WO2016202165A1 - 一种钞票的多光谱识别装置及方法 - Google Patents
一种钞票的多光谱识别装置及方法 Download PDFInfo
- Publication number
- WO2016202165A1 WO2016202165A1 PCT/CN2016/083909 CN2016083909W WO2016202165A1 WO 2016202165 A1 WO2016202165 A1 WO 2016202165A1 CN 2016083909 W CN2016083909 W CN 2016083909W WO 2016202165 A1 WO2016202165 A1 WO 2016202165A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- module
- banknote
- images
- spectral
- image
- 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
Links
Images
Definitions
- Embodiments of the present invention relate to a contactless optical detection technology, and in particular, to a multi-spectral identification device and method for banknotes.
- Optical color changing inks are mostly used on banknotes to make counterfeit coins easy to distinguish. From different angles, the ink will show different colors, thus ensuring that the color is not copied. China has used optical color-changing ink as the anti-counterfeiting feature for the first time in the new RMB 100 and 50 yuan.
- the prior art optical color changing ink recognition device comprises a frame, a lens, two light sources with different illumination angles on both sides of the lens, and the light color ink anti-counterfeiting mark on the surface of the real banknote when the light is irradiated at different angles, different incident angles
- the spectral information of the light reflected by the color changing ink is different.
- the electrical signals converted by different spectral information are also different.
- the reflected light is received by the contact image sensor, and the contact image sensor analyzes the received two kinds of light. Compare and then compare the obtained information with the relevant information of the real banknote to achieve the purpose of authenticity.
- a disadvantage of the prior art is that the identification device of the optical color-changing ink cannot be inferior to the image of the respective spectral segments of the banknote.
- Embodiments of the present invention provide a multi-spectral identification device for banknotes, which is used to acquire images of various spectral segments of a banknote and compare them with preset conditions to improve recognition accuracy.
- Wide field of view imaging module tunable wavelength module, photodetection module, and authentication module
- the wide field of view imaging module is configured to acquire a reflection image of a banknote to be detected
- the tunable wavelength module is configured to extract, from the reflected image, n reflective sub-images under n spectral segments, where n is a natural number greater than 2;
- the photodetection module is configured to perform photoelectric signal conversion on the n reflective sub-images, and acquire n converted images corresponding to the n reflective sub-images;
- the authentication module is configured to identify the authenticity of the banknote to be detected by using the n converted images.
- the tunable wavelength module includes an upper reflective layer, a lower reflective layer, and an adjustable component for adjusting a relative position of the upper reflective layer and the lower reflective layer.
- the multi-spectral identification device of the banknote further includes a signal generating module, and the signal generating module is configured to intermittently transmit the voltage signals S 1 , S 2 , S 3 , . . . S n to the adjustable wavelength module.
- the adjustable member is an electrostatic brake, the brake electrostatic voltage for receiving the signal S 1, S 2, S 3 , ... S n;
- the upper reflective layer is mounted on the upper dielectric plate
- the lower reflective layer is mounted on the lower dielectric plate
- the electrostatic actuator is electrically connected to the upper dielectric plate and the lower dielectric plate, respectively.
- the multi-spectral identification device of the banknote further includes an image separation module, wherein the image separation module is configured to separate n optical color-changing ink conversion images from the n converted images;
- the authentication module is configured to identify the authenticity of the banknote to be detected by the n optical color changing ink conversion images.
- the multi-spectral identification device of the banknote further includes a left light source module and a right light source module, and the left light source module and the right light source module are respectively located on the left side and the right side of the multi-spectral identification device of the banknote.
- the embodiment of the invention further provides a multi-spectral identification method for banknotes, comprising:
- the S2 specifically includes:
- the method further comprises: separating the acquired n converted images, and separating the n optical color changing ink conversion images.
- the S4 specifically includes: performing authenticity identification of the banknote to be detected by converting the image by the n optical color changing inks.
- the wide field of view imaging module receives the reflected image of the banknote to be detected, and then the tunable wavelength module extracts the n reflected sub-images of the n spectral segments from the reflected image, and then the photodetecting module performs the photoelectric signals on the n reflective sub-images. Converting and acquiring n converted images corresponding to the n reflective sub-images, and then the authentication module identifies the authenticity of the banknote to be detected through the n converted images, according to the above technical solution, since the reflected image is divided into n spectral segments For n reflection sub-images, the more the spectral segments, the finer the discrimination, so the discrimination accuracy of banknotes can be improved.
- FIG. 1 is a schematic structural view of a multi-spectral identification device for banknotes according to an embodiment of the present invention
- FIG. 2 is a control relationship diagram of a multi-spectral identification device for banknotes according to an embodiment of the present invention
- FIG. 3 is a schematic diagram showing the principle of a Fabry-Perot interferometer of a multi-spectral identification device for banknotes according to an embodiment of the present invention
- FIG. 4 is a schematic diagram of a tunable wavelength module of a multi-spectral identification device for banknotes according to an embodiment of the present invention
- FIG. 5 is a flowchart of a multi-spectral identification method for banknotes according to an embodiment of the present invention.
- the specific technical solution of the multi-spectral identification device of the banknote according to the embodiment of the present invention is described in detail below.
- the multi-spectral identification device of the banknote can be applied to different fields without specific limitation.
- a technical solution of a multi-spectral identification device for banknotes includes:
- a wide field of view imaging module 5 a tunable wavelength module 7, a photodetection module 8, an authentication module 11;
- a wide field of view imaging module 5 for acquiring a reflection image of the banknote to be detected
- the tunable wavelength module 7 is configured to extract the n reflected sub-images in the n spectral segments of the reflected image, where n is a natural number greater than 2;
- the photodetection module 8 is configured to perform photoelectric signal conversion on the n reflective sub-images, and acquire n converted images corresponding to the n reflective sub-images;
- the authentication module 11 identifies the authenticity of the banknote to be detected by the n converted images.
- the wide field of view imaging module receives the reflected image of the banknote to be detected, and then the tunable wavelength module extracts the n reflected sub-images of the n spectral segments from the reflected image, and then the photodetecting module reflects n The sub-image performs photoelectric signal conversion, and acquires n converted images corresponding to the n reflective sub-images, and then the authentication module identifies the authenticity of the banknote to be detected through the n converted images.
- the reflected image is divided into n The n reflection sub-images under the spectral segment, the more the spectral segments, the finer the discrimination, so the discrimination accuracy of the banknote can be improved.
