WO2018152977A1 - Procédé, terminal de réduction de bruit d'image, et support de stockage informatique - Google Patents

Procédé, terminal de réduction de bruit d'image, et support de stockage informatique Download PDF

Info

Publication number
WO2018152977A1
WO2018152977A1 PCT/CN2017/085512 CN2017085512W WO2018152977A1 WO 2018152977 A1 WO2018152977 A1 WO 2018152977A1 CN 2017085512 W CN2017085512 W CN 2017085512W WO 2018152977 A1 WO2018152977 A1 WO 2018152977A1
Authority
WO
WIPO (PCT)
Prior art keywords
noise reduction
current frame
lens
image
frame image
Prior art date
Application number
PCT/CN2017/085512
Other languages
English (en)
Chinese (zh)
Inventor
张文娜
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2018152977A1 publication Critical patent/WO2018152977A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/80—Camera processing pipelines; Components thereof
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/80—Camera processing pipelines; Components thereof
    • H04N23/81—Camera processing pipelines; Components thereof for suppressing or minimising disturbance in the image signal generation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H04N5/21—Circuitry for suppressing or minimising disturbance, e.g. moiré or halo
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G06T2207/20032—Median filtering

Definitions

  • the present disclosure relates to image processing technologies in the field of electronic applications, and more particularly to an image noise reduction method and terminal and computer storage medium.
  • Image noise image noise
  • image noise reduction technology has become a problem that cannot be ignored.
  • the image denoising algorithm for image denoising is used to denoise the image, such as median filtering algorithm, adaptive Wiener filtering algorithm, wavelet denoising algorithm and other image denoising algorithms.
  • the image is uniformly processed by an image denoising algorithm according to the noise characteristics and the noise size of the entire image.
  • noise reduction processing due to the imaging characteristics of the lens, the brightness of the image is attenuated from the center of the image to the periphery of the image.
  • the image after noise reduction has the following two extremes: (1) ), the image uses a strong noise reduction processing method, the image has a better noise reduction effect around the image, and the image is clean, but the image center detail sacrifices a lot, the image resolution decreases, and the image is blurred; (2) the image adopts a weaker drop.
  • the noise processing method retains more details of the image as a whole, but the noise around the image is large and the image is not clean. In short, when the image is denoised by the existing noise reduction processing method, the image quality will be insufficient.
  • Embodiments of the present invention are directed to an image denoising method, a terminal, and a computer storage medium, which are capable of differentiating image denoising processing, and improving image quality after image denoising.
  • An embodiment of the present invention provides an image noise reduction method, including:
  • the determining, according to the first lens parameter, the first noise reduction system of the current frame image Number including:
  • N third noise reduction coefficients respectively corresponding to the N first images, and determining the N third noise reduction coefficients as the first corresponding to the current frame image a noise reduction coefficient, wherein the N first images are in one-to-one correspondence with the N third noise reduction coefficients.
  • the determining the N third noise reduction coefficients respectively corresponding to the N first images according to the first lens uniformity characteristic parameter and the first lens parameter comprises:
  • the performing noise reduction processing on the current frame image according to the first noise reduction coefficient comprises:
  • the determining, according to the first lens parameter, the first noise reduction coefficient of the current frame image includes:
  • the first lens parameter comprises: a first lens uniformity characteristic parameter and a first lens uniformity compensation parameter
  • the embodiment of the invention provides a terminal, including:
  • An acquiring unit configured to acquire a current frame image; and acquire a first lens parameter corresponding to the current frame image;
  • a determining unit configured to determine a first noise reduction coefficient of the current frame image according to the first lens parameter
  • noise reduction unit configured to perform noise reduction processing on the current frame image according to the first noise reduction coefficient.
  • the terminal further includes: a dividing unit;
  • the dividing unit is configured to divide the current frame image into N first images corresponding to N regions, where N is a natural number greater than 1;
  • the determining unit is configured to determine, according to the first lens parameter, N third noise reduction coefficients respectively corresponding to the N first images, and determine the N third noise reduction coefficients as the current The first noise reduction coefficient corresponding to the frame image, wherein the N first images are in one-to-one correspondence with the N third noise reduction coefficients.
  • the terminal further includes: a calculating unit and a converting unit;
  • the calculating unit is configured to calculate a first distance between a center point of each first image and a center point of the current frame image
  • the determining unit is further configured to determine, according to the first distance and the first lens parameter, each first gain value corresponding to each of the first images;
  • the converting unit is configured to convert the third noise reduction coefficient corresponding to each of the first images according to each of the first gain values until the N third noise reduction coefficients are obtained.
  • the noise reduction unit is configured to separately perform noise reduction processing on the N first images according to the N third noise reduction coefficients, where the N third noise reductions are performed.
  • the coefficients are in one-to-one correspondence with the N first images.
  • the terminal further includes: a determining unit;
  • the acquiring unit is further configured to acquire a second lens uniformity characteristic parameter corresponding to the image of the previous frame;
  • the determining unit is configured to determine whether there is a difference between the first lens uniformity characteristic parameter and the second lens uniformity characteristic parameter according to a preset tolerance threshold;
  • the first lens parameter includes: a first lens uniformity Characteristic parameters and first lens uniformity compensation parameters;
  • the determining unit is configured to determine, according to the first lens uniformity compensation parameter, a first noise reduction coefficient of the current frame image, or determine if there is no difference, if the difference is determined And determining, by the second noise reduction coefficient corresponding to the image of the previous frame, the second noise reduction coefficient as the first noise reduction coefficient.
  • the embodiment of the invention further provides a computer storage medium, wherein the computer storage medium stores computer executable instructions, and the computer executable instructions are used to perform at least one of the foregoing methods.
  • An embodiment of the present invention provides an image denoising method and a terminal, which acquires a current frame image, acquires a first lens parameter corresponding to a current frame image, and determines a first noise reduction coefficient of the current frame image according to the first lens parameter;
  • the first noise reduction coefficient performs noise reduction processing on the current frame image.
  • FIG. 1 is a flowchart 1 of an image noise reduction method according to an embodiment of the present invention
  • FIG. 2 is a schematic structural diagram of a pixel arrangement manner according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of CMOS photon sensing according to an embodiment of the present invention.
  • FIG. 4 is an LS graph of an exemplary 3-frame image according to an embodiment of the present invention.
  • FIG. 5 is an image diagram of an exemplary original RAW data according to an embodiment of the present invention.
  • FIG. 6 is an exemplary LSC characteristic curve corresponding to FIG. 5 according to an embodiment of the present invention.
  • FIG. 7 is a second flowchart of an image noise reduction method according to an embodiment of the present invention.
  • FIG. 8 is a schematic structural diagram 1 of a terminal according to an embodiment of the present disclosure.
  • FIG. 9 is a schematic structural diagram 2 of a terminal according to an embodiment of the present disclosure.
  • FIG. 10 is a schematic structural diagram 3 of a terminal according to an embodiment of the present disclosure.
  • FIG. 11 is a schematic structural diagram 4 of a terminal according to an embodiment of the present disclosure.
  • FIG. 12 is a schematic structural diagram 5 of a terminal according to an embodiment of the present invention.
  • An embodiment of the present invention provides an image noise reduction method. As shown in FIG. 1 , the method may include:
  • the terminal when the terminal photographs the target object by using the photographing function, when the terminal starts the photographing function, the terminal can preview the image corresponding to the target object through the display screen, so the terminal can acquire the current frame image.
  • the current frame image refers to a preview image corresponding to the target object acquired by the terminal in the current frame.
  • the terminal acquires a current frame image corresponding to the target object, and the current frame image is the current frame, in the process of the image capturing or photographing by the terminal.
  • Raw RAW data In the embodiment of the present invention, the acquisition of the original RAW data of the current frame is implemented by an image sensor provided in the terminal. Since the image sensor can acquire the preview image of the target object in real time, that is, the different frame preview images are always obtained during the preview, the embodiment of the present invention describes the current frame image acquired by the image sensor, and the principle of each frame image is the same. .
  • the image sensor may include a charge coupled device (CCD) image sensor and a Color Filter Array (CMOS) filter element of a complementary metal oxide semiconductor (CMOS) image sensor.
  • CMOS complementary metal oxide semiconductor
  • the image sensor is a two-dimensional matrix composed of photosensitive elements densely arranged in two directions, forming a Bayer mode RGB pixel dot arrangement, and (a), (b), which are commonly shown in FIG. (c), (d) Four Bayer patterns are arranged, and each CCD corresponds to one pixel. Among them, R induces red light, G senses green light, B induces blue light, and in Bayer mode, G is twice as large as R and B (because human eyes are more sensitive to green).
  • each CCD or CMOS in the above two-dimensional matrix is only used to sense the energy of photons, and each CCD or CMOS generates a corresponding proportion of charges due to the intensity of incoming light, and each CCD or CMOS will The charge information is collected and amplified and stored.
  • the CMOS sensor photons shown in the CMOS photon sensing diagram shown in FIG. 3 finally store the corresponding proportion of charges (blocks shown in FIG. 3), and RAW records each.
  • the charge value of the pixel position, which does not record any color information, that is, the original RAW data of the target object is only a grayscale file.
  • the terminal can perform shooting of the target object through the camera disposed thereon, and in the shooting process, due to the imaging characteristics of the lens, as shown in FIG. 4, different frame images acquired by the terminal (FIG. 4 is
  • the brightness of the three brightness curves of the three-frame image has different degrees of attenuation from the center of the image to the ideal brightness curve around the image.
  • lens uniformity that is, a lens shading curve.
  • the original RAW data when presented in the form of an image, there is no color in a grayscale image mode. In the image formed by the original RAW data as shown in FIG. 5, the image formed by the original RAW data can be seen as black and white. And there is also a mosaic phenomenon (ie noise).
  • the terminal needs to perform image processing (ie, image signal processing) on the RAW data for luminance compensation and noise reduction (image signal processing).
  • image processing ie, image signal processing
  • the terminal can perform image signal processing (ISP, Image Signal Processing) through an image processor.
  • the operation of the terminal on the target object in the embodiment of the present invention may be understood as a process in which an image processing application in the terminal applies a preview image of the image to be processed.
  • the terminal When the terminal performs the image preview, after acquiring the current frame image, the RAW data of the current frame image acquired by the terminal needs to perform image signal processing. Specifically, the terminal may calibrate the brightness attenuation of the current frame image, and adopt a uniform lens.
  • the LSS Lins Shading Correction
  • the first lens parameter includes: a first lens uniformity characteristic parameter and a first lens uniformity compensation parameter.
  • the brightness compensation is performed on different frame images (which can be understood as current frame images at different times), that is, the brightness is performed.
  • Gain processing as shown in FIG. 6, the terminal has a small brightness gain value for the image center of the current frame image (for example, a gain of 1 time), and the gain value around the image of the current frame image is large (for example, a gain of 2 times). Therefore, when the terminal performs ISP processing, the terminal can acquire the first lens uniform characteristic parameter of the current frame image before the LSC is performed, and the first lens uniformity compensation parameter after the LSC is performed. Specifically, the terminal can be obtained by using a lens shading characteristic curve and an LSC characteristic curve (as shown in FIGS. 5 and 6).
  • the terminal in the LSC characteristic curve, can set the ratio of the uniformity of the periphery of the image and the ratio of the uniformity of the image center and the position of the curve inflection point by the parameter, that is, the image after the calibration is not necessarily the central brightness and the four corners.
  • the brightness is exactly the same, but it can also be adjusted. Therefore, the terminal needs to record the lens shading characteristic curve before and after calibration, and obtain the LSC characteristic curve by comparing the lens shading characteristic curve; or directly obtain the LSC characteristic curve from the ISP LSC algorithm, thereby making
  • the first lens uniformity characteristic parameter and the first lens uniformity compensation parameter corresponding to the current frame image are obtained by the terminal in the embodiment of the present invention.
  • the terminal may determine the first noise reduction coefficient of the current frame image in consideration of the first lens parameter.
  • the process of determining, by the terminal, the first noise reduction coefficient of the current frame image according to the first lens parameter may be: the terminal dividing the current frame image into N first images corresponding to the N regions, Wherein, N is a natural number greater than 1; the terminal determines N third noise reduction coefficients corresponding to the N first images according to the first lens parameter (specifically, the first lens uniformity compensation parameter), and the N third descending The noise coefficient is determined as a first noise reduction coefficient corresponding to the current frame image, wherein the N first images are in one-to-one correspondence with the N third noise reduction coefficients.
  • the process of determining, according to the first lens parameter (specifically, the first lens uniformity compensation parameter), the N third noise reduction coefficients corresponding to the N first images respectively is: the terminal calculates the center point of each first image. a first distance from a center point of the current frame image; then, the terminal determines each of the first distance and the first lens parameter Each first gain value corresponding to the first image; finally, the terminal may convert the third noise reduction coefficient corresponding to each first image according to each first gain value until N third noise reduction coefficients are obtained.
  • the LSC characteristic curve of the current frame image indicates the distance between the gain value and the different position of the current frame image from the center of the current frame image.
  • the terminal will be current. After the frame image is divided into N regions, the position of the center point of each region is used as the position of each region in the current frame image. Therefore, the terminal can compensate for the first lens uniformity represented by the first distance and the LSC characteristic curve. a parameter, determining each first gain value corresponding to each first image corresponding to each region, the terminal may obtain N first gain values, each first gain value corresponding to each first image .
  • the terminal divides the image into m*n regions, calculates a first distance between a center point of each region and a center point of the current frame image, and finds a corresponding value in the LSC curve of the current frame image by using a table lookup or a function calculation method.
  • Each first distance is incremented by each first gain value; as follows:
  • the terminal divides the entire image into m*n areas, and the specific areas are as follows:
  • Ration 0,0 ,Ration 0,1 ,Ration 0,2 ,...,Ration m-1,n-1 ,m and n can be set by themselves; calculate the center of the image center of the current frame image The first distance of the center of the area from the center of the image is recorded as: Dis.Ration 0,0 , Dis.Ration 0,1 , Dis.Ration 0,2 ,...,Dis.Ration m-1,n- 1;
  • the terminal finds the gain value added by each region in the LSC algorithm in the LSC curve of the current frame image according to the first distance: gain.Ration 0,0 ,gain.Ration 0,1 ,gain .Ration 0,2 ,...,gain.Ration m-1,n-1 .
  • the look-up table method first sample the LSC curve, obtain a two-dimensional chart (two-dimensional chart of distance and gain values), and then find the corresponding gain.Region in the chart according to each Dis.Region; When Dis.Region is in the middle of two DISs, linear interpolation is used to obtain the gain corresponding to the Dis.
  • the terminal converts the gain value into a noise reduction coefficient: ratio.Ration 0,0 , ratio.Ration 0,1 , ratio.Ration 0,2 ,..., ratio.Ration m-1,n-1 ;
