Image denoising method, image denoising device, terminal and storage medium

Abstract

The present disclosure relates to an image denoising method, an image denoising device, a terminal and a storage medium. The image denoising method comprises: obtaining a first image to be denoised; downsampling the first image to obtain at least one second image with a higher blur than the first image; denoising the first image and at least one second image respectively to obtain a denoised image; fusing the denoised first image and at least one second image to obtain a denoised first image, wherein the resolution of the second image is lower than or equal to the first image, so that the second image has image details different from those of the first image. When denoising the first image and the second image, denoising of different degrees between the images can be performed according to the different image contents, so as to retain as many image details as possible. Compared with directly performing one-time intensity denoising on the first image, this method can retain as many image details as possible and improve image quality.

Description

Technical Field

[0001] The present disclosure relates to the technical field of stroboscopic detection, and in particular to an image noise reduction method, an image noise reduction device, a terminal and a storage medium. Background Art

[0002] In the process of image generation and transmission, images are often degraded due to interference and influence of various noises. In particular, in the process of image acquisition, image sensors will introduce various noises due to the influence of material properties, working environment, electronic components and circuit structure, including thermal noise caused by resistors, channel thermal noise of field effect transistors, photon noise, dark current noise, etc. Noise often appears in the image as an isolated pixel or pixel block that causes a strong visual effect. It appears in the form of useless information and disrupts the observable information of the image, which will have an adverse effect on the subsequent image processing and image visual effects. Summary of the invention

[0003] The present disclosure provides an image noise reduction method, an image noise reduction device, a terminal and a storage medium.

[0004] According to a first aspect of an embodiment of the present disclosure, there is provided an image noise reduction method, comprising:

[0005] Acquire a first image to be denoised;

[0006] Downsampling the first image to obtain at least one second image with a higher blur than that of the first image;

[0007] Respectively performing noise reduction on the first image and at least one of the second images to obtain noise-reduced images;

[0008] The first denoised image and at least one of the second images are fused to obtain a denoised first image, wherein a resolution of the second image is lower than or equal to that of the first image.

[0009] In some embodiments, downsampling the first image to obtain at least one second image having a higher blur than the first image includes:

[0010] Downsampling the first image to obtain a first blurred image with a higher blurriness than the first image;

[0011] The first blurred image is downsampled to obtain a second blurred image with a higher blurriness than the first blurred image; wherein the first blurred image and the second blurred image are both the second image.

[0012] In some embodiments, the denoising the first image and at least one of the second images respectively comprises:

[0013] Performing noise reduction on the first image, the first blurred image, and the second blurred image;

[0014] The fusing the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction includes:

[0015] Acquire a first detail based on the first image after noise reduction;

[0016] Acquire a second detail based on the first blurred image after noise reduction;

[0017] Acquire a third detail based on the second blurred image after noise reduction;

[0018] The first detail, the second detail, and the third detail are fused to obtain the first image after noise reduction.

[0019] In some embodiments, the first image and at least one of the second images are both images to be denoised; and denoising the first image and at least one of the second images respectively to obtain the denoised images includes:

[0020] Dividing each of the images to be denoised into a plurality of target blocks according to noise levels of different image areas in the images to be denoised;

[0021] Determining similar blocks of each target block based on similarity of pixel values ​​within the image region;

[0022] The target block is denoised based on the pixel values ​​of the similar blocks to obtain a denoised target block, wherein an image composed of the denoised target blocks is the denoised image.

[0023] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the method includes:

[0024] Obtaining the variance value of the pixel points of each target block;

[0025] Based on each of the variance values, obtaining an adjustment value of the value of the pixel point;

[0026] The pixel values ​​of each target block are corrected based on the adjustment value to obtain a corrected target block; wherein the pixel values ​​of the similar blocks are used to reduce noise of the corrected target block.

[0027] In some embodiments, the denoising the target block based on the pixel values ​​of the similar blocks to obtain the denoised target block includes:

[0028] If there are multiple similar blocks to a target block, based on the pixel values ​​of the similar blocks and the pixel values ​​of the target block to be denoised, the sum of the absolute pixel differences between each of the similar blocks and the target block to be denoised is obtained;

[0029] Obtaining a weight corresponding to each of the similar blocks based on the sum of absolute pixel differences between each of the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised;

[0030] Based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, the denoised pixel values ​​of the target block to be denoised are obtained.

[0031] In some embodiments, after obtaining the denoised image, the method further includes:

[0032] Based on the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, the value of each pixel in the denoised image is adjusted.

[0033] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the method further comprises:

[0034] Acquire a noise residual distribution after noise reduction of at least one frame of image before the first image;

[0035] Based on the noise residual distribution, the value of each pixel in the corrected target block is adjusted.

[0036] According to a second aspect of the embodiments of the present disclosure, there is provided an image noise reduction device, comprising:

[0037] A first processing unit, used for acquiring a first image to be denoised;

[0038] A second processing unit, configured to downsample the first image to obtain at least one second image with a higher blur than the first image;

[0039] A third processing unit, configured to perform noise reduction on the first image and at least one of the second images respectively to obtain noise-reduced images;

[0040] The fourth processing unit is used to fuse the first image after denoising and at least one of the second images to obtain the first image after denoising, wherein the resolution of the second image is lower than or equal to the first image.

[0041] In some embodiments, the second processing unit is used to downsample the first image to obtain at least one second image with a higher blur than the first image, including:

[0042] The second processing unit is specifically configured to downsample the first image to obtain a first blurred image with a higher blur than that of the first image;

[0043] The first blurred image is downsampled to obtain a second blurred image with a higher blurriness than the first blurred image; wherein the first blurred image and the second blurred image are both the second image.

[0044] In some embodiments, the third processing unit, for performing noise reduction on the first image and at least one of the second images, comprises:

[0045] The third processing unit is specifically configured to perform noise reduction on the first image, the first blurred image, and the second blurred image;

[0046] The fourth processing unit is used to fuse the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction, including:

[0047] The fourth processing unit is specifically configured to obtain a first detail based on the first image after noise reduction;

[0048] Acquire a second detail based on the first blurred image after noise reduction;

[0049] Acquire a third detail based on the second blurred image after noise reduction;

[0050] The first detail, the second detail, and the third detail are fused to obtain the first image after noise reduction.

[0051] In some embodiments, the first image and at least one of the second images are both images to be denoised; the third processing unit is used to perform denoising on the first image and at least one of the second images respectively to obtain denoised images, including:

[0052] The third processing unit is specifically configured to divide each of the images to be denoised into a plurality of target blocks according to noise levels of different image regions in the images to be denoised;

[0053] Determining similar blocks of each target block based on similarity of pixel values ​​within the image region;

[0054] The target block is denoised based on the pixel values ​​of the similar blocks to obtain a denoised target block, wherein an image composed of the denoised target blocks is the denoised image.