- the tunable wavelength module 7 comprises: an upper reflective layer, a lower reflective layer, an adjustable component, and an adjustable component for adjusting the relative positions of the upper reflective layer and the lower reflective layer.
- the multi-spectral identification device of the banknote further comprises a signal generating module 6 for transmitting the voltage signals S 1 , S 2 , S 3 , . . . S n to the tunable wavelength module 7 at intervals.
- the adjustable component is an electrostatic brake
- the electrostatic brake is used to receive voltage signals S 1 , S 2 , S 3 , . . . S n ;
- the upper reflective layer is mounted on the upper dielectric plate, and the lower reflective layer is mounted on the lower dielectric plate.
- the brakes are respectively connected to the upper dielectric plate and the lower dielectric plate.
- the multi-spectral identification device of the banknote further comprises an image separation module 10 for separating n optical color-changing ink conversion images from the n converted images; the authentication module 11 is configured to convert through n optical color-changing inks The image identifies the authenticity of the banknote to be detected.
- the light source module may further include a left light source module 3, a right light source module 4, which is not specifically limited herein), a wide field of view imaging module 5, a signal generation module 6, and an adjustable
- the banknote module 1 is located on the left and right sides of the transparent cover glass 2, and is used for receiving the banknote to be detected, and is electrically connected to the information processing module 9 (connected through the bus A). It should be noted that the present invention In the embodiment, the banknote to be detected may also be received by other means, which is not specifically limited herein;
- the transparent protective glass 2 is located directly above the structural frame 12 and can be used for keeping the internal module clean.
- the transparent protective glass 2 can be disposed according to the actual application environment or can be covered with a transparent plastic material. , there is no specific limit here;
- the light source module, the light source module may be used to illuminate the banknote to be detected, and the multi-spectral identification device of the banknote in this embodiment may include a left light source module 3 and a right light source module 4, and the left light source module 3 and the right light source module 4 may be respectively placed at The left and right inner sides of the structural frame 12, the left light source module 3 and the right light source module 4 can emit white light or other color light, and are electrically connected to the information processing module 9 (connected through the bus A), where the structure and the light source of the light source module are The placement mode of the module and the color of the light source module are not specifically limited;
- a wide field of view imaging module 5 which may be a wide field of view imaging lens module, the wide market imaging module 5 may be located below the banknote in order to obtain a full reflection image of the banknote to be detected;
- the wide field of view imaging module 5 may be located in the transparent protective glass 2 In the lower part (in order to facilitate the collection of all the reflected images of the banknotes to be detected, the wide field of view imaging module 5 may be located directly below the transparent cover glass 2, not specifically defined), the left light source module 3 and the right light source module 4 The emitted light is irradiated to different angles of the banknote, and the image reflected by the banknote to be detected contains different spectral information, and then the wide field of view imaging module 5 acquires the reflected image of the banknote to be detected;
- the tunable wavelength module 7 can be a micro-mechanical tunable wavelength module, and the tunable wavelength module 7 can be located below the wide-field imaging module 5 (in order to more accurately identify the banknote to be detected) The authenticity, the tunable wavelength module 7 can be located directly below the wide field of view imaging module 5, without limitation, for extracting n reflected sub-images under n spectral segments from the reflected image, the tunable wavelength module 7 is electrically connected to the signal generating module 6 through the bus B, and the adjustable wavelength module 7 and the authentication module 11 are electrically connected through the bus D;
- the photodetection module 8 can be located below the tunable wavelength module 7 (not specifically defined) for photoelectric conversion of n reflective sub-images of the banknote to be detected, and obtained by photoelectric signal conversion and n reflections
- the photo-detection module 8 and the image separation module 10 are electrically connected to each other through the bus C;
- a signal generating module 6 for transmitting one-dimensional voltage signals S 1 , S 2 , S 3 , . . . S n to the tunable wavelength module 7 , and the signal generating module 6 and the tunable wavelength module 7 passing through the bus B Electrical connection
- the signal processing module 9 can be located below the photodetection module 8 (specifically not limited), and the signal processing module 9 is used for voltage signal control of the signal generating module 6, the left light source module 3, and the right light source module 4. Controlling, electrically connected to the banknote module 1, the left light source module 3, the right light source module 4, and the signal generating module 6 via the bus A;
- the image separation module 10 may be located on the right side of the information processing module 9 (not specifically defined), and the image separation module 10 is used for separating and separating the optical color ink regions of the n converted images of the banknotes to be detected.
- the image separation module 10 is electrically connected to the photodetection module 8 via the bus C, and the image separation module 10 and the authentication module 11 are electrically connected through the bus D;
- An authentication module 11 which may be located on the right side of the image separation module 10 (specifically not limited), for identifying the authenticity of the banknote to be detected by n converted images, and multi-spectral recognition of the banknote in the embodiment of the present invention
- the authentication module 11 identifies the authenticity of the banknote to be detected by the n optical color changing ink conversion images, and the authentication module 11 and the image separating module 10 are connected by the bus E;
- the banknote module 1 connected to the financial self-service device transports the banknote to be detected, and the banknote module 1 triggers the information processing module 9 when the banknote to be detected completely covers the transparent cover glass 2, and the information processing module 9 activates the left light source.
- the light emitted by the module 3 and the right light source module 4, the left light source module 3 and the right light source module 4 is irradiated to the surface of the banknote to be detected through the transparent cover glass 2;
- the information processing module 9 triggers the signal generation module 6 while activating the left light source module 3 and the right light source module 4, and the signal generation module 6 generates the interval one-dimensional voltage signals S 1 , S according to the voltage interval ⁇ V within the range of the maximum voltage V max . 2, S 3, ... S n , wherein one-dimensional signal voltage S 1, S 2, 3, ... S n S of length N, the time period between adjacent voltage signal T 1 (Note that the one-dimensional The transmission intervals of the voltage signals S 1 , S 2 , S 3 , ... S n are not necessarily equal), N ⁇ 3;
- the wide field of view imaging module 5 receives the reflected image of the banknote to be detected transmitted by the transparent cover glass 2, and converges the reflected image of the banknote to be detected, and then reaches the photodetection module 8 through the tunable wavelength module 7, in order to facilitate the optical of the banknote to be detected.
- the collection of the reflection information in different directions of the color changing ink, the wide field of view imaging module 5 may be an imaging lens or a fisheye lens having an imaging field angle of more than M degrees, and the number of lenses of the wide field of view imaging module 5 is set to N, Where M ⁇ 160, N ⁇ 3;
- the tunable wavelength module 7 receives the one-dimensional voltage signals S 1 , S 2 , S 3 , . . . S n transmitted by the signal generating module 6 , and each voltage signal corresponds to one spectral segment, and thus may have n spectral segments; Module 7 will extract n reflection sub-images under n spectral segments according to time period T 1 , where n is a natural number greater than 2;
- the tunable wavelength module 7 can be composed of a Fabry-Perot interferometer and a micro-mechanical control device.
- the Fabry-Perot interferometer is composed of two reflective layers having the same characteristics (which can be divided into an upper reflective layer and a lower reflective layer according to the positional relationship) and a spacer layer therebetween, and the spacer layer
- the corresponding Fabry-Perot interferometer transmits the dominant wavelength ⁇ 1 as:
- ⁇ 1 2nd; where n is the refractive index of the Fabry-Perot interferometer spacer layer.
- the Fabry-Perot interferometer has a high transmittance for light of a specific wavelength associated with the optical thickness of the spacer layer and the reflection passband of the reflective film, while the transmittance of light of other wavelengths is low, exhibiting band-pass characteristics. .
- the micromechanical control device is composed of an electrostatic brake, an upper dielectric plate and a lower dielectric plate, wherein the upper reflective layer of the Fabry-Perot interferometer is attached to the inner side of the upper dielectric plate, and the lower reflective layer is attached to the inner side of the lower dielectric plate.
- the reflective layer on the upper and lower dielectric plates is a spacer layer corresponding to the Fabry-Perot interferometer.
- the electrostatic brake of the tunable wavelength module 7 receives the voltage signals S 1 , S 2 , S 3 , ... S n generated by the signal generating module 6 , and the electrostatic brake drives the upper and lower dielectric plates to move the reflective layer to change according to the input voltage signal.
- the thickness d 1 of the spacer layer, then d 1 is:
- V max is the maximum voltage that the electrostatic brake can withstand
- V is the voltage signal generated by the signal generator
- d is the initial distance of the spacer layer.
- the transmission dominant wavelength ⁇ 2 corresponding to the tunable wavelength module 7 is:
- the one-dimensional voltage signals S 1 , S 2 , S 3 , . . . S n transmitted by the electrostatic brake receiving signal generating module 6 of the tunable wavelength module 7 will drive the upper and lower dielectric plates to be sequentially moved n times according to the time period T 1 .
- the reflected image of the banknote to be detected is adjusted to n reflection sub-images in the range of visible light (400 nm - 760 nm) and near infrared (800-1100 nm).
- the movement of the upper and lower dielectric plates will drive the n-time movement of the inner reflective layer, and the thickness of the spacer layer of the Fabry-interferometer will also change, so that the corresponding wavelength of the tunable wavelength module 7 will also change accordingly. n times.
- Photodetection module 8 can acquire n converted images in n spectral segments
- the n converted images corresponding to the n spectral segments are uploaded to the image separating module 10, and the image separating module 10 separates and obtains n optical color changing inks corresponding to the spectral segments transmitted under the n voltage signals from the obtained n converted images.
- image The specific process is as follows:
- the image separation module 10 first sets an optical color-changing ink approximate region R, and intercepts n optical color-changing ink-converted image images of the region R.
- k 1, 2, ..., n ⁇ .
- the image separation module 10 uses the classical OTSU binarization algorithm to obtain accurate n optical color-changing ink conversion images.
- Image separation module 10 converts n optical color changing inks into images Uploaded to the authentication module 11, the authentication module 11 sequentially calculates a difference image between the optically color-changing ink-converted images of adjacent time periods T 1 (adjacent time periods T 1 generating adjacent wavelengths) And calculating a histogram corresponding to the n-1 difference image, and the discriminating module 11 associates the histogram corresponding to the n-1 difference image into the feature vector H in the chronological order, and substitutes the feature vector H into the SVM (Support Vector Machine) method. Identifying the authenticity of the optical color-changing ink in the determined decision threshold, the discriminant formula is:
- the present invention adopts a pattern recognition method to extract features from the image, and then classifies the features using a classifier to obtain authenticity discrimination.
- the present invention takes a histogram of the difference image as a feature and combines the histograms of all the difference images as the final classification feature.
- the N-1 differential image D i corresponds to a histogram, wherein the gray level of the histogram is 511, and the histograms of the N-1 differential images are connected end to end to form the feature vector H.
- n i is the number of pixels in the image ⁇ P that take the value i
- n is the total number of pixels.
- H [H ist ( ⁇ P 1 ),H ist ( ⁇ P 2 ),...H ist ( ⁇ P n )]
- the discriminating module 11 performs final authenticity discrimination according to the authenticity feature vector H, and the discriminant formula is;
- the pseudo-feature vector H used in the present invention has a high dimensionality, and the learning training SVM classifier does not depend on the dimension of the sample, which is very suitable for high-dimensional pattern recognition.
- the present invention uses the SVM classifier to perform authenticity on the pseudo-feature vector H. Pseudo-criticism.
- the specific process of the SVM classifier to determine the weight parameters w and b is as follows:
- the identification device collects a large number of samples of true and false banknotes and composes a training sample set M is the number of samples of the training sample set.
- I m is the image of the collected banknote
- y m ⁇ -1,1 ⁇ is the sample label
- -1 indicates that the banknote is a counterfeit banknote
- +1 indicates a genuine banknote.
- the final result of the training is based on the training sample set
- the SVM classifier optimal weight parameters w * , b * are calculated.
- the identification device extracts feature composition training sets for each banknote according to the above process Solve the following formula through the classic SMO optimization algorithm:
- H i is a feature vector of the i-th banknote image I i
- y i is a sample tag of the i-th banknote
- L(a) is the energy function to be optimized.
- w * , b * are the optimal weight parameters.
- Sequential minimal optimization is an algorithm used to solve the optimization problem generated in the training process of SVM. It is widely used in the training process of SVM and is in the popular SVM library. Get implemented. After the training process is completed, the identification device extracts the feature H for each image of the banknote to be identified according to the above steps, and then according to the formula
- an embodiment of the present invention further provides a multi-spectral identification method for a banknote, including:
- S2 specifically includes:
- Extracting n reflected sub-images in n spectral segments from the acquired preset voltage signals S 1 , S 2 , S 3 , . . . , S n , preset voltage signals S 1 , S 2 , S 3 , ...S n corresponds one-to-one with n spectral segments.
- the method further comprises: separating the acquired n converted images, and separating the n optical color changing ink conversion images.
- the step S4 specifically comprises: performing authenticity identification of the banknote to be detected by converting the image by the n optical color changing inks.
