CN104601961B - A kind of method of work of the video monitoring system based on Optical Fiber Transmission - Google Patents
A kind of method of work of the video monitoring system based on Optical Fiber Transmission Download PDFInfo
- Publication number
- CN104601961B CN104601961B CN201510056556.7A CN201510056556A CN104601961B CN 104601961 B CN104601961 B CN 104601961B CN 201510056556 A CN201510056556 A CN 201510056556A CN 104601961 B CN104601961 B CN 104601961B
- Authority
- CN
- China
- Prior art keywords
- image
- noise
- video
- pixel
- frame
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 239000013307 optical fiber Substances 0.000 title claims abstract description 31
- 238000012544 monitoring process Methods 0.000 title claims abstract description 26
- 230000005540 biological transmission Effects 0.000 title abstract description 13
- 238000004891 communication Methods 0.000 claims abstract description 16
- 230000000295 complement effect Effects 0.000 claims description 22
- 230000000873 masking effect Effects 0.000 claims description 22
- 239000006185 dispersion Substances 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 5
- 238000009826 distribution Methods 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 abstract description 23
- 239000000284 extract Substances 0.000 abstract description 4
- 238000001914 filtration Methods 0.000 description 20
- 238000010586 diagram Methods 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000007704 transition Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000000877 morphologic effect Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000010339 dilation Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Image Processing (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a kind of video monitoring system and its method of work based on Optical Fiber Transmission, this video monitoring system includes：The first optical fiber communication modules that video acquisition module is connected with the video acquisition module, first optical fiber communication modules are connected by optical fiber with the second optical fiber communication modules, second optical fiber communication modules are connected with vision signal modular converter, and the vision signal modular converter is connected by video image denoising module with display module；The present invention can effectively extract the nearly all graininess noise target of video image in video monitoring system, and movable information is not lost in well keeping original video, will not introduce new noise and original picture sharpness will not be reduced with reliability and real-time higher, and present invention image after noise reduction.
Description
Technical field
Patent of the present invention is related to the monitoring field processing technology field based on Optical Fiber Transmission vision signal, in particular for regarding
Frequency sequence noise reduction and enhanced technical field.
Background technology
Video monitoring have the advantages that intuitively, convenience, reliability, the information content enrich, therefore be widely used in finance, business
The occasions such as industry, traffic, house, community, are that the environmental monitoring and safety precaution in these fields serve the effect that can not ignore.Light
Fine Transmission system has the advantages such as long transmission distance, information capacity big, strong antijamming capability, good confidentiality, has turned into logical at present
The important means of letter.Can effectively ensure that view data will not be stolen using optical fiber technology transmission image, while having Gao Ling
The features such as sensitivity, high-resolution.People embody a concentrated reflection of the aspects such as the picture and high-quality vision of high-quality to its demand.
However, monitor video is in acquisition, transmission, storage and reproduction process, noise pollution, noise can be unavoidably subject to
Presence can not only influence the sensory experience of video image, can also reduce its information content.In addition, video image denoising is also to regard
The treatment of frequency image later stage is such as compressed, coding, the pretreatment stage of target identification.Therefore, the research of vedio noise reduction algorithm has
Very important meaning.Existing video image denoising algorithm mainly has following sorting technique：
(1) pixel domain noise reduction algorithm and Transformation Domain noise reduction algorithm two can be divided into according to the processing region of video image denoising
Class：Pixel-domain video image noise reduction algorithm is directly to process noise in space-time three-dimensional (3D) space that video image is constituted, this
Class algorithm occurs relatively early, develops comparative maturity, and its advantage is that amount of calculation is small, noise reduction is preferable；Transformation Domain video image denoising
Algorithm is first to be changed the content of video image, noise reduction process is carried out in Transformation Domain (such as wavelet field), then by anti-
Conversion obtains final de-noising video image, this kind of algorithm than pixel domain algorithm excellent noise reduction effect, but amount of calculation is larger, it is past
Toward the real-time demand that can not meet Video processing.
(2) according to wave filter support filter range can be divided into time-domain filtering (1D filtering), airspace filter (2D filtering),
Space-time filtering (3D filtering) three classes.Compared to two dimensional image, the correlation of video image information is present in one-dimensional time-domain simultaneously
(1D) and two-dimensional space domain (2D), 1D filtering only make use of correlation of the video sequence in time-domain, and 2D filtering only make use of
Correlation of the video image in spatial domain, and 3D filtering make use of video image related in time-domain to spatial domain simultaneously
Property carry out noise reduction, therefore, the noise reduction of 3D filtering is better than other two class wave filters.In space-time filtering, time-domain filtering
Device and spatial filter mainly have two kinds of combinations：The first, first carry out airspace filter carries out time-domain filtering again；Second,
Analysis result according to image switches on time-domain filtering and airspace filter, and combines image analysis result and noise estimation size
Filtering strength is set to control time domain and airspace filter.