[0055] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the third processing unit is further configured to obtain a variance value of a pixel point of each of the target blocks;

[0056] Based on each of the variance values, obtaining an adjustment value of the value of the pixel point;

[0057] The pixel values ​​of each target block are corrected based on the adjustment value to obtain a corrected target block; wherein the pixel values ​​of the similar blocks are used to reduce noise of the corrected target block.

[0058] In some embodiments, the third processing unit is used to reduce noise on the target block based on the pixel values ​​of the similar blocks to obtain the target block after noise reduction, including:

[0059] The third processing unit is specifically configured to obtain, if there are multiple similar blocks to a target block, a sum of absolute pixel differences between each of the similar blocks and the target block to be denoised based on pixel values ​​of the similar blocks and pixel values ​​of the target block to be denoised;

[0060] Obtaining a weight corresponding to each of the similar blocks based on the sum of absolute pixel differences between each of the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised;

[0061] Based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, the denoised pixel values ​​of the target block to be denoised are obtained.

[0062] In some embodiments, after obtaining the denoised image, the third processing unit is further used to adjust the value of each pixel in the denoised image based on the Euclidean distance between each pixel in the image to be denoised and the center pixel of the image.

[0063] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the third processing unit is further configured to:

[0064] Acquire a noise residual distribution after noise reduction of at least one frame of image before the first image; and

[0065] Based on the noise residual distribution, the value of each pixel in the corrected target block is adjusted.

[0066] According to a third aspect of an embodiment of the present disclosure, a terminal is provided, comprising: a processor and a memory for storing a computer program that can be run on the processor, wherein the processor, when used to run the computer program, executes the steps of the method described in the first aspect of the above embodiment.

[0067] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the method described in the first aspect of the above embodiment are implemented.

[0068] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects:

[0069] The image denoising method provided by the embodiment of the present disclosure downsamples the first image to be denoised to obtain at least one second image with a higher blur than the first image; denoises the first image and at least one second image respectively to obtain a denoised image; and fuses the denoised first image and at least one second image to obtain a denoised first image, wherein the resolution of the second image is lower than or equal to the first image, so that the second image has image details different from those of the first image. When denoising the first image and the second image, denoising of different degrees between images can be performed according to different image contents, so that denoising of different degrees of denoising between images is performed in different levels to retain as many image details as possible, and then the denoised image is obtained by fusing the denoised first image and at least one denoised downsampled image. Compared with directly performing one-time intensity denoising on the first image, this method can retain as many image details as possible and improve image quality.

[0070] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0072] Figure 1 The figure is a flow chart of an image noise reduction method according to an exemplary embodiment.

[0073] Figure 2 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 2 .

[0074] Figure 3 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 3 .

[0075] Figure 4 is a curve diagram showing a corresponding relationship between a variance value and an adjustment value in an image noise reduction method according to an exemplary embodiment.

[0076] Figure 5 It is a schematic diagram showing characteristic display of a radius graph according to an exemplary embodiment.

[0077] Figure 6 The figure is a curve diagram showing the corresponding relationship between pixel values ​​and adjustment coefficients in multiple frames of images after noise reduction according to an exemplary embodiment.

[0078] Figure 7 It is a data flow chart of multi-scale NLM denoising of RGB domain images according to an exemplary embodiment.

[0079] Figure 8 The figure is a schematic structural diagram of an image noise reduction device according to an exemplary embodiment.

[0080] Fig. 9 The present invention is a block diagram of a terminal device according to an exemplary embodiment. DETAILED DESCRIPTION

[0081] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0082] The present disclosure provides an image noise reduction method. Figure 1 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 1 .like Figure 1 As shown, the image denoising method comprises:

[0083] Step 10: Acquire the first image to be denoised;

[0084] Step 11, downsampling the first image to obtain at least one second image with a higher blur than the first image;

[0085] Step 12: performing noise reduction on the first image and at least one of the second images respectively to obtain noise-reduced images;

[0086] Step 13: Fusing the first denoised image and at least one of the second images to obtain a first denoised image, wherein the resolution of the second image is lower than or equal to the first image.

[0087] In the disclosed embodiment, the first image to be denoised may be a noisy image captured by an image processing device such as a CCD or CMOS. Blur refers to the degree of blurriness resulting from the loss of details after mean denoising of the image. The blurriness of the second image is higher than that of the first image. The second image may include multiple images with different blurriness. The blurriness of each second image is higher than that of the first image. At the same time, the resolution of each second image is lower than or equal to that of the first image. When denoising the first image and multiple second images, denoising of different degrees between images may be performed according to the different image contents. For example, for an image with more detailed features, the denoising intensity may be appropriately weakened to retain as many detailed features as possible. For an image with fewer detailed features, the denoising intensity may be appropriately enhanced to reflect the denoising effect.

[0088] In the disclosed embodiment, the first image to be denoised is downsampled to obtain at least one second image with a higher blur than the first image; the first image and at least one second image are denoised respectively to obtain a denoised image; the denoised first image and at least one second image are fused to obtain a denoised first image, wherein the resolution of the second image is lower than or equal to the first image, so that the second image has image details different from those of the first image. When denoising the first image and the second image, different denoising degrees can be performed between the images according to the different image contents, so that the denoising is performed at different levels between the images to retain as many image details as possible, and then the denoised image is obtained by fusing the denoised first image and at least one denoised downsampled image. Compared with directly performing one-time intensive denoising on the first image, this method can retain as many image details as possible and improve image quality.

[0089] In some embodiments, downsampling the first image to obtain at least one second image having a higher blur than the first image includes:

[0090] Downsampling the first image to obtain a first blurred image with a higher blurriness than the first image;

[0091] The first blurred image is downsampled to obtain a second blurred image with a higher blurriness than the first blurred image; wherein the first blurred image and the second blurred image are both the second image.

[0092] In the disclosed embodiment, the first image is subjected to mean filtering using a 2*2 window, and the resulting image may be a first blurred image, that is, a medium blurred image. The first blurred image is subjected to mean filtering using a 2*2 window again, and the resulting image may be a second blurred image, that is, a highly blurred image. The first blurred image and the second blurred image are both second images, and have different blurriness. The blurriness of the second blurred image is higher than that of the first blurred image, and the blurriness of the first blurred image is higher than that of the first image. Among them, the first image can be understood as a low-blurred image; the first blurred image can be understood as a medium-blurred image; and the second blurred image can be understood as a highly blurred image. The blurriness of the highly blurred image is higher than that of the medium-blurred image, and the blurriness of the medium-blurred image is higher than that of the low-blurred image. The low-blurred image is subjected to mean filtering using a 2*2 window, and the resulting image is a medium-blurred image. The low-blurred image is subjected to mean filtering using a 4*4 window, and the resulting image is a highly blurred image. Figure 2 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 2 .like Figure 2 As shown in the figure, the first image is downsampled once to obtain a medium blur image, and the first image is subsampled to obtain a high blur image. The medium blur image and the high blur image are both the second images obtained by downsampling. At the same time, the image scale of the medium blur image is 1 / 2*1 / 2 of the low blur image (original image). The image scale of the high blur image is 1 / 4*1 / 4 of the low blur image (original image).

[0093] The image resolution of a low blur image is higher than that of a medium blur image, and the image resolution of a medium blur image is higher than that of a high blur image. For example, the image resolution of a low blur image can be 3000*4000, the image resolution of a medium blur image can be 1500*2000, and the image resolution of a high blur image can be 750*1000. A low blur image can present fine details of an image with high resolution, such as hair, pores, tile texture and other detailed features. A medium blur image can only present details of coarse edges of an image with medium resolution, such as building edge features (windows, building sides), etc. A high blur image can only present a very flat and unclear underlying image.

[0094] In some embodiments, the denoising the first image and at least one of the second images respectively comprises:

[0095] Performing noise reduction on the first image, the first blurred image, and the second blurred image;

[0096] The fusing the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction includes:

[0097] Acquire a first detail based on the first image after noise reduction;

[0098] Acquire a second detail based on the first blurred image after noise reduction;

[0099] Acquire a third detail based on the second blurred image after noise reduction;

[0100] The first detail, the second detail, and the third detail are fused to obtain the first image after noise reduction.

[0101] In the disclosed embodiment, the first details obtained based on the first image after noise reduction may be high-frequency details extracted from the low-blur image; the second details obtained based on the first blurred image after noise reduction may be medium-frequency details extracted from the medium-blur image; the third details obtained based on the second blurred image after noise reduction may be low-frequency details extracted from the high-blur image. The medium-frequency details are the pixels in the medium-blur image minus the pixels in the high-blur image, the high-frequency details are the pixels in the low-blur image minus the pixels in the medium-blur image, and the low-frequency details are the pixels in the high-blur image.

[0102] High-frequency details can be extracted from the low-blur image after denoising, medium-frequency details can be extracted from the medium-blur image, and low-frequency details can be extracted from the high-blur image. When the first details, the second details, and the third details are fused to obtain the first image after denoising, the low-blur image, the medium-blur image, and the high-blur image after denoising can be fused to obtain the denoised image, which can be specifically achieved by: extracting low-frequency details from the high-blur image and adding them to the low-blur image to supplement the low-frequency details, extracting medium-frequency details from the medium-blur image and adding them to the low-blur image to supplement the medium-frequency details, and outputting the fused denoised image; or, extracting high-frequency details from the low-blur image and adding them to the high-blur image to supplement the high-frequency details, extracting medium-frequency details from the medium-blur image and adding them to the high-blur image to supplement the medium-frequency details, and outputting the fused denoised image.

[0103] In some embodiments, the first image and at least one of the second images are both images to be denoised; and denoising the first image and at least one of the second images respectively to obtain the denoised images includes:

[0104] Dividing each of the images to be denoised into a plurality of target blocks according to noise levels of different image areas in the images to be denoised;

[0105] Determining similar blocks of each target block based on similarity of pixel values ​​within the image region;

[0106] The target block is denoised based on the pixel values ​​of the similar blocks to obtain a denoised target block, wherein an image composed of the denoised target blocks is the denoised image.

[0107] In the disclosed embodiment, Figure 3 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 3 .like Figure 3 As shown, the first image and all the second images can be images that need noise reduction. At the same time, the noise levels (noise levels) of different image areas of any image that needs noise reduction are different. Different noise levels represent different noise information contained in the area. The higher the noise level, the more noise information contained in the area. According to different noise levels, the image can be divided into flat area, weak texture area, and strong edge area. Specifically, the variance or mean of the target pixel in each different area can be compared with the size of the preset threshold to determine the flat area, weak texture area, and strong edge area.

[0108] For example, if the variance of the target pixel in the region is less than the first preset threshold, the corresponding image region is a flat region; if the variance of the target pixel in the region is greater than the first preset threshold but less than the second preset threshold, the corresponding image region is a weak texture region; if the variance of the target pixel in the region is greater than the second preset threshold, the corresponding image region is a strong edge region. Figure 4 As shown, x1 is a first preset threshold, and x2 is a second preset threshold. The second preset threshold is greater than the first preset threshold.

[0109] In the disclosed embodiment, the pixel value is a value assigned by a computer when the original image is digitized, and it represents the average brightness information of a small square of the original, or the average reflection density information of the small square.

[0110] In an embodiment of the present disclosure, similar blocks of each target block are determined based on the similarity of pixel values ​​within the image area, including: based on the similarity of pixel values ​​within the image area, similar blocks corresponding to each target block in a plurality of target blocks are determined, wherein each target block corresponds to one or more similar blocks.

[0111] Each of the target blocks corresponds to one or more similar blocks, including:

[0112] The target block with texture corresponds to a first number of similar blocks;

[0113] The target block in the flat area corresponds to a second number of similar blocks; the first number is smaller than the second number.

[0114] In order to protect the texture and avoid excessive noise reduction intensity, which may cause loss of details, the number of similar blocks can be appropriately reduced during noise reduction. When the target block in the flat area is denoised, the number of similar blocks can be appropriately increased to increase the noise reduction intensity. That is, the noise reduction intensity of the target block can be adjusted by controlling the selection of similar blocks. The more similar blocks are fused to the target block, the greater the noise reduction intensity.

[0115] In the disclosed embodiment, similar blocks of each target block can also be determined based on the similarity of pixel information other than pixel values ​​in the image region. The pixel information includes: pixel coordinates and / or color values, etc. For example, similar blocks of the target block are determined based on the similarity of pixel colors in the region.

[0116] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the method includes:

[0117] Obtaining the variance value of the pixel points of each target block;

[0118] Based on each of the variance values, obtaining an adjustment value of the value of the pixel point;

[0119] The pixel values ​​of each target block are corrected based on the adjustment value to obtain a corrected target block; wherein the pixel values ​​of the similar blocks are used to reduce noise of the corrected target block.

[0120] In the embodiment of the present disclosure, the variance value of the pixel points of each target block refers to the variance value of the pixel values ​​of the pixel points of each target block. Figure 3 As shown, based on the image to be denoised, the variance value corresponding to each target block in the image to be denoised is obtained. Based on each of the variance values, the adjustment value of the pixel value is obtained, including obtaining the adjustment value corresponding to each variance value according to the corresponding relationship between the variance value and the adjustment value. The pixel value in the image with original noise is adjusted by the adjustment value to obtain a corrected noise image. Figure 4 is a curve diagram showing the corresponding relationship between the variance value and the adjustment value in an image noise reduction method according to an exemplary embodiment. Figure 4 As shown, the size of the adjustment value on the vertical axis corresponding to the horizontal axis variance value can be obtained through the curve correspondence between the horizontal axis variance value and the vertical axis adjustment value. After obtaining the adjustment value corresponding to the variance value, the value of the pixel point of the corrected target block is obtained by multiplying the original value of the pixel point of the target block by the adjustment value. For example, Adjust noise value = noise value * k, where Adjust noise value is the value of the pixel point of the target block after correction, noise value is the original value of the pixel point of the target block, and k is the adjustment value.

[0121] In some embodiments, the denoising the target block based on the pixel values ​​of the similar blocks to obtain the denoised target block includes:

[0122] If there are multiple similar blocks to a target block, based on the pixel values ​​of the similar blocks and the pixel values ​​of the target block to be denoised, the sum of the absolute pixel differences between each of the similar blocks and the target block to be denoised is obtained;

[0123] Obtaining a weight corresponding to each of the similar blocks based on the sum of absolute pixel differences between each of the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised;

[0124] Based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, the denoised pixel values ​​of the target block to be denoised are obtained.

[0125] In the embodiment of the present disclosure, when the sum of the absolute pixel differences between each similar block and the target block to be denoised is obtained based on the pixel values ​​of the similar blocks and the pixel values ​​of the target block to be denoised, the sum of the absolute pixel differences between the similar blocks and the target block to be denoised can be obtained based on the pixel values ​​of the most central pixels in the target block and the similar blocks. The sum of the absolute pixel differences is the sum of the absolute differences of the pixels, that is, Figure 3 The pixel value of the target block to be denoised can be the value of the pixel point of the corrected target block, that is, Figure 3 The value of the pixel in the corrected noise map.

[0126] Based on the sum of the absolute differences of the pixels between the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised, the weight corresponding to each similar block is obtained, specifically:

[0127] weight=1–param1*(SAD / Adjust noise value–param2), where weight is the weight corresponding to the similar block, SAD is the sum of the absolute pixel differences between the similar block and the target block to be denoised, Adjustnoise value is the value of the pixel of the corrected target block, and param1 and param2 are adjustment parameters set in advance corresponding to the target block to be denoised. For different target blocks, param1 and param2 may be different. Wherein, the value range of param1 is 0 to 4.0. The value range of param2 is 0 to 1.0. Based on the weights corresponding to each of the similar blocks and the pixel values ​​of each of the similar blocks, the denoised pixel value of the target block to be denoised is obtained, including:

[0128] denoise=(S1*weight1+S2*weight2+…+S n *weight n ) / (weight1+weight2+…+weight n ).

[0129] Among them, S n is the pixel value of the nth similar block, weight nis the weight corresponding to the nth similar block, denoising the pixel value of the target block to be denoised, and n is a positive integer greater than or equal to 1.

[0130] In the embodiment of the present disclosure, the above method can be used to obtain the pixel value after noise reduction corresponding to each target block. The noise-reduced image is obtained by splicing each target block after noise reduction, that is, Figure 3 The denoised image after denoising.

[0131] The present invention divides each image to be denoised into multiple target blocks according to the noise levels of different image areas, and performs separate denoising with different degrees of denoising on different target blocks to achieve balanced overall image noise reduction. Compared with the denoising method with uniform denoising intensity, the present invention can reduce the phenomenon that details are lost in some areas due to excessive denoising intensity, and the phenomenon that denoising effect is unsatisfactory in some areas due to low denoising intensity.

[0132] In some embodiments, after obtaining the denoised image, the method further includes:

[0133] Based on the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, the value of each pixel in the denoised image is adjusted.

[0134] In the embodiment of the present disclosure, the value of each pixel in the denoised image, especially the value of the pixel in the edge area, can be adjusted according to the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, including:

[0135] The adjustment coefficient is determined according to the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, and the value of each pixel in the denoised image is adjusted based on the adjustment coefficient, specifically:

[0136] output=(1-x)*denoise+x*original image, where output is the value of the pixel in the image adjusted by the adjustment coefficient, denoise is the pixel value of the target block to be denoised after denoising, original image is the value of the pixel in the original image, and x is the adjustment coefficient, and the value range of x is 0~1. The smaller x is, the greater the ratio of the denoised image will be, and the stronger the denoising effect of the output image will be. By adjusting the adjustment coefficient, the denoising strength of the image edge can be effectively increased. Figure 5 FIG. 1 is a schematic diagram showing the characteristic display of a radius graph according to an exemplary embodiment. Figure 5 As shown in the figure, the center of the radius graph pixel is very bright (pixel value is large) and the surrounding is very dark (pixel value is small). Euclidean distance usually adopts the distance definition, which represents the real distance between two points in m-dimensional space. In this application, it is used to represent the degree of deviation between each pixel in the image to be denoised and the center pixel of the image, that is, Figure 3Meaning of the data indicated in the radius graph.

[0137] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the method further comprises:

[0138] Acquire a noise residual distribution after noise reduction of at least one frame of image before the first image;

[0139] Based on the noise residual distribution, the value of each pixel in the corrected target block is adjusted.

[0140] In the embodiment of the present disclosure, according to the noise residual distribution after multiple frames of the previous images are denoised (i.e. Figure 3 The noise residual image obtained after denoising the multi-frame denoised images shown in FIG. 1 ) is used to adjust the value of each pixel point of the corrected target block. Figure 6 is a graph showing the corresponding relationship between pixel values ​​and adjustment coefficients in multiple frames of denoised images according to an exemplary embodiment. Figure 6 As shown in FIG. 1 , the horizontal axis is the value of each pixel in the image after noise reduction, and the vertical axis is the adjustment coefficient corresponding to the pixel value. The pixel value of the target block after adjustment and correction is the product of the value of each pixel in the image after noise reduction and the adjustment coefficient.

[0141] Sigma map is data provided by multi-frame noise reduction. Generally, the noise reduction method disclosed in the present invention can be connected to the back of multi-frame noise reduction to achieve the purpose of further noise reduction. In multi-frame noise reduction, multiple frames are fused to reduce noise, but if there are moving objects in multiple frames, the effect of fusion noise reduction will be worse. Therefore, Sigma map is equivalent to a residual information map of noise, indicating which parts of this picture (after multi-frame noise reduction) have more obvious residual noise (generally moving areas) and which parts have less, so as to directionally guide the subsequent noise reduction in the area with more residual noise, and enhance the noise reduction strength to achieve a balanced noise reduction effect.