Landscapes
- Inspection Of Paper Currency And Valuable Securities (AREA)
Abstract
一种钞票的多光谱识别装置及识别方法，用于获取钞票的各个光谱段的图像并与预设条件进行比对，提高识别精度。识别装置包括：用于获取待检测钞票的反射图像的宽视场成像模块(5)；用于对反射图像进行n个光谱段下的n个反射子图像的提取的可调波长模块(7)，n为大于2的自然数；用于对反射子图像进行光电信号转换，并获取相对应的转换图像的光电探测模块(8)；用于通过转换图像识别待检测钞票真伪的鉴别模块(11)。
Description
本申请要求于2015年06月19日提交中国专利局、申请号为201510347268.7、发明名称为“一种钞票的多光谱识别装置及方法”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
本发明实施例涉及一种无接触式光学检测技术,尤其涉及一种钞票的多光谱识别装置及方法。
光学变色油墨，简称光变油墨，目前大多被应用于钞票上使假币容易被辨别出来。从不同角度观察，油墨会显现出不同的颜色，从而保证了颜色不被复制，我国在新版人民币100元和50元上首次使用了光学变色油墨作为防伪特征。
现有技术中的光学变色油墨的识别装置，包括框架、透镜、位于透镜两侧发光角度不同的两个光源，当光线以不同的角度照射真钞表面的变色油墨防伪标识，不同的入射角度的光线被变色油墨反射后的光谱信息就不同，不同的光谱信息转换出的电信号也不同，经过反射后的光线被接触式图像传感器接收，接触式图像传感器对接收到的两种光线进行分析处理比较，然后将得到的信息与真钞的相关信息比对，以达到鉴别真伪的目的。
现有技术的缺陷是，光学变色油墨的识别装置不能比对钞票的各个光谱段的图像，识别精度差。
发明内容
本发明实施例提供一种钞票的多光谱识别装置，用于获取钞票的各个光谱段的图像并与预设条件进行比对，提高识别精度。
本发明实施例一种钞票的多光谱识别装置的技术方案包括：
宽视场成像模块、可调波长模块、光电探测模块、鉴别模块；
所述宽视场成像模块用于获取待检测钞票的反射图像；
所述可调波长模块用于对所述反射图像进行n个光谱段下的n个反射子图像的提取，其中n为大于2的自然数；
所述光电探测模块用于对所述n个反射子图像进行光电信号转换，并获取与所述n个反射子图像相对应的n个转换图像；
所述鉴别模块用于通过所述n个转换图像识别所述待检测钞票的真伪。
优选的，
所述可调波长模块包括：上反射层、下反射层、可调节部件，所述可调节部件用于调节所述上反射层和所述下反射层的相对位置。
优选的，
所述的钞票的多光谱识别装置还包括信号发生模块，所述信号发生模块用于向所述可调波长模块间隔发送电压信号S1、S2、S3、…Sn。
优选的，
所述可调节部件为静电制动器，所述静电制动器用于接收所述电压信号S1、S2、S3、…Sn；
所述上反射层安装在上电介质板上，所述下反射层安装在下电介质板上，所述静电制动器分别与所述上电介质板、所述下电介质板电性连接。
优选的，
所述钞票的多光谱识别装置还包括图像分离模块，所述图像分离模块用于从所述n个转换图像中分离出n个光学变色油墨转换图像；
所述鉴别模块用于通过所述n个光学变色油墨转换图像识别所述待检测钞票的真伪。
优选的，
所述钞票的多光谱识别装置还包括左光源模块和右光源模块，所述左光源模块、所述右光源模块各分别位于所述钞票的多光谱识别装置的左侧、右侧。
本发明实施例还提供一种钞票的多光谱识别方法，包括：
S1：获取待检测钞票的反射图像；
S2：对获取的所述反射图像进行n个光谱段下的n个反射子图像的提取，其中n为大于2的自然数；
S3：对提取的所述n个反射子图像进行光电信号转换，并获取与所述n个反射子图像相对应的n个转换图像；
S4：通过获取的所述n个转换图像识别所述待检测钞票的真伪。
优选的，
所述S2具体包括：
获取预置电压信号S1、S2、S3、…Sn；
根据获取的所述预置电压信号S1、S2、S3、…Sn对所述反射图像进行n个光谱段下的n个反射子图像的提取，所述预置电压信号S1、S2、S3、…Sn与所述n个光谱段一一对应。
优选的，
所述S3之后、所述S4之前还包括：对获取的所述n个转换图像进行分离，分离出n个光学变色油墨转换图像。
优选的，
所述S4具体包括：通过所述n个光学变色油墨转换图像进行所述待检测钞票的真伪识别。
本发明实施例的有益效果是：
首先宽视场成像模块接收待检测钞票的反射图像，然后可调波长模块对反射图像进行n个光谱段下的n个反射子图像的提取，然后光电探测模块将n个反射子图像进行光电信号转换，并获取与n个反射子图像相对应的n个转换图像，然后鉴别模块通过n个转换图像识别待检测钞票的真伪，采用上述技术方案，由于反射图像被分成n个光谱段下的n个反射子图像，由于光谱段越多，鉴别越精细，因此可以提高钞票的鉴别精度。
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。
图1为本发明实施例一种钞票的多光谱识别装置的结构示意图；
图2为本发明实施例一种钞票的多光谱识别装置的控制关系图；
图3为本发明实施例一种钞票的多光谱识别装置的法布里-珀罗干涉仪原理示意图；
图4为本发明实施例一种钞票的多光谱识别装置的可调波长模块示意图；
图5为本发明实施例一种钞票的多光谱识别方法的流程图。
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做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。
下面以待检测介质为钞票详细说明本发明实施例一种钞票的多光谱识别装置的具体技术方案，需要说明的是，该钞票的多光谱识别装置可应用于不同领域，不做具体限定。
请参阅图1至图4，本发明实施例一种钞票的多光谱识别装置的技术方案包括：
宽视场成像模块5、可调波长模块7、光电探测模块8、鉴别模块11；
宽视场成像模块5，用于获取待检测钞票的反射图像；
可调波长模块7，用于对反射图像进行n个光谱段下的n个反射子图像的提取，其中n为大于2的自然数；
光电探测模块8，用于将n个反射子图像进行光电信号转换，并获取与n个反射子图像相对应的n个转换图像；
鉴别模块11，通过n个转换图像识别待检测钞票的真伪。
采用上述技术方案，首先宽视场成像模块接收待检测钞票的反射图像，然后可调波长模块对反射图像进行n个光谱段下的n个反射子图像的提取，然后光电探测模块将n个反射子图像进行光电信号转换，并获取与n个反射子图像相对应的n个转换图像，然后鉴别模块通过n个转换图像识别待检测钞票的真伪，采用上述技术方案，由于反射图像被分成n个光谱段下的n个反射子图像，由于光谱段越多，鉴别越精细，因此可以提高钞票的鉴别精度。
优选的，可调波长模块7包括：上反射层、下反射层、可调节部件，可调节部件用于调节上反射层和下反射层的相对位置。
优选的，钞票的多光谱识别装置还包括信号发生模块6，信号发生模块6用于向可调波长模块7间隔发送电压信号S1、S2、S3、…Sn。
优选的，可调节部件为静电制动器，静电制动器用于接收电压信号S1、S2、S3、…Sn；上反射层安装在上电介质板上，下反射层安装在下电介质板上，静电制动器分别与上电介质板、下电介质板连接。
优选的，钞票的多光谱识别装置还包括图像分离模块10，图像分离模块10用于从n个转换图像中分离出n个光学变色油墨转换图像；鉴别模块11用于通过n个光学变色油墨转换图像识别待检测钞票的真伪。
本发明实施例中的一种钞票的多光谱识别装置可以具体包括：
走钞模块1、透明保护玻璃2、光源模块(光源模块可以进一步包括左光源模块3、右光源模块4，此处不做具体限定)、宽视场成像模块5、信号发生模块6、可调波长模块7、光电探测模块8、信息处理模块9、图像分离模块10、鉴别模块11、结构框架12、结构框架13和结构框架14。
下面对一种钞票的多光谱识别装置的各模块进行详细说明：
走钞模块1，该走钞模块1位于透明保护玻璃2的左右两侧，用于接收待检测的钞票，与信息处理模块9电性连接(通过总线A连接)，需要说明的是，本发明实施例中也可以通过其他方式接收待检测的钞票，此处不做具体限定；
透明保护玻璃2，该透明保护玻璃2位于结构框架12的正上方，可以用于保持内部模块的清洁，此处的透明保护玻璃2可根据实际应用环境设置，也可使用透明塑料类材质进行遮盖，此处不做具体限定；
光源模块，光源模块可以用于照射待检测的钞票，本实施例中的钞票的多光谱识别装置可以包括左光源模块3和右光源模块4，左光源模块3和右光源模块4可以分别放置在结构框架12的左、右内侧，左光源模块3和右光源模块4可以发出白光或其他颜色光，与信息处理模块9电性连接(通过总线A连接)，此处对于光源模块的结构、光源模块的放置方式、光源模块的发光颜色均不做具体限定；
宽视场成像模块5，该宽视场成像模块5可以为宽视场成像透镜模块，该宽市场成像模块5为了获取待检测钞票的全部反射图像，可以位于钞票的下方；
当本发明实施例中的钞票的多光谱识别装置包括透明保护玻璃2和左光源模块3和右光源模块4时，该宽视场成像模块5可以位于透明保护玻璃2
的下方(需要说明的是，为了便于采集待检测钞票的全部反射图像，宽视场成像模块5可以位于透明保护玻璃2的正下方，不做具体限定)，左光源模块3和右光源模块4发出的光照射到钞票的不同角度，待检测钞票反射出的图像中包含不同的光谱信息，然后宽视场成像模块5获取待检测钞票的反射图像；
可调波长模块7，需要说明的是，可调波长模块7可以为微机械可调波长模块，可调波长模块7可以位于宽视场成像模块5的下方(为了更准确的识别待检测钞票的真伪，可调波长模块7可以位于宽视场成像模块5的正下方，不做具体限定)，用于对反射图像进行n个光谱段下的n个反射子图像的提取，可调波长模块7通过总线B与信号发生模块6电性连接，可调波长模块7与鉴别模块11通过总线D电性连接；
光电探测模块8，该光电探测模块8可以位于可调波长模块7的下方(不做具体限定)，用于待检测钞票的n个反射子图像的光电转换，通过光电信号转换获取与n个反射子图像相对应的n个转换图像，光电探测模块8与图像分离模块10通过总线C电性连接；
信号发生模块6，该信号发生模块6用于向可调波长模块7间隔发送一维电压信号S1、S2、S3、…Sn，信号发生模块6与可调波长模块7通过总线B电性连接；
信号处理模块9，该信号处理模块9可以位于光电探测模块8的下方(具体不做限定)，信号处理模块9用于信号发生模块6的电压信号控制、左光源模块3、右光源模块4的控制，与走钞模块1、左光源模块3、右光源模块4和信号发生模块6通过总线A电性连接；
图像分离模块10，该图像分离模块10可以位于信息处理模块9的右侧(具体不做限定)，图像分离模块10用于对待检测钞票的n个转换图像的光学变色油墨区域的分离，分离出n个光学变色油墨转换图像，图像分离模块10通过总线C与光电探测模块8电性连接，图像分离模块10与鉴别模块11通过总线D电性连接；
鉴别模块11，该鉴别模块可以位于图像分离模块10的右侧(具体不做限定)，用于通过n个转换图像识别待检测钞票的真伪，当本发明实施例中的钞票的多光谱识别装置包括图像分离模块10时，鉴别模块11通过n个光学变色油墨转换图像识别待检测钞票的真伪，鉴别模块11与图像分离模块10通过总线E连接；
采用上述技术方案，与金融自助设备连接的走钞模块1输送待检测的钞票，走钞模块1在待检测的钞票完全覆盖透明保护玻璃2时触发信息处理模块9，信息处理模块9启动左光源模块3和右光源模块4，左光源模块3和右光源模块4发出的光通过透明保护玻璃2照射到待检测钞票的表面；
信息处理模块9在启动左光源模块3和右光源模块4的同时触发信号发生模块6，信号发生模块6在最大电压Vmax的范围内按照电压间距ΔV产生间隔的一维电压信号S1、S2、S3、…Sn，其中一维电压信号S1、S2、S3、…Sn的长度为N，相邻电压信号间的时间周期为T1(需要说明的是，一维电压信号S1、S2、S3、…Sn的发送间隔时间不一定相等)，N≥3；
宽视场成像模块5接收由透明保护玻璃2透射的待检测钞票的反射图像，并将待检测钞票的反射图像会聚后通过可调波长模块7到达光电探测模块8，为了便于待检测钞票的光学变色油墨的不同方向的反射信息的收集，宽视场成像模块5可以为成像视场角度达到M度以上的成像镜头或鱼眼镜头，宽视场成像模块5的透镜数量设定为N个，其中M≥160，N≥3；
可调波长模块7接收信号发生模块6间隔发送的一维电压信号S1、S2、S3、…Sn，每一个电压信号对应一个光谱段，因此可以有n个光谱段；可调波长模块7将按照时间周期T1提取n个光谱段下的n个反射子图像，其中n为大于2的自然数；
可调波长模块7可由法布里-珀罗干涉仪和微机械控制设备两部分组成。
其中，如图3所示，法布里–珀罗干涉仪由两个特性相同的反射层(可以根据位置关系分为上反射层、下反射层)及其中间的间隔层构成，间隔
层的厚度设置为d，则对应的法布里-珀罗干涉仪透射主波长λ1为：
λ1＝2nd；其中，n是法布里–珀罗干涉仪间隔层对应的折射率。
法布里–珀罗干涉仪只有对与间隔层光学厚度和反射膜反射通带相关的特定波长的光具有较高的透射率，而其它波长光的透射率则很低，表现出带通特性。
其中，微机械控制装置由静电制动器、上电介质板和下电介质板组成，其中法布里-珀罗干涉仪的上反射层分别附着在上电介质板的内侧，下反射层附着在下电介质板的内侧，上、下电介质板上的反射层之间为法布里-珀罗干涉仪所对应的间隔层。
可调波长模块7的静电制动器接收信号发生模块6间隔产生的电压信号S1、S2、S3、…Sn，静电制动器驱动上、下电介质板带动反射层移动，按照输入的电压信号改变间隔层的厚度d1，则d1为：
此时可调波长模块7所对应的透射主波长λ2为：
可调波长模块7的静电制动器接收信号发生模块6传输来的一维电压信号S1、S2、S3、…Sn，将会带动上、下电介质板按照时间周期T1依次移动n次，将待检测钞票的反射图像在可见光(400nm-760nm)和近红外(800-1100nm)范围内调整为n个反射子图像。上、下电介质板的移动将会带动内侧的反射层的n次移动，法布里-干涉仪的间隔层厚度也随之改变，因而所对应的可调波长模块7透射主波长也相应的改变n次。
光电探测模块8可以获取在n个光谱段的n个转换图像