(3) can be divided into based on two kinds of filtering methods of estimation and Motion Adaptive according to estimation mode：It is based on
The filtering method of estimation is the correlation in order to make full use of sequence of video images in time domain, is first transported before filtering
It is dynamic to estimate, use time domain filtering as far as possible in the case where motion " smear " is not caused, verified the method is in most of feelings
Noise reduction can be improved under condition；The filtering method of Motion Adaptive is filtered in time domain, but can be using certain certainly
Adaptation mechanism moves the time domain non-stationary for causing to reduce.The algorithm makes the visual effect after noise reduction be significantly improved, but by
In introduce motion estimation part and cause noise reduction algorithm to calculate quantitative change is big, and effect after video image denoising is to a certain degree
The upper degree of accuracy depending on estimation.
As can be seen here, above-mentioned existing optical fiber monitoring video sequence noise reduction technology, currently also without targetedly high-performance
Algorithm.Traditional filtering easily causes the reduction of image sharpness, and newest noise reduction technology operand is huge, without good
Real-time.Therefore, for such issues that, be badly in need of have high robust, the video image noise reducing method of high real-time.
The content of the invention
It is an object of the invention to provide a kind of video monitoring system and its method of work based on Optical Fiber Transmission, to solve filter
Except the technical problem of video noise, and then guarantee video information is completely and on the basis of acutance, the fluency for maintaining video to show.
In order to solve the above-mentioned technical problem, the invention provides a kind of video monitoring system, including：Video acquisition module,
The first optical fiber communication modules being connected with the video acquisition module, first optical fiber communication modules pass through optical fiber and the second optical fiber
Communication module is connected, and second optical fiber communication modules are connected with vision signal modular converter, the vision signal modular converter
It is connected with display module by video image denoising module.
Preferably, in order to improve noise reduction, the video image denoising module includes：Difference image acquiring unit, obtains
The difference image of two field picture must be spaced；Mask generation unit, is connected with difference image acquiring unit, with the difference image for extracting
In the noise region that includes, and generate corresponding mask image；And the pixel filling unit being connected with mask generation unit, its
It is suitable to the pixel of the correspondence position of middle two field picture be filled out the noise location of pixels of the mask image on first two field picture
Fill, to obtain noise-reduced image.
Further, the mask image is complementary masking image, and the mask generation unit is further adapted for leading in noise region
Cross binary conversion treatment and be converted to Real-valued image, and by the Real-valued image through gauss low frequency filter mask image, and
The mask image for asking difference to be negated；The pixel filling unit is suitable to use the noise location of pixels of the complementary masking image
The pixel of the correspondence position of middle two field picture is filled on first two field picture.
Another aspect, the present invention is additionally provided described on the basis of the video monitoring system based on Optical Fiber Transmission
The method of work of video monitoring system.
The method of work of the video monitoring system, including the video image denoising module video image denoising side
Method, that is, obtain the difference image of interval two field picture, extracts the noise region included in the difference image, and generate corresponding mask
Image；And by the pixel of the correspondence position of the middle two field picture of noise location of pixels use of the mask image on first two field picture
It is filled, to obtain noise-reduced image.
Further, the preparation method of the difference image includes：The absolute value of neighbor frame difference is obtained using frame difference method, with
To the difference image；I.e.
G (x, y)=| gn(x,y)-gn+2(x,y)|；
In formula, gn(x, y) and gn+2(x, y) is respectively continuous FnFrame and Fn+2The gray scale at two field picture position (x, y) place
Value, n is positive integer；Wherein, FnFrame is set to first frame, Fn+1Frame is set to intermediate frame, Fn+2Frame is set to last frame.
Further, the extracting method in the noise region includes：The difference image is filtered, Dynamic Extraction is discrete
Off-limits pixel is spent, as doubtful noise；And by the doubtful noise after morphology obtains doubtful noise region,
False noise is filtered by grader again, to obtain the noise region.
Further, the difference image is filtered, the method for the off-limits pixel of Dynamic Extraction dispersion includes：
Construction neighborhood window W, calculates pixel average m (x, y) and standard deviation d (x, y) in neighborhood window, and the neighborhood is calculated after being weighted
The discrete tolerance threshold e (x, y) of pixel in window；And pixel grey scale distribution allowