CN109377450B - Edge protection denoising method - Google Patents
Edge protection denoising method Download PDFInfo
- Publication number
- CN109377450B CN109377450B CN201810971999.2A CN201810971999A CN109377450B CN 109377450 B CN109377450 B CN 109377450B CN 201810971999 A CN201810971999 A CN 201810971999A CN 109377450 B CN109377450 B CN 109377450B
- Authority
- CN
- China
- Prior art keywords
- pixel
- pixels
- image
- gray
- filtering
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000001914 filtration Methods 0.000 claims abstract description 74
- 230000003044 adaptive effect Effects 0.000 claims abstract description 5
- 230000002146 bilateral effect Effects 0.000 claims description 11
- 230000014509 gene expression Effects 0.000 claims description 2
- 238000009499 grossing Methods 0.000 claims description 2
- 239000011159 matrix material Substances 0.000 claims description 2
- 239000000126 substance Substances 0.000 claims 2
- 238000012545 processing Methods 0.000 description 16
- 230000000694 effects Effects 0.000 description 8
- 238000003708 edge detection Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 235000002566 Capsicum Nutrition 0.000 description 4
- 239000006002 Pepper Substances 0.000 description 4
- 241000722363 Piper Species 0.000 description 4
- 235000016761 Piper aduncum Nutrition 0.000 description 4
- 235000017804 Piper guineense Nutrition 0.000 description 4
- 235000008184 Piper nigrum Nutrition 0.000 description 4
- 238000007796 conventional method Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 150000003839 salts Chemical class 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008094 contradictory effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The invention provides a denoising method for edge protection, which comprises the following steps: step 1, performing pixel classification on an input image according to the gray domain relevance of a central pixel and a neighborhood pixel of the image, and dividing the pixel into three types of pixels of an image homogeneous region or an edge region, a single-point noise pixel and a noise block region; and 2, adopting space domain and gray domain combined adaptive filtering under a multi-structure template for the 3 types of pixels with different attributes.
Description
Technical Field
The invention relates to an image processing technology, in particular to a denoising method for edge protection.
Background
The image quality is reduced due to the interference of various noise sources in the image acquisition and transmission process, so that the detection and identification capability of human eyes or machines on targets in the image is weakened, and meanwhile, the edge information of the image is detected as a target detection and tracking mode, so that in the digital image processing, the noise is filtered, the edge is reserved, and the method has important significance for improving the image definition and improving the visual effect. Low-pass filtering is often used to smooth the image, remove noise, and high-pass filtering is used to sharpen the image and enhance edges, which belong to two contradictory processing directions.
In the traditional image preprocessing method, such as median filtering, edges are blurred when Gaussian filtering removes noise and smoothes images, and in recent years, a plurality of edge-preserving filtering modes are proposed, such as guiding filtering indicates edge and non-edge areas through a local linear relation existing between a guiding image and a filtering output image, the non-edge areas are smoothed, and edge information is preserved; the bilateral filtering calculates the weight by simultaneously considering the information of the spatial domain and the pixel value domain, thereby achieving the filtering effect of protecting the edge; the least square filtering adjusts the regular term parameters through the gradient of the input image in the horizontal direction and the vertical direction, the weight of a smooth term is constrained, the smaller the gradient is, the smoother the output image is, the larger the gradient is, and the more complete the structure information of the output image is kept. However, the high-frequency noise is wrongly judged as an edge to be reserved, the noise processing capability is greatly limited, and the salt and pepper noise and the like cannot be effectively filtered.
Disclosure of Invention
The invention aims to provide a denoising method for edge protection, which can carry out adaptive filtering on an image, remove noise and simultaneously keep edge information.
The technical scheme for realizing the purpose of the invention is as follows: a denoising method for edge protection comprises the following steps:
step 1, performing pixel classification on an input image according to the gray domain relevance of a central pixel and a neighborhood pixel of the image, and dividing the pixel into three types of pixels of an image homogeneous region or an edge region, a single-point noise pixel and a noise block region;
and 2, adopting space domain and gray domain combined adaptive filtering under a multi-structure template for the 3 types of pixels with different attributes.