  • the noise reduction level is set according to the degree of noise corresponding to the different gain, and a two-dimensional chart (a two-dimensional chart of gain and ratio) is generated, because the LSC can be used in different brightness environments.
  • Set different correction levels, so 2D charts can be divided into three cases: bright light, normal light and lowlight or more detailed classification.
  • the terminal can find the corresponding ratio.Region in the two-dimensional graph according to each gain.Region; when a ratio.Region is in the middle of the two ratios, the terminal uses linear interpolation to obtain the ratio corresponding to the gain.
  • the terminal can determine the first noise reduction coefficient of the current frame image according to the first lens uniformity compensation parameter.
  • Table 1 is a combination chart of a two-dimensional graph of distance and gain values and a two-dimensional graph of gain and ratio when the gain is 1 time.
  • the terminal may perform noise reduction processing on the current frame image according to the first noise reduction coefficient.
  • the terminal may separately perform noise reduction processing on the N first images according to the N third noise reduction coefficients, wherein the N third noise reduction coefficients are in one-to-one correspondence with the N first images.
  • the terminal may use the obtained first noise reduction coefficient to adjust the noise reduction level of each region corresponding to the current frame image, that is, determine a third noise reduction coefficient of each region, and the noise reduction level may be expressed as: noise. Ration 0,0 , noise.Ration 0,1 , noise.Ration 0,2 &noise.Ration m-1,n-1 .
  • the terminal performs noise reduction processing according to the noise reduction data corresponding to the noise reduction level.
  • the noise reduction coefficient of the current frame image may be adjusted according to the change of the previous frame image, thereby realizing the differentiation of different frame images. Image denoising processing is performed, and image quality is improved by image noise reduction.
  • An embodiment of the present invention provides an image noise reduction method. As shown in FIG. 7, the method may include:
  • the process of “the terminal acquires the current frame image” is the same as that of S101 in the foregoing embodiment. The description is consistent and will not be described here.
  • the process of “the terminal acquiring the first lens parameter corresponding to the current frame image” is consistent with the description of S102 in the foregoing embodiment, and details are not described herein again.
  • S204 Determine, according to a preset tolerance threshold, whether there is a difference between the first lens uniformity characteristic parameter and the second lens uniformity characteristic parameter, where the first lens parameter comprises: a first lens uniformity characteristic parameter and a first lens uniformity compensation parameter.
  • the terminal acquires a second lens uniformity characteristic parameter corresponding to the previous frame image, and determines the current frame according to the preset tolerance threshold, the second lens uniformity characteristic parameter, the first lens uniformity characteristic parameter, and the first lens uniformity compensation parameter.
  • the first noise reduction factor of the image is the first noise reduction factor of the image.
  • the terminal After acquiring the first lens uniformity characteristic parameter and the first lens uniformity compensation parameter (ie, the first lens parameter) corresponding to the current frame image, the terminal acquires the current frame image in real time, and The processing mode of each frame is the same. Therefore, when the terminal acquires the latest current frame image, the terminal can obtain the second lens uniformity characteristic parameter corresponding to the previous frame image in the same manner. In this way, the terminal can determine whether there is a difference between the previous frame image and the current frame image by using the first lens uniformity characteristic parameter, the second lens uniformity characteristic parameter, and the preset tolerance threshold.
  • the terminal may determine, according to the result of determining the difference, whether the first frame noise reduction coefficient is used to obtain the first frame noise reduction coefficient, or the noise reduction coefficient of the previous frame image is used as the first noise reduction coefficient. Noise reduction is performed on the current frame image.
  • the terminal may determine whether there is a difference between the first lens uniformity characteristic parameter and the second lens uniformity characteristic parameter according to the preset tolerance threshold. If it is determined that there is a difference, the terminal determines the first noise reduction coefficient of the current frame image according to the first lens uniformity compensation parameter.
  • the process of determining, by the terminal according to the first lens uniformity compensation parameter, the first noise reduction coefficient of the current frame image may be: the terminal divides the current frame image into N corresponding to the N regions. a first image, wherein N is a natural number greater than 1; the terminal determines N third noise reduction coefficients respectively corresponding to the N first images according to the first lens uniformity compensation parameter, and determines N third noise reduction coefficients as a first noise reduction coefficient corresponding to the current frame image, wherein the N first images are in one-to-one correspondence with the N third noise reduction coefficients.
  • the terminal determines, according to the first lens uniformity compensation parameter, N third noise reduction coefficients corresponding to the N first images respectively
  • the process is specifically: the terminal calculates a first distance between a center point of each first image and a center point of the current frame image; and then, the terminal determines, according to the first distance and the first lens uniformity compensation parameter, that each first image corresponds to Each of the first gain values; finally, the terminal may convert the third noise reduction coefficient corresponding to each of the first images according to each of the first gain values until N third noise reduction coefficients are obtained.
  • the LSC characteristic curve of the current frame image indicates the distance between the gain value and the different position of the current frame image from the center of the current frame image.
  • the terminal will be current. After the frame image is divided into N regions, the position of the center point of each region is used as the position of each region in the current frame image. Therefore, the terminal can compensate for the first lens uniformity represented by the first distance and the LSC characteristic curve. a parameter, determining each first gain value corresponding to each first image corresponding to each region, the terminal may obtain N first gain values, each first gain value corresponding to each first image .
  • the terminal divides the image into m*n regions, calculates a first distance between a center point of each region and a center point of the current frame image, and finds a corresponding value in the LSC curve of the current frame image by using a table lookup or a function calculation method.
  • Each first distance is incremented by each first gain value; as follows:
  • the terminal divides the entire image into m*n areas, and the specific areas are as follows:
  • Ration 0,0 ,Ration 0,1 ,Ration 0,2 ,...,Ration m-1,n-1 ,m and n can be set by themselves; calculate the center of the image center of the current frame image The first distance of the center of the area from the center of the image is recorded as: Dis.Ration 0,0 , Dis.Ration 0,1 , Dis.Ration 0,2 ,...,Dis.Ration m-1,n- 1 ;
  • the terminal finds the gain value added by each region in the LSC algorithm in the LSC curve of the current frame image according to the first distance: gain.Ration 0,0 ,gain.Ration 0,1 ,gain .Ration 0,2 ,...,gain.Ration m-1,n-1 .
  • the look-up table method first sample the LSC curve, obtain a two-dimensional chart (two-dimensional chart of distance and gain values), and then find the corresponding gain.Region in the chart according to each Dis.Region; When Dis.Region is in the middle of two DISs, linear interpolation is used to obtain the gain corresponding to the Dis.
  • the terminal converts the gain value into a noise reduction coefficient: ratio.Ration 0,0 , ratio.Ration 0,1 , ratio.Ration 0,2 ,..., ratio.Ration m-1,n-1 ;
  • the noise reduction level is set according to the degree of noise corresponding to the different gain, and a two-dimensional chart (a two-dimensional chart of gain and ratio) is generated, because the LSC can be used in different brightness environments.
  • Set different correction levels, so 2D charts can be divided into three cases: bright light, normal light and lowlight or more detailed classification.
  • the terminal can find the corresponding ratio.Region in the two-dimensional graph according to each gain.Region; when a ratio.Region is in the middle of the two ratios, the terminal uses linear interpolation to obtain the ratio corresponding to the gain.
  • the terminal can determine the first noise reduction coefficient of the current frame image according to the first lens uniformity compensation parameter.
  • Table 1 is a combination chart of a two-dimensional graph of distance and gain values and a two-dimensional graph of gain and ratio when the gain is 1 time.
  • the terminal may obtain the second lens uniformity compensation parameter corresponding to the previous frame image or obtain the second lens uniformity directly by the algorithm according to the second lens uniformity characteristic parameter.
  • the compensation parameter therefore, the terminal can also determine whether there is a difference between the previous frame image and the current frame image by using the first lens uniformity compensation parameter, the second lens uniformity compensation parameter, and the preset tolerance threshold.
  • the embodiment of the present invention does not limit the manner of determining whether there is a difference between the previous frame image and the current frame image.
  • the terminal may ignore the slight difference or difference of the brightness of the previous frame image and the current frame image. Therefore, setting the preset tolerance threshold to make the difference between the previous frame image and the current frame image is obvious. Case.
  • the preset tolerance threshold may be preset, or may be adjusted according to user requirements, and the preset tolerance threshold may be obtained through experimental results, and may not be set too large or too small, due to If the preset tolerance threshold is set too large, the new noise reduction coefficient will not be determined in real time when the characteristic curve corresponding to the adjacent two frames changes to a certain extent; if the preset tolerance threshold is set too small, the adjacent two frames of data A slight change between them will also re-update the noise reduction coefficient, making the image processing data larger and longer.
  • the terminal After acquiring the first lens parameter corresponding to the current frame image, the terminal acquires in real time The current frame image is processed in the same manner for each frame of the image. Therefore, when the terminal acquires the latest current frame image, the terminal can obtain the second lens uniformity corresponding to the previous frame image in the same manner. Characteristic parameters. In this way, the terminal can determine whether there is a difference between the previous frame image and the current frame image by using the first lens uniformity characteristic parameter, the second lens uniformity characteristic parameter, and the preset tolerance threshold. Then, the terminal may determine, according to the result of determining the difference, whether the first frame noise reduction coefficient is used to obtain the first frame noise reduction coefficient, or the noise reduction coefficient of the previous frame image is used as the first noise reduction coefficient. Noise reduction is performed on the current frame image.
  • the terminal may determine, according to the preset tolerance threshold, whether there is a difference between the first lens uniformity characteristic parameter and the second lens uniformity characteristic parameter. If it is determined that there is no difference, it indicates that the noise reduction coefficient currently used by the terminal, that is, the noise reduction coefficient of the image of the previous frame can be continuously used, that is, the terminal can acquire the second noise reduction coefficient corresponding to the image of the previous frame. And determining the second noise reduction coefficient as the first noise reduction coefficient.
  • the terminal may separately perform noise reduction processing on the N first images according to the N third noise reduction coefficients, wherein the N third noise reduction coefficients and the N first The images correspond one by one.
  • the terminal may use the obtained first noise reduction coefficient to adjust the noise reduction level of each region corresponding to the current frame image, that is, determine a third noise reduction coefficient of each region, and the noise reduction level may be expressed as: noise. Ration 0,0 , noise.Ration 0,1 , noise.Ration 0,2 &noise.Ration m-1,n-1 .
  • the terminal performs noise reduction processing according to the noise reduction data corresponding to the noise reduction level.
  • the noise reduction coefficient of the current frame image may be adjusted according to the change of the previous frame image, thereby realizing the differentiation of different frame images. Image denoising processing is performed, and image quality is improved by image noise reduction.
  • the embodiment of the present invention provides a terminal 1, which may include:
  • the acquiring unit 10 is configured to acquire a current frame image, and acquire a first lens parameter corresponding to the current frame image;
  • the determining unit 11 is configured to determine a first noise reduction coefficient of the current frame image according to the first lens parameter.
  • the noise reduction unit 12 is configured to perform noise reduction processing on the current frame image according to the first noise reduction coefficient.
  • the terminal 1 further includes: a determining unit 13.
  • the acquiring unit is further configured to acquire a second lens uniformity characteristic parameter corresponding to the image of the previous frame.
  • the determining unit 13 is configured to determine whether there is a difference between the first lens uniformity characteristic parameter and the second lens uniformity characteristic parameter according to a preset tolerance threshold, where the first lens parameter comprises: the first lens is uniform sexual characteristic parameters and first lens uniformity compensation parameters.
  • the determining unit 11 is specifically configured to: if it is determined that there is a difference, or if it is determined that there is no difference, acquire a second noise reduction coefficient corresponding to the image of the previous frame, and determine the second noise reduction coefficient Is the first noise reduction coefficient.
  • the terminal 1 further includes: a dividing unit 14.
  • the dividing unit 14 is configured to divide the current frame image into N first images corresponding to N regions.
  • the determining unit 11 is further configured to determine, according to the first lens parameter, N third noise reduction coefficients respectively corresponding to the N first images, and determine the N third noise reduction coefficients as The first noise reduction coefficient corresponding to the current frame image, wherein the N first images are in one-to-one correspondence with the N third noise reduction coefficients.
  • the terminal 1 further includes: a computing unit 15 and a converting unit 16.
  • the calculating unit 15 is configured to calculate a first distance between a center point of each first image and a center point of the current frame image.
  • the determining unit 11 is further configured to determine, according to the first distance and the first lens parameter, each first gain value corresponding to each of the first images.
  • the converting unit 16 is configured to convert the third noise reduction coefficient corresponding to each of the first images according to each of the first gain values until the N third noise reduction coefficients are obtained.
  • the noise reduction unit 12 is configured to separately perform noise reduction processing on the N first images according to the N third noise reduction coefficients, where the N third noise reduction coefficients are The N first images are in one-to-one correspondence.
  • the above-mentioned obtaining unit 10, determining unit 11, noise reducing unit 12, judging unit 13, dividing unit 14, computing unit 15, and converting unit 16 may be processor 17 located on the first terminal.
  • the implementation is specifically implemented by a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA).
  • the first terminal further includes: a storage medium 18, The storage medium 18 can be coupled to the processor 17 via a system bus 19 for storing executable program code, the program code including computer operating instructions, the storage medium 18 may include high speed RAM memory, and may also include non-easy Loss memory, for example, at least one disk storage.
  • the noise reduction coefficient of the current frame image may be adjusted according to the change of the previous frame image, thereby realizing the differentiation of different frame images. Image denoising processing is performed, and image quality is improved by image noise reduction.
  • embodiments of the present invention can be provided as a method, system, or computer program product.
  • embodiments of the invention may take the form of a hardware embodiment, a software embodiment, or a combination of software and hardware aspects.
  • embodiments of the invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • the technical solution provided by the embodiment of the present invention can be applied to the field of image processing.
  • the current frame image is acquired; the first lens parameter corresponding to the current frame image is acquired; and the first noise reduction coefficient of the current frame image is determined according to the first lens parameter;
  • the first noise reduction coefficient performs noise reduction processing on the current frame image.
  • the above scheme is used to implement the technology, because the terminal can calculate the image of each frame based on the lens parameters. Considering the selection of different noise reduction coefficients, the difference of the noise reduction processing for selecting appropriate noise reduction parameters for different frame images is realized, and the image denoising processing is realized by differentiating the image quality.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Studio Devices (AREA)
  • Image Processing (AREA)

Abstract

L'invention concerne un procédé et un terminal de réduction de bruit d'image, le procédé consistant à : acquérir une trame actuelle d'une image (S101) ; acquérir des premiers paramètres de lentille qui correspondent à la trame actuelle de l'image (S102) ; selon les premiers paramètres de lentille, déterminer un premier coefficient de réduction de bruit de la trame actuelle de l'image (S103) ; selon le premier coefficient de réduction de bruit, effectuer une réduction de bruit sur la trame actuelle de l'image (S104).
PCT/CN2017/085512 2017-02-27 2017-05-23 Procédé, terminal de réduction de bruit d'image, et support de stockage informatique WO2018152977A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710106574.0 2017-02-27
CN201710106574.0A CN108513043A (zh) 2017-02-27 2017-02-27 一种图像降噪方法及终端

Publications (1)

Publication Number Publication Date
WO2018152977A1 true WO2018152977A1 (fr) 2018-08-30

Family

ID=63253490

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/085512 WO2018152977A1 (fr) 2017-02-27 2017-05-23 Procédé, terminal de réduction de bruit d'image, et support de stockage informatique

Country Status (2)

Country Link
CN (1) CN108513043A (fr)
WO (1) WO2018152977A1 (fr)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109727211A (zh) * 2018-12-21 2019-05-07 厦门美图之家科技有限公司 一种图像去噪方法、装置、计算设备及介质
CN109639932A (zh) * 2019-02-28 2019-04-16 努比亚技术有限公司 图像处理方法、移动终端及计算机可读存储介质
CN110290289B (zh) * 2019-06-13 2021-07-09 Oppo广东移动通信有限公司 图像降噪方法、装置、电子设备以及存储介质
CN111131716B (zh) * 2019-12-31 2021-06-15 联想(北京)有限公司 图像处理方法以及电子设备
CN111372014B (zh) * 2020-03-17 2022-04-22 展讯通信（上海）有限公司 镜头阴影补偿方法及装置、存储介质、终端

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006295807A (ja) * 2005-04-14 2006-10-26 Canon Inc 撮像装置、画像処理方法、プログラム、及び記憶媒体
CN103308184A (zh) * 2013-05-13 2013-09-18 浙江大立科技股份有限公司 红外焦平面阵列探测器单元、红外成像***及校正方法
CN101669142B (zh) * 2007-04-24 2013-10-16 帝欧希数字光学科技（欧洲）有限公司 用于调整将核应用于信号的效果以在信号上取得预期效果的技术
CN104869287A (zh) * 2015-05-18 2015-08-26 成都平行视野科技有限公司 基于移动设备gpu和角速度传感器的视频拍摄降噪方法
CN106231211A (zh) * 2016-08-31 2016-12-14 宁波优而雅电器有限公司 一种红外摄像方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8259198B2 (en) * 2009-10-20 2012-09-04 Apple Inc. System and method for detecting and correcting defective pixels in an image sensor
CN105005973B (zh) * 2015-06-30 2018-04-03 广东欧珀移动通信有限公司 一种图像快速去噪的方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006295807A (ja) * 2005-04-14 2006-10-26 Canon Inc 撮像装置、画像処理方法、プログラム、及び記憶媒体
CN101669142B (zh) * 2007-04-24 2013-10-16 帝欧希数字光学科技（欧洲）有限公司 用于调整将核应用于信号的效果以在信号上取得预期效果的技术
CN103308184A (zh) * 2013-05-13 2013-09-18 浙江大立科技股份有限公司 红外焦平面阵列探测器单元、红外成像***及校正方法
CN104869287A (zh) * 2015-05-18 2015-08-26 成都平行视野科技有限公司 基于移动设备gpu和角速度传感器的视频拍摄降噪方法
CN106231211A (zh) * 2016-08-31 2016-12-14 宁波优而雅电器有限公司 一种红外摄像方法

Also Published As

Publication number Publication date
CN108513043A (zh) 2018-09-07

Similar Documents

Publication Publication Date Title
WO2018152977A1 (fr) Procédé, terminal de réduction de bruit d'image, et support de stockage informatique
JP5113171B2 (ja) イメージ情報にフィルタをかけるための適応空間イメージフィルタ
US8208039B2 (en) Image processing apparatus and computer-readable medium
US8363131B2 (en) Apparatus and method for local contrast enhanced tone mapping
EP3308534A1 (fr) Module de mise à l'échelle de réseau de filtres colorés
US8018504B2 (en) Reduction of position dependent noise in a digital image
US20170134634A1 (en) Photographing apparatus, method of controlling the same, and computer-readable recording medium
WO2009064537A1 (fr) Procédé et système de gestion du bruit pour un traitement spatial dans des systèmes de saisie de vidéos/images numériques
WO2020107995A1 (fr) Procédé et appareil d'imagerie, dispositif électronique et support d'informations lisible par ordinateur
JP2011003048A (ja) 画像処理装置、及び画像処理プログラム
CN110852956A (zh) 一种高动态范围图像的增强方法
US8942477B2 (en) Image processing apparatus, image processing method, and program
JP2014027403A (ja) 画像処理装置
JP6904788B2 (ja) 画像処理装置、画像処理方法、及びプログラム
JP7463640B2 (ja) 空間多重露光のための方法、装置及び記憶媒体
JP6492452B2 (ja) 制御システム、撮像装置、制御方法およびプログラム
US20200351417A1 (en) Image processing
CN114331893A (zh) 一种获取图像噪声的方法、介质和电子设备
WO2016200480A1 (fr) Module de mise à l'échelle de réseau de filtres colorés
JP2016040870A (ja) 画像処理装置、像形成方法およびプログラム
JP5627730B2 (ja) 画像処理装置及び画像処理方法
JP2012178191A (ja) 画像処理装置及び画像処理方法
TWI388201B (zh) 利用遮罩加速濾除雜訊之影像處理裝置、影像處理方法及數位相機
JP2016170637A (ja) 画像処理装置および画像処理方法
WO2013127631A1 (fr) Génération d'image virtuelle

Legal Events

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

Ref document number: 17897457

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17897457

Country of ref document: EP

Kind code of ref document: A1