将n个光谱段对应的n个转换图像上传到图像分离模块10，图像分离模块10从获取的n
个转换图像中分离获得n个电压信号下透射的光谱段所对应的n个光学变色油墨转换图像
具体过程如下：
由于n个光谱段照射时，待检测钞票位置固定，所以这n个转换图像
是严格对齐的，由于光学变色油墨在待检测钞票的同一区域。每张钞票经过采集模块时位置会有一定变动，但是变动不会太大。为此，图像分离模块10首先设定一个光学变色油墨大致区域R，并截取区域R的n个光学变色油墨转换图像图像
光学变色油墨区域在若干光谱段图像中较黑，图像分离模块10取最小图像Rmin，其中Rmin的每一点(x，y)的值为Rmin(x,y)＝min{Rk(x,y)|k＝1,2,...,n}。图像分离模块10采用经典的OTSU二值化算法得到精确的n个光学变色油墨转换图像
钞票图像容易受到污迹、褶皱等影响，但是这些影响会在所有波长的图像中同时显现，为了消除这些影响，本发明可以采用差分图像来提高鉴别模块11的鲁棒性。图像分离模块10将n个光学变色油墨转换图像
上传到鉴别模块11，鉴别模块11依次计算相邻时间周期T1(相邻时间周期T1产生相邻波长)的光学变色油墨转换图像之间的差分图像
并计算n-1幅差分图像对应的直方图，鉴别模块11将算n-1幅差分图像对应的直方图按照时间次序首尾相连组成特征向量H，将特征向量H代入SVM(Support Vector Machine)方法确定的判定阈值中鉴别光学变色油墨的真伪，判别公式为：
为了从差分图像得到待检测钞票的真伪判别，本发明采用模式识别方法，从图像中提取特征，然后使用分类器对特征进行分类得到真伪判别。
由于真伪钞票光学变色油墨区域的主要差别在于对各波长的光的反射分布差异，反映到差分图像上主要表现为强度分布差异，并且所有的差分图像的直方图特征都对最后的真伪判别有贡献，所以本发明采用差分图像的直方图作为特征，并将所有的差分图像的直方图联合起来作为最后的分
类特征。具体为计算N-1幅差分图像Di对应直方图，其中直方图的灰度级别是511级，并将N-1幅差分图像的直方图按照次序首尾相连组成特征向量H。具体过程如下：
差分图像
计算每张差分图像的直方图Hist(ΔPk),k＝1,2,...N，直方图为一个511维的向量，计算公式为：
Hist(ΔPk)＝ni/n,i＝-255,-254,...0,...,255
其中ni为图像ΔP中取值为i的像素点的个数，n为像素点的总个数。最后将N个直方图连接起来组成一个N*511维的向量，成为最后的鉴伪特征向量，即
H＝[Hist(ΔP1),Hist(ΔP2),...Hist(ΔPn)]
鉴别模块11根据鉴伪特征向量H进行最后真伪判别，判别公式为；
根据s的值做出判断s≥0，此钞票的光学变色油墨为真；s＜0，此钞票的光学变色油墨为假。其中wi,i＝1,2,...,n*511；b为待训练的权重参数。本发明中采用的鉴伪特征向量H的维数较高，学习训练SVM分类器不依赖样本的维数，非常适合高维模式识别问题，本发明采用SVM分类器对鉴伪特征向量H进行真伪判别。SVM分类器确定权重参数w和b的具体过程如下：
识别装置收集大量真假钞票样本，并组成训练样本集
M为训练样本集的样本个数。其中Im为采集的钞票图像，ym∈{-1,1}为样本标签，-1表示该张钞票为伪钞，+1表示为真钞。
训练的最终结果是根据训练样本集
计算得到SVM分类器最优权重参数w*，b*。
识别装置对每张钞票按照上述过程提取特征组成特征训练集
通过经典的SMO优化算法求解下式：
最小化：
约束条件：
得到最优解a*为一个向量
然后根据公式求出计算得到SVM分类器最优权重参数w*，b*：
其中，其中Hi为第i张钞票图像Ii的特征向量，yi为第i张钞票的样本标签，a为拉格朗日乘子向量a＝(a1,a2,...aM)，L(a)为待优化的能量函数。w*，b*为最优权重参数。
序列最小优化算法(英语：Sequential minimal optimization,SMO)是一种用于解决支持向量机训练过程中所产生优化问题的算法，目前被广泛使用于SVM的训练过程中，并在通行的SVM库中得到实现。在训练过程完毕后，识别装置对每张待鉴别钞票图像I根据上述步骤提取特征H，然后根据公式
请参阅图5，本发明实施例还提供一种钞票的多光谱识别方法，包括：
S1：获取待检测钞票的反射图像；
S2：对获取的反射图像进行n个光谱段下的n个反射子图像的提取，其中n为大于2的自然数；
S3：对提取的n个反射子图像进行光电信号转换，并获取与n个反射子图像相对应的n个转换图像；
S4：通过获取的n个转换图像识别待检测钞票的真伪。
优选的，S2具体包括：
获取预置电压信号S1、S2、S3、…Sn；
根据获取的预置电压信号S1、S2、S3、…Sn对反射图像进行n个光谱段下的n个反射子图像的提取，预置电压信号S1、S2、S3、…Sn与n个光谱段一一对应。
优选的，步骤S3之后、S4之前还包括：对获取的n个转换图像进行分离，分离出n个光学变色油墨转换图像。
优选的，步骤S4具体包括：通过n个光学变色油墨转换图像进行待检测钞票的真伪识别。
以上实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。
Claims (10)
- 一种钞票的多光谱识别装置，其特征在于，包括：宽视场成像模块、可调波长模块、光电探测模块、鉴别模块；所述宽视场成像模块用于获取待检测钞票的反射图像；所述可调波长模块用于对所述反射图像进行n个光谱段下的n个反射子图像的提取，其中n为大于2的自然数；所述光电探测模块用于对所述n个反射子图像进行光电信号转换，并获取与所述n个反射子图像相对应的n个转换图像；所述鉴别模块用于通过所述n个转换图像识别所述待检测钞票的真伪。
- 根据权利要求1所述的钞票的多光谱识别装置，其特征在于，所述可调波长模块包括：上反射层、下反射层、可调节部件，所述可调节部件用于调节所述上反射层和所述下反射层的相对位置。
- 根据权利要求2所述的钞票的多光谱识别装置，其特征在于，所述的钞票的多光谱识别装置还包括信号发生模块，所述信号发生模块用于向所述可调波长模块间隔发送电压信号S1、S2、S3、…Sn。
- 根据权利要求3所述的钞票的多光谱识别装置，其特征在于，所述可调节部件为静电制动器，所述静电制动器用于接收所述电压信号S1、S2、S3、…Sn；所述上反射层安装在上电介质板上，所述下反射层安装在下电介质板上，所述静电制动器分别与所述上电介质板、所述下电介质板电性连接。
- 根据权利要求4所述的钞票的多光谱识别装置，其特征在于，所述钞票的多光谱识别装置还包括图像分离模块，所述图像分离模块用于从所述n个转换图像中分离出n个光学变色油墨转换图像；所述鉴别模块用于通过所述n个光学变色油墨转换图像识别所述待检 测钞票的真伪。
- 根据权利要求1所述的钞票的多光谱识别装置，其特征在于，所述钞票的多光谱识别装置还包括左光源模块和右光源模块，所述左光源模块、所述右光源模块各分别位于所述钞票的多光谱识别装置的左侧、右侧。
- 一种钞票的多光谱识别方法，其特征在于，包括：S1：获取待检测钞票的反射图像；S2：对获取的所述反射图像进行n个光谱段下的n个反射子图像的提取，其中n为大于2的自然数；S3：对提取的所述n个反射子图像进行光电信号转换，并获取与所述n个反射子图像相对应的n个转换图像；S4：通过获取的所述n个转换图像识别所述待检测钞票的真伪。
- 根据权利要求7所述的钞票的多光谱识别方法，其特征在于，所述S2具体包括：获取预置电压信号S1、S2、S3、…Sn；根据获取的所述预置电压信号S1、S2、S3、…Sn对所述反射图像进行n个光谱段下的n个反射子图像的提取，所述预置电压信号S1、S2、S3、…Sn与所述n个光谱段一一对应。
- 根据权利要求8所述的钞票的多光谱识别方法，其特征在于，所述S3之后、所述S4之前还包括：对获取的所述n个转换图像进行分离，分离出n个光学变色油墨转换图像。
- 根据权利要求9所述的钞票的多光谱识别方法，其特征在于，所述S4具体包括：通过所述n个光学变色油墨转换图像进行所述待检测钞票的真伪识别。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510347268.7 | 2015-06-19 | ||
| CN201510347268.7A CN104916036A (zh) | 2015-06-19 | 2015-06-19 | 一种钞票的多光谱识别装置及方法 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2016202165A1 true WO2016202165A1 (zh) | 2016-12-22 |
Family
ID=54085071
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2016/083909 Ceased WO2016202165A1 (zh) | 2015-06-19 | 2016-05-30 | 一种钞票的多光谱识别装置及方法 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN104916036A (zh) |
| WO (1) | WO2016202165A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240643A (zh) * | 2021-05-14 | 2021-08-10 | 广州广电运通金融电子股份有限公司 | 一种基于多光谱图像的钞票质量检测方法、系统及介质 |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104916036A (zh) * | 2015-06-19 | 2015-09-16 | 广州广电运通金融电子股份有限公司 | 一种钞票的多光谱识别装置及方法 |
| CN105277498B (zh) * | 2015-11-17 | 2017-03-08 | 广东南天司法鉴定所 | 一种检测文件墨迹种类的方法和系统 |
| JP6615014B2 (ja) * | 2016-03-15 | 2019-12-04 | グローリー株式会社 | 紙葉類識別装置および紙葉類識別方法 |
| CN111950317B (zh) * | 2020-08-07 | 2024-05-14 | 量子云码(福建)科技有限公司 | 一种微观编码图像提取装置及提取图像后鉴别真伪的方法 |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101083000A (zh) * | 2006-05-31 | 2007-12-05 | Mei公司 | 验钞的方法和设备 |
| CN101290665A (zh) * | 2008-06-06 | 2008-10-22 | 梓昆科技（杭州）有限公司 | 多光谱点钞仪及其点钞控制方法 |
| CN101308540A (zh) * | 2008-05-04 | 2008-11-19 | 梓昆科技（杭州）有限公司 | 接触式多光谱图像传感器 |
| CN101605199A (zh) * | 2008-06-11 | 2009-12-16 | 三菱电机株式会社 | 图像读取装置 |
| CN102236931A (zh) * | 2010-04-27 | 2011-11-09 | 北京中科金财科技股份有限公司 | 票据防伪光谱分析信息数字化采集与识别方法及设备 |
| CN103903326A (zh) * | 2014-03-03 | 2014-07-02 | 广州科珥光电科技有限公司 | 多光谱伪钞检测系统及其检测方法 |
| CN104916036A (zh) * | 2015-06-19 | 2015-09-16 | 广州广电运通金融电子股份有限公司 | 一种钞票的多光谱识别装置及方法 |
| EP3018636A1 (en) * | 2014-11-06 | 2016-05-11 | Aristotle University Of Thessaloniki-Research Committee | Multi signature anti-counterfeiting process and system |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0954849A (ja) * | 1995-08-14 | 1997-02-25 | Shibaura Eng Works Co Ltd | 紙幣識別装置 |
| CN2831273Y (zh) * | 2005-09-23 | 2006-10-25 | 中国印钞造币总公司 | 一种纸币鉴别仪滤光片自动转换装置 |
| US7441704B2 (en) * | 2006-03-03 | 2008-10-28 | Ncr Corporation | System and method for identifying a spatial code |
| CN101458836B (zh) * | 2007-12-14 | 2010-11-03 | 广州广电运通金融电子股份有限公司 | 有价票据真伪鉴别方法及系统 |
| GB0913798D0 (en) * | 2009-08-07 | 2009-09-16 | Innovative Technology Ltd | Banknote validator |
| JP5531832B2 (ja) * | 2010-07-06 | 2014-06-25 | セイコーエプソン株式会社 | 光フィルター、光フィルターモジュール、分光測定器および光機器 |
| JP2013030100A (ja) * | 2011-07-29 | 2013-02-07 | Canon Components Inc | イメージセンサユニットおよび紙葉類識別装置 |
| CN102592347A (zh) * | 2012-02-24 | 2012-07-18 | 湖南丰汇银佳科技有限公司 | 基于光谱分析技术的钞票鉴伪方法和装置 |
| WO2014005085A1 (en) * | 2012-06-29 | 2014-01-03 | De La Rue North America Inc. | Systems for capturing images of a document |
| CN102759441A (zh) * | 2012-07-20 | 2012-10-31 | 西北工业大学 | 多波段偏振成像防伪元件辨识系统 |
-
2015
- 2015-06-19 CN CN201510347268.7A patent/CN104916036A/zh active Pending
-
2016
- 2016-05-30 WO PCT/CN2016/083909 patent/WO2016202165A1/zh not_active Ceased
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101083000A (zh) * | 2006-05-31 | 2007-12-05 | Mei公司 | 验钞的方法和设备 |
| CN101308540A (zh) * | 2008-05-04 | 2008-11-19 | 梓昆科技（杭州）有限公司 | 接触式多光谱图像传感器 |
| CN101290665A (zh) * | 2008-06-06 | 2008-10-22 | 梓昆科技（杭州）有限公司 | 多光谱点钞仪及其点钞控制方法 |
| CN101605199A (zh) * | 2008-06-11 | 2009-12-16 | 三菱电机株式会社 | 图像读取装置 |
| CN102236931A (zh) * | 2010-04-27 | 2011-11-09 | 北京中科金财科技股份有限公司 | 票据防伪光谱分析信息数字化采集与识别方法及设备 |
| CN103903326A (zh) * | 2014-03-03 | 2014-07-02 | 广州科珥光电科技有限公司 | 多光谱伪钞检测系统及其检测方法 |
| EP3018636A1 (en) * | 2014-11-06 | 2016-05-11 | Aristotle University Of Thessaloniki-Research Committee | Multi signature anti-counterfeiting process and system |
| CN104916036A (zh) * | 2015-06-19 | 2015-09-16 | 广州广电运通金融电子股份有限公司 | 一种钞票的多光谱识别装置及方法 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240643A (zh) * | 2021-05-14 | 2021-08-10 | 广州广电运通金融电子股份有限公司 | 一种基于多光谱图像的钞票质量检测方法、系统及介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104916036A (zh) | 2015-09-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111046703B (zh) | 人脸防伪检测方法、装置及多目相机 | |
| US20220114834A1 (en) | Face recognition monitoring system based on spectrum and multi-band fusion and recognition method using same | |
| WO2016202165A1 (zh) | 一种钞票的多光谱识别装置及方法 | |
| US20190005307A1 (en) | Method and system for biometric recognition | |
| WO2016078455A1 (zh) | 纸币褶皱的识别装置和方法 | |
| CN108229362A (zh) | 一种基于门禁系统的双目人脸识别活体检测方法 | |
| US11620849B2 (en) | Spectral imaging chip and apparatus, information processing method, fingerprint living body identification device and fingerprint module | |
| CN114912598B (zh) | 光人工神经网络智能芯片及制备方法 | |
| CN1627317A (zh) | 利用主动光源获取人脸图像的方法 | |
| WO2012083711A1 (zh) | 一种纸币检测方法及装置 | |
| CN111582197A (zh) | 一种基于近红外和3d摄像技术的活体以及人脸识别系统 | |
| WO2013131407A1 (zh) | 双验证人脸防伪方法及装置 | |
| CN102339494A (zh) | 接触式图像传感器、多角度光学特征检测方法及装置 | |
| CN117994624B (zh) | 一种基于可见光和高光谱图像信息融合的目标识别方法 | |
| CN111539348A (zh) | 一种基于偏振成像的人脸活体检测方法 | |
| WO2012083860A1 (zh) | 一种人脸识别装置和方法 | |
| KR101797953B1 (ko) | 신경망 학습기법을 이용한 위조지문 판별장치 및 그 방법 | |
| CN206075299U (zh) | 一种带伪指纹识别的活体指纹采集装置 | |
| CN107437293A (zh) | 一种基于票据全局特征的票据防伪鉴别方法 | |
| CN112232152B (zh) | 非接触式指纹识别方法、装置、终端和存储介质 | |
| Burge et al. | Multispectral iris fusion for enhancement, interoperability, and cross wavelength matching | |
| CN105550668B (zh) | 一种活体生物特征采集装置及识别活体生物特征的方法 | |
| CN107393121A (zh) | 一种纸币鉴伪方法 | |
| CN205750988U (zh) | 一种纸币鉴伪装置 | |
| Apoloni et al. | Philippine currency counterfeit detector using image processing |
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: 16810906 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: 16810906 Country of ref document: EP Kind code of ref document: A1 |