WO2020124873A1 - 图像处理方法 - Google Patents

图像处理方法 Download PDF

Info

Publication number
WO2020124873A1
WO2020124873A1 PCT/CN2019/081610 CN2019081610W WO2020124873A1 WO 2020124873 A1 WO2020124873 A1 WO 2020124873A1 CN 2019081610 W CN2019081610 W CN 2019081610W WO 2020124873 A1 WO2020124873 A1 WO 2020124873A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
processed
pixels
area
value
Prior art date
Application number
PCT/CN2019/081610
Other languages
English (en)
French (fr)
Inventor
陈云娜
金羽锋
Original Assignee
深圳市华星光电半导体显示技术有限公司
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 深圳市华星光电半导体显示技术有限公司 filed Critical 深圳市华星光电半导体显示技术有限公司
Publication of WO2020124873A1 publication Critical patent/WO2020124873A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • 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
    • G06T5/30—Erosion or dilatation, e.g. thinning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/73—Deblurring; Sharpening
    • 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
    • G06T2207/20028—Bilateral filtering

Definitions

  • the present invention relates to the field of display technology, and in particular to an image processing method.
  • Block-based discrete cosine transform (BDCT) coding has a wide range of applications in the field of compression.
  • Common image and video compression standards including JPEG and H264, use BDCT coding.
  • BDCT coding ignores the correlation of neighboring blocks, and a discontinuous phenomenon occurs at the boundary of the block, resulting in noise.
  • the image compressed by the BDCT encoding method will be subjected to noise reduction through global noise reduction, so as to eliminate the noise caused by the discontinuous boundary of the image block to improve The quality of the image.
  • a common method is to perform noise reduction on the image through bilateral filtering and noise reduction or use a sparse representation method of a pre-trained dictionary to perform noise reduction on the image.
  • the noise in the image is in different areas, and the human eye has different sensitivities. For relatively flat areas, the human eye is more sensitive to noise, and for textured areas, texture has a masking effect on noise. Not easy to detect.
  • the entire image adopts a uniform distance attenuation parameter and grayscale attenuation parameter, that is, a unified bilateral filtering formula is used to process the entire image.
  • An object of the present invention is to provide an image processing method, which can reduce the noise of the image while eliminating the blurring of the texture area caused by the noise reduction, and improve the display quality of the image.
  • the present invention first provides an image processing method, including the following steps:
  • Step S1 Provide the image to be processed
  • Step S2 dividing the image to be processed into a flat area, a texture area and an edge area;
  • Step S3. Use different bilateral filter functions to perform noise reduction processing on the flat area, the texture area and the edge area, thereby forming a processed image.
  • the image to be processed is a grayscale image; the image to be processed includes a plurality of pixels arranged in an array; each pixel has a corresponding original grayscale value;
  • the original gray value of each pixel of the image to be processed is also normalized to form a normalized gray value
  • the step S2 specifically includes:
  • Step S21 Calculate the gradient information of the image to be processed using the normalized gray values of multiple pixels, the gradient information includes gradient values corresponding to the multiple pixels; define pixels whose gradient values are less than the first threshold as flat pixels, Define pixels whose gradient value is greater than or equal to the first threshold value and less than or equal to the second threshold value as texture pixels, and define pixels whose gradient value is greater than or equal to the second threshold value as edge pixels; the first threshold value is less than the second threshold value;
  • Step S22 Divide the to-be-processed image into a plurality of blocks arranged in sequence, and calculate the number of flat pixels, texture pixels, and edge pixels in each block;
  • Step S23 Set one of the multiple blocks as the block to be analyzed. If the number of texture pixels in the block to be analyzed is greater than the number of flat pixels and greater than the number of edge pixels, the block to be analyzed Perform morphological expansion operations, otherwise keep the block to be analyzed unchanged;
  • Step S24 Repeat the above step S23 until multiple blocks have completed the operation of step S23;
  • Step S25 Divide the image to be processed in step S24 into a flat area, an edge area, and a texture area, where the gradient value of the pixels in the flat area is less than the first threshold, and the gradient value of the pixels in the texture area is greater than or equal to the first threshold Less than or equal to the second threshold, the gradient value of the pixels in the edge area is greater than the second threshold.
  • the step S3 is specifically: using a preset first bilateral filter function to convert the original gray value of each pixel in the flat area into a processing gray value, and using a preset second bilateral filter function to convert each pixel in the texture area The original gray value is converted into the processed gray value, and the preset third bilateral filter function is used to convert the original gray value of each pixel in the edge area into the processed gray value;
  • the first bilateral filter function is a
  • N(x) represents the area of a preset filter window in the image to be processed
  • y represents the position of a pixel in the filter window
  • x represents the position of the currently processed pixel
  • I(x) represents the currently processed pixel
  • the original gray value of, I(y) represents the original gray value of the pixel with y position
  • d1 is the preset first distance attenuation rate
  • r1 is the preset first gray attenuation rate
  • the second bilateral filter function is the first bilateral filter function
  • d2 is the preset second distance attenuation rate
  • r2 is the preset second gray attenuation rate
  • the third bilateral filter function is the third bilateral filter function
  • d3 is the preset third distance attenuation rate
  • r3 is the preset third gray attenuation rate
  • the first distance attenuation rate is greater than the second distance attenuation rate, and the second distance attenuation rate is greater than the third distance attenuation rate;
  • the first gray attenuation rate is greater than the second gray attenuation rate, and the second gray attenuation rate is greater than the third gray attenuation rate.
  • the image to be processed is obtained by extracting the luminance channel in the YCbCr data of the color image;
  • the image to be processed is an image after compression processing.
  • the normalized gray value of a plurality of pixels is used to calculate the gradient information of the image to be processed using an edge detection method.
  • the first threshold is 0.08, and the second threshold is 0.6.
  • the invention also provides an image processing method, including the following steps:
  • Step S1 Provide the image to be processed
  • Step S2 dividing the image to be processed into a flat area, a texture area and an edge area;
  • Step S3 Use different bilateral filter functions to perform noise reduction processing on the flat area, the texture area and the edge area to form a processed image
  • the image to be processed is a grayscale image; the image to be processed includes a plurality of pixels arranged in an array; each pixel has a corresponding original grayscale value;
  • the original gray value of each pixel of the image to be processed is also normalized to form a normalized gray value
  • the step S2 specifically includes:
  • Step S21 Calculate the gradient information of the image to be processed using the normalized gray values of multiple pixels, the gradient information includes gradient values corresponding to the multiple pixels; define pixels whose gradient values are less than the first threshold as flat pixels, Define pixels whose gradient value is greater than or equal to the first threshold value and less than or equal to the second threshold value as texture pixels, and define pixels whose gradient value is greater than or equal to the second threshold value as edge pixels; the first threshold value is less than the second threshold value;
  • Step S22 Divide the to-be-processed image into a plurality of blocks arranged in sequence, and calculate the number of flat pixels, texture pixels, and edge pixels in each block;
  • Step S23 Set one of the multiple blocks as the block to be analyzed. If the number of texture pixels in the block to be analyzed is greater than the number of flat pixels and greater than the number of edge pixels, the block to be analyzed Perform morphological expansion operations, otherwise keep the block to be analyzed unchanged;
  • Step S24 Repeat the above step S23 until multiple blocks have completed the operation of step S23;
  • Step S25 Divide the image to be processed in step S24 into a flat area, an edge area, and a texture area, where the gradient value of the pixels in the flat area is less than the first threshold, and the gradient value of the pixels in the texture area is greater than or equal to the first threshold Less than or equal to the second threshold, and the gradient value of the pixels in the edge area is greater than the second threshold;
  • the image to be processed is obtained by extracting the luminance channel in the YCbCr data of the color image;
  • the image to be processed is an image after compression processing.
  • the invention also provides an image processing method, including the following steps:
  • Step S1' providing an image to be processed
  • Step S2' dividing the image to be processed into a flat area and a texture area
  • Step S3' Perform sparse representation processing based on the pre-trained dictionary on the original gray value of each pixel in the flat area, and keep the original gray value of each pixel in the texture area unchanged, thereby forming a processed image.
  • the image to be processed is a grayscale image; the image to be processed includes a plurality of pixels arranged in an array; each pixel has a corresponding original grayscale value;
  • step S1' after the image to be processed is provided, the original gray value of each pixel of the image to be processed is also normalized to form a normalized gray value;
  • the step S2' specifically includes:
  • Step S21' using the normalized gray values of multiple pixels to calculate gradient information of the image to be processed, the gradient information including gradient values corresponding to the multiple pixels respectively; defining pixels whose gradient value is less than the first threshold as flat pixels , Define pixels whose gradient value is greater than or equal to the first threshold as texels;
  • Step S22' dividing the image to be processed into multiple blocks arranged in sequence, and calculating the number of flat pixels and texture pixels in each block;
  • Step S23' One of the multiple blocks is set as the block to be analyzed. If the number of texels in the block to be analyzed is greater than the number of flat pixels, it is determined that the block to be analyzed is a texture area. Otherwise, it is determined that the block to be analyzed is a flat area;
  • Step S24' Repeat the above step S23' until multiple blocks have completed the operation of step S23'.
  • the image to be processed is obtained by extracting the luminance channel in the YCbCr data of the color image;
  • the image to be processed is an image after compression processing.
  • the normalized gray value of a plurality of pixels is used to calculate the gradient information of the image to be processed using an edge detection method
  • the first threshold is 0.08.
  • the image to be processed is divided into flat regions, texture regions and edge regions, and different bilateral filter functions are used to perform noise reduction processing on the flat regions, texture regions and edge regions for
  • the distance attenuation rate and grayscale attenuation rate of the bilateral filter function for processing the flat area, the bilateral filter function for processing the texture area, and the bilateral filter function for processing the edge area are gradually reduced, thereby forming a processed image
  • the phenomenon of blurring the texture area caused by noise improves the display quality of the image.
  • FIG. 1 is a flowchart of a first embodiment of the image processing method of the present invention
  • step S2 is a flowchart of step S2 of the first embodiment of the image processing method of the present invention
  • FIG. 3 is a flowchart of a second embodiment of the image processing method of the present invention.
  • Fig. 4 is a flowchart of step S2' of the second embodiment of the image processing method of the present invention.
  • the first embodiment of the image processing method of the present invention includes the following steps:
  • Step S1 Provide an image to be processed.
  • the image to be processed is a grayscale image.
  • the image to be processed includes a plurality of pixels arranged in an array. Each pixel has a corresponding original gray value.
  • the image to be processed is obtained by extracting the luminance channel in the YCbCr data of the color image.
  • the image to be processed is an image that has undergone compression processing, for example, may be an image that has undergone compression processing using BDCT encoding.
  • step S1 after providing the image to be processed, the original gray value of each pixel of the image to be processed is also normalized to form a normalized gray value, and the normalized gray value of each pixel
  • the value range is 0-1.
  • Step S2 the image to be processed is divided into a flat area, a texture area and an edge area.
  • the step S2 specifically includes:
  • Step S21 Calculate gradient information of the image to be processed using the normalized gray values of multiple pixels, where the gradient information includes gradient values corresponding to the multiple pixels, respectively.
  • a pixel whose gradient value is less than the first threshold is defined as a flat pixel
  • a pixel whose gradient value is greater than or equal to the first threshold and less than or equal to the second threshold is defined as a texture pixel
  • a pixel whose gradient value is greater than or equal to the second threshold is defined as an edge pixel.
  • the first threshold is less than the second threshold.
  • the normalized gray value of a plurality of pixels is used to calculate the gradient information of the image to be processed by using an edge detection method such as a Sobel operator.
  • the range of the gradient value corresponding to each pixel is 0-1.
  • the first threshold is 0.08
  • the second threshold is 0.6.
  • Step S22 Divide the image to be processed into a plurality of blocks arranged in sequence, and calculate the number of flat pixels, texture pixels, and edge pixels in each block.
  • the method of dividing the image to be processed into blocks in step S22 may be selected according to actual needs, for example, it may be divided into multiple blocks arranged in 8 rows and 8 columns.
  • Step S23 Set one of the multiple blocks as the block to be analyzed. If the number of texture pixels in the block to be analyzed is greater than the number of flat pixels and greater than the number of edge pixels, the block to be analyzed Perform the morphological dilation operation, otherwise keep the block to be analyzed unchanged, thereby eliminating small holes in the texture and preventing additional noise from affecting the image quality.
  • Step S24 Repeat the above step S23 until multiple blocks have completed the operation of step S23.
  • Step S25 Divide the image to be processed in step S24 into a flat area, an edge area, and a texture area, where the gradient value of the pixels in the flat area is less than the first threshold, and the gradient value of the pixels in the texture area is greater than or equal to the first threshold Less than or equal to the second threshold, the gradient value of the pixels in the edge area is greater than the second threshold.
  • Step S3. Use different bilateral filter functions to perform noise reduction processing on the flat area, the texture area and the edge area, thereby forming a processed image.
  • the step S3 is specifically: using a preset first bilateral filter function to convert the original gray value of each pixel in the flat area into a processing gray value, and using a preset second bilateral filter function to convert the The original gray value of each pixel is converted into a processed gray value, and a preset third bilateral filter function is used to convert the original gray value of each pixel in the edge area into a processed gray value.
  • the first bilateral filter function is a
  • N(x) represents the area of a preset filter window in the image to be processed
  • y represents the position of a pixel in the filter window
  • x represents the position of the currently processed pixel
  • I(x) represents the currently processed pixel
  • the original gray value of, I(y) represents the original gray value of the pixel with y position
  • d1 is the preset first distance attenuation rate
  • r1 is the preset first gray attenuation rate.
  • the second bilateral filter function is the first bilateral filter function
  • d2 is the preset second distance attenuation rate
  • r2 is the preset second gray attenuation rate
  • the third bilateral filter function is the third bilateral filter function
  • d3 is the preset third distance attenuation rate
  • r3 is the preset third gray attenuation rate
  • the first distance attenuation rate is greater than the second distance attenuation rate, and the second distance attenuation rate is greater than the third distance attenuation rate.
  • the first gray attenuation rate is greater than the second gray attenuation rate, and the second gray attenuation rate is greater than the third gray attenuation rate.
  • the image to be processed is divided into flat regions, texture regions, and edge regions, and the flat regions, texture regions, and edge regions are reduced using different bilateral filter functions
  • Noise processing the distance attenuation rate and grayscale attenuation rate of the bilateral filter function for processing the flat area, the bilateral filter function for processing the texture area and the bilateral filter function for processing the edge area are gradually reduced
  • it can avoid the problem of using the same bilateral filter function to perform noise reduction on the entire image to blur the image in the texture area, and improve the display quality of the image.
  • the second embodiment of the image processing method of the present invention includes the following steps:
  • Step S1' providing an image to be processed.
  • the image to be processed is a grayscale image.
  • the image to be processed includes a plurality of pixels arranged in an array. Each pixel has a corresponding original gray value.
  • the image to be processed is obtained by extracting the luminance channel in the YCbCr data of the color image.
  • the image to be processed is an image that has undergone compression processing, for example, may be an image that has undergone compression processing using BDCT encoding.
  • the original gray value of each pixel of the image to be processed is also normalized to form a normalized gray value, and the normalized gray value of each pixel
  • the range of degrees is 0-1.
  • Step S2' the image to be processed is divided into a flat area and a texture area.
  • the step S2' specifically includes:
  • Step S21' Calculate the gradient information of the image to be processed using the normalized gray values of multiple pixels, where the gradient information includes gradient values corresponding to the multiple pixels, respectively.
  • Pixels with gradient values less than the first threshold are defined as flat pixels, and pixels with gradient values greater than or equal to the first threshold are defined as texture pixels.
  • the normalized gray value of a plurality of pixels is used to calculate the gradient information of the image to be processed using an edge detection method such as a Sobel operator.
  • the range of the gradient value corresponding to each pixel is 0-1.
  • the first threshold is 0.08.
  • step S22' the image to be processed is divided into a plurality of blocks arranged in sequence, and the number of flat pixels and texture pixels in each block is calculated.
  • step S22' the size of the divided block of the image to be processed is selected according to the size of the sparse representation processing window when the sparse representation processing is performed subsequently.
  • Step S23' One of the multiple blocks is set as the block to be analyzed. If the number of texels in the block to be analyzed is greater than the number of flat pixels, it is determined that the block to be analyzed is a texture area. Otherwise, it is determined that the block to be analyzed is a flat area.
  • Step S24' Repeat the above step S23' until multiple blocks have completed the operation of step S23'.
  • Step S3' Perform sparse representation processing based on the pre-trained dictionary on the original gray value of each pixel of the flat area, and keep the original gray value of each pixel of the texture area unchanged, thereby forming a processed image.
  • step S3' the original gray value of each pixel in the flat area is subjected to sparse representation processing based on a pre-trained dictionary to obtain multiple sets of calculation results corresponding to each pixel in the flat area, and calculating each group The result is averaged to obtain the processed gray value corresponding to each pixel of the flat area.
  • the image to be processed is divided into a flat area and an edge area, and the flat area is subjected to sparse representation processing based on a pre-trained dictionary and the texture area is kept unchanged, thereby Forming a processed image can ensure that the image has a good noise reduction effect, while avoiding the problem of blurring the image in the texture area due to the sparse representation processing of the texture area, and improving the display quality of the image.
  • the image to be processed is divided into a flat area, a texture area and an edge area, and different bilateral filter functions are used to perform noise reduction processing on the flat area, the texture area and the edge area.
  • the distance decay rate and gray decay rate of the bilateral filter function for processing the flat area, the bilateral filter function for processing the texture area, and the bilateral filter function for processing the edge area gradually decrease, thereby forming a processed image,
  • divide the image to be processed into a flat area and an edge area perform sparse representation processing based on the pre-trained dictionary on the flat area and keep the texture area unchanged, thereby forming a processed image, which can eliminate noise due to noise reduction while reducing the image.
  • the resulting blurring of texture areas improves the display quality of the image.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)
  • Facsimile Image Signal Circuits (AREA)