Compared with the prior art, the invention has the following advantages: compared with the traditional filtering method which is difficult to meet the defects of image denoising and edge protection, the method can carry out self-adaptive filtering, noise filtering and edge maintenance on the noise image, and is an effective filtering method which synthesizes smooth denoising of the image and can protect the edge.
The invention is further described below with reference to the accompanying drawings.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a comparison graph of the results of filtering processing on a lens graph added with 0.05 intensity noise using the method of the present invention and the conventional technique, respectively.
Fig. 3 is a comparison graph of the results of filtering processing on a lens graph added with 0.25-intensity noise using the method of the present invention and the conventional technique, respectively.
Fig. 4 is a comparison graph of the results of edge detection after filtering the lena graph added with 0.05 intensity noise using the method of the present invention and the conventional technique.
FIG. 5 is a comparison graph of the results of edge detection after filtering noisy microimages using the method of the present invention and the conventional technique.
Detailed Description
With reference to fig. 1, a denoising method for edge protection includes the following steps:
step 1, performing pixel classification on an input image according to the gray level correlation between pixels in the image and neighborhood pixels, and dividing the pixels into three types of image homogeneous region or edge region pixels, single-point noise pixels and noise block region pixels;
step 2, filtering and denoising the classified pixels and the templates in a corresponding mode, comprising: carrying out bilateral filtering processing on pixels in the image homogeneous region and pixels at the edge of the image; carrying out eight-neighborhood median filtering processing on the single-point noise pixel; and (5) carrying out special-shaped homogeneous filtering processing on the noise block area.
The pixels in the step 1 refer to all pixels in the image, but for each template, only the attribute of the central pixel can be judged each time, and the attributes of all pixel points can be obtained or obtained through the movement adjustment of the position of the template in the image.
The template in step 2 refers to a matrix block area (in a special case, some area blocks selected according to a certain rule may also be selected) in the image, and the image processing is generally performed in this area.
The specific process of the step 1 is as follows:
step 1.1, setting a gray level correlation rule among pixels: the size of the template for judging the gray level correlation is m multiplied by m, i is a central pixel point, and the gray level of the pixel under each template is ixAnd t represents the distance between the neighborhood pixel and the center pixel, ktIs a gray level difference threshold value, the larger t is, the threshold value ktThe larger; defining the expression that the central pixel has complete relevance with the neighborhood pixel as:
|i-ix|＜＝kt (1)
wherein k istThe values may be set according to the actual scene and imaging device performance parameters.
Step 1.2, calculating the number z of neighborhood pixels which have complete relevance with a central pixel in the template, wherein the z has the following conditions:
z＝0 (3)
step 1.3, judging pixel attributes according to the result obtained in step 1.2: when z satisfies formula (2), indicating that the gray levels of the neighborhood pixels at least on one side of the template are similar to the gray level of the central pixel, wherein the pixels are pixels in the image homogeneous region or pixels at the edge of the image; when z satisfies the formula (3), indicating that the neighborhood pixels in all directions do not satisfy the relevance with the central pixel, and judging the pixel as a single-point noise pixel; when z satisfies the formula (4), the central pixel point is in a noise block area and an image arc edge or an angular point, and the specific attribute of the pixel is judged by combining the gray level change trend;
firstly, finding out a pixel point closest to the gray level of a central pixel on a 3-order template (namely a region with the size of 3 x 3, namely a small region with the size of 3 x 3 is divided in a large image) and marking as igThen, finding out the outer layer pixel point on the extension line of the central pixel and the pixel point, and marking as ihI is tohComparing the gray value with the gray value of the central pixel point, wherein the existing conditions are as follows:
i≈ig&|i-ih|<＝kt (5)
i≈ig&|i-ih|>kt (6)
equation (5) indicates that pixels i and ihIf the correlation is satisfied, the central pixel is located on the arc edge or the angular point, the formula (6) indicates that the central pixel is located in the noise block region, and the noise pollution direction is the direction in which the central pixel points to the pixel point closest to the gray level of the central pixel, and the pixel in the direction is the noise point more likely at the momentAnd if the pixel is large, selecting the pixel with the reverse direction as the pixel participating in the filtering to perform the filtering operation on the central point.
The operation of step 1 is completed, and the pixel classification result is obtained as follows: the pixels satisfying the formulas (2) and (5) are pixels in the homogeneous region or edge pixels of the image; the pixel satisfying the formula (3) is a single-point noise pixel; the pixel satisfying the equation (6) is a noise block region pixel.
In step 2, filtering processing is performed on the 3 types of pixels with different attributes in different manners. The method specifically comprises the following steps:
and 2.1, for pixels in a homogeneous region or pixels in an edge region of the image, weakening the smooth intensity and strengthening the edge protection, taking pixels similar to the gray scale region in the neighborhood as a filtering template, distributing weights according to spatial relevance, and filtering the image, wherein bilateral filtering is a typical method with the effect and is taken as a filtering mode of the image. The method comprises the following two aspects: the method comprises the following steps: (1) spatial filtering: carrying out weighted average on adjacent pixels on the pixel point space; (2) value domain filtering: and carrying out weighted average on pixels adjacent to the pixel values of the pixel points. The spatial filtering and the value domain filtering together form bilateral filtering, which is expressed as multiplication of weights, namely, only neighborhood pixels which are adjacent in a spatial domain and similar in a gray domain can have great influence on a filtering result.
Let the image to be processed be I, where the coordinate point p ═ x, y is the position of the pixel point to be processed, and the gray value is IpLet the filtered image be U and the gray scale at coordinate point p ═ x, y be Up. S is a region with a p point as a center and N as a neighborhood radius, q is (u, v) is a neighborhood pixel point of a pixel p in the S region, and the gray value is Iq. The bilateral filtering output image is:
wherein:
in the above formula, Cd,rTo normalize the factor, wdIs a spatial proximity factor, wrIs a gray level similarity factor, σdAs a spatial distance weight coefficient parameter, σrIs a luminance weight coefficient parameter.
Step 2.2: the number of pixels in the neighborhood which are associated with the gray level is 0, in order to remove noise and minimize the edge blurring effect, the pixels are processed in a small space domain, and the typical mode is low-order template median filtering. For single point noise pixels, low order template median filtering is used. When the pixel is determined to be a single-point noise pixel, the median filtering of the low-order template can protect the pixel at the edge of the image as much as possible, and the pixel is prevented from being subjected to excessive smooth blurring when the pixel is at the edge of the image. The single-point noise pixel median filtering output image is:
if＝med(ik-N,ik-N+1,…,ik+N-1,ik+N) (11)
wherein 2N +1 is m × m, which is the number of pixels in the filtering template, i represents the gray value of the pixel, and the subscript of i represents different positions in the template, and the combination is the pixels at different positions in the template. Within the template, by taking the value of k, a range of pixel numbers for median filtering can be defined. .
Step 2.3: for the pixels in the noise block area, one or more pixels with the gray levels different from the gray levels of the pollution-free image exist in the pixel neighborhood, if simple median filtering or mean filtering is adopted, the output result is influenced by other noise points in the neighborhood and is easy to deviate from an expected processing value, so that the abnormal homogeneous filtering is carried out by judging the noise variation trend, selecting the reverse pixels in the noise spreading direction as the filtering template pixels. The noise block region pixel filtering output is:
wherein
For normalizing the coefficients, n represents the number of weight classes, SnIndicating the area of the pixel at each weight level, NSnRepresenting the number of pixels at each weight level; k is a radical ofnRepresenting the weight coefficients. The weight coefficient is inversely proportional to the distance from the center pixel.
Dividing the pixels into n regions according to the weight level, i of each partpRepresent the gray value of the pixel corresponding to this region (the pixel of a region, not a single pixel, is used for weighted summation).
According to the distance judgment of the pixels in the template from the noise block area, the pixels can be divided into several distance grades according to the requirements (more grades are divided when the precision is high, and less grades are divided when the precision is low), and the pixels with the same distance are divided into one area.
To demonstrate the advantages of the present invention, several sets of tests were performed:
test 1: the method of the invention and the traditional technology are used for respectively carrying out filtering processing on the lena graph added with 0.05 intensity noise, and an effect graph 2 is obtained.
Test 2: the method of the invention and the traditional technology are used for respectively carrying out filtering processing on the lena graph added with 0.25 intensity noise, and an effect graph 3 is obtained.
Test 3: the method and the traditional technology are used for respectively carrying out the edge extraction experiment on the filtering result graph in the experiment 1 to obtain an edge extraction effect graph 4.
Experiment 4: the method and the traditional technology are respectively used for carrying out an edge extraction experiment on the low-light-level image with noise to obtain an edge extraction effect graph 5.
The filtering results shown in fig. 2 and 3 show that: the noise image is processed by a plurality of filtering methods, the median filtering and the method are used for obviously filtering the salt and pepper noise, and the guided filtering and the bilateral filtering can not effectively filter the salt and pepper noise.
Fig. 4 and 5 show that: for the Lena image added with the salt and pepper noise with the intensity of 0.05 and the noise-containing low-light-level image, the original image subjected to smoothing processing is not added with filtering, and more noise points are mistakenly detected as edges; median filtering can filter noise but causes serious loss of edge information, edge breakage and information loss; the edge information of the image denoised by the guided filtering and bilateral filtering algorithms is relatively complete, but the image denoised by the guided filtering and bilateral filtering algorithms is limited by the denoising performance, and more parts which falsely detect noise as edges exist in the edge detection result; the image denoised by the filtering of the algorithm has the advantages that the influence of noise on the edge detection result is minimum, the edge information is rich, the breakpoint is few, the contour is complete, the image contour has more meticulous and accurate image contour compared with the guided filtering and the bilateral filtering, and the edge detection effect is ideal.
In summary, with the aid of the above technical solutions of the present invention, the edge-protection denoising method of the present invention implements a processing method combining image classification and adaptive filtering processing. Good experimental results were obtained. Compared with the traditional method, the method can effectively filter noise, smooth the image and protect the image edge. After the image containing noise is filtered and output by the method, the noise error detection and the edge missing detection are obviously inhibited in the edge extraction result, and the obtained image profile is relatively ideal.
Claims (4)
1. A denoising method for edge protection is characterized by comprising the following steps:
step 1, performing pixel classification on an input image according to the gray domain relevance of a central pixel and a neighborhood pixel of the image, and dividing the pixel into three types of pixels of an image homogeneous region or an edge region, a single-point noise pixel and a noise block region;
step 2, adopting space domain and gray domain combined adaptive filtering under a multi-structure template for the 3 types of pixels with different attributes;
the specific process of the step 1 is as follows:
step 1.1, the gray difference between a neighborhood pixel and a center pixel in a template is in a certain range, and if the gray difference indicates that the neighborhood pixel and the center pixel have gray domain relevance; wherein the template is a matrix square area in the image;
step 1.2, calculating the number z of neighborhood pixels which are associated with the gray level of the central pixel in the template,
z＝0 (2)
wherein, m × m is the size of the template for judging the gray level correlation;
step 1.3, finding out a pixel point closest to the gray level of the central pixel on a 3-order template, and marking as ig；
Finding the central pixel and the pixel point igOuter layer pixel point on the connecting line extension line is marked as ih；
Will ihComparing the gray value with the gray value of the central pixel point, the following two conditions exist
i≈ig&|i-ih|<＝kt (4)
i≈ig&|i-ih|>kt (5)
Wherein k istIs the gray level difference threshold, i is the center pixel;
step 1.4, if z simultaneously satisfies the formulas (1) and (4), the pixel is an image homogeneous region or an edge region pixel;
if z satisfies formula (2), the pixel is a single-point noise pixel;
if z satisfies both the expressions (3) and (5), the pixel is a noise block region pixel.
2. The method according to claim 1, wherein the step 2 is to perform smoothing and denoising on the pixels in the homogeneous region or the pixels in the edge region of the image by using bilateral filtering, and the specific process is as follows:
step 2.1.1, let the image to be processed be I, where the coordinate point p ═ x, y is the position of the pixel point to be processed, and the gray value is Ip(ii) a Let the filtered image be U and the gray scale at coordinate point p ═ x, y be Up(ii) a In addition, S is a region with a p point as a center and N as a neighborhood radius, q is (u, v) is a neighborhood pixel point of a pixel p in the S region, and the gray value is Iq；
Step 2.1.2, obtaining a bilateral filtering output image U (p) as follows:
wherein the content of the first and second substances,
Cd,rto normalize the factor, wdIs a spatial proximity factor, wrIs a gray level similarity factor, σdAs a spatial distance weight coefficient parameter, σrIs a luminance weight coefficient parameter.
3. Method according to claim 1, characterized in that the output image i after filtering for single point noise pixels in step 2fComprises the following steps:
if＝med(ik-N,ik-N+1,L,ik+N-1,ik+N) (11)
where 2N +1 is m × m, which is the number of pixels in the filtering template, i is the gray value of the pixel, and the subscript of i represents different positions in the template.
4. The method of claim 1, wherein in step 2, for the noise block region pixels, the pixels are first divided into n regions, and the filtered output image is
Wherein the content of the first and second substances,
in order to normalize the coefficients of the coefficients,
iprepresenting the gray value of the pixel corresponding to one of the n regions,
n represents the number of weight classes, SnIndicating the area where the pixels of each weight level are located,
representing the number of pixels at each weight level;
knrepresenting the weight coefficients.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810971999.2A CN109377450B (en) | 2018-08-24 | 2018-08-24 | Edge protection denoising method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810971999.2A CN109377450B (en) | 2018-08-24 | 2018-08-24 | Edge protection denoising method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109377450A CN109377450A (en) | 2019-02-22 |
| CN109377450B true CN109377450B (en) | 2021-12-10 |
Family
ID=65403929
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810971999.2A Active CN109377450B (en) | 2018-08-24 | 2018-08-24 | Edge protection denoising method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109377450B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110827577B (en) * | 2019-04-11 | 2021-10-26 | 重庆西高科技有限公司 | Early warning system based on parameter identification |
| CN111882504B (en) * | 2020-08-05 | 2022-08-16 | 展讯通信（上海）有限公司 | Method and system for processing color noise in image, electronic device and storage medium |
| CN113191961B (en) * | 2021-03-19 | 2022-10-25 | 聚融医疗科技(杭州)有限公司 | Ultrasonic denoising method and system based on bilateral filtering and feature recognition |
| CN113222853B (en) * | 2021-05-26 | 2022-07-12 | 武汉博宇光电系统有限责任公司 | Progressive infrared image noise reduction method based on noise estimation |
| CN115661135B (en) * | 2022-12-09 | 2023-05-05 | 山东第一医科大学附属省立医院(山东省立医院) | Lesion area segmentation method for cardiovascular and cerebrovascular angiography |
| CN115994870B (en) * | 2023-02-23 | 2023-12-19 | 苏州梅曼智能科技有限公司 | Image processing method for enhancing denoising |
| CN116681703B (en) * | 2023-08-03 | 2023-10-10 | 杭州鸿世电器股份有限公司 | Intelligent switch quality rapid detection method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009078717A2 (en) * | 2007-12-17 | 2009-06-25 | Nederlanse Organisatie Voor Toegepast-Natuurwetenschappelijk Onderzoek Tno | Image processing apparatus and method with contrast correction |
| CN101860667A (en) * | 2010-05-06 | 2010-10-13 | 中国科学院西安光学精密机械研究所 | Method of quickly eliminating composite noise in images |
| CN101950414A (en) * | 2010-09-02 | 2011-01-19 | 西安电子科技大学 | Non-local mean de-noising method for natural image |
| CN104376540A (en) * | 2014-12-01 | 2015-02-25 | 北京思比科微电子技术股份有限公司 | Bayer image denoising method |
| CN105787902A (en) * | 2016-03-22 | 2016-07-20 | 天津大学 | Image noise reduction method which utilizes partitioning ordering to detect noise |
| CN105913396A (en) * | 2016-04-11 | 2016-08-31 | 湖南源信光电科技有限公司 | Noise estimation-based image edge preservation mixed de-noising method |
| CN106023204A (en) * | 2016-05-20 | 2016-10-12 | 陕西师范大学 | Method and system for removing mosquito noise based on edge detection algorithm |
| CN106651938A (en) * | 2017-01-17 | 2017-05-10 | 湖南优象科技有限公司 | Depth map enhancement method blending high-resolution color image |
| CN106846281A (en) * | 2017-03-09 | 2017-06-13 | 广州四三九九信息科技有限公司 | image beautification method and terminal device |
| CN108109113A (en) * | 2017-11-09 | 2018-06-01 | 齐鲁工业大学 | Single image to the fog method and device based on bilateral filtering and medium filtering |
-
2018
- 2018-08-24 CN CN201810971999.2A patent/CN109377450B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009078717A2 (en) * | 2007-12-17 | 2009-06-25 | Nederlanse Organisatie Voor Toegepast-Natuurwetenschappelijk Onderzoek Tno | Image processing apparatus and method with contrast correction |
| CN101860667A (en) * | 2010-05-06 | 2010-10-13 | 中国科学院西安光学精密机械研究所 | Method of quickly eliminating composite noise in images |
| CN101950414A (en) * | 2010-09-02 | 2011-01-19 | 西安电子科技大学 | Non-local mean de-noising method for natural image |
| CN104376540A (en) * | 2014-12-01 | 2015-02-25 | 北京思比科微电子技术股份有限公司 | Bayer image denoising method |
| CN105787902A (en) * | 2016-03-22 | 2016-07-20 | 天津大学 | Image noise reduction method which utilizes partitioning ordering to detect noise |
| CN105913396A (en) * | 2016-04-11 | 2016-08-31 | 湖南源信光电科技有限公司 | Noise estimation-based image edge preservation mixed de-noising method |
| CN106023204A (en) * | 2016-05-20 | 2016-10-12 | 陕西师范大学 | Method and system for removing mosquito noise based on edge detection algorithm |
| CN106651938A (en) * | 2017-01-17 | 2017-05-10 | 湖南优象科技有限公司 | Depth map enhancement method blending high-resolution color image |
| CN106846281A (en) * | 2017-03-09 | 2017-06-13 | 广州四三九九信息科技有限公司 | image beautification method and terminal device |
| CN108109113A (en) * | 2017-11-09 | 2018-06-01 | 齐鲁工业大学 | Single image to the fog method and device based on bilateral filtering and medium filtering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109377450A (en) | 2019-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109377450B (en) | Edge protection denoising method | |
| CN108921800B (en) | Non-local mean denoising method based on shape self-adaptive search window | |
| CN108765325B (en) | Small unmanned aerial vehicle blurred image restoration method | |
| CN108389175B (en) | Image defogging method integrating variation function and color attenuation prior | |
| CN111080661B (en) | Image-based straight line detection method and device and electronic equipment | |
| CN105913396A (en) | Noise estimation-based image edge preservation mixed de-noising method | |
| CN111833366A (en) | Edge detection method based on Canny algorithm | |
| CN108038833B (en) | Image self-adaptive sharpening method for gradient correlation detection and storage medium | |
| CN109003233B (en) | Image denoising method based on self-adaptive weight total variation model | |
| CN105139391B (en) | A kind of haze weather traffic image edge detection method | |
| Soni et al. | An improved image dehazing technique using CLAHE and guided filter | |
| CN107292834B (en) | Infrared image detail enhancement method | |
| CN116468641A (en) | Infrared blood vessel image enhancement processing method | |
| CN111861896A (en) | UUV-oriented underwater image color compensation and recovery method | |
| CN103971345B (en) | A kind of image de-noising method based on improvement bilateral filtering | |
| Yu et al. | Image and video dehazing using view-based cluster segmentation | |
| Bao et al. | An edge-preserving filtering framework for visibility restoration | |
| Chen et al. | Improve transmission by designing filters for image dehazing | |
| Wu et al. | An improved Canny edge detection algorithm with iteration gradient filter | |
| CN105913391A (en) | Defogging method based on shape variable morphological reconstruction | |
| CN106611419B (en) | The extracting method in image road surface region | |
| Kusnik et al. | On the robust technique of mixed Gaussian and impulsive noise reduction in color digital images | |
| CN115829967A (en) | Industrial metal surface defect image denoising and enhancing method | |
| CN115034985A (en) | Underwater image enhancement method | |
| Jin et al. | An approach to remove impulse noise from a corrupted image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |