CN103871026A - Image denoising method and system - Google Patents
Image denoising method and system Download PDFInfo
- Publication number
- CN103871026A CN103871026A CN201210531822.3A CN201210531822A CN103871026A CN 103871026 A CN103871026 A CN 103871026A CN 201210531822 A CN201210531822 A CN 201210531822A CN 103871026 A CN103871026 A CN 103871026A
- Authority
- CN
- China
- Prior art keywords
- pixel
- matrix
- noise reduction
- image
- region
- 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.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention provides an image denoising method. The method comprises performing pixel point area specific segmentation on images to be denoised, performing specific selection on segmented pixel point areas, performing specific matrix transformation and filtering processing on a pixel value matrix corresponding to each selected pixel point area, and solving an average value of the pixel values of specific pixel points which are subjected to the specific matrix transformation and the filtering processing such that while the denoising effect is ensured, the denoising computation amount is effectively reduced, and the denoising efficiency is also improved. The invention also provides an image denoising system.
Description
Technical field
The present invention relates to a kind of image processing techniques, particularly a kind of image denoising method and system.
Background technology
Image is often subject to the impact such as imaging device and external environmental noise interference in digitizing and transmitting procedure, is called noisy image or noise image, and the process that reduces noise in digital picture is called image noise reduction.
Noise is the major reason of image disruption.May there are various noises in piece image, these noises may produce in actual applications in transmission, also may in quantizing to wait processing, produce.Can be divided into three kinds of forms (f (x, y) represents given original image, g (x, y) presentation video signal, n (x, y) represents noise) according to the relation of noise and signal:
1) additive noise, this noise like and received image signal are irrelevant, and noisy image can be expressed as f (x; y)=g (x; y)+n (x, y), the noise producing when the camera-scanning image of interchannel noise and vidicon just belongs to this noise like;
2) multiplicative noise, this noise like is relevant with picture signal, noisy image can be expressed as f (x, y)=g (x, y)+n (x, y) × g (x, y), for example, noise when flying-spot scanner scan image, coherent noise in television image, the grain noise in film just belongs to this noise like;
3) quantizing noise, this noise like and received image signal are irrelevant, are that quantizing process exists quantization error, then are reflected to receiving end and produce.
Therefore, how fast and accurately image to be carried out to the large problem that noise reduction is industry.Along with informationalized development, the application of great amount of images has become a kind of normality, but, the Image Denoising that people adopt at present has been difficult to satisfy the demands, for example, existing noise reduction technology is more for the overlapping region of each pixel of image, and operand is larger, and the system calculation resources taking when computing is more.
Summary of the invention
Fundamental purpose of the present invention is to provide a kind of image denoising method, with in the situation that ensureing noise reduction, effectively reduces noise reduction operand, improves noise reduction efficacy.
In addition, also provide a kind of image noise reduction system, with in the situation that ensureing noise reduction, effectively reduce noise reduction operand, improve noise reduction efficacy.
A kind of image denoising method, is applicable to data processing equipment, is characterised in that, the method comprising the steps of: A, obtain the image for the treatment of noise reduction, treat noise reduction image and carry out pixel Region Segmentation according to default step-length, matrix line number and columns; B, selecting the pixel region cut apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number, is that each pixel region of selecting generates a pixel value matrix; C, each the pixel value matrix generating is carried out to matrixing and filtration treatment to generate pixel value filtered matrix, pixel value to corresponding same pixel point in all pixel value filtered matrix is averaged, and pixel value after noise reduction using average as this same pixel point.
Whether further,, after step C, the method also comprises: analyze and have the pixel region of cutting apart not selected; In the time having the pixel region of cutting apart not selected, proceed to execution step B, or, while being all selected in the pixel region of cutting apart, process ends.
Further, describedly each the pixel value matrix generating is carried out to matrixing and filtration treatment comprise with the step that generates pixel value filtered matrix: E1, each the pixel value matrix generating is converted to generate corresponding pixel frequency spectrum matrix by specific mapping algorithm; E2, according to default threshold values, the element value in each pixel frequency spectrum matrix is filtered, to generate corresponding pixel frequency spectrum filtered matrix; E3, to generate each pixel frequency spectrum filtered matrix carry out inverse transformation by specific inverse transformation algorithm, to generate corresponding pixel value filtered matrix.
Further, described specific mapping algorithm is dct algorithm, and described specific inverse transformation algorithm is inverse discrete cosine transformation algorithm.
Further, described step e 2 is: the element value that is less than or equal to default threshold values in each pixel frequency spectrum matrix is carried out to zero clearing processing, to generate corresponding pixel frequency spectrum filtered matrix.
A kind of image noise reduction system, is applied to data processing equipment.This system comprises: Region Segmentation module, for obtaining the image for the treatment of noise reduction, treat noise reduction image and carry out pixel Region Segmentation according to default step-length, matrix line number and columns, select the pixel region of cutting apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number; Matrixing module, is used to each pixel region of selection to generate a pixel value matrix, and each the pixel value matrix generating is carried out to matrixing and filtration treatment to generate pixel value filtered matrix; And noise reduction computing module, for the pixel value of the corresponding same pixel point of all pixel value filtered matrix is averaged, and pixel value after noise reduction using average as this same pixel point.
Further, whether described noise reduction computing module also has the pixel region of cutting apart not selected for analyzing; Described Region Segmentation module, in the time having the pixel region of cutting apart not selected, selects the pixel region of cutting apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number.
Further, described matrixing module is used for: each the pixel value matrix generating is converted to generate corresponding pixel frequency spectrum matrix by specific mapping algorithm; According to default threshold values, the element value in each pixel frequency spectrum matrix is filtered, to generate corresponding pixel frequency spectrum filtered matrix; Each the pixel frequency spectrum filtered matrix generating is carried out to inverse transformation by specific inverse transformation algorithm, to generate corresponding pixel value filtered matrix.
Further, described specific mapping algorithm is dct algorithm, and described specific inverse transformation algorithm is inverse discrete cosine transformation algorithm.
Further, described matrixing module is used for: the element value that each pixel frequency spectrum matrix is less than or equal to default threshold values carries out zero clearing processing, to generate corresponding pixel frequency spectrum filtered matrix.
Compare prior art, the present invention is undertaken by treating noise reduction image that pixel region is specific to be cut apart, specific selection is carried out in the pixel region of cutting apart, pixel value matrix corresponding to each pixel region of selecting carried out to particular matrix conversion and filtration treatment, and the pixel value of the specific pixel point through particular matrix conversion and filtration treatment is averaged, when having ensured noise reduction, effectively reduce noise reduction operand, improved noise reduction efficacy.
Brief description of the drawings
Fig. 1 is the operation Organization Chart of image noise reduction system of the present invention preferred embodiment.
Fig. 2 is the functional block diagram of image noise reduction system in Fig. 1.
Fig. 3 is the exemplary plot that in Fig. 1, image noise reduction system is carried out pixel Region Segmentation to image.
Fig. 4 is the concrete implementing procedure figure of image denoising method preferred embodiment of the present invention.
Realization, functional characteristics and the advantage of the object of the invention, in connection with embodiment, are described further with reference to accompanying drawing.
Embodiment
Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.
As shown in Figure 1, be the operation Organization Chart of image noise reduction system of the present invention preferred embodiment.This image noise reduction system 11 runs in data processing equipment 1.Described data processing equipment 1 can be mobile phone, panel computer, personal digital assistant (Personal Digital Assistant, PDA) or other any suitable data processing equipments.Described data processing equipment 1 comprises storage unit 13 and processing unit 10.
This storage unit 13, for storing this image noise reduction system 11, and the service data of this image noise reduction system 11.
