Method, apparatus, and system for image processing

Abstract

The present application discloses a method, apparatus and system for image processing. The method for image processing includes: collecting a background pattern image within a predetermined area; determining the aperture edge of the background pattern image to divide the predetermined area into a valid area and an invalid area; collecting a target image within the predetermined area; and updating the target image located in the invalid area to a background image and retaining the target image located in the valid area. This method can avoid introducing the target image in the invalid area, thereby improving the recognition rate of the target image.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and more particularly, to a method, apparatus, and system for image processing. Background Art

[0002] With the development of biotechnology, living fingerprints are increasingly applied to identity recognition systems, such as fingerprints, palm prints, faces, irises, etc. Among them, fingerprint recognition technology is highly feasible, secure, and low-cost, and has been applied in multiple fields such as mobile terminals, homes, and finance.

[0003] With the continuous development of fingerprint recognition technology, under-screen fingerprint recognition technology has been favored by many manufacturers for its unique advantages. Among them, optical under-screen fingerprint recognition utilizes the refraction and reflection of light, which can largely avoid the interference of ambient light and obtain better stability in extreme environments. Taking optical under-screen fingerprint recognition in smartphones as an example, when a finger presses the screen, the screen emits light to illuminate the finger area, and the reflected light that illuminates the finger area returns through the gaps between the screen pixels to the optical fingerprint module closely attached under the screen. After the optical fingerprint module collects an image based on the reflected light, it can determine whether the fingerprint in the collected image matches the pre-registered fingerprint, thereby realizing fingerprint recognition. The key to optical fingerprint image recognition lies in comparing the features of the sample and the template and giving a matching score based on the similarity of the features.

[0004] However, as Figure 1a and 1b shown, the area of the photosensitive region of the optical fingerprint module is relatively large, and the light source illumination often cannot cover the entire screen, and the fingerprint pressing area is also not fixed; considering the light source center, finger offset, and strong ambient light, etc., the area of the effective fingerprint is sometimes less than half of the screen area, and the signals outside the effective fingerprint will cause interference during the image processing and recognition processes, seriously affecting the recognition rate and increasing the usage risk.

[0005] After analysis, it is found that among the incorrect matches and rejections in optical under-screen fingerprint recognition, some are caused by the incorrect features of the fingerprint boundary and outside the boundary. As Figures 2a - 2c shown, it can be seen that the boundary of the fingerprint and the part outside the boundary region introduce invalid information (as shown in the dotted box) after filtering. During subsequent fingerprint recognition, this invalid information will introduce incorrect features, resulting in fingerprint recognition failure.

[0006] Therefore, it is desirable to provide an improved method for image processing to solve the above problems. Summary of the Invention

[0007] In view of the above problems, the object of the present invention is to provide a method, apparatus, and system for image processing, so as to minimize the incorrect features in the image.

[0008] According to a first aspect of the present invention, there is provided a method for image processing, including:

[0009] Collect a background pattern image within a predetermined area;

[0010] Determine the aperture edge of the background pattern image to divide the predetermined area into a valid area and an invalid area;

[0011] Collect a target image within the predetermined area; and

[0012] Update the target image located in the invalid area to a background image, and retain the target image located in the valid area.

[0013] Optionally, determining the aperture edge of the background pattern image includes:

[0014] Obtain a first matrix of the background pattern image;

[0015] Determine a threshold of the aperture edge according to a change in signal intensity on the diagonal of the background pattern image;

[0016] Perform binary segmentation on the first matrix according to the threshold to obtain a 0-1 matrix, where 0 elements in the 0-1 matrix represent the invalid area, and 1 elements represent the valid area.

[0017] Optionally, determining the threshold of the aperture edge includes:

[0018] Perform median filtering on the first matrix to obtain a second matrix;

[0019] Calculate the normalized vector and the first-order gradient vector from the center of the second matrix to the four corners respectively;

[0020] Select a plurality of elements in the second matrix that meet a preset condition, and set the median or average value of the plurality of elements that meet the preset condition as the threshold.

[0021] Optionally, the preset condition includes:

[0022] The element is located on the diagonal of the second matrix;

[0023] The ratio of the value of the element to the value of the center of the second matrix does not exceed a first value;

[0024] The gradient of the element with its adjacent third point does not exceed a second value.

[0025] Optionally, it further includes: obtaining the neighborhood of each element in the 0-1 matrix, and updating the elements with a ratio lower than a predetermined value in the neighborhood to another element.

[0026] Optionally, it further includes: determining an overexposed area according to the signal intensity of the background pattern image and / or the target image,

[0027] wherein, when the signal intensity in a partial peripheral area of the background pattern image and / or the target image exceeds the central intensity, the partial peripheral area is determined as the overexposed area, and the target image located within the overexposed area is deleted.

[0028] Optionally, it further includes: identifying, testing, comparing or verifying the target image located within the effective area.

[0029] Optionally, after acquiring the target image within the predetermined area, it further includes:

[0030] performing preprocessing on the target image;

[0031] counting the directional intensity and the number of ridge-valley lines within the neighborhood of each element in the first matrix of the target image, and taking the set of multiple elements with directional intensity and multiple ridge-valley lines within the neighborhood as the fingerprint area; and

[0032] retaining the target image that is located within the effective area and within the fingerprint area.

[0033] Optionally, calculating the variance of vectors with a fixed length in eight directions within the neighborhood of each element, and taking the direction with the minimum variance as the direction of the element, and the directional intensity is the most frequently occurring direction within the neighborhood divided by the area of the neighborhood.

[0034] Optionally, performing smoothing processing on the neighborhood vectors perpendicular to the direction, and calculating the number of first-order gradient sign changes to obtain the number of ridge-valley lines.

[0035] Optionally, the preprocessing includes: obtaining a background-pattern-removed target image according to the target image and the background pattern image.

[0036] According to a second aspect of the present invention, there is provided an image processing apparatus, including: an image processing module, after receiving a background pattern image and a target image within a predetermined area, the image processing module determines the aperture edge of the background pattern image to divide the predetermined area into an effective area and an ineffective area, updates the target image located within the ineffective area to a background image, and retains the target image located within the effective area.

[0037] Optionally, the image processing module includes:

[0038] a first unit for acquiring the first matrix of the background pattern image;

[0039] The second unit determines the threshold of the aperture edge according to the change in the signal intensity on the diagonal of the shading image;

[0040] The third unit performs binary segmentation on the first matrix according to the threshold to obtain a 0-1 matrix, where the 0 elements in the 0-1 matrix represent the invalid regions and the 1 elements represent the valid regions.

[0041] Optionally, the second unit includes:

[0042] A filtering unit performs median filtering on the first matrix to obtain a second matrix;

[0043] A calculation unit calculates the normalized vector and the first-order gradient vector from the center of the second matrix to the four corners respectively;

[0044] A selection unit selects multiple elements in the second matrix that meet the preset conditions, and sets the median or mean value of the multiple elements that meet the preset conditions as the threshold.

[0045] Optionally, the preset conditions of the selection unit include:

[0046] The element is located on the diagonal of the second matrix;

[0047] The ratio of the value of the element to the value of the center of the second matrix does not exceed a first value;

[0048] The gradient of the element with its adjacent third point does not exceed a second value.

[0049] Optionally, it further includes: a fourth unit, which obtains the neighborhood of each element in the 0-1 matrix and updates the elements with a ratio lower than a predetermined value in the neighborhood to another element.

[0050] Optionally, the image processing module further includes a fifth unit, which determines the overexposed region according to the signal intensity of the shading image and / or the target image,

[0051] wherein, when the signal intensity in a partial peripheral region of the shading image and / or the target image exceeds the central intensity, the partial peripheral region is determined as the overexposed region, and the target image located in the overexposed region is deleted.

[0052] Optionally, the image processing module further includes a sixth unit, and the sixth unit includes:

[0053] A preprocessing unit preprocesses the target image after collecting the target image in the predetermined region;

[0054] A statistical unit that calculates the directional intensity and the number of ridge-valley lines in the neighborhood of each element in the first matrix of the target image, and uses the set of multiple elements with directional intensity and multiple ridge-valley lines in the neighborhood as the fingerprint area; and

[0055] A processing unit that retains the target image that is within the effective area and within the fingerprint area.

[0056] Optionally, the statistical unit calculates the variance of vectors with a fixed length in eight directions in the neighborhood of each element, and uses the direction with the minimum variance as the direction of the element. The directional intensity is the most frequently occurring direction in the neighborhood divided by the area of the neighborhood.

[0057] The statistical unit smooths the neighborhood vectors perpendicular to the direction and calculates the number of first-order gradient sign changes to obtain the number of ridge-valley lines.

[0058] Optionally, the preprocessing unit obtains a target image without background pattern based on the target image and the background pattern image.

[0059] According to a third aspect of the present invention, there is provided an image processing system, including:

[0060] An image acquisition module that respectively acquires a background pattern image and a target image in a predetermined area; and

[0061] The image processing module as described above, determines the aperture edge of the background pattern image to divide the predetermined area into an effective area and an invalid area, updates the target image within the invalid area to a background image, and retains the target image within the effective area.

[0062] The image processing method, device and system provided by the present invention utilize the background pattern image to determine the effective area and the invalid area of image acquisition, remove the target image within the invalid area, and retain the target image within the effective area. Therefore, it avoids introducing target images with incorrect features within the invalid area, thereby improving the recognition rate of the target image and facilitating reducing the false acceptance rate and the false recognition rate in subsequent image recognition processes.

[0063] Furthermore, this technical solution determines the exposure area and deletes the target image within the exposure area, which can further avoid introducing incorrect features.

[0064] Furthermore, this technical solution can determine the fingerprint area according to the fingerprint change rule included in the target image and delete the image outside the fingerprint area, thereby further avoiding introducing incorrect features. Description of the Drawings

[0065] Through the following description of the embodiments of the present invention with reference to the accompanying drawings, the above and other objects, features, and advantages of the present invention will become clearer. In the drawings:

[0066] Figure 1a and 1b show a schematic diagram of the photosensitive area of an optical fingerprint module according to the prior art;

[0067] Figures 2a - 2c respectively show fingerprint images obtained after filtering the fingerprints collected by an optical fingerprint module according to the prior art;

[0068] Figure 3 shows a flowchart of a method for image processing according to an embodiment of the present invention;

[0069] Figure 4a and 4b respectively show schematic diagrams of the background pattern before and after processing according to an embodiment of the present invention;

[0070] Figure 5a , 5b and 5c respectively show the original image, the image after preprocessing, and the image after determining the exposure area of a target image according to an embodiment of the present invention;

[0071] Figure 6a , 6b and 6c respectively show the original image, the image after preprocessing, and the image after determining the fingerprint area of a target image according to an embodiment of the present invention;

[0072] Figure 7 shows a system block diagram of image processing according to an embodiment of the present invention. Detailed Embodiments

[0073] The following describes the present invention based on embodiments, but the present invention is not limited to these embodiments. In the following detailed description of the present invention, some specific details are described in detail. Those skilled in the art can fully understand the present invention without the description of these details. In order to avoid obscuring the essence of the present invention, well-known methods, processes, and procedures are not described in detail. Additionally, the drawings are not necessarily drawn to scale.

[0074] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions, and operations of the system, method, and device according to the embodiments of the present invention. The boxes in the flowcharts and block diagrams may represent a module, a program segment, or just a piece of code, and the modules, program segments, and codes are all executable instructions for implementing the specified logical functions. It should also be noted that the executable instructions for implementing the specified logical functions can be recombined to generate new modules and program segments. Therefore, the boxes in the accompanying drawings and the order of the boxes are only used to better illustrate the processes and steps of the embodiments, and should not be used as a limitation to the invention itself.

[0075] Figure 3 Shows a flowchart of a method for image processing according to an embodiment of the present invention; Figure 4a And 4b Respectively show schematic diagrams of the background pattern before and after processing according to an embodiment of the present invention; Figure 5a 、 5b And 5c respectively show the original image of the target image, the image after preprocessing, and the image after determining the exposure area according to an embodiment of the present invention; Figure 6a 、 6b And 6c respectively show the original image of the target image, the image after preprocessing, and the image after determining the fingerprint area according to an embodiment of the present invention.

[0076] In step S1, a background pattern image is collected within a predetermined area. In this step, for example, an image acquisition module is used to collect a background pattern image within a predetermined area, and the predetermined area is an image acquisition area set in the image acquisition module, such as a glass cover plate with a light source. However, in the image acquisition module, the light source illumination often cannot cover the entire screen. As Figure 4a Shown, the light source of the collected background pattern image is uneven, gradually becoming darker from the center of the light source to the edge, and the edge is relatively blurred, and there are even other interference noises mixed together. If the background pattern image is not processed, then during the subsequent acquisition and recognition process of the target image, the background pattern image and the interference noises will interfere with the recognition, resulting in an increase in the false acceptance rate (FAR) and false rejection rate (FRR) of the image. Therefore, it is necessary to perform certain processing on the background pattern image to be processed.

[0077] In step S2, the aperture edge of the background pattern image is determined to divide the predetermined area into an effective area and an ineffective area.

[0078] In this step, determining the aperture edge of the shading image includes: obtaining the first matrix of the shading image; determining the threshold of the aperture edge according to the change of the signal intensity (which can represent the brightness) on the diagonal of the shading image; performing binary segmentation on the first matrix according to the threshold to obtain a 0-1 matrix, where the 0 elements in the 0-1 matrix represent invalid regions and the 1 elements represent valid regions.

[0079] In this embodiment, determining the threshold of the aperture edge includes: performing median filtering on the first matrix to obtain a second matrix; calculating the normalized vectors and first-order gradient vectors from the center of the second matrix to the four corners; selecting multiple elements in the second matrix that meet the preset conditions, and setting the median or mean of the multiple elements that meet the preset conditions as the threshold. Optionally, the preset conditions include: the element is located on the diagonal of the second matrix; the ratio of the value of the element to the value of the center of the second matrix does not exceed a first value; the gradient of the element to its adjacent third point does not exceed a second value. Among them, the ratio of the value of the element to the value of the center of the second matrix not exceeding the first value indicates that the signal intensity of the element is within a preset range, and the gradient of the element to its adjacent third point not exceeding the second value indicates that the signal intensity change rate of the element is within a preset range. It should be understood that the signal intensity change rate of the element can also be judged by calculating the gradient of the element to its adjacent first point, second point, fourth point or more points to determine whether it is within the preset range.

[0080] As an example, obtain the first matrix A n*n (i.e., the pixel matrix) of the shading image, use a ij to represent the signal value of the i-th row and j-th column of the shading image, and use a center to represent the brightness of the 20*20 area at the center of the image; perform median filtering on the first matrix A n*n to obtain a second matrix B n*n , and the change rule of the second matrix is clearer than that of the first matrix; then, calculate the normalized vectors M1, M2, M3, M4 from the center of the second matrix B n*n to the four corners, and the first-order gradient vectors N1, N2, N3, N4. Optionally, when calculating the gradient with the adjacent third point, the obtained first-order gradient vector will be more stable; select multiple elements in the second matrix that meet the preset conditions, and set the median or mean of the multiple elements that meet the preset conditions as the threshold. Specifically, M 1ii = b ii / mean(b center ), N 1ii = M 1i+3,i+3 - M 1ii , where i = 1 to floor(n / 2), and the range of i is from 1 to the integer part of n / 2. For example, based on empirical data, set the first value to 0.4 and the second value to 0.05. If the element bii The ratio of the value to the value at the center of the second matrix satisfies M 1ii ≤0.4, and the gradient of element b ii with its third adjacent point satisfies N 1ii ≤0.05, then the first threshold δ1 = a ii , where a ii is the element in the first matrix corresponding to element b ii ; and so on, the second threshold δ2, the third threshold δ3, and the fourth threshold δ4 can be obtained; optionally, one of the first threshold δ1, the second threshold δ2, the third threshold δ3, and the fourth threshold δ4 is selected as the threshold δ, or the finally obtained threshold δ takes the median or mean value greater than 0 among the above four thresholds.

[0081] In this example, after obtaining the threshold δ, the pixel matrix A of the shading image is binarized by the threshold δ n*n to obtain a 0-1 matrix mask1 with the same size as the shading.

[0082] In some alternative embodiments, it further includes: obtaining the neighborhood of each element in the 0-1 matrix, and updating the elements with a proportion lower than a predetermined value in the neighborhood to another element. For example, if most of the elements (e.g., 80% of the elements) in the neighborhood of a 0 element in the 0-1 matrix are 1 elements, then update the 0 element to a 1 element. That is, after obtaining the 0-1 matrix mask1, a basic neighborhood judgment is made on each element in the 0-1 matrix mask1 to eliminate possible isolated points (elements with too low a proportion of the same value in the neighborhood), as shown in Figure 4b the figure.

[0083] Optionally, when the shading is updated each time, the 0-1 matrix mask1 area is also updated, so as to adapt to problems such as firmware offset caused by long-term use of the device.

[0084] In step S3, a target image is collected within a predetermined area. The target image is, for example, a fingerprint image. For example, in optical fingerprint acquisition technology, when a finger is placed on the fingerprint acquisition area, a light-emitting layer such as an Organic Light-Emitting Diode (OLED) under the glass cover plate serves as the light source provided during fingerprint acquisition. Utilizing the difference in refractive index between the finger's skin and the refractive index in the medium (such as air), fingerprint acquisition is performed in the area where total internal reflection occurs in the medium but not in the finger's skin, and the light signal is converted into an electrical signal through a conversion layer on the side of the light-emitting layer facing the glass cover plate, and the image information is uploaded to the image processing module.

[0085] In step S4, the target images within the invalid region are updated to background images, and the target images within the valid region are retained. In this step, as Figure 4b shown, since the target images within the invalid region are removed and the target images within the valid region are retained, the introduction of target images with incorrect features within the invalid region is avoided, thereby improving the recognition rate of the target images and facilitating the reduction of the false acceptance rate and false recognition rate in the subsequent image recognition process.

[0086] In step S5, according to the signal intensity of the shading image and / or the target image, an overexposed region is determined. Among them, when the signal intensity of a partial peripheral region of the shading image and / or the target image exceeds the central intensity, the partial peripheral region is determined as the overexposed region, and the target images within the overexposed region are deleted.

[0087] In this step, the area and position where the original image (as Figure 5a shown) exceeds the strong light threshold are statistically analyzed, and an overexposed region marking model is established. The signal value of the original image should decrease from the center of the aperture outward, approximately following a normal distribution; while the signal intensity of the overexposed region exceeds the signal intensity of the aperture center, and the signal intensity value of the overexposed region is basically stable. The overexposed region will bring incorrect features in the subsequent recognition of the target image, as Figure 5b shown; in this embodiment, whether there is an overexposed region can be judged through the distribution of the signal quantity, and the overexposed region mask2 is marked, as Figure 5c shown.

[0088] As an example, use a ij to represent the signal value of the i-th row and j-th column of the image, and a center to represent the brightness of the 20*20 region at the center of the image; then the determination method of the overexposed region mask2 can be expressed as: a ij >a center , and mean(Neighb(a ij ))>a center . If a ij satisfies the above conditions, then mask a ij =0, that is, the target images within the overexposed region are deleted, or the target images within the overexposed region are unified as background images.

[0089] In step S6, after collecting the target images within the predetermined region, the fingerprint region is determined, and the target images within the fingerprint region are retained.

[0090] In this step, preprocess the target image; count the directional intensity and the number of ridge-valley lines in the neighborhood of each element in the first matrix of the target image, and take the set of multiple elements with directional intensity and multiple ridge-valley lines in the neighborhood as the fingerprint area; and retain the target image that is within the valid area and within the fingerprint area. Optionally, the preprocessing includes: obtaining a target image with the background removed based on the target image and the background image, as Figure 6b shown.

[0091] As an example, the method for counting the directional intensity and the number of ridge-valley lines includes: calculating the variance of the vectors with a fixed length in eight directions in the neighborhood of each element, taking the direction with the minimum variance as the direction of the element, and the directional intensity is the most frequently occurring direction in the neighborhood divided by the area of the neighborhood; performing smoothing processing on the neighborhood vectors perpendicular to the direction, and calculating the number of sign changes of the first-order gradient to obtain the number of ridge-valley lines.

[0092] In this example, the original image is as Figure 6a shown. For example, use a ij to represent the signal value at the i-th row and j-th column of the image, and a ij_neighbor represents the 9*9 neighborhood of a ij . It should be understood that the size of the neighborhood can be changed according to actual needs. First, calculate the direction field image_dirfield of all points in the image: simplify all directions into eight direction labels 1-8, calculate the variance of the vectors with a fixed length of 5 in eight directions in the neighborhood of each point, and take the direction with the minimum variance as the direction of the point (the minimum variance indicates that the data in this direction fluctuates less and the probability of ridge lines and valley lines is high).

[0093] For example, the variance of the horizontal direction 1 of point a ij is: dir1 = [image0(i, j - 4), image0(i, j - 2), image0(i, j), image0(i, j + 2), image0(i, j + 4)]; var1 = Σ(mean(dir1) - dir1(i)) 2 .

[0094] The directional intensity δ1 in the neighborhood: After obtaining the direction field image_dirfield, the directional intensity δ1 of a ij is the most frequently occurring direction among the eight directions in a ij_neighbor divided by the neighborhood size^2.

[0095] The number of ridge-valley lines δ2: Take the 1*21 neighborhood vector vector ij perpendicular to the texture direction of a ij in image_new (the neighborhood size is selected according to the image DPI), and perform smoothing processing to obtain smooth(vector ij) Calculate the number of sign changes in the first-order gradient, which is the number of ridge-valley lines δ2 in the region.

[0096] In this example, if δ 1ij ≤ 0.25 and δ 2ij ≤ 1; then mask3 ij = 0. That is, the image located outside the fingerprint area is deleted, as Figure 6c shown.

[0097] Optionally, when selecting the neighborhood, the neighborhood size needs to be determined according to the image size (Dots Per Inch, DPI) to ensure that the neighborhood can contain 4 - 5 ridge-valley lines; the judgment thresholds of δ1 and δ2 are, for example, selected within the theoretical range according to the test results (using the fingerprint area in the image processing process and conducting recognition test and comparison verification with the fingerprint data test set).

[0098] Optionally, after obtaining the target image in any of the above steps, it further includes: identifying, testing, comparing, or verifying the target image located within the valid area and / or fingerprint area.

[0099] The image processing method according to the present invention can be deployed on a single or multiple servers. For example, different modules can be respectively deployed on different servers to form dedicated servers. Or, the same functional units, modules, or systems can be distributedly deployed on multiple servers to reduce the load pressure. The servers include, but are not limited to, multiple PC machines, PC servers, blade servers, supercomputers, etc. within the same local area network and connected through the Internet.

[0100] Figure 7 Shows a system block diagram of image processing according to an embodiment of the present invention.

[0101] As Figure 7 shown, the image processing system 100 includes an image acquisition module 110 and an image processing module 120.

[0102] The image acquisition module 110 respectively acquires the background image and the target image within a predetermined area; the image processing module 120 determines the aperture edge of the background image to divide the predetermined area into a valid area and an invalid area, updates the target image located within the invalid area to the background image, and retains the target image located within the valid area.

[0103] As an example, taking fingerprint image acquisition as an example, the image acquisition module 110 (such as a panel) located in the display panel includes, from the contact surface with the acquisition object downward, for example, in sequence: a glass cover plate, an optical path layer, a conversion layer, and a light-emitting layer. Among them, the optical path layer is located below the glass cover plate and can be a medium that transmits light (such as air, a collimator, and water, etc.) to form a light transmission path; the conversion layer is located below the optical path layer and can be a device for photoelectric conversion (such as a photodiode, etc.) to realize photoelectric conversion to form fingerprint information; the light-emitting layer is located below the conversion layer and can be a light-emitting device (such as an LED lamp, a light-emitting diode, etc.) to provide a light source for fingerprint acquisition.

[0104] When fingerprint acquisition is performed, the light-emitting layer emits light, which passes through the optical path layer and reaches the glass cover plate. There is an acquisition object (such as a finger) on it. Since the light reflection coefficients of the finger valleys and ridges are different, fingerprint information is formed. The reflected light can pass through the optical path layer again and reach the conversion layer to convert the reflected light into an electrical signal, and at the same time, the electrical signal is transmitted to the image processing module 120.

[0105] As an example, the image processing module 120 includes a first unit 121, a second unit 122, and a third unit 123. The first unit 121 obtains the first matrix of the background image; the second unit 122 determines the threshold of the aperture edge according to the change in the signal intensity on the diagonal of the background image; the third unit 123 performs binary segmentation on the first matrix according to the threshold to obtain a 0-1 matrix. The 0 elements in the 0-1 matrix represent invalid regions, and the 1 elements represent valid regions.

[0106] In this example, the second unit 122 includes: a filtering unit that performs median filtering on the first matrix to obtain a second matrix; a calculation unit that calculates the normalized vectors and first-order gradient vectors from the center of the second matrix to the four corners respectively; a selection unit that selects multiple elements in the second matrix that meet the preset conditions and sets the median or average value of the multiple elements that meet the preset conditions as the threshold.

[0107] Optionally, the preset conditions of the selection unit include: the element is located on the diagonal of the second matrix; the ratio of the value of the element to the value of the center of the second matrix does not exceed a first value; the gradient of the element to its adjacent third point does not exceed a second value.

[0108] Optionally, the image processing module 120 further includes a fourth unit 124. The fourth unit 124 obtains the neighborhood of each element in the 0-1 matrix and updates the elements with a ratio lower than a predetermined value in the neighborhood to another element.

[0109] Optionally, the image processing module 120 further includes a fifth unit 125. The fifth unit 125 determines an overexposed area according to the signal intensity of the background image and / or the target image. When the signal intensity of a partial peripheral area of the background image and / or the target image exceeds the central intensity, the partial peripheral area is determined as the overexposed area, and the target image located within the overexposed area is deleted.

[0110] Optionally, the image processing module 120 further includes a sixth unit 126. The sixth unit 126 includes: a preprocessing unit that preprocesses the target image after collecting the target image within a predetermined area; a statistical unit that statistically calculates the directional intensity and the number of ridge-valley lines within the neighborhood of each element in the first matrix of the target image, and uses the set of multiple elements with directional intensity and multiple ridge-valley lines within the neighborhood as the fingerprint area; and a processing unit that retains the target image located within the effective area and within the fingerprint area.

[0111] Optionally, the statistical unit calculates the variance of vectors with a fixed length in eight directions within the neighborhood of each element, uses the direction with the minimum variance as the direction of the element, calculates the directional intensity as the direction that appears most frequently within the neighborhood divided by the area of the neighborhood, and the statistical unit performs smoothing processing on the neighborhood vectors perpendicular to the direction and calculates the number of first-order gradient sign changes to obtain the number of ridge-valley lines.

[0112] Optionally, the preprocessing unit obtains a background-removed target image according to the target image and the background image.

[0113] As used herein, the term "module" may refer to, be part of, or include the following: an Application Specific Integrated Circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or grouped) and / or a memory (shared, dedicated, or grouped) that executes one or more software or firmware programs, combinational logic circuits, and / or other suitable components that provide the described functionality.

[0114] Those skilled in the art can understand that each module or unit of the image processing system according to the present invention can be implemented by hardware, firmware, or software. Software includes, for example, encoded programs formed in various programming languages such as JAVA, C / C++ / C#, SQL, etc. Although the steps and the order of the steps of the embodiments of the present invention are given in the method and the method diagram, the executable instructions for implementing the specified logical functions of the steps can be recombined to generate new steps. The order of the steps should not be limited only to the order of the steps in the method and the method diagram, and can be adjusted at any time according to the needs of the function. For example, some of the steps can be executed in parallel or in the reverse order.

[0115] Meanwhile, those of ordinary skill in the art can realize that, in combination with the structures and methods of the examples described in the embodiments disclosed herein, different configuration methods or adjustment methods can be used to implement the described functions by reasonably deforming each structure or the structure. However, such implementation should not be considered to exceed the scope of this application. Moreover, it should be understood that the connection relationships between the various components of the amplifier in the aforementioned figures in the embodiments of this application are illustrative examples and do not impose any limitations on the embodiments of this application.

[0116] The foregoing are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, various modifications and changes can be made to the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for image processing, characterized in that, Comprising: Collecting a background pattern image within a predetermined area; Determining the aperture edge of the background pattern image to divide the predetermined area into a valid area and an invalid area; Collecting a target image within the predetermined area; And Updating the target image located in the invalid area to a background image and retaining the target image located in the valid area, wherein determining the aperture edge of the background pattern image includes: Obtaining a first matrix of the background pattern image; Determining a threshold of the aperture edge according to the change of signal intensity on the diagonal of the background pattern image; Performing binary segmentation on the first matrix according to the threshold to obtain a 0-1 matrix, where 0 elements in the 0-1 matrix represent the invalid area and 1 elements represent the valid area.

2. The method according to claim 1, characterized in that Determining the threshold of the aperture edge includes: Performing median filtering on the first matrix to obtain a second matrix; Calculating the normalized vector and the first-order gradient vector from the center of the second matrix to the four corners respectively; Selecting a plurality of elements in the second matrix that meet a preset condition and setting the median or mean value of the plurality of elements that meet the preset condition as the threshold.

3. The method according to claim 2, wherein The preset condition includes: The element is located on the diagonal of the second matrix; The ratio of the value of the element to the value of the center of the second matrix does not exceed a first value; The gradient of the element to its adjacent third point does not exceed a second value.

4. The method according to claim 2, characterized in that, Also comprising: Obtaining the neighborhood of each element in the 0-1 matrix and updating the elements with a ratio lower than a predetermined value in the neighborhood to another element.

5. The method according to claim 1, characterized in that, Also comprising: Determining an overexposed area according to the signal intensity of the background pattern image and / or the target image, wherein when the signal intensity of a partial peripheral area of the background pattern image and / or the target image exceeds the central intensity, the partial peripheral area is determined as the overexposed area and the target image located in the overexposed area is deleted.

6. The method according to claim 1, characterized in that, Also comprising: Identifying, testing, comparing or verifying the target image located in the valid area.

7. The method according to claim 1, wherein After collecting the target image within the predetermined area, it also includes: Performing preprocessing on the target image; Counting the direction intensity and the number of ridge-valley lines in the neighborhood of each element in the first matrix of the target image, and taking the set of a plurality of elements with direction intensity and multiple ridge-valley lines in the neighborhood as the fingerprint area; and Retaining the target image located in the valid area and within the fingerprint area.

8. The method according to claim 7, wherein Calculating the variance of vectors with a fixed length in eight directions in the neighborhood of each element, and taking the direction with the minimum variance as the direction of the element, and the direction intensity is the number of times the most frequently occurring direction in the neighborhood divided by the area of the neighborhood.

9. The method according to claim 7, characterized in that, Smoothing the neighborhood vector perpendicular to the direction and calculating the number of times of the first-order gradient sign change to obtain the number of ridge-valley lines.

10. The method according to claim 7, wherein The preprocessing includes: obtaining a target image without the background pattern according to the target image and the background pattern image.

11. An apparatus for image processing, characterized in that, Comprising: An image processing module, after receiving the background pattern image and the target image within a predetermined area, determines the aperture edge of the background pattern image to divide the predetermined area into a valid area and an invalid area, updates the target image located in the invalid area to a background image, and retains the target image located in the valid area. Wherein, the image processing module includes: A first unit that obtains a first matrix of the background pattern image; A second unit that determines a threshold of the aperture edge according to the change in the signal intensity on the diagonal of the background pattern image; A third unit that performs binary segmentation on the first matrix according to the threshold to obtain a 0-1 matrix, where the 0 elements in the 0-1 matrix represent the invalid area, and the 1 elements represent the valid area.

12. The device according to claim 11, characterized in that, The second unit includes: A filtering unit that performs median filtering on the first matrix to obtain a second matrix; A calculation unit that calculates the normalized vectors and the first-order gradient vectors from the center of the second matrix to the four corners respectively; A selection unit that selects a plurality of elements in the second matrix that meet a preset condition, and sets the median or mean value of the plurality of elements that meet the preset condition as the threshold.

13. The device according to claim 12, characterized in that, The preset condition of the selection unit includes: The element is located on the diagonal of the second matrix; The ratio of the value of the element to the value of the center of the second matrix does not exceed a first value; The gradient of the element to its adjacent third point does not exceed a second value.

14. The device according to claim 11, characterized in that, It further includes: A fourth unit that obtains the neighborhood of each element in the 0-1 matrix and updates the elements with a ratio lower than a predetermined value in the neighborhood to another element.

15. The device according to claim 11, characterized in that, The image processing module further includes a fifth unit that determines an overexposed area according to the signal intensity of the background pattern image and / or the target image. Wherein, when the signal intensity of a partial peripheral area of the background pattern image and / or the target image exceeds the central intensity, the partial peripheral area is determined as the overexposed area, and the target image located in the overexposed area is deleted.

16. The device according to claim 11, wherein The image processing module further includes a sixth unit, and the sixth unit includes: A preprocessing unit that preprocesses the target image after collecting the target image within the predetermined area; A statistical unit that statistically calculates the direction intensity and the number of ridge-valley lines in the neighborhood of each element in the first matrix of the target image, and takes the set of a plurality of elements with direction intensity and multiple ridge-valley lines in the neighborhood as the fingerprint area; and A processing unit that retains the target image located in the valid area and within the fingerprint area.

17. The device according to claim 16, wherein: The statistical unit calculates the variance of vectors with a fixed length in eight directions in the neighborhood of each element, takes the direction with the smallest variance as the direction of the element, and the direction intensity is the number of times the most frequently occurring direction in the neighborhood divided by the area of the neighborhood. The statistical unit performs smoothing processing on the neighborhood vectors perpendicular to the direction and calculates the number of times of the first-order gradient sign change to obtain the number of ridge-valley lines.

18. The device according to claim 16, characterized in that, The preprocessing unit obtains a target image without the background pattern according to the target image and the background pattern image.

19. An image processing system, characterized in that, It includes: An image acquisition module that respectively acquires a background pattern image and a target image within a predetermined area; and An image processing module as described in any one of claims 11-18, determines the aperture edge of the background pattern image to divide the predetermined area into a valid area and an invalid area, updates the target image located within the invalid area to a background image, and retains the target image located within the valid area.

Images