A palm feature recognition device, a recognition method thereof, and a storage medium

Abstract

The application provides a palm feature recognition device and a recognition method thereof and a storage medium. The palm feature recognition device comprises a light source, a multispectral camera and a control and processor; wherein: the light source is used for projecting a first light beam with a wavelength in a visible light band and a second light beam in an infrared light band to a palm to be recognized; the multispectral camera is used for collecting the first light beam and the second light beam reflected back by the palm to be recognized, and generating a multispectral image of the palm to be recognized; the control and processor are used for extracting a channel image of each channel from the multispectral image, and according to all the channel images and preset palmprint image reconstruction coefficients, reconstructing a palmprint image of the palm to be recognized and recognizing the palmprint, and according to all the channel images and preset palm vein image reconstruction coefficients, reconstructing a palm vein image of the palm to be recognized and recognizing the palm vein. The application can obtain the palmprint image and the palm vein image, and has high recognition accuracy.

Description

[Technical Field] This application relates to the field of biometric technology, and in particular to a palm feature recognition device, recognition method, and storage medium thereof. [Background Technology] In recent years, palm biometric features such as palm prints and palm veins have emerged as new biometric identification methods, generally achieved through imaging the palm. Related technologies typically include three types of palm feature recognition devices: First, those that use a color camera to image the palm and identify corresponding palm print information, but lack liveness detection functionality; second, those that use an infrared camera to image the palm and identify corresponding palm vein information, which, while possessing liveness detection capabilities, provides less palm vein information than palm print information; and third, those that use both a color camera and an infrared camera to image the palm and identify corresponding palm print information, which can simultaneously identify both palm print and palm vein information and also have liveness detection functionality, but require two imaging components—a color camera and an infrared camera—resulting in higher manufacturing costs. Furthermore, the palm images captured by the color camera and infrared camera exhibit parallax, necessitating registration. [Summary of the Invention] This application provides a palm feature recognition device, recognition method, and storage medium thereof, aiming to solve at least one problem existing in palm feature recognition devices in related technologies.

[0001] To address the aforementioned technical problems, a first aspect of this application provides a palm feature recognition device, including a light source, a multispectral camera, and a controller and processor. The light source projects a first light beam with a wavelength in the visible light band and a second light beam with a wavelength in the infrared light band onto the palm to be recognized. The multispectral camera acquires the first light beam and the second light beam reflected back from the palm to be recognized, generating a multispectral image of the palm to be recognized. The controller and processor are used to: extract channel images of each channel from the multispectral image; reconstruct a palmprint image of the palm to be recognized and recognize the palmprint based on all channel images and preset palmprint image reconstruction coefficients; and reconstruct a palm vein image of the palm to be recognized and recognize the palm vein based on all channel images and preset palm vein image reconstruction coefficients.

[0002] In some embodiments, the wavelength of the first beam is between 400 nm and 580 nm, and the wavelength of the second beam is between 800 nm and 980 nm. In some embodiments, the light source is a ring light source, and the multispectral camera is located at the center of the ring light source. The ring light source includes multiple first light sources for generating the first beam and multiple second light sources for generating the second beam. The multiple first light sources and multiple second light sources are arranged alternately. The multiple first light sources are used to generate first beams of one or more wavelengths, and the multiple second light sources are used to generate second beams of one or more wavelengths. In some embodiments, the palmprint image reconstruction coefficients include palmprint image reconstruction sub-coefficients corresponding to each channel image, and the palm vein image reconstruction coefficients include palm vein image reconstruction sub-coefficients corresponding to each channel image. The control and processor are specifically used to: multiply each channel image by its corresponding palmprint image reconstruction sub-coefficient and sum them to obtain a palmprint image; multiply each channel image by its corresponding palm vein image reconstruction sub-coefficient and sum them to obtain a palm vein image.

[0003] A second aspect of this application provides a palm feature recognition method, comprising: projecting a first light beam with a wavelength in the visible light band and a second light beam with a wavelength in the infrared light band onto a palm to be recognized; acquiring the first light beam and the second light beam reflected back from the palm to be recognized, and generating a multispectral image of the palm to be recognized; extracting channel images of each channel from the multispectral image; reconstructing a palmprint image of the palm to be recognized based on all channel images and preset palmprint image reconstruction coefficients, and recognizing the palmprint based on the palmprint image; and reconstructing a palm vein image of the palm to be recognized based on all channel images and preset palm vein image reconstruction coefficients, and recognizing the palm vein based on the palm vein image.

[0004] In some embodiments, the palm feature recognition method further includes: acquiring the brightness curve of a light source and the multispectral response curve of a multispectral camera, wherein the light source is used to emit a first beam and a second beam, and the multispectral camera is used to acquire multispectral images; generating the actual response curve of the multispectral camera to the first beam and the second beam based on the brightness curve and the multispectral response curve; acquiring the target response curve of a preset target channel, and calculating the palmprint image reconstruction coefficient and the palm vein image reconstruction coefficient based on the target response curve and the actual response curve. In some embodiments, the actual response curve includes multiple actual sub-response curves of multiple channels of the multispectral camera to the light source, and the target response curve includes a first target sub-response curve of a first target channel and a second target sub-response curve of a second target channel; calculating the palmprint image reconstruction coefficient and the palm vein image reconstruction coefficient based on the target response curve and the actual response curve includes: calculating a first fitting coefficient for fitting multiple actual sub-response curves to the first target sub-response curve, and using the first fitting coefficient as the palmprint image reconstruction coefficient; calculating a second fitting coefficient for fitting multiple actual sub-response curves to the second target sub-response curve, and using the second fitting coefficient as the palm vein image reconstruction coefficient. In some embodiments, the wavelength of the first beam is a first wavelength, the wavelength of the second beam is a second wavelength, the center wavelength of the first target channel is a first wavelength, and the center wavelength of the second target channel is a second wavelength.

[0005] In some embodiments, the palmprint image reconstruction coefficients include palmprint image reconstruction sub-coefficients corresponding to each channel image, and the palm vein image reconstruction coefficients include palm vein image reconstruction sub-coefficients corresponding to each channel image; reconstructing the palmprint image of the hand to be identified based on all channel images and preset palmprint image reconstruction coefficients includes: multiplying each channel image by its corresponding palmprint image reconstruction sub-coefficient and summing the results to obtain the palmprint image; reconstructing the palm vein image of the hand to be identified based on all channel images and preset palm vein image reconstruction coefficients includes: multiplying each channel image by its corresponding palm vein image reconstruction sub-coefficient and summing the results to obtain the palm vein image.

[0006] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the palm feature recognition method described in the second aspect of this application.

[0007] As can be seen from the above description, compared with related technologies, the beneficial effects of this application are as follows: By combining a light source, a multispectral camera, and a control and processor, a single imaging component can be used to identify palm prints and palm veins, avoiding the parallax problem between images acquired by two imaging components in palm feature recognition devices in related technologies. Furthermore, palm print and palm vein images are reconstructed based on multispectral images after obtaining them, which improves the accuracy of palm print and palm vein recognition compared to palm print and palm vein recognition based on multispectral images. [Attached Image Description] To more clearly illustrate the related technologies or the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the related technologies or the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application, and not all embodiments. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic diagram of the frame of the palm feature recognition device provided in the embodiments of this application; Figure 2 A schematic diagram showing the relative positions of the light source and the multispectral camera provided in an embodiment of this application; Figure 3 Example diagram of the brightness curve of the light source provided in the embodiments of this application; Figure 4 Example diagram of the multispectral response curve of a 9-channel multispectral camera provided in the embodiments of this application; Figure 5 Example graph of actual response curve provided for embodiments of this application; Figure 6 Example diagram of target response curve provided for embodiments of this application; Figure 7 This is an example diagram of a fitted response curve obtained by fitting the actual response curve to the target response curve, provided in an embodiment of this application. Figure 8 Example diagram of channel images of 9 channels extracted from a 9-channel multispectral image, provided for embodiments of this application; Figure 9 Example image of the reconstructed palm print provided in the embodiments of this application; Figure 10 Example image of reconstructed palm vein provided in the embodiments of this application; Figure 11 A flowchart illustrating the palm feature recognition method provided in this application embodiment; Figure 12 This is a schematic flowchart illustrating the calculation of palm print and palm vein image reconstruction coefficients provided in an embodiment of this application.

Detailed Implementation Methods

[0009] Figure 1 This is a schematic diagram of the framework of a palm feature recognition device provided in an embodiment of this application. The palm feature recognition device is used to acquire palm print images and palm vein images of a hand 100 to be recognized, and to perform palm print recognition based on the acquired palm print images, and palm vein recognition based on the acquired palm vein images. The palm feature recognition device includes a controller and processor 10, a light source 20, and a multispectral camera 30. The controller and processor 10 is connected to the light source 20 and the multispectral camera 30, and can control the light source 20 and the multispectral camera 30, for example, controlling the light source 20 to emit light and controlling the multispectral camera 30 to acquire image data.

[0010] In this embodiment, the light source 20 generates a first light beam with a wavelength in the visible light band and a second light beam with a wavelength in the infrared light band. During palm feature recognition, the palm to be recognized 100 is placed in the optical path of the light source 20, with the palm surface facing the light source 20. The first and second light beams emitted by the light source 20 are projected onto the palm to be recognized 100 and reflected by the palm to be recognized to the multispectral camera 30. The multispectral camera 30 is used to acquire the first and second light beams reflected back from the palm to be recognized 100 and generate a multispectral image of the palm to be recognized 100 based on the acquired beams. Since the multispectral camera 30 has multiple channels, the controller and processor 10 can extract the channel image of each channel from the multispectral image and reconstruct the palmprint image of the palm to be recognized 100 based on all channel images and preset palmprint image reconstruction coefficients for palmprint recognition. The controller and processor 10 also reconstructs the palm vein image of the palm to be recognized 100 based on all channel images and preset palm vein image reconstruction coefficients for palm vein recognition. The reconstruction coefficients of the palm print image and the palm vein image were pre-calculated, and the specific calculation process will be described in detail below.

[0011] In this embodiment, the palm feature recognition device, in addition to including the light source 20, the multispectral camera 30, and the control and processor 10, may also include a memory (not shown in the figure) connected to the control and processor 10. This memory stores all data during the palmprint and palm vein recognition process, such as the multispectral image of the palm 100 to be recognized and the channel image of each channel, palmprint image reconstruction coefficients, palm vein image reconstruction coefficients, and the palmprint and palm vein images of the palm 100 to be recognized. The memory also stores a computer program that can be invoked and executed by the control and processor 10. This computer program specifies the procedure for the control and processor 10 to perform palmprint and palm vein recognition on the palm 100 to be recognized, including but not limited to controlling the light source 20 to generate a first beam and a second beam, controlling the multispectral camera 30 to generate a multispectral image of the palm 100 to be recognized, reconstructing the palmprint and palm vein images of the palm 100 to be recognized from the multispectral image, and recognizing the palmprint and palm veins.

[0012] As can be seen from the above, compared with palm feature recognition devices that only set up color cameras or infrared cameras, or that set up both color cameras and infrared cameras, the palm feature recognition device of this application embodiment, through the cooperation of light source 20, multispectral camera 30 and control and processor 10, realizes that a single imaging component (i.e., multispectral camera 30) can complete the recognition of palm prints and palm veins. Furthermore, after obtaining the multispectral image of the palm to be recognized 100, the palm print image and palm vein image are reconstructed, and then the palm print is recognized based on the palm print image and the palm vein is recognized based on the palm vein image. This is more accurate than directly using multispectral images for palm print and palm vein recognition. In addition, it is less expensive than setting up both color cameras and infrared cameras at the same time, and avoids the problem of parallax between palm print images and palm vein images.

[0013] Experimental verification revealed that when the wavelength of the light beam illuminating the palm 100 to be identified is less than 580nm, the acquired palm print image is relatively clear; when the wavelength of the light beam illuminating the palm 100 to be identified is greater than 580nm, the acquired palm vein image becomes increasingly clear, and when it is greater than 800nm, the clarity of the acquired palm vein image is even clearer. In some embodiments, the wavelength of the first light beam generated by the light source 20 is ≤580nm and the wavelength of the second light beam is ≥800nm, resulting in clearer reconstructed palm print and palm vein images. In one implementation, the wavelength of the first beam is between 400nm and 580nm, such as 400nm, 450nm, 500nm, 525nm, 550nm, 580nm, etc.; the wavelength of the second beam is between 800nm ​​and 980nm, such as 800nm, 850nm, 880nm, 900nm, 940nm, 980nm, etc.; thus, the palm print image and palm vein image obtained subsequently based on multispectral image reconstruction are clearer.

[0014] In some embodiments, the light source 20 is a ring light source, and the ring shape presented may include, but is not limited to, a circular ring, an elliptical ring, a rectangular ring, a trapezoidal ring, a triangular ring, and a polygonal ring. Figure 2 This is a schematic diagram showing the relative positions of the light source 20 and the multispectral camera 30 provided in an embodiment of this application. The light source 20 includes multiple first light sources 21 for generating a first light beam and multiple second light sources 22 for generating a second light beam. The multiple first light sources 21 and multiple second light sources 22 are arranged alternately, that is, any two adjacent light sources include one first light source 21 and one second light source 22. In this way, the distribution of the first light beam and the second light beam projected by the ring light source onto the palm 100 to be identified is more uniform. Among them, the light beam emitted by the first light source 21 is mainly used to acquire palm print data, and the light beam emitted by the second light source 22 is mainly used to acquire palm vein data.

[0015] In one implementation, the multispectral camera 30 is located at the center of the ring light source. For example, when the ring light source presents a circular shape, the multispectral camera 30 is located at the center of the ring. In this way, the first beam and the second beam reflected by the palm 100 to be identified, collected by the multispectral camera 30, are more uniform and more abundant, which is beneficial to improving the imaging accuracy of subsequent multispectral images.

[0016] In one implementation, a plurality of first light sources 21 emit first light beams of one or more wavelengths (within the visible light band); for example, all of the plurality of first light sources 21 emit first light beams of the same wavelength; or, for another example, the plurality of first light sources 21 emit first light beams of at least two wavelengths, i.e., all or some of the first light sources 21 emit first light beams of different wavelengths. A plurality of second light sources 22 emit second light beams of one or more wavelengths (within the infrared light band); for example, all of the plurality of second light sources 22 emit second light beams of the same wavelength; or, for another example, the plurality of second light sources 22 emit second light beams of at least two wavelengths, i.e., some or all of the second light sources 22 emit second light beams of different wavelengths.

[0017] The multispectral camera 30 includes a multispectral image sensor, which comprises a multispectral filter array and a photosensitive chip array. The multispectral filter array includes multiple filters with different center wavelengths to allow light beams of different wavelengths to pass through, thus giving the multispectral camera 30 multiple channels. These multiple channels include a visible light channel and an infrared light channel, enabling the acquisition of a first light beam and a second light beam reflected from the palm 100 to be identified. Since ambient light beams are also reflected by the palm 100, the multispectral camera 30 also acquires the reflected ambient light beams, and then generates a multispectral image.

[0018] The controller and processor 10 receives multispectral images acquired by the multispectral camera 30 and extracts data from each channel of each pixel in the multispectral image, thereby obtaining channel images corresponding to each channel. Due to current manufacturing limitations, each channel in the multispectral camera 30 is relatively wide, and each channel may receive both the first and second light beams. Each channel image contains both palmprint and palm vein data, lacking channel images with only pure palmprint and palm vein data. Therefore, the controller and processor 10 fuses all channel images according to preset palmprint image reconstruction coefficients to reconstruct the palmprint image of the hand 100 to be identified, and fuses all channel images according to preset palm vein image reconstruction coefficients to reconstruct the palm vein image of the hand 100 to be identified. In this way, the palmprint and palm vein images are less likely to be mixed up, facilitating separate palmprint and palm vein recognition, resulting in higher accuracy for both recognition.

[0019] The palmprint image reconstruction coefficients and palm vein image reconstruction coefficients each consist of multiple palm vein image reconstruction sub-coefficients, each corresponding to multiple channel images. Each channel image is multiplied by its corresponding palmprint image reconstruction sub-coefficient, and then summed to obtain the palmprint image; similarly, each channel image is multiplied by its corresponding palm vein image reconstruction sub-coefficient, and then summed to obtain the palm vein image. For example, if a multispectral camera 30 has n channels, then n channel images can be extracted, where n ≥ 4; the palmprint image reconstruction coefficients and palm vein image reconstruction coefficients each consist of n palm vein image reconstruction sub-coefficients.

[0020] The palmprint image reconstruction coefficients and palm vein image reconstruction coefficients can be pre-calculated and stored in memory. First, the spectrum of light source 20 is tested using a spectrometer to obtain the brightness curve of light source 20. The brightness curve includes the brightness curves corresponding to the first beam and the second beam. The brightness curve actually indicates the relative intensity of the first beam and the second beam, that is, the relative intensity of the first light source 21 and the second light source 22. Simultaneously, the multispectral response curve of multispectral camera 30 is acquired. The multispectral response curve includes the spectral response curve corresponding to each channel. The multispectral response curve can be pre-calibrated and stored in memory. Then, based on the brightness curve of light source 20 and the multispectral response curve of multispectral camera 30, the actual response curve of multispectral camera 20 to light source 20 is generated. For example, the actual response curve is obtained by multiplying the brightness curve and the multispectral response curve. The actual response curve includes the actual sub-response curve of each channel to light source 20. Target channels and target response curves of target channels can be pre-designed for the wavelengths of the first beam and the second beam. The palmprint image reconstruction coefficients and palm vein image reconstruction coefficients are calculated based on the actual response curve and the target response curve.