[0142] The disclosed embodiment also provides a multi-scale NLM denoising method for RGB domain images. Figure 7 FIG. 1 is a data flow chart of multi-scale NLM denoising of RGB domain images according to an exemplary embodiment. Figure 7 As shown in the figure, the original image is first decomposed using Laplace filtering. Secondly, the NLM method is applied to each level of decomposition for noise reduction, the processing detail coefficients of each decomposition level and the approximate coefficients of the highest decomposition level are determined, and finally Laplace synthesis is performed to obtain a denoised image.

[0143] The embodiment of the present disclosure also provides an image noise reduction device. Figure 8 FIG. 1 is a schematic diagram showing the structure of an image noise reduction device according to an exemplary embodiment. Figure 8 As shown, the image noise reduction device comprises:

[0144] A first processing unit 81, configured to obtain a first image to be denoised;

[0145] A second processing unit 82 is used to downsample the first image to obtain at least one second image with a higher blur than the first image;

[0146] A third processing unit 83 is used to perform noise reduction on the first image and at least one of the second images to obtain noise-reduced images;

[0147] The fourth processing unit 84 is used to fuse the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction, wherein the resolution of the second image is lower than or equal to the first image.

[0148] In the disclosed embodiment, the first image to be denoised may be a noisy image captured by an image processing device such as a CCD or CMOS. Blur refers to the degree of blurriness resulting from the loss of details after mean denoising of the image. The blurriness of the second image is higher than that of the first image. The second image may include multiple images with different blurriness. The blurriness of each second image is higher than that of the first image. At the same time, the resolution of each second image is lower than or equal to that of the first image. When denoising the first image and multiple second images, denoising of different degrees between images may be performed according to the different image contents. For example, for an image with more detailed features, the denoising intensity may be appropriately weakened to retain as many detailed features as possible. For an image with fewer detailed features, the denoising intensity may be appropriately enhanced to reflect the denoising effect.

[0149] In the disclosed embodiment, the first image to be denoised is downsampled to obtain at least one second image with a higher blur than the first image; the first image and at least one second image are denoised respectively to obtain a denoised image; the denoised first image and at least one second image are fused to obtain a denoised first image, wherein the resolution of the second image is lower than or equal to the first image, so that the second image has image details different from those of the first image. When denoising the first image and the second image, different denoising degrees can be performed between the images according to the different image contents, so that the denoising is performed at different levels between the images to retain as many image details as possible, and then the denoised image is obtained by fusing the denoised first image and at least one denoised downsampled image. Compared with directly performing one-time intensive denoising on the first image, this method can retain as many image details as possible and improve image quality.

[0150] In some embodiments, the second processing unit is used to downsample the first image to obtain at least one second image with a higher blur than the first image, including:

[0151] The second processing unit is specifically configured to downsample the first image to obtain a first blurred image with a higher blur than that of the first image;

[0152] The first blurred image is downsampled to obtain a second blurred image with a higher blurriness than the first blurred image; wherein the first blurred image and the second blurred image are both the second image.

[0153] In the disclosed embodiment, the first image is subjected to mean filtering using a 2*2 window, and the resulting image may be a first blurred image, that is, a medium blurred image. The first blurred image is subjected to mean filtering using a 2*2 window again, and the resulting image may be a second blurred image, that is, a highly blurred image. The first blurred image and the second blurred image are both second images, and have different blurriness. The blurriness of the second blurred image is higher than that of the first blurred image, and the blurriness of the first blurred image is higher than that of the first image. Among them, the first image can be understood as a low-blurred image; the first blurred image can be understood as a medium-blurred image; and the second blurred image can be understood as a highly blurred image. The blurriness of the highly blurred image is higher than that of the medium-blurred image, and the blurriness of the medium-blurred image is higher than that of the low-blurred image. The low-blurred image is subjected to mean filtering using a 2*2 window, and the resulting image is a medium-blurred image. The low-blurred image is subjected to mean filtering using a 4*4 window, and the resulting image is a highly blurred image. Figure 2 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 2 .like Figure 2 As shown in the figure, the first image is downsampled once to obtain a medium blur image, and the first image is subsampled to obtain a high blur image. The medium blur image and the high blur image are both the second images obtained by downsampling. At the same time, the image scale of the medium blur image is 1 / 2*1 / 2 of the low blur image (original image). The image scale of the high blur image is 1 / 4*1 / 4 of the low blur image (original image).

[0154] The image resolution of a low blur image is higher than that of a medium blur image, and the image resolution of a medium blur image is higher than that of a high blur image. For example, the image resolution of a low blur image can be 3000*4000, the image resolution of a medium blur image can be 1500*2000, and the image resolution of a high blur image can be 750*1000. A low blur image can present fine details of an image with high resolution, such as hair, pores, tile texture and other detailed features. A medium blur image can only present details of coarse edges of an image with medium resolution, such as building edge features (windows, building sides), etc. A high blur image can only present a very flat and unclear underlying image.

[0155] In some embodiments, the third processing unit, for performing noise reduction on the first image and at least one of the second images, comprises:

[0156] The third processing unit is specifically configured to perform noise reduction on the first image, the first blurred image, and the second blurred image;

[0157] The fourth processing unit is used to fuse the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction, including:

[0158] The fourth processing unit is specifically configured to obtain a first detail based on the first image after noise reduction;

[0159] Acquire a second detail based on the first blurred image after noise reduction;

[0160] Acquire a third detail based on the second blurred image after noise reduction;

[0161] The first detail, the second detail, and the third detail are fused to obtain the first image after noise reduction.

[0162] In the disclosed embodiment, the first details obtained based on the first image after noise reduction may be high-frequency details extracted from the low-blur image; the second details obtained based on the first blurred image after noise reduction may be medium-frequency details extracted from the medium-blur image; the third details obtained based on the second blurred image after noise reduction may be low-frequency details extracted from the high-blur image. The medium-frequency details are the pixels in the medium-blur image minus the pixels in the high-blur image, the high-frequency details are the pixels in the low-blur image minus the pixels in the medium-blur image, and the low-frequency details are the pixels in the high-blur image.

[0163] High-frequency details can be extracted from the low-blur image after denoising, medium-frequency details can be extracted from the medium-blur image, and low-frequency details can be extracted from the high-blur image. When the first details, the second details, and the third details are fused to obtain the first image after denoising, the low-blur image, the medium-blur image, and the high-blur image after denoising can be fused to obtain the denoised image, which can be specifically achieved by: extracting low-frequency details from the high-blur image and adding them to the low-blur image to supplement the low-frequency details, extracting medium-frequency details from the medium-blur image and adding them to the low-blur image to supplement the medium-frequency details, and outputting the fused denoised image; or, extracting high-frequency details from the low-blur image and adding them to the high-blur image to supplement the high-frequency details, extracting medium-frequency details from the medium-blur image and adding them to the high-blur image to supplement the medium-frequency details, and outputting the fused denoised image.

[0164] In some embodiments, the first image and at least one of the second images are both images to be denoised; the third processing unit is used to perform denoising on the first image and at least one of the second images respectively to obtain denoised images, including:

[0165] The third processing unit is specifically configured to divide each of the images to be denoised into a plurality of target blocks according to noise levels of different image regions in the images to be denoised;

[0166] Determining similar blocks of each target block based on similarity of pixel values ​​within the image region;

[0167] The target block is denoised based on the pixel values ​​of the similar blocks to obtain a denoised target block, wherein an image composed of the denoised target blocks is the denoised image.

[0168] In the disclosed embodiment, Figure 3 The following is a flow chart of an image noise reduction method according to an exemplary embodiment. Figure 3 .like Figure 3 As shown, the first image and all the second images can be images that need noise reduction. At the same time, the noise levels (noise levels) of different image areas of any image that needs noise reduction are different. Different noise levels represent different noise information contained in the area. The higher the noise level, the more noise information contained in the area. According to different noise levels, the image can be divided into flat area, weak texture area, and strong edge area. Specifically, the variance or mean of the target pixel in each different area can be compared with the size of the preset threshold to determine the flat area, weak texture area, and strong edge area.

[0169] For example, if the variance of the target pixel in the region is less than the first preset threshold, the corresponding image region is a flat region; if the variance of the target pixel in the region is greater than the first preset threshold but less than the second preset threshold, the corresponding image region is a weak texture region; if the variance of the target pixel in the region is greater than the second preset threshold, the corresponding image region is a strong edge region. Figure 4 As shown, x1 is a first preset threshold, and x2 is a second preset threshold. The second preset threshold is greater than the first preset threshold.

[0170] In the disclosed embodiment, the pixel value is a value assigned by a computer when the original image is digitized, and it represents the average brightness information of a small square of the original, or the average reflection density information of the small square.

[0171] In an embodiment of the present disclosure, similar blocks of each target block are determined based on the similarity of pixel values ​​within the image area, including: based on the similarity of pixel values ​​within the image area, similar blocks corresponding to each target block in a plurality of target blocks are determined, wherein each target block corresponds to one or more similar blocks.

[0172] Each of the target blocks corresponds to one or more similar blocks, including:

[0173] The target block with texture corresponds to a first number of similar blocks;

[0174] The target block in the flat area corresponds to a second number of similar blocks; the first number is smaller than the second number.

[0175] In order to protect the texture during noise reduction, the number of corresponding similar blocks can be appropriately reduced. In order to increase the noise reduction strength, the number of corresponding similar blocks can be appropriately increased during noise reduction of the target block in the flat area. That is, the noise reduction strength of the target block can be adjusted by controlling the selection of similar blocks. The more similar blocks are fused to the target block, the greater the noise reduction strength.

[0176] In the disclosed embodiment, similar blocks of each target block can also be determined based on the similarity of pixel information other than pixel values ​​in the image region. Pixel information includes: pixel coordinates and / or color values, etc. For example, similar blocks of target blocks are determined based on the similarity of pixel colors in the region. In some embodiments, before denoising the multiple target blocks based on the multiple similar blocks to obtain the denoised multiple target blocks, the third processing unit is also used to obtain the variance value of the pixel points of each target block;

[0177] Based on each of the variance values, obtaining an adjustment value of the value of the pixel point;

[0178] The pixel values ​​of each target block are corrected based on the adjustment value to obtain a corrected target block; wherein the pixel values ​​of the similar blocks are used to reduce noise of the corrected target block.

[0179] In the embodiment of the present disclosure, the variance value of the pixel points of each target block refers to the variance value of the pixel values ​​of the pixel points of each target block. Figure 3 As shown, based on the image to be denoised, the variance value corresponding to each target block in the image to be denoised is obtained. Based on each of the variance values, the adjustment value of the pixel value is obtained, including obtaining the adjustment value corresponding to each variance value according to the corresponding relationship between the variance value and the adjustment value. The pixel value in the image with original noise is adjusted by the adjustment value to obtain a corrected noise image. Figure 4 is a curve diagram showing the corresponding relationship between the variance value and the adjustment value in an image noise reduction method according to an exemplary embodiment. Figure 4As shown, the size of the adjustment value on the vertical axis corresponding to the horizontal axis variance value can be obtained through the curve correspondence between the horizontal axis variance value and the vertical axis adjustment value. After obtaining the adjustment value corresponding to the variance value, the value of the pixel point of the corrected target block is obtained by multiplying the original value of the pixel point of the target block by the adjustment value. For example, Adjust noise value = noise value * k, where Adjust noise value is the value of the pixel point of the target block after correction, noise value is the original value of the pixel point of the target block, and k is the adjustment value.

[0180] In some embodiments, the third processing unit is used to reduce noise on the target block based on the pixel values ​​of the similar blocks to obtain the target block after noise reduction, including:

[0181] The third processing unit is specifically configured to obtain, if there are multiple similar blocks to a target block, a sum of absolute pixel differences between each of the similar blocks and the target block to be denoised based on pixel values ​​of the similar blocks and pixel values ​​of the target block to be denoised;

[0182] Obtaining a weight corresponding to each of the similar blocks based on the sum of absolute pixel differences between each of the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised;

[0183] Based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, the denoised pixel values ​​of the target block to be denoised are obtained.

[0184] In the embodiment of the present disclosure, when the sum of the absolute pixel differences between each similar block and the target block to be denoised is obtained based on the pixel values ​​of the similar blocks and the pixel values ​​of the target block to be denoised, the sum of the absolute pixel differences between the similar blocks and the target block to be denoised can be obtained based on the pixel values ​​of the most central pixels in the target block and the similar blocks. The sum of the absolute pixel differences is the sum of the absolute differences of the pixels, that is, Figure 3 The pixel value of the target block to be denoised can be the value of the pixel point of the corrected target block, that is, Figure 3 The value of the pixel in the corrected noise map.

[0185] Based on the sum of the absolute differences of the pixels between the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised, the weight corresponding to each similar block is obtained, specifically:

[0186] weight=1–param1*(SAD / Adjust noise value–param2), where weight is the weight corresponding to the similar block, SAD is the sum of the absolute differences between the similar block and the target block to be denoised, Adjustnoise value is the value of the pixel of the corrected target block, param1 and param2 are both adjustment parameters set in advance corresponding to the target block to be denoised. For different target blocks, param1 and param2 may be different.

[0187] Obtaining the denoised pixel value of the target block to be denoised based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, including:

[0188] denoise=(S1*weight1+S2*weight2+…+S n *weight n ) / (weight1+weight2+…+weight n ).

[0189] Among them, S n is the pixel value of the nth similar block, weight n is the weight corresponding to the nth similar block, denoising the pixel value of the target block to be denoised, and n is a positive integer greater than or equal to 1.

[0190] In the embodiment of the present disclosure, the above method can be used to obtain the pixel value after noise reduction corresponding to each target block. The noise-reduced image is obtained by splicing each target block after noise reduction, that is, Figure 3 The denoised image after denoising.

[0191] The present invention divides each image to be denoised into multiple target blocks according to the noise levels of different image areas, and performs separate denoising with different degrees of denoising on different target blocks to achieve balanced overall image noise reduction. Compared with the denoising method with uniform denoising intensity, the present invention can reduce the phenomenon that details are lost in some areas due to excessive denoising intensity, and the phenomenon that denoising effect is unsatisfactory in some areas due to low denoising intensity.

[0192] In some embodiments, after obtaining the denoised image, the third processing unit is further used to adjust the value of each pixel in the denoised image based on the Euclidean distance between each pixel in the image to be denoised and the center pixel of the image.

[0193] In the embodiment of the present disclosure, the value of each pixel in the denoised image, especially the value of the pixel in the edge area, can be adjusted according to the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, including:

[0194] The adjustment coefficient is determined according to the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, and the value of each pixel in the denoised image is adjusted based on the adjustment coefficient, specifically:

[0195] output=(1-x)*denoise+x*original image, where output is the value of the pixel in the image adjusted by the adjustment coefficient, denoise is the pixel value of the target block to be denoised after denoising, original image is the value of the pixel in the original image, and x is the adjustment coefficient, and the value range of x is 0~1. The smaller x is, the greater the ratio of the denoised image will be, and the stronger the denoising effect of the output image will be. By adjusting the adjustment coefficient, the denoising strength of the image edge can be effectively increased. Figure 5 FIG. 1 is a schematic diagram showing the characteristic display of a radius graph according to an exemplary embodiment. Figure 5 As shown in the figure, the center of the radius graph pixel is very bright (pixel value is large) and the surrounding is very dark (pixel value is small). Euclidean distance usually adopts the distance definition, which represents the real distance between two points in m-dimensional space. In this application, it is used to represent the degree of deviation between each pixel in the image to be denoised and the center pixel of the image, that is, Figure 3 Meaning of the data indicated in the radius graph.

[0196] In some embodiments, before denoising the plurality of target blocks based on the plurality of similar blocks to obtain the plurality of denoised target blocks, the third processing unit is further configured to:

[0197] Acquire a noise residual distribution after noise reduction of at least one frame of image before the first image; and

[0198] Based on the noise residual distribution, the value of each pixel in the corrected target block is adjusted.

[0199] In the embodiment of the present disclosure, according to the noise residual distribution after multiple frames of the previous images are denoised (i.e. Figure 3 The noise residual image obtained after denoising the multi-frame denoised images shown in FIG. 1 ) is used to adjust the value of each pixel point of the corrected target block. Figure 6 is a graph showing the corresponding relationship between pixel values ​​and adjustment coefficients in multiple frames of denoised images according to an exemplary embodiment. Figure 6 As shown in FIG. 1 , the horizontal axis is the value of each pixel in the image after noise reduction, and the vertical axis is the adjustment coefficient corresponding to the pixel value. The pixel value of the target block after adjustment and correction is the product of the value of each pixel in the image after noise reduction and the adjustment coefficient.

[0200] Sigma map is data provided by multi-frame noise reduction. Generally, the noise reduction method disclosed in the present invention can be connected to the back of multi-frame noise reduction to achieve the purpose of further noise reduction. In multi-frame noise reduction, multiple frames are fused to reduce noise, but if there are moving objects in multiple frames, the effect of fusion noise reduction will be worse. Therefore, Sigma map is equivalent to a residual information map of noise, indicating which parts of this picture (after multi-frame noise reduction) have more obvious residual noise (generally moving areas) and which parts have less, so as to directionally guide the subsequent noise reduction in the area with more residual noise, and enhance the noise reduction strength to achieve a balanced noise reduction effect.

[0201] An embodiment of the present disclosure further provides a terminal, comprising: a processor and a memory for storing a computer program that can be run on the processor, wherein the processor is used to execute the steps of the method described in the above embodiment when running the computer program.

[0202] The present disclosure also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the method described in the above embodiment when executed by a processor.

[0203] Fig. 9 1 is a block diagram of a terminal device according to an exemplary embodiment. For example, the terminal device may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0204] Reference Fig. 9 The terminal device may include one or more of the following components: a processing component 802 , a memory 804 , a power component 806 , a multimedia component 808 , an audio component 810 , an input / output (I / O) interface 812 , a sensor component 814 , and a communication component 816 .

[0205] The processing component 802 generally controls the overall operation of the terminal device, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above method. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0206] The memory 804 is configured to store various types of data to support operations on the terminal device. Examples of such data include instructions for any application or method operating on the terminal device, contact data, phone book data, messages, pictures, videos, etc. The memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0207] The power component 806 provides power to various components of the terminal device. The power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the terminal device.

[0208] The multimedia component 808 includes a screen that provides an output interface between the terminal device and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front camera and / or a rear camera. When the terminal device is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0209] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC), and when the terminal device is in an operation mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 804 or sent via the communication component 816. In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.

[0210] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0211] The sensor assembly 814 includes one or more sensors for providing various aspects of status assessment for the terminal device. For example, the sensor assembly 814 can detect the open / closed state of the terminal device, the relative positioning of components, such as the display and keypad of the terminal device, and the sensor assembly 814 can also detect the position change of the terminal device or a component of the terminal device, the presence or absence of user contact with the terminal device, the orientation or acceleration / deceleration of the terminal device, and the temperature change of the terminal device. The sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 814 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0212] The communication component 816 is configured to facilitate wired or wireless communication between the terminal device and other devices. The terminal device can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0213] In an exemplary embodiment, the terminal device may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0214] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in this disclosure. The specification and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.

[0215] It should be understood that the present invention is not limited to the exact construction that has been described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. An image denoising method, characterized in that: include: Acquire a first image to be denoised; Downsampling the first image to obtain at least one second image with a higher blur than that of the first image; Respectively performing noise reduction on the first image and at least one of the second images to obtain noise-reduced images; fusing the first image after denoising and at least one of the second images to obtain a first image after denoising, wherein the resolution of the second image is lower than or equal to the first image; The step of downsampling the first image to obtain at least one second image with a higher blur than that of the first image comprises: Downsampling the first image to obtain a first blurred image with a higher blurriness than the first image; The first blurred image is downsampled to obtain a second blurred image with a higher blurriness than the first blurred image; wherein the first blurred image and the second blurred image are both the second image.

2. The image denoising method according to claim 1, characterized in that: The performing noise reduction on the first image and at least one of the second images respectively comprises: Performing noise reduction on the first image, the first blurred image, and the second blurred image; The fusing the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction includes: Acquire a first detail based on the first image after noise reduction; Acquire a second detail based on the first blurred image after noise reduction; Acquire a third detail based on the second blurred image after noise reduction; The first detail, the second detail, and the third detail are fused to obtain the first image after noise reduction.

3. The image denoising method according to claim 1, characterized in that: in, The first image and at least one of the second images are both images to be denoised; The step of respectively reducing noise on the first image and at least one of the second images to obtain a noise-reduced image includes: Dividing each of the images to be denoised into a plurality of target blocks according to noise levels of different image areas in the images to be denoised; Determining similar blocks of each target block based on similarity of pixel values ​​within the image region; The target block is denoised based on the pixel values ​​of the similar blocks to obtain a denoised target block, wherein an image composed of the denoised target blocks is the denoised image.

4. The image denoising method according to claim 3, characterized in that: Before performing noise reduction on the target block based on the pixel values ​​of the similar blocks to obtain the noise reduced target block, the method includes: Obtaining the variance value of the pixel points of each target block; Based on each of the variance values, obtaining an adjustment value of the value of the pixel point; The pixel values ​​of each target block are corrected based on the adjustment value to obtain a corrected target block; wherein the pixel values ​​of the similar blocks are used to reduce noise of the corrected target block.

5. The image denoising method according to claim 3, characterized in that: The step of reducing noise on the target block based on the pixel values ​​of the similar blocks to obtain the target block after reducing noise includes: If there are multiple similar blocks to a target block, based on the pixel values ​​of the similar blocks and the pixel values ​​of the target block to be denoised, the sum of the absolute pixel differences between each of the similar blocks and the target block to be denoised is obtained; Obtaining a weight corresponding to each of the similar blocks based on the sum of absolute pixel differences between each of the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised; Based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, the denoised pixel values ​​of the target block to be denoised are obtained.

6. The image denoising method according to any one of claims 3 to 5, characterized in that: After obtaining the denoised image, the method further includes: Based on the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image, the value of each pixel in the denoised image is adjusted.

7. The image denoising method according to claim 4, characterized in that: Before performing noise reduction on the target block based on the pixel values ​​of the similar blocks to obtain the noise reduced target block, the method further includes: Acquire a noise residual distribution after noise reduction of at least one frame of image before the first image; Based on the noise residual distribution, the value of each pixel in the corrected target block is adjusted.

8. An image noise reduction device, characterized in that: include: A first processing unit, used for acquiring a first image to be denoised; A second processing unit, configured to downsample the first image to obtain at least one second image with a higher blur than the first image; The second processing unit is used to downsample the first image to obtain at least one second image with a higher blur than the first image, including: the second processing unit is specifically used to downsample the first image to obtain a first blurred image with a higher blur than the first image; downsample the first blurred image to obtain a second blurred image with a higher blur than the first blurred image; wherein the first blurred image and the second blurred image are both the second image; A third processing unit, configured to perform noise reduction on the first image and at least one of the second images respectively to obtain noise-reduced images; The fourth processing unit is used to fuse the first image after denoising and at least one of the second images to obtain the first image after denoising, wherein the resolution of the second image is lower than or equal to the first image.

9. The image noise reduction device according to claim 8, characterized in that: The third processing unit, used for performing noise reduction on the first image and at least one of the second images respectively, comprises: The third processing unit is specifically configured to perform noise reduction on the first image, the first blurred image, and the second blurred image; The fourth processing unit is used to fuse the first image after noise reduction and at least one of the second images to obtain the first image after noise reduction, including: The fourth processing unit is specifically configured to obtain a first detail based on the first image after noise reduction; Acquire a second detail based on the first blurred image after noise reduction; Acquire a third detail based on the second blurred image after noise reduction; The first detail, the second detail, and the third detail are fused to obtain the first image after noise reduction.

10. The image noise reduction device according to claim 8, characterized in that: The first image and at least one of the second images are both images to be denoised; The third processing unit is used to perform noise reduction on the first image and at least one of the second images to obtain noise-reduced images, including: The third processing unit is specifically configured to divide each of the images to be denoised into a plurality of target blocks according to noise levels of different image regions in the images to be denoised; Determining similar blocks of each target block based on similarity of pixel values ​​within the image region; The target block is denoised based on the pixel values ​​of the similar blocks to obtain a denoised target block, wherein an image composed of the denoised target blocks is the denoised image.

11. The image noise reduction device according to claim 10, characterized in that: Before performing noise reduction on the target block based on the pixel values ​​of the similar blocks to obtain the noise-reduced target block, the third processing unit is further used to obtain the variance value of the pixel points of each of the target blocks; Based on each of the variance values, obtaining an adjustment value of the value of the pixel point; The pixel values ​​of each target block are corrected based on the adjustment value to obtain a corrected target block; wherein the pixel values ​​of the similar blocks are used to reduce noise of the corrected target block.

12. The image noise reduction device according to claim 10, characterized in that: The third processing unit is used to reduce noise on the target block based on the pixel values ​​of the similar blocks to obtain a target block after noise reduction, including: The third processing unit is specifically configured to obtain, if there are multiple similar blocks to a target block, a sum of absolute pixel differences between each of the similar blocks and the target block to be denoised based on pixel values ​​of the similar blocks and pixel values ​​of the target block to be denoised; Obtaining a weight corresponding to each of the similar blocks based on the sum of absolute pixel differences between each of the similar blocks and the target block to be denoised, and the pixel value of the target block to be denoised; Based on the weights corresponding to the similar blocks and the pixel values ​​of the similar blocks, the denoised pixel values ​​of the target block to be denoised are obtained.

13. The image noise reduction device according to any one of claims 10 to 12, characterized in that: After obtaining the denoised image, the third processing unit is further used to adjust the value of each pixel in the denoised image based on the Euclidean distance between each pixel in the image to be denoised and the central pixel of the image.

14. The image noise reduction device according to claim 11, characterized in that: Before performing noise reduction on the target block based on the pixel values ​​of the similar blocks to obtain the noise-reduced target block, the third processing unit is further used to: Acquire a noise residual distribution after noise reduction of at least one frame of image before the first image; and Based on the noise residual distribution, the value of each pixel in the corrected target block is adjusted.

15. A terminal, characterized in that: include: A processor and a memory for storing a computer program that can be run on the processor, wherein the processor executes the steps of the method according to any one of claims 1 to 7 when running the computer program.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images