源前端设置光线 调节装置, 将所述调制光源发出的不同波段的光线变为均匀光线。
9、 一种人脸识别方法, 包括:
针对调制光源发射的光线的不同波段, 通过至少一个摄像头采集相应的人脸 图像信息; 以及
对采集的人脸图像信息进行图像处理, 并得到人脸识别结果。
10、 根据权利要求 9所述的人脸识别方法, 还包括: 在人脸信息的录入阶段, 针对每一张人脸, 对所述至少一个摄像头针对所述不同波段采集的人脸图像信息 进行图像融合, 以得到基准图像并存储所述基准图像。
11、根据权利要求 10所述的人脸识别方法, 其中所述对采集到的人脸图像信 息进行图像处理, 并得到人脸识别结果, 包括: 将所述至少一个摄像头针对不同 波段现场采集到的人脸图像信息进行图像融合, 将得到的融合图像与预先存储的 所有基准图像进行比对, 将与所述融合图像最接近的一幅基准图像作为人脸识别 结果; 和 /或, 将所述至少一个摄像头针对不同波段现场采集到的人脸图像信息所 对应的人脸图像分别与预先存储的所有基准图像进行比对, 将最接近的一对人脸 图像和基准图像中的那一幅基准图像作为人脸识别结果。
12、 根据权利要求 9至 1 1任一项所述的人脸识别方法, 其中, 所述调制光源 为处于太阳光光谱辐照度及光谱光子辐照度较低的波段内的红外光。
PCT/CN2011/084408 2010-12-24 2011-12-22 一种人脸识别装置和方法 WO2012083860A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2010106048383A CN102567706A (zh) 2010-12-24 2010-12-24 一种人脸识别装置和方法
CN201010604838.3 2010-12-24

Publications (1)

Publication Number Publication Date
WO2012083860A1 true WO2012083860A1 (zh) 2012-06-28

Family

ID=46313178

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/084408 WO2012083860A1 (zh) 2010-12-24 2011-12-22 一种人脸识别装置和方法

Country Status (2)

Country Link
CN (1) CN102567706A (zh)
WO (1) WO2012083860A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107993339A (zh) * 2017-12-28 2018-05-04 联合华通(天津)科技有限公司 带有人脸识别功能的智能ic卡终端机

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105517680B (zh) * 2015-04-28 2020-03-10 北京旷视科技有限公司 用于识别人脸的装置、系统和方法
CN105869226A (zh) * 2016-06-02 2016-08-17 南京安智易达智能科技有限公司 基于人脸识别的监所自动点名系统及其点名方法
CN106022321A (zh) * 2016-07-18 2016-10-12 北京红马传媒文化发展有限公司 一种采集设备和基于采集设备的人脸识别设备终端
CN106446860A (zh) * 2016-10-10 2017-02-22 上海成业智能科技股份有限公司 在光扰条件下清晰采集人脸识别图像的方法
CN107038776A (zh) * 2017-03-24 2017-08-11 新疆习美人工智能科技有限公司 一种门禁系统及门禁控制方法
CN107463924A (zh) * 2017-08-31 2017-12-12 江苏青诚信息科技有限公司 一种人脸识别系统
CN110222624B (zh) * 2019-05-31 2023-04-28 深圳群安电子科技有限公司 一种动态人脸检测装置
CN111145244A (zh) * 2019-12-20 2020-05-12 万翼科技有限公司 房间面积获取方法及相关装置
CN112115925A (zh) * 2020-11-18 2020-12-22 鹏城实验室 人脸识别方法、装置及计算机可读存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1924888A (zh) * 2005-08-31 2007-03-07 上海银晨智能识别科技有限公司 一种获取人脸图像的装置及方法
CN100576231C (zh) * 2007-01-15 2009-12-30 中国科学院自动化研究所 图像采集装置及应用该装置的人脸识别系统和方法
CN101650775A (zh) * 2008-08-15 2010-02-17 深圳市飞瑞斯科技有限公司 基于均匀光源的人脸识别装置
CN101404060B (zh) * 2008-11-10 2010-06-30 北京航空航天大学 一种基于可见光与近红外Gabor信息融合的人脸识别方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1343107A3 (en) * 2002-03-04 2005-03-23 Samsung Electronics Co., Ltd. Method and apparatus for recognising faces using principal component analysis and second order independent component analysis on parts of the image faces
CN200986643Y (zh) * 2006-12-25 2007-12-05 上海非思智能科技股份有限公司 一种适应于多种光线环境的图像传感装置
CN101493884B (zh) * 2008-01-24 2012-05-23 中国科学院自动化研究所 多光谱图像采集装置和方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1924888A (zh) * 2005-08-31 2007-03-07 上海银晨智能识别科技有限公司 一种获取人脸图像的装置及方法
CN100576231C (zh) * 2007-01-15 2009-12-30 中国科学院自动化研究所 图像采集装置及应用该装置的人脸识别系统和方法
CN101650775A (zh) * 2008-08-15 2010-02-17 深圳市飞瑞斯科技有限公司 基于均匀光源的人脸识别装置
CN101404060B (zh) * 2008-11-10 2010-06-30 北京航空航天大学 一种基于可见光与近红外Gabor信息融合的人脸识别方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107993339A (zh) * 2017-12-28 2018-05-04 联合华通(天津)科技有限公司 带有人脸识别功能的智能ic卡终端机
CN107993339B (zh) * 2017-12-28 2024-01-05 联合华通(天津)科技有限公司 带有人脸识别功能的智能ic卡终端机

Also Published As

Publication number Publication date
CN102567706A (zh) 2012-07-11

Similar Documents

Publication Publication Date Title
WO2012083860A1 (zh) 一种人脸识别装置和方法
JP6461406B1 (ja) 車両検査セキュリティシステムにおける改善された顔検出および認識のための装置、システムおよび方法
CN100361135C (zh) 一种获取人脸图像的方法及人脸识别方法与系统
US8754934B2 (en) Dual-camera face recognition device and method
KR101286454B1 (ko) 눈 영상의 특성을 이용한 모조얼굴 식별장치 및 방법
KR100939856B1 (ko) 개인인증방법 및 개인인증장치
KR101537808B1 (ko) 즉석에서 손가락 지문을 인식하기 위한 장치
CN102622588A (zh) 双验证人脸防伪方法及装置
CN103106401A (zh) 具有人机交互机制的移动终端虹膜识别装置和方法
KR20110094037A (ko) 비디오 적외선 망막 이미지 스캐너
CN112232163B (zh) 指纹采集方法及装置、指纹比对方法及装置、设备
CN107239772A (zh) 掌纹掌脉图像采集装置及通关闸机
KR20090086891A (ko) 적외선을 이용한 얼굴인식 시스템 및 방법
CN104992163A (zh) 一种人脸图像与虹膜图像同时采集光学装置及其身份认证方法
CN103336950A (zh) 一种人脸识别方法及系统
Fletcher et al. Development of mobile-based hand vein biometrics for global health patient identification
JP2004030564A (ja) 個人認証方法、個人認証装置および撮影装置
CN108388858B (zh) 虹膜防伪方法与装置
CN103632135B (zh) 一种基于双图像传感器的虹膜图像捕捉装置
CN108513661A (zh) 身份鉴权方法、身份鉴权装置、和电子设备
US20170024603A1 (en) Biometric image optimization using light fields
He et al. Contactless autofeedback iris capture design
JP2008021072A (ja) 撮影システム、撮影装置およびそれを用いた照合装置ならびに撮影方法
US11341224B2 (en) Handheld multi-sensor biometric imaging device and processing pipeline
US20210192205A1 (en) Binding of selfie face image to iris images for biometric identity enrollment

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: 11850302

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11850302

Country of ref document: EP

Kind code of ref document: A1