This processing unit 10, for calling and carry out this image noise reduction system 11, to realize the noise reduction to image.
As shown in Figure 2, be the functional block diagram of image noise reduction system in Fig. 1.This image noise reduction system 11 comprises Region Segmentation module 110, matrixing module 111 and noise reduction computing module 112.
This Region Segmentation module 110, for obtaining the image for the treatment of noise reduction.In the present embodiment, this Region Segmentation module 110 is obtained the image for the treatment of noise reduction from this storage unit 13; In other embodiments of the invention, this Region Segmentation module 110 is obtained the image for the treatment of noise reduction from other any suitable devices.
This Region Segmentation module 110, also carries out pixel Region Segmentation for treating noise reduction image according to default step-length, matrix line number and columns.In the present embodiment, described step-length refers to the pixel number (step-length is taking 4 as example) at adjacent pixel cut zone interval, described matrix line number equals the number (matrix line number is taking 8 as example) of each each row pixel of pixel cut zone, and described matrix columns equals the number (matrix columns is taking 8 as example) of the every a line pixel of each pixel cut zone.Example as shown in Figure 3, the number of each row pixel of pixel cut zone O-O2-B2-B, A-A2-C2-C, O1-O3-B3-B1 or A1-A3-C3-C1 be 8 and the number of every a line pixel be 8, the pixel number that adjacent pixel cut zone O-O2-B2-B and A-A2-C2-C are separated by is 4, and the pixel number that adjacent pixel cut zone O-O2-B2-B and O1-O3-B3-B1 are separated by is 4.In other embodiments of the invention, described default step-length, matrix line number and columns can also be other any suitable numbers.
This Region Segmentation module 110, also, in the time having the pixel region of cutting apart not selected, selects the pixel region of cutting apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number.In the present embodiment, described default number is 4.As shown in Figure 3, this Region Segmentation module 110 is selected O-O2-B2-B, A-A2-C2-C, O1-O3-B3-B1 and A1-A3-C3-C1 region to example, and the pixel region of these 4 selections all comprises the region A1-A2-B2-B1 of part same pixel point.In other embodiments of the invention, described default number can also be other any suitable numbers.
This matrixing module 111, is used to each pixel region of selection to generate a pixel value matrix, and each the pixel value matrix generating is converted to generate corresponding pixel frequency spectrum matrix by specific mapping algorithm.In the present embodiment, described specific mapping algorithm is DCT(Discrete Cosine Transform, discrete cosine transform) algorithm.In other embodiments of the invention, described specific mapping algorithm is other any suitable algorithms.
This matrixing module 111, also for according to default threshold values, the element value of each pixel frequency spectrum matrix being filtered, to generate corresponding pixel frequency spectrum filtered matrix.In the present embodiment, this matrixing module 111 is filtered the element value that is less than or equal to default threshold values in each pixel frequency spectrum matrix, i.e. zero clearing processing.In other embodiments of the invention, this matrixing module 111 is carried out other any suitable filtration treatment according to default threshold values to the element value in each pixel frequency spectrum matrix.In the present embodiment, described default threshold values is unification, identical; In other embodiments of the invention, described default threshold values has multiple, is respectively used to filter specific element value in each pixel frequency spectrum matrix.
This matrixing module 111, also for each the pixel frequency spectrum filtered matrix generating is carried out to inverse transformation by specific inverse transformation algorithm, to generate corresponding pixel value filtered matrix.In the present embodiment, described specific inverse transformation algorithm is IDCT(Inverse Discrete Cosine Transform, inverse discrete cosine transformation) algorithm.In other embodiments of the invention, described specific inverse transformation algorithm is other any suitable algorithms.
This noise reduction computing module 112, for the pixel value of the corresponding same pixel point of all pixel value filtered matrix is averaged, and pixel value after noise reduction using average as this same pixel point.For example, each pixel in the A1-A2-B2-B1 of pixel region is to there being a described average.
Whether this noise reduction computing module 112, also have the pixel region of cutting apart not selected by this Region Segmentation module 110 for analyzing.
As shown in Figure 4, be the concrete implementing procedure figure of image denoising method preferred embodiment of the present invention.
It is emphasized that: process flow diagram shown in Fig. 4 is only a preferred embodiment, those skilled in the art is when knowing, any embodiment building around inventive concept should not depart from the scope containing in following technical scheme:
Obtain the image for the treatment of noise reduction; Treat noise reduction image and carry out pixel Region Segmentation according to default step-length, matrix line number and columns; In the time having the pixel region of cutting apart not selected, select the pixel region of cutting apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number; For each pixel region of selecting generates a pixel value matrix; Each the pixel value matrix generating is carried out to matrixing and filtration treatment to generate pixel value filtered matrix; Pixel value to corresponding same pixel point in all pixel value filtered matrix is averaged, and pixel value after noise reduction using average as this same pixel point.
It is below the noise reduction process that progressively realizes image in conjunction with the present embodiment.
Step S10, obtains the image for the treatment of noise reduction.In the present embodiment, from this storage unit 13, obtain the image for the treatment of noise reduction; In other embodiments of the invention, from other any suitable devices, obtain the image for the treatment of noise reduction.
Step S11, treats noise reduction image and carries out pixel Region Segmentation according to default step-length, matrix line number and columns.
Step S12, selects the pixel region of cutting apart to make the pixel region of all selections comprise the region of part same pixel point according to default number.In the present embodiment, described default number is 4.Example as shown in Figure 3, is selected O-O2-B2-B, A-A2-C2-C, O1-O3-B3-B1 and A1-A3-C3-C1 region, and the pixel region of these 4 selections all comprises the region A1-A2-B2-B1 of part same pixel point.
Step S13, for each pixel region of selecting generates a pixel value matrix, and converts to generate corresponding pixel frequency spectrum matrix to each the pixel value matrix generating by specific mapping algorithm.In the present embodiment, described specific mapping algorithm is DCT(Discrete Cosine Transform, discrete cosine transform) algorithm.
Step S15, filters the element value in each pixel frequency spectrum matrix according to default threshold values, to generate corresponding pixel frequency spectrum filtered matrix.In the present embodiment, the element value that is less than (or being less than or equal to) default threshold values in each pixel frequency spectrum matrix is filtered to i.e. zero clearing processing.In other embodiments of the invention, according to default threshold values, the element value in each pixel frequency spectrum matrix is carried out to other any suitable filtration treatment.In the present embodiment, described default threshold values is unification, identical; In other embodiments of the invention, described default threshold values has multiple, is respectively used to filter specific element value in each pixel frequency spectrum matrix.
Step S16, carries out inverse transformation to each the pixel frequency spectrum filtered matrix generating by specific inverse transformation algorithm, to generate corresponding pixel value filtered matrix.In the present embodiment, described specific inverse transformation algorithm is IDCT(Inverse Discrete Cosine Transform, inverse discrete cosine transformation) algorithm.In other embodiments of the invention, described specific inverse transformation algorithm is other any suitable algorithms.
Step S17, averages to the pixel value of corresponding same pixel point in all pixel value filtered matrix, and pixel value after noise reduction using average as this same pixel point.For example, each pixel in the A1-A2-B2-B1 of pixel region is to there being a described average.
Whether step S18, analyze and have the pixel region of cutting apart not selected.
In the time having the pixel region of cutting apart not selected, proceed to and carry out above-mentioned steps S12, or while being all selected in the pixel region of cutting apart, flow process finishes.
These are only the preferred embodiments of the present invention; not thereby limit the scope of the claims of the present invention; every equivalent structure or conversion of equivalent flow process that utilizes instructions of the present invention and accompanying drawing content to do; or be directly or indirectly used in other relevant technical fields, be all in like manner included in scope of patent protection of the present invention.
Claims (10)
1. an image denoising method, is characterized in that, the method comprising the steps of:
A, obtain the image for the treatment of noise reduction, treat noise reduction image and carry out pixel Region Segmentation according to default step-length, matrix line number and columns;
B, selecting the pixel region cut apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number, is that each pixel region of selecting generates a pixel value matrix;
C, each the pixel value matrix generating is carried out to matrixing and filtration treatment to generate pixel value filtered matrix, pixel value to corresponding same pixel point in all pixel value filtered matrix is averaged, and pixel value after noise reduction using average as this same pixel point.
2. image denoising method as claimed in claim 1, is characterized in that, after step C, the method also comprises:
Analyze and whether have the pixel region of cutting apart not selected;
In the time having the pixel region of cutting apart not selected, proceed to execution step B.
3. image denoising method as claimed in claim 1 or 2, is characterized in that, described each pixel value matrix to generation carries out matrixing and filtration treatment comprises with the step that generates pixel value filtered matrix:
E1, to generate each pixel value matrix convert to generate corresponding pixel frequency spectrum matrix by specific mapping algorithm;
E2, according to default threshold values, the element value in each pixel frequency spectrum matrix is filtered, to generate corresponding pixel frequency spectrum filtered matrix;
E3, to generate each pixel frequency spectrum filtered matrix carry out inverse transformation by specific inverse transformation algorithm, to generate corresponding pixel value filtered matrix.
4. image denoising method as claimed in claim 3, is characterized in that, described specific mapping algorithm is dct algorithm, and described specific inverse transformation algorithm is inverse discrete cosine transformation algorithm.
5. image denoising method as claimed in claim 3, is characterized in that, described step e 2 is: the element value that is less than or equal to default threshold values in each pixel frequency spectrum matrix is carried out to zero clearing processing, to generate corresponding pixel frequency spectrum filtered matrix.
6. an image noise reduction system, is characterized in that, this system comprises:
Region Segmentation module, for obtaining the image for the treatment of noise reduction, treat noise reduction image and carry out pixel Region Segmentation according to default step-length, matrix line number and columns, select the pixel region of cutting apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number;
Matrixing module, is used to each pixel region of selection to generate a pixel value matrix, and each the pixel value matrix generating is carried out to matrixing and filtration treatment to generate pixel value filtered matrix; And
Noise reduction computing module, for the pixel value of the corresponding same pixel point of all pixel value filtered matrix is averaged, and pixel value after noise reduction using average as this same pixel point.
7. image noise reduction system as claimed in claim 6, is characterized in that:
Whether described noise reduction computing module, also have the pixel region of cutting apart not selected for analyzing;
Described Region Segmentation module, in the time having the pixel region of cutting apart not selected, selects the pixel region of cutting apart to make the pixel region of all selections all comprise the region of part same pixel point according to default number.
8. the image noise reduction system as described in claim 6 or 7, is characterized in that, described matrixing module is used for:
Each the pixel value matrix generating is converted to generate corresponding pixel frequency spectrum matrix by specific mapping algorithm;
According to default threshold values, the element value in each pixel frequency spectrum matrix is filtered, to generate corresponding pixel frequency spectrum filtered matrix;
Each the pixel frequency spectrum filtered matrix generating is carried out to inverse transformation by specific inverse transformation algorithm, to generate corresponding pixel value filtered matrix.
9. image noise reduction system as claimed in claim 8, is characterized in that, described specific mapping algorithm is dct algorithm, and described specific inverse transformation algorithm is inverse discrete cosine transformation algorithm.
10. image noise reduction system as claimed in claim 8, is characterized in that, described matrixing module is used for: the element value that each pixel frequency spectrum matrix is less than or equal to default threshold values carries out zero clearing processing, to generate corresponding pixel frequency spectrum filtered matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210531822.3A CN103871026A (en) | 2012-12-11 | 2012-12-11 | Image denoising method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210531822.3A CN103871026A (en) | 2012-12-11 | 2012-12-11 | Image denoising method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103871026A true CN103871026A (en) | 2014-06-18 |
Family
ID=50909535
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210531822.3A Pending CN103871026A (en) | 2012-12-11 | 2012-12-11 | Image denoising method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103871026A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101582984A (en) * | 2009-04-14 | 2009-11-18 | 公安部物证鉴定中心 | Method and device for eliminating image noise |
| US20090327386A1 (en) * | 2008-06-25 | 2009-12-31 | Joel Warren Schoenblum | Combined deblocking and denoising filter |
| CN102184523A (en) * | 2011-04-13 | 2011-09-14 | 展讯通信（上海）有限公司 | Digital image noise reduction method and system based on multiresolution and overcomplete transformation |
-
2012
- 2012-12-11 CN CN201210531822.3A patent/CN103871026A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090327386A1 (en) * | 2008-06-25 | 2009-12-31 | Joel Warren Schoenblum | Combined deblocking and denoising filter |
| CN101582984A (en) * | 2009-04-14 | 2009-11-18 | 公安部物证鉴定中心 | Method and device for eliminating image noise |
| CN102184523A (en) * | 2011-04-13 | 2011-09-14 | 展讯通信（上海）有限公司 | Digital image noise reduction method and system based on multiresolution and overcomplete transformation |
Non-Patent Citations (1)
| Title |
|---|
| 芮挺 等: "采用离散余弦变换的小波图像去噪方法", 《光电工程》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8285068B2 (en) | Combined deblocking and denoising filter | |
| TWI596573B (en) | Image processing device for reducing image noise and the method thereof | |
| US9811882B2 (en) | Method and apparatus for processing super resolution image using adaptive preprocessing filtering and/or postprocessing filtering | |
| US8457194B2 (en) | Processing real-time video | |
| US20140241645A1 (en) | Image processing apparatus, image processing method, and memory device in which image processing program is stored | |
| CN110428382B (en) | Efficient video enhancement method and device for mobile terminal and storage medium | |
| US20180114296A1 (en) | Systems and methods for image enhancement | |
| US10706507B2 (en) | Hybrid denoising of images and videos based on interest metrics | |
| JP2014146988A (en) | Dynamic image encoder | |
| KR20100101463A (en) | Apparatus and method for eliminating noise | |
| Baljozović et al. | Mixed noise removal filter for multi‐channel images based on halfspace deepest location | |
| CN103745443A (en) | Method and equipment for improving image quality | |
| CN101552922B (en) | Signal processing device and method | |
| WO2009110673A1 (en) | Method and apparatus for adaptive update of directional wavelet transform based on lifting, and recording medium therefor | |
| CN112085667A (en) | Deblocking effect removing method and device based on pseudo-analog video transmission | |
| Kim et al. | Vaguelette-wavelet decomposition for frequency adaptive image restoration using directional wavelet bases | |
| CN111815535B (en) | Image processing method, apparatus, electronic device, and computer readable medium | |
| CN103871026A (en) | Image denoising method and system | |
| JP6195404B2 (en) | Processing system, pre-processing device, post-processing device, pre-processing program and post-processing program | |
| CN110910439B (en) | Image resolution estimation method and device and terminal | |
| CN116366852A (en) | Video coding and decoding method, device, equipment and medium for machine vision task | |
| US20120236932A1 (en) | Detecting Artifacts in Quantization Noise in Images Compresses Using Discrete Cosine Transforms | |
| Yang et al. | Signal denoising based on the adaptive shrinkage function and neighborhood characteristics | |
| CN111918061A (en) | Coding block dividing method and device, storage medium and electronic equipment | |
| CN116456087A (en) | Parameter determination method of adaptive loop filtering and related equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140618 |