Abstract

本发明提供一种图像处理方法。本发明的图像处理方法中将待处理图像划分为平坦区、纹理区及边缘区，利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，用于对平坦区进行处理的双边滤波函数、用于对纹理区进行处理的双边滤波函数及用于对边缘区进行处理的双边滤波函数的距离衰减率及灰度衰减率逐渐减小，从而形成处理图像，或者将待处理图像划分为平坦区及边缘区，对平坦区进行基于预训练字典的稀疏表示处理并保持纹理区不变，从而形成处理图像，能够在对图像进行降噪的同时消除由于降噪产生的纹理区模糊的现象，提升图像的显示品质。

Description

图像处理方法 技术领域
本发明涉及显示技术领域，尤其涉及一种图像处理方法。
背景技术
随着人们生活水平的提高，对电子产品的画面显示效果追求越来越高。现有技术中为了提高画面的显示效果，通常会在画面显示时进行图像处理，以改善显示效果。
基于块的离散余弦变换(BDCT)编码在压缩领域具有广泛的应用，包括JPEG及H264在内的常见的图像和视频压缩标准均采用了BDCT编码。然而，BDCT编码由于忽略了相邻块的相关性，会在块的边界出现不连续的现象，产生噪声。为解决这一问题，现有的图像处理技术中会通过全局降噪的方式对采用BDCT编码的方式压缩后的图像进行降噪处理，以消除图像的块的边界不连续而产生的噪声从而提升图像的质量。常用的做法是通过双边滤波降噪的方式对图像进行降噪处理或者采用预训练字典的稀疏表示方法对图像进行降噪处理。
根据人眼视觉***(HVS)，图像中的噪声处于不同的区域，人眼感受度不同，对于相对平坦的区域，人眼对噪声比较敏感，对于纹理区域，纹理对噪声具有掩盖作用，人眼不易察觉。现有技术在通过双边滤波降噪的方式对原始图像进行降噪处理时，整幅图像采用统一的距离衰减参数及灰度衰减参数，也即使用统一的双边滤波公式对整幅图像进行处理，虽然能够将平坦区的噪声进行去除，具有保边去噪的功能，但是会使得纹理区的弱边缘变得模糊，使得纹理区的模糊程度增加，影响了纹理区的图像效果，使得图像整体的显示质量下降。而现有技术采用预训练字典的稀疏表示方法对图像进行降噪处理时，对整幅图像均采用稀疏表示，也会使得纹理区变得模糊。
发明内容
本发明的目的在于提供一种图像处理方法，能够在对图像进行降噪的同时消除由于降噪产生的纹理区模糊的现象，提升图像的显示品质。
为实现上述目的，本发明首先提供一种图像处理方法，包括如下步骤：
步骤S1、提供待处理图像；
步骤S2、将待处理图像划分为平坦区、纹理区及边缘区；
步骤S3、利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，从而形成处理图像。
所述待处理图像为灰度图像；所述待处理图像包括呈阵列式排布的多个像素；每一像素具有对应的原始灰度值；
所述步骤S1在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值；
所述步骤S2具体包括：
步骤S21、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值；定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值小于等于第二阈值的像素为纹理像素，定义梯度值大于第二阈值的像素为边缘像素；第一阈值小于第二阈值；
步骤S22、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素、纹理像素及边缘像素的个数；
步骤S23、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数同时大于边缘像素的个数时，对待分析区块进行形态学膨胀操作，否则保持待分析区块不变；
步骤S24、重复上述步骤S23，直至多个区块均完成步骤S23的操作；
步骤S25、将完成步骤S24的待处理图像划分为平坦区、边缘区及纹理区，其中，平坦区中的像素的梯度值小于第一阈值，纹理区中的像素的梯度值大于等于第一阈值小于等于第二阈值，边缘区中的像素的梯度值大于第二阈值。
所述步骤S3具体为：利用预设的第一双边滤波函数将平坦区中各个像素的原始灰度值转换为处理灰度值，利用预设的第二双边滤波函数将纹理区中各个像素的原始灰度值转换为处理灰度值，利用预设的第三双边滤波函数将边缘区中各个像素的原始灰度值转换为处理灰度值；
所述第一双边滤波函数为
Figure PCTCN2019081610-appb-000001
其中，
Figure PCTCN2019081610-appb-000002
N(x)表示位于待处理图像中的一预设的滤波窗口所在的区域，y表示所述滤波窗口中的一像素的位置，x表示当前处理像素的位置，I(x)表示当前处理像素的原始灰度值，I(y)表示具有y位置的像素的原始灰度值，
Figure PCTCN2019081610-appb-000003
表示当前处理像素的处理灰度值， d1为预设的第一距离衰减率，r1为预设的第一灰度衰减率；
所述第二双边滤波函数为
Figure PCTCN2019081610-appb-000004
其中，
Figure PCTCN2019081610-appb-000005
d2为预设的第二距离衰减率，r2为预设的第二灰度衰减率；
所述第三双边滤波函数为
Figure PCTCN2019081610-appb-000006
其中，
Figure PCTCN2019081610-appb-000007
d3为预设的第三距离衰减率，r3为预设的第三灰度衰减率；
所述第一距离衰减率大于第二距离衰减率，所述第二距离衰减率大于第三距离衰减率；
所述第一灰度衰减率大于第二灰度衰减率，所述第二灰度衰减率大于第三灰度衰减率。
所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得；
所述待处理图像为经过压缩处理后的图像。
所述步骤S21中，利用多个像素的归一化灰度值采用边缘侦测方法计算所述待处理图像的梯度信息。
所述第一阈值为0.08，所述第二阈值为0.6。
本发明还提供一种图像处理方法，包括如下步骤：
步骤S1、提供待处理图像；
步骤S2、将待处理图像划分为平坦区、纹理区及边缘区；
步骤S3、利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，从而形成处理图像；
所述待处理图像为灰度图像；所述待处理图像包括呈阵列式排布的多个像素；每一像素具有对应的原始灰度值；
所述步骤S1在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值；
所述步骤S2具体包括：
步骤S21、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值；定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值小于等于第二阈值的 像素为纹理像素，定义梯度值大于第二阈值的像素为边缘像素；第一阈值小于第二阈值；
步骤S22、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素、纹理像素及边缘像素的个数；
步骤S23、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数同时大于边缘像素的个数时，对待分析区块进行形态学膨胀操作，否则保持待分析区块不变；
步骤S24、重复上述步骤S23，直至多个区块均完成步骤S23的操作；
步骤S25、将完成步骤S24的待处理图像划分为平坦区、边缘区及纹理区，其中，平坦区中的像素的梯度值小于第一阈值，纹理区中的像素的梯度值大于等于第一阈值小于等于第二阈值，边缘区中的像素的梯度值大于第二阈值；
所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得；
所述待处理图像为经过压缩处理后的图像。
本发明还提供一种图像处理方法，包括如下步骤：
步骤S1’、提供待处理图像；
步骤S2’、将待处理图像划分为平坦区及纹理区；
步骤S3’、对平坦区的各个像素的原始灰度值进行基于预训练字典的稀疏表示处理，并保持纹理区的各个像素的原始灰度值不变，从而形成处理图像。
所述待处理图像为灰度图像；所述待处理图像包括呈阵列式排布的多个像素；每一像素具有对应的原始灰度值；
所述步骤S1’在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值；
所述步骤S2’具体包括：
步骤S21’、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值；定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值的像素为纹理像素；
步骤S22’、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素及纹理像素的个数；
步骤S23’、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数，则判定该待分析区块为纹理区，否则判定该待分析区块为平坦区；
步骤S24’、重复上述步骤S23’，直至多个区块均完成步骤S23’的操作。
所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得；
所述待处理图像为经过压缩处理后的图像。
所述步骤S21’中，利用多个像素的归一化灰度值采用边缘侦测方法计算所述待处理图像的梯度信息；
所述第一阈值为0.08。
本发明的有益效果：本发明的图像处理方法中将待处理图像划分为平坦区、纹理区及边缘区，利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，用于对平坦区进行处理的双边滤波函数、用于对纹理区进行处理的双边滤波函数及用于对边缘区进行处理的双边滤波函数的距离衰减率及灰度衰减率逐渐减小，从而形成处理图像，或者将待处理图像划分为平坦区及边缘区，对平坦区进行基于预训练字典的稀疏表示处理并保持纹理区不变，从而形成处理图像，能够在对图像进行降噪的同时消除由于降噪产生的纹理区模糊的现象，提升图像的显示品质。
附图说明
为了能更进一步了解本发明的特征以及技术内容，请参阅以下有关本发明的详细说明与附图，然而附图仅提供参考与说明用，并非用来对本发明加以限制。
附图中，
图1为本发明的图像处理方法的第一实施例的流程图；
图2为本发明的图像处理方法的第一实施例的步骤S2的流程图；
图3为本发明的图像处理方法的第二实施例的流程图；
图4为本发明的图像处理方法的第二实施例的步骤S2’的流程图。
具体实施方式
为更进一步阐述本发明所采取的技术手段及其效果，以下结合本发明的优选实施例及其附图进行详细描述。
请参阅图1，本发明的图像处理方法的第一实施例包括如下步骤：
步骤S1、提供待处理图像。
具体地，所述待处理图像为灰度图像。所述待处理图像包括呈阵列式排布的多个像素。每一像素具有对应的原始灰度值。
具体地，所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道 进行提取获得。
具体地，所述待处理图像为经过压缩处理后的图像，例如可以为利用BDCT编码进行压缩处理后的图像。
具体地，所述步骤S1在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值，每一像素的归一化灰度值的取值范围均为0-1。
步骤S2、将待处理图像划分为平坦区、纹理区及边缘区。
具体地，请参阅图2，所述步骤S2具体包括：
步骤S21、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值。定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值小于等于第二阈值的像素为纹理像素，定义梯度值大于第二阈值的像素为边缘像素。第一阈值小于第二阈值。
具体地，所述步骤S21中，利用多个像素的归一化灰度值采用边缘侦测方法例如Sobel算子计算所述待处理图像的梯度信息。
具体地，每一像素对应的梯度值的取值范围为0-1。
优选地，所述第一阈值为0.08，所述第二阈值为0.6。
步骤S22、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素、纹理像素及边缘像素的个数。
具体地，所述步骤S22中将待处理图像划分区块的方式可依据实际需求进行选择，例如可以划分为8行8列排布的多个区块。
步骤S23、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数同时大于边缘像素的个数时，对待分析区块进行形态学膨胀操作，否则保持待分析区块不变，从而消除纹理中小的孔洞，防止额外的噪声影响图像质量。
步骤S24、重复上述步骤S23，直至多个区块均完成步骤S23的操作。
步骤S25、将完成步骤S24的待处理图像划分为平坦区、边缘区及纹理区，其中，平坦区中的像素的梯度值小于第一阈值，纹理区中的像素的梯度值大于等于第一阈值小于等于第二阈值，边缘区中的像素的梯度值大于第二阈值。
步骤S3、利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，从而形成处理图像。
具体地，所述步骤S3具体为：利用预设的第一双边滤波函数将平坦区中各个像素的原始灰度值转换为处理灰度值，利用预设的第二双边滤波函 数将纹理区中各个像素的原始灰度值转换为处理灰度值，利用预设的第三双边滤波函数将边缘区中各个像素的原始灰度值转换为处理灰度值。
所述第一双边滤波函数为
Figure PCTCN2019081610-appb-000008
其中，
Figure PCTCN2019081610-appb-000009
N(x)表示位于待处理图像中的一预设的滤波窗口所在的区域，y表示所述滤波窗口中的一像素的位置，x表示当前处理像素的位置，I(x)表示当前处理像素的原始灰度值，I(y)表示具有y位置的像素的原始灰度值，
Figure PCTCN2019081610-appb-000010
表示当前处理像素的处理灰度值，d1为预设的第一距离衰减率，r1为预设的第一灰度衰减率。
所述第二双边滤波函数为
Figure PCTCN2019081610-appb-000011
其中，
Figure PCTCN2019081610-appb-000012
d2为预设的第二距离衰减率，r2为预设的第二灰度衰减率。
所述第三双边滤波函数为
Figure PCTCN2019081610-appb-000013
其中，
Figure PCTCN2019081610-appb-000014
d3为预设的第三距离衰减率，r3为预设的第三灰度衰减率。
所述第一距离衰减率大于第二距离衰减率，所述第二距离衰减率大于第三距离衰减率。所述第一灰度衰减率大于第二灰度衰减率，所述第二灰度衰减率大于第三灰度衰减率。
需要说明的是，本发明的图像处理方法的第一实施例中，将待处理图像划分为平坦区、纹理区及边缘区，利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，用于对平坦区进行处理的双边滤波函数、用于对纹理区进行处理的双边滤波函数及用于对边缘区进行处理的双边滤波函数的距离衰减率及灰度衰减率逐渐减小，从而形成处理图像，能够在保证对图像具有较好的降噪效果的同时，避免利用同一双边滤波函数对图像整体进行降噪处理使得纹理区图像模糊的问题，提升图像的显示质量。
请参阅图3，本发明的图像处理方法的第二实施例包括如下步骤：
步骤S1’、提供待处理图像。
具体地，所述待处理图像为灰度图像。所述待处理图像包括呈阵列式排布的多个像素。每一像素具有对应的原始灰度值。
具体地，所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得。
具体地，所述待处理图像为经过压缩处理后的图像，例如可以为利用BDCT编码进行压缩处理后的图像。
具体地，所述步骤S1’在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值，每一像素的归一化灰度值的取值范围均为0-1。
步骤S2’、将待处理图像划分为平坦区及纹理区。
具体地，请参阅图4，所述步骤S2’具体包括：
步骤S21’、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值。定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值的像素为纹理像素。
具体地，所述步骤S21’中，利用多个像素的归一化灰度值采用边缘侦测方法例如Sobel算子计算所述待处理图像的梯度信息。
具体地，每一像素对应的梯度值的取值范围为0-1。
优选地，所述第一阈值为0.08。
步骤S22’、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素及纹理像素的个数。
具体地，所述步骤S22’中将待处理图像划分区块的尺寸依据后续进行稀疏表示处理时的稀疏表示处理窗口的尺寸进行选择。
步骤S23’、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数，则判定该待分析区块为纹理区，否则判定该待分析区块为平坦区。
步骤S24’、重复上述步骤S23’，直至多个区块均完成步骤S23’的操作。
步骤S3’、对平坦区的各个像素的原始灰度值进行基于预训练字典的稀疏表示处理，并保持纹理区的各个像素的原始灰度值不变，从而形成处理图像。
具体地，所述步骤S3’，对平坦区的各个像素的原始灰度值进行基于预训练字典的稀疏表示处理后得到分别与平坦区的各个像素对应的多组计算结果，将每一组计算结果取平均值，得到分别与平坦区的各个像素对应的处理灰度值。
需要说明的是，本发明的图像处理方法的第二实施例中，将待处理图像划分为平坦区及边缘区，对平坦区进行基于预训练字典的稀疏表示处理并保持纹理区不变，从而形成处理图像，能够在保证对图像具有较好的降 噪效果的同时，避免由于对纹理区进行稀疏表示处理进行降噪使得纹理区图像模糊的问题，提升图像的显示质量。
综上所述，本发明的图像处理方法中将待处理图像划分为平坦区、纹理区及边缘区，利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，用于对平坦区进行处理的双边滤波函数、用于对纹理区进行处理的双边滤波函数及用于对边缘区进行处理的双边滤波函数的距离衰减率及灰度衰减率逐渐减小，从而形成处理图像，或者将待处理图像划分为平坦区及边缘区，对平坦区进行基于预训练字典的稀疏表示处理并保持纹理区不变，从而形成处理图像，能够在对图像进行降噪的同时消除由于降噪产生的纹理区模糊的现象，提升图像的显示品质。
以上所述，对于本领域的普通技术人员来说，可以根据本发明的技术方案和技术构思作出其他各种相应的改变和变形，而所有这些改变和变形都应属于本发明权利要求的保护范围。

Claims (14)

  1. 一种图像处理方法，包括如下步骤：
    步骤S1、提供待处理图像；
    步骤S2、将待处理图像划分为平坦区、纹理区及边缘区；
    步骤S3、利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降噪处理，从而形成处理图像。
  2. 如权利要求1所述的图像处理方法，其中，所述待处理图像为灰度图像；所述待处理图像包括呈阵列式排布的多个像素；每一像素具有对应的原始灰度值；
    所述步骤S1在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值；
    所述步骤S2具体包括：
    步骤S21、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值；定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值小于等于第二阈值的像素为纹理像素，定义梯度值大于第二阈值的像素为边缘像素；第一阈值小于第二阈值；
    步骤S22、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素、纹理像素及边缘像素的个数；
    步骤S23、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数同时大于边缘像素的个数时，对待分析区块进行形态学膨胀操作，否则保持待分析区块不变；
    步骤S24、重复上述步骤S23，直至多个区块均完成步骤S23的操作；
    步骤S25、将完成步骤S24的待处理图像划分为平坦区、边缘区及纹理区，其中，平坦区中的像素的梯度值小于第一阈值，纹理区中的像素的梯度值大于等于第一阈值小于等于第二阈值，边缘区中的像素的梯度值大于第二阈值。
  3. 如权利要求2所述的图像处理方法，其中，所述步骤S3具体为：利用预设的第一双边滤波函数将平坦区中各个像素的原始灰度值转换为处理灰度值，利用预设的第二双边滤波函数将纹理区中各个像素的原始灰度值转换为处理灰度值，利用预设的第三双边滤波函数将边缘区中各个像素的原始灰度值转换为处理灰度值；
    所述第一双边滤波函数为
    Figure PCTCN2019081610-appb-100001
    其中，
    Figure PCTCN2019081610-appb-100002
    N(x)表示位于待处理图像中的一预设的滤波窗口所在的区域，y表示所述滤波窗口中的一像素的位置，x表示当前处理像素的位置，I(x)表示当前处理像素的原始灰度值，I(y)表示具有y位置的像素的原始灰度值，
    Figure PCTCN2019081610-appb-100003
    表示当前处理像素的处理灰度值，d1为预设的第一距离衰减率，r1为预设的第一灰度衰减率；
    所述第二双边滤波函数为
    Figure PCTCN2019081610-appb-100004
    其中，
    Figure PCTCN2019081610-appb-100005
    d2为预设的第二距离衰减率，r2为预设的第二灰度衰减率；
    所述第三双边滤波函数为
    Figure PCTCN2019081610-appb-100006
    其中，
    Figure PCTCN2019081610-appb-100007
    d3为预设的第三距离衰减率，r3为预设的第三灰度衰减率；
    所述第一距离衰减率大于第二距离衰减率，所述第二距离衰减率大于第三距离衰减率；
    所述第一灰度衰减率大于第二灰度衰减率，所述第二灰度衰减率大于第三灰度衰减率。
  4. 如权利要求1所述的图像处理方法，其中，所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得；
    所述待处理图像为经过压缩处理后的图像。
  5. 如权利要求2所述的图像处理方法，其中，所述步骤S21中，利用多个像素的归一化灰度值采用边缘侦测方法计算所述待处理图像的梯度信息。
  6. 如权利要求2所述的图像处理方法，其中，所述第一阈值为0.08，所述第二阈值为0.6。
  7. 一种图像处理方法，包括如下步骤：
    步骤S1、提供待处理图像；
    步骤S2、将待处理图像划分为平坦区、纹理区及边缘区；
    步骤S3、利用不同的双边滤波函数对平坦区、纹理区及边缘区进行降 噪处理，从而形成处理图像；
    其中，所述待处理图像为灰度图像；所述待处理图像包括呈阵列式排布的多个像素；每一像素具有对应的原始灰度值；
    所述步骤S1在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值；
    所述步骤S2具体包括：
    步骤S21、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值；定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值小于等于第二阈值的像素为纹理像素，定义梯度值大于第二阈值的像素为边缘像素；第一阈值小于第二阈值；
    步骤S22、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素、纹理像素及边缘像素的个数；
    步骤S23、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数同时大于边缘像素的个数时，对待分析区块进行形态学膨胀操作，否则保持待分析区块不变；
    步骤S24、重复上述步骤S23，直至多个区块均完成步骤S23的操作；
    步骤S25、将完成步骤S24的待处理图像划分为平坦区、边缘区及纹理区，其中，平坦区中的像素的梯度值小于第一阈值，纹理区中的像素的梯度值大于等于第一阈值小于等于第二阈值，边缘区中的像素的梯度值大于第二阈值；
    其中，所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得；
    所述待处理图像为经过压缩处理后的图像。
  8. 如权利要求7所述的图像处理方法，其中，所述步骤S3具体为：利用预设的第一双边滤波函数将平坦区中各个像素的原始灰度值转换为处理灰度值，利用预设的第二双边滤波函数将纹理区中各个像素的原始灰度值转换为处理灰度值，利用预设的第三双边滤波函数将边缘区中各个像素的原始灰度值转换为处理灰度值；
    所述第一双边滤波函数为
    Figure PCTCN2019081610-appb-100008
    其中，
    Figure PCTCN2019081610-appb-100009
    N(x)表示位于待处理图像中的一预设的滤波窗口所在的区域，y表示所述滤波窗口中的一像素的位置， x表示当前处理像素的位置，I(x)表示当前处理像素的原始灰度值，I(y)表示具有y位置的像素的原始灰度值，
    Figure PCTCN2019081610-appb-100010
    表示当前处理像素的处理灰度值，d1为预设的第一距离衰减率，r1为预设的第一灰度衰减率；
    所述第二双边滤波函数为
    Figure PCTCN2019081610-appb-100011
    其中，
    Figure PCTCN2019081610-appb-100012
    d2为预设的第二距离衰减率，r2为预设的第二灰度衰减率；
    所述第三双边滤波函数为
    Figure PCTCN2019081610-appb-100013
    其中，
    Figure PCTCN2019081610-appb-100014
    d3为预设的第三距离衰减率，r3为预设的第三灰度衰减率；
    所述第一距离衰减率大于第二距离衰减率，所述第二距离衰减率大于第三距离衰减率；
    所述第一灰度衰减率大于第二灰度衰减率，所述第二灰度衰减率大于第三灰度衰减率。
  9. 如权利要求7所述的图像处理方法，其中，所述步骤S21中，利用多个像素的归一化灰度值采用边缘侦测方法计算所述待处理图像的梯度信息。
  10. 如权利要求7所述的图像处理方法，其中，所述第一阈值为0.08，所述第二阈值为0.6。
  11. 一种图像处理方法，包括如下步骤：
    步骤S1’、提供待处理图像；
    步骤S2’、将待处理图像划分为平坦区及纹理区；
    步骤S3’、对平坦区的各个像素的原始灰度值进行基于预训练字典的稀疏表示处理，并保持纹理区的各个像素的原始灰度值不变，从而形成处理图像。
  12. 如权利要求11所述的图像处理方法，其中，所述待处理图像为灰度图像；所述待处理图像包括呈阵列式排布的多个像素；每一像素具有对应的原始灰度值；
    所述步骤S1’在提供待处理图像之后还对所述待处理图像的各个像素的原始灰度值进行归一化处理形成归一化灰度值；
    所述步骤S2’具体包括：
    步骤S21’、利用多个像素的归一化灰度值计算待处理图像的梯度信息，所述梯度信息包括分别与多个像素对应的梯度值；定义梯度值小于第一阈值的像素为平坦像素，定义梯度值大于等于第一阈值的像素为纹理像素；
    步骤S22’、将待处理图像划分依次设置的多个区块，计算每一区块中平坦像素及纹理像素的个数；
    步骤S23’、将多个区块中的一个设定为待分析区块，若待分析区块内的纹理像素的个数大于平坦像素的个数，则判定该待分析区块为纹理区，否则判定该待分析区块为平坦区；
    步骤S24’、重复上述步骤S23’，直至多个区块均完成步骤S23’的操作。
  13. 如权利要求11所述的图像处理方法，其中，所述待处理图像通过对彩色图像的YCbCr数据中的亮度通道进行提取获得；
    所述待处理图像为经过压缩处理后的图像。
  14. 如权利要求12所述的图像处理方法，其中，所述步骤S21’中，利用多个像素的归一化灰度值采用边缘侦测方法计算所述待处理图像的梯度信息；
    所述第一阈值为0.08。
PCT/CN2019/081610 2018-12-19 2019-04-04 图像处理方法 WO2020124873A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811558267.7A CN109584185B (zh) 2018-12-19 2018-12-19 图像处理方法
CN201811558267.7 2018-12-19

Publications (1)

Publication Number Publication Date
WO2020124873A1 true WO2020124873A1 (zh) 2020-06-25

Family

ID=65930142

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/081610 WO2020124873A1 (zh) 2018-12-19 2019-04-04 图像处理方法

Country Status (2)

Country Link
CN (1) CN109584185B (zh)
WO (1) WO2020124873A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117686096A (zh) * 2024-01-30 2024-03-12 大连云智信科技发展有限公司 一种基于目标检测的畜禽动物体温检测方法
CN118037734A (zh) * 2024-04-12 2024-05-14 济宁丰玺生物材料有限公司 一种全降解塑料制品质量检测方法

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111986095B (zh) * 2019-05-22 2024-03-19 上海哔哩哔哩科技有限公司 基于边缘提取的图像处理方法及图像处理装置
CN113450264A (zh) * 2020-03-27 2021-09-28 上海途擎微电子有限公司 去振铃效应的方法及其相关装置
CN111429383B (zh) * 2020-04-13 2022-09-27 展讯通信（上海）有限公司 图像降噪方法及装置、计算机可读存储介质
CN113628119B (zh) * 2020-05-06 2024-05-03 北京君正集成电路股份有限公司 一种图像去噪过程中识别待去噪块的方法
CN112365413A (zh) * 2020-10-30 2021-02-12 湖北锐世数字医学影像科技有限公司 图像处理方法、装置、设备、***及计算机可读存储介质
CN113256534B (zh) * 2021-06-16 2022-01-07 湖南兴芯微电子科技有限公司 一种图像增强方法、装置及介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103116875A (zh) * 2013-02-05 2013-05-22 浙江大学 自适应双边滤波图像去噪方法
CN104680485A (zh) * 2013-11-27 2015-06-03 展讯通信（上海）有限公司 一种基于多分辨率的图像去噪方法及装置
CN107358586A (zh) * 2017-07-06 2017-11-17 中山大学 一种图像增强方法、装置及设备

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7599569B2 (en) * 2006-01-13 2009-10-06 Ati Technologies, Ulc Method and apparatus for bilateral high pass filter
TW200828983A (en) * 2006-12-27 2008-07-01 Altek Corp Method of eliminating image noise
CN106991652A (zh) * 2017-03-03 2017-07-28 杭州电子科技大学 稀疏度约束和字典原子大小自适应的彩色图像修复方法
CN108335264B (zh) * 2018-01-30 2021-08-03 武汉大学 一种基于混合分辨率稀疏字典学习的图像超分辨率方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103116875A (zh) * 2013-02-05 2013-05-22 浙江大学 自适应双边滤波图像去噪方法
CN104680485A (zh) * 2013-11-27 2015-06-03 展讯通信（上海）有限公司 一种基于多分辨率的图像去噪方法及装置
CN107358586A (zh) * 2017-07-06 2017-11-17 中山大学 一种图像增强方法、装置及设备

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117686096A (zh) * 2024-01-30 2024-03-12 大连云智信科技发展有限公司 一种基于目标检测的畜禽动物体温检测方法
CN117686096B (zh) * 2024-01-30 2024-04-16 大连云智信科技发展有限公司 一种基于目标检测的畜禽动物体温检测方法
CN118037734A (zh) * 2024-04-12 2024-05-14 济宁丰玺生物材料有限公司 一种全降解塑料制品质量检测方法

Also Published As

Publication number Publication date
CN109584185A (zh) 2019-04-05
CN109584185B (zh) 2020-10-16

Similar Documents

Publication Publication Date Title
WO2020124873A1 (zh) 图像处理方法
US11127122B2 (en) Image enhancement method and system
Lee et al. Contrast enhancement based on layered difference representation of 2D histograms
Wang et al. A fast single-image dehazing method based on a physical model and gray projection
US9196024B2 (en) Method and apparatus for enhancing color
CN109325922B (zh) 一种图像自适应增强方法、装置及图像处理设备
US20200058102A1 (en) Background suppression method and detecting device in automatic optical detection of display panel
US10650500B2 (en) Image local contrast enhancement method
CN111598791B (zh) 一种基于改进动态大气散射系数函数的图像去雾方法
CN109919859B (zh) 一种户外场景图像去雾增强方法、计算设备及其存储介质
WO2021114564A1 (zh) 一种低对比度红外图像的增强方法
CN111612741B (zh) 一种基于失真识别的精确无参考图像质量评价方法
CN108280836B (zh) 一种图像处理方法及装置
TW201308997A (zh) 用於顯示裝置之影像品質提升方法
WO2022143421A1 (zh) 噪声强度估计方法、装置及电子设备
WO2016008259A1 (zh) 获取直方图的方法、亮度动态调整的方法及图像处理装置
Nevriyanto et al. Image enhancement using the image sharpening, contrast enhancement, and Standard Median Filter (Noise Removal) with pixel-based and human visual system-based measurements
WO2021128498A1 (zh) 图像自适应降噪方法及装置
CN115393216A (zh) 基于偏振特性与大气传输模型的图像去雾方法和装置
CN109064413B (zh) 图像对比度增强方法及采用其的图像采集医疗设备
Ding et al. Color image enhancement with a human visual system based adaptive filter
CN112825189B (zh) 一种图像去雾方法及相关设备
He et al. A night low‐illumination image enhancement model based on small probability area filtering and lossless mapping enhancement
CN109636749B (zh) 图像处理方法
CN117252773A (zh) 基于自适应颜色校正和导向滤波的图像增强方法及***

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: 19897776

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: 19897776

Country of ref document: EP

Kind code of ref document: A1