[0021] The target channels include a first target channel that receives only the first beam and a second target channel that receives only the second beam. The center wavelengths of the first and second target channels are the same as the wavelengths of the first light source 21 and the second light source 22, respectively. It is necessary to obtain the first target sub-response curve of the first target channel and the second target sub-response curve of the second target channel. The first fitting coefficient for fitting all actual sub-response curves to the first target sub-response curve is calculated and used as the palm print image reconstruction coefficient. Similarly, the second fitting coefficient for fitting all actual sub-response curves to the second target sub-response curve is calculated and used as the palm vein image reconstruction coefficient. This can be obtained through fitting tools or algorithms. The number of first and second target channels is not limited to one. When the wavelengths of the first beams generated by all first light sources 21 are multiple, the number of first target channels is multiple, and each of these multiple first target channels has a different center wavelength, corresponding to the wavelengths of the various first beams. The same applies to the second target channels.

[0022] Taking a palm feature recognition device as an example, where the wavelength of the first beam generated by all first light sources 21 is 525nm, the wavelength of the second beam generated by all second light sources 22 is 850nm, and the multispectral camera 30 has 9 channels, the center wavelength of the first target channel is 525nm, and the center wavelength of the second target channel is 850nm. The brightness curve of light source 20 can be found in [reference needed]. Figure 3 The given example diagram shows the brightness curves of the light source. The brightness curves of the first beam and the second beam are m1 and m2, respectively. Figure 3 In the diagram, m1 and m2 form a brightness curve. The vertical axis of this brightness curve is represented by the matrix Illu(1,band), and the horizontal axis is the wavelength. The data in Illu(1,band) are the normalized brightness data of light source 20. The multispectral response curve includes spectral response curves for 9 channels, totaling 9 spectral response curves. For more information on the multispectral response curve, please refer to [reference needed]. Figure 4The given example diagram shows the multispectral response curves of a 9-channel multispectral camera. The spectral response curves for the 9 channels are n1, n2, n3, n4, n5, n6, n7, n8, and n9. The horizontal axis of the spectral response curves for the 9 channels is represented by wavelength, and the vertical axis is represented by matrices QE1(1,band), QE2(1,band), QE3(1,band), QE4(1,band), QE5(1,band), QE6(1,band), QE7(1,band), QE8(1,band), and QE9(1,band), respectively. The data in each matrix represents the transmittance of the corresponding channel to each wavelength. Multiplying the brightness curve of light source 20 with the 9 spectral response curves of multispectral camera 30 yields the actual sub-response curves of the 9 channels to light source 20. The combination of the 9 actual sub-response curves constitutes the actual response curve of multispectral camera 30 to light source 20, as shown below. Figure 5 The provided example graphs of actual response curves are o1, o2, o3, o4, o5, o6, o7, o8, and o9. The vertical axes Res1 to Res9 of these nine actual sub-response curves are represented as follows: Res1=Illu(1,band).*QE1(1,band); Res2=Illu(1,band).*QE2(1,band); Res3=Illu(1,band).*QE3(1,band); Res4=Illu(1,band).*QE4(1,band); Res5=Illu(1,band).*QE5(1,band); Res6=Illu(1,band).*QE6(1,band); Res7=Illu(1,band).*QE7(1,band); Res8=Illu(1,band).*QE8(1,band); Res9=Illu(1,band).*QE9(1,band); The operator “.*” represents dot product, which is the product of corresponding data in two matrices.

[0023] Corresponding to Figure 4 The spectral response curves of the nine channels of the multispectral camera 30. Figure 5The actual sub-response curves of the nine channels of the multispectral camera 30 to the light source 20 all include the responses to the first beam at 525nm and the second beam at 850nm. That is to say, palm print data and palm vein data are mixed together in each channel. In order to distinguish between palm print data and palm vein data, this embodiment reconstructs separate responses to the first beam at 525nm (i.e., the first target channel) and separate responses to the second beam at 850nm (the second target channel).

[0024] The center wavelength of the first target channel is 525nm, and the center wavelength of the second target channel is 850nm. (See reference...) Figure 6 The example graph of the target response curve is given. Figure 6 The first target sub-response curve and the second target sub-response curve are p1 and p2, respectively. The ordinate of the first target sub-response curve is represented by matrix Tgt1(1,band), and the ordinate of the second target sub-response curve is represented by matrix Tgt2(1,band).

[0025] Multiple actual sub-response curves are fitted to a first target sub-response curve to obtain the palmprint image reconstruction coefficients (i.e., the first fitting coefficients when fitting to the first target sub-response curve); multiple actual sub-response curves are fitted to a second target sub-response curve to obtain the palm vein image reconstruction coefficients (i.e., the second fitting coefficients when fitting to the second target sub-response curve). Specifically, the palmprint image reconstruction coefficients and palm vein image reconstruction coefficients can be calculated using fitting tools such as MATLAB or fitting algorithms such as least squares.

[0026] The process of fitting multiple actual sub-response curves into the first target sub-response curve and the second target sub-response curve can be represented by the following two relationships: Tgt1(1,band)=a1*Res1+b1*Res2+c1*Res3+d1*Res4+e1*Res5+f1*Res6+g1*Res7+h1*Res8+i1*Res9; Tgt2(1,band)=a2*Res1+b2*Res2+c2*Res3+d2*Res4+e2*Res5+f2*Res6+g2*Res7+h2*Res8+i2*Res9; The first fitting coefficients (a1, b1, c1, d1, e1, f1, g1, h1, i1) and the second fitting coefficients (a2, b2, c2, d2, e2, f2, g2, h2, i2) can be calculated through fitting and other methods. The nine first fitting coefficients (a1 to i1) represent the palmprint image reconstruction coefficients corresponding to the nine channel images extracted from the multispectral image, forming the palmprint image reconstruction coefficients. Similarly, the nine second fitting coefficients (a2 to i2) represent the palm vein image reconstruction coefficients corresponding to the nine channel images extracted from the multispectral image, forming the palm vein image reconstruction coefficients. For example... Figure 7 As shown, the first and second fitted target channel curves obtained by fitting with the first and second fitting coefficients are q1 and q2, respectively. It can be seen that the fitted first and second fitted target channel curves q1 and q2 are highly similar to the first target sub-response curve p1 and the second target sub-response curve p2.

[0027] like Figure 8 The image shown is an example of extracting 9 channel images corresponding to 9 channels from a multispectral image. These 9 channel images are represented by gray1, gray2, gray3, gray4, gray5, gray6, gray7, gray8, and gray9, respectively. The 9 channel images gray1, gray2, gray3, gray4, gray5, gray6, gray7, gray8, and gray9 are multiplied by the corresponding palmprint image reconstruction coefficients a1, b1, c1, d1, e1, f1, g1, h1, and i1, respectively, and then summed to obtain the result shown below. Figure 9 The palmprint image of the hand to be identified (100) is shown below; where the palmprint image Palmprint of the hand to be identified (100) is represented as: Palmprint = a1*gray1 + b1*gray2 + c1*gray3 + d1*gray4 + e1*gray5 + f1*gray6 + g1*gray7 + h1*gray8 + i1*gray9. The nine channel images gray1, gray2, gray3, gray4, gray5, gray6, gray7, gray8, and gray9 are multiplied by the corresponding palm vein image reconstruction coefficients (a2, b2, c2, d2, e2, f2, g2, h2, i2), and then summed to obtain the result as shown below. Figure 10The image shown is a palm vein image of the hand to be identified 100; wherein, the palm vein image Palmvein of the hand to be identified 100 is represented as: Palmvein=a2*gray1+b2*gray2+c2*gray3+d2*gray4+e2*gray5+f2*gray6+g2*gray7+h2*gray8+i2*gray9.

[0028] from Figure 4 It can be seen that the nine channels of the multispectral camera 30 are all relatively wide, meaning they can receive beams across a wide wavelength range. Each channel acquires a beam including a first beam and a second beam. The multispectral image contains palmprint and palm vein data, and even after extracting the images from each channel, this data remains, hindering subsequent palmprint and palm vein recognition. The target channels (i.e., the first target channel and the second target channel) required in this embodiment correspond to narrow wavelengths; that is, both the first target sub-response curve and the second target sub-response curve are narrow filter curves. This embodiment... Figure 4 Fitting the filter curves from 9 wide-band wavelengths Figure 6 The response of the narrow-band filter curve shown is used to obtain the image of the ideal target channel. This results in a clearer reconstructed palmprint and palm vein image, improving the accuracy of palmprint and palm vein recognition. In other words, the original 9 channels of the multispectral camera 30 do not have pure 525nm and 850nm channels. Current manufacturing processes also make it difficult to produce such narrow-band channels. This embodiment, combined with the design of the light source 20, fits pure 525nm and 850nm channels from the original 9 wide channels of the multispectral camera 30, thereby reconstructing pure palmprint and palm vein images and improving the accuracy of palmprint and palm vein recognition.

[0029] Based on the aforementioned palm feature recognition device, this application embodiment also provides a palm feature recognition method. Figure 11 This is a flowchart illustrating a palm feature recognition method provided in an embodiment of this application. The palm feature recognition method includes the following steps 1101 to 1105.

[0030] Step 1101: Project a first light beam with a wavelength in the visible light band and a second light beam with a wavelength in the infrared light band onto the palm to be identified.

[0031] Step 1102: Acquire the first and second beams reflected back from the palm to be identified, and generate a multispectral image of the palm to be identified.

[0032] Step 1103: Extract the channel image of each channel from the multispectral image.

[0033] Step 1104: Based on all channel images and preset palmprint image reconstruction coefficients, reconstruct the palmprint image of the hand to be identified, and identify the palmprint based on the palmprint image.

[0034] Step 1105: Based on all channel images and preset palm vein image reconstruction coefficients, reconstruct the palm vein image of the hand to be identified, and identify the palm vein based on the palm vein image.

[0035] In some implementations, the palm feature recognition method described above further includes a process of pre-establishing palmprint image reconstruction coefficients and palm vein image reconstruction coefficients. Specifically, Figure 12 This is a schematic flowchart of the process for calculating the reconstruction coefficients of palm print and palm vein images provided in an embodiment of this application. The process for calculating the reconstruction coefficients of palm print and palm vein images includes the following steps 1201 to 1204.

[0036] Step 1201: Obtain the brightness curve of the light source and the multispectral response curve of the multispectral camera.

[0037] Step 1202: Based on the brightness curve and the multispectral response curve, generate the actual response curve of the multispectral camera to the light source.

[0038] Step 1203: Obtain the target response curve of the preset target channel.

[0039] Step 1204: Calculate the palm print image reconstruction coefficient and palm vein image reconstruction coefficient based on the target response curve and the actual response curve.

[0040] In some embodiments, the target channel includes a first target channel and a second target channel, and the target response curve includes a first target sub-response curve and a second target sub-response curve. Step 1204 includes the steps of: calculating a first fitting coefficient for fitting multiple actual sub-response curves to the first target sub-response curve, and using the first fitting coefficient as the palm print image reconstruction coefficient; calculating a second fitting coefficient for fitting multiple actual sub-response curves to the second target sub-response curve, and using the second fitting coefficient as the palm vein image reconstruction coefficient. In some embodiments, the wavelength of the first beam is a first wavelength, the wavelength of the second beam is a second wavelength, the center wavelength of the first target channel is a first wavelength, and the center wavelength of the second target channel is a second wavelength.

[0041] In some implementations, step 1104 includes: multiplying each channel image by its corresponding palmprint image reconstruction sub-coefficient and summing the results to obtain a palmprint image. Step 1105 includes: multiplying each channel image by its corresponding palm vein image reconstruction sub-coefficient and summing the results to obtain a palm vein image.

[0042] The specific details of each step in the above-described palm feature recognition method can be found in the description of the palm feature recognition device, and will not be repeated here. In other words, the relevant content of the above-described palm feature recognition device can be applied to the above-described palm feature recognition method.

[0043] The above-described embodiments are merely preferred implementations of the present application and are not the only limitation on the described content; those skilled in the art can make flexible settings based on the actual application scenarios on the basis of the embodiments of the present application.

[0044] In summary, this application provides a palm feature recognition device and a corresponding recognition method. By combining the characteristics of the multispectral camera 30 and the design of the light source 20, this application enables the reconstruction of independent and clear palm print and palm vein images from a single multispectral image acquired from the palm 100 to be recognized. This allows for simultaneous recognition of palm prints and palm veins, as well as liveness detection. This not only improves the accuracy of palm print and palm vein recognition but also avoids the parallax problem between color images and infrared images in related palm feature recognition schemes.

[0045] Furthermore, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a control processor, causes the control processor to execute the aforementioned palm feature recognition method provided in embodiments of this application.

[0046] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0047] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., a solid-state drive).

[0048] It should be noted that the various embodiments in this application are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For product-related embodiments, since they are similar to method-related embodiments, the descriptions are relatively simple, and relevant parts can be referred to the descriptions of the method-related embodiments.

[0049] It should also be noted that, in this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0050] The above description of the disclosed embodiments enables those skilled in the art to implement or use the content of this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined in this application may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A palm feature recognition device, characterized in that, include: A light source is used to project a first beam of light with a wavelength in the visible light band and a second beam of light with a wavelength in the infrared light band onto the palm of the hand to be identified. A multispectral camera is used to acquire the first light beam and the second light beam reflected back from the palm to be identified, and to generate a multispectral image of the palm to be identified; A controller and processor are configured to: acquire the brightness curve of a light source and the multispectral response curve of a multispectral camera, wherein the light source is used to emit a first light beam and a second light beam, and the multispectral camera is used to acquire the multispectral image; generate the actual response curve of the multispectral camera to the first light beam and the second light beam based on the brightness curve and the multispectral response curve; acquire the target response curve of a preset target channel, and calculate the palmprint image reconstruction coefficient and the palm vein image reconstruction coefficient based on the target response curve and the actual response curve; extract the channel image of each channel from the multispectral image; reconstruct the palmprint image of the hand to be identified and identify the palmprint based on all the channel images and the palmprint image reconstruction coefficient; and reconstruct the palm vein image of the hand to be identified and identify the palm vein based on all the channel images and the palm vein image reconstruction coefficient.

2. The palm feature recognition device as described in claim 1, characterized in that, The wavelength of the first beam is between 400nm and 580nm, and the wavelength of the second beam is between 800nm ​​and 980nm.

3. The palm feature recognition device as described in claim 1, characterized in that, The light source is a ring light source, and the multispectral camera is located at the center of the ring light source. The ring light source includes a plurality of first light sources for generating the first light beam and a plurality of second light sources for generating the second light beam. The plurality of first light sources and the plurality of second light sources are arranged alternately. The plurality of first light sources are used to generate the first light beam of one or more wavelengths, and the plurality of second light sources are used to generate the second light beam of one or more wavelengths.

4. The palm feature recognition device as described in claim 1, characterized in that, The palmprint image reconstruction coefficients include palmprint image reconstruction sub-coefficients corresponding to each of the channel images, and the palm vein image reconstruction coefficients include palm vein image reconstruction coefficients corresponding to each of the channel images; the control and processor are specifically used for: Each of the aforementioned channel images is multiplied by the corresponding palmprint image reconstruction sub-coefficient, and the results are summed to obtain the palmprint image. Each of the aforementioned channel images is multiplied by the corresponding palm vein image reconstruction coefficient, and the results are summed to obtain the palm vein image.

5. A method for palm feature recognition, characterized in that, include: A first beam with a wavelength in the visible light band and a second beam with a wavelength in the infrared light band are projected onto the palm of the hand to be identified; The first light beam and the second light beam reflected back from the palm to be identified are collected to generate a multispectral image of the palm to be identified; The brightness curve of the light source and the multispectral response curve of the multispectral camera are obtained. The light source is used to emit the first beam and the second beam, and the multispectral camera is used to acquire the multispectral image. Based on the brightness curve and the multispectral response curve, the actual response curves of the multispectral camera to the first beam and the second beam are generated; Obtain the target response curve of the preset target channel, and calculate the palm print image reconstruction coefficient and palm vein image reconstruction coefficient based on the target response curve and the actual response curve; Extract the channel image of each channel from the multispectral image; Based on all the channel images and the palmprint image reconstruction coefficients, the palmprint image of the hand to be identified is reconstructed, and the palmprint is identified based on the palmprint image; Based on all the channel images and the palm vein image reconstruction coefficients, a palm vein image of the hand to be identified is reconstructed, and the palm vein is identified based on the palm vein image.

6. The palm feature recognition method as described in claim 5, characterized in that, The actual response curves include multiple actual sub-response curves of the multiple channels of the multispectral camera to the light source, and the target response curves include the first target sub-response curve of the first target channel and the second target sub-response curve of the second target channel. The step of calculating the palmprint image reconstruction coefficient and the palm vein image reconstruction coefficient based on the target response curve and the actual response curve includes: Calculate the first fitting coefficient of the multiple actual sub-response curves to fit the first target sub-response curve, and use the first fitting coefficient as the palmprint image reconstruction coefficient; Calculate the second fitting coefficients of the multiple actual sub-response curves to fit the second target sub-response curve, and use the second fitting coefficients as the reconstruction coefficients of the palm vein image.

7. The palm feature recognition method as described in claim 6, characterized in that, The center wavelength of the first target channel is the same as the wavelength of the first beam, and the center wavelength of the second target channel is the same as the wavelength of the second beam.

8. The palm feature recognition method as described in claim 5, characterized in that, The palmprint image reconstruction coefficients include palmprint image reconstruction sub-coefficients corresponding to each of the channel images, and the palm vein image reconstruction coefficients include palm vein image reconstruction sub-coefficients corresponding to each of the channel images; The step of reconstructing the palmprint image of the hand to be identified based on all the channel images and preset palmprint image reconstruction coefficients includes: Each of the aforementioned channel images is multiplied by the corresponding palmprint image reconstruction sub-coefficient, and the results are summed to obtain the palmprint image. The step of reconstructing the palm vein image of the hand to be identified based on all the channel images and preset palm vein image reconstruction coefficients includes: Each of the aforementioned channel images is multiplied by the corresponding palm vein image reconstruction coefficient, and the results are summed to obtain the palm vein image.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the palm feature recognition method as described in any one of claims 5 to 8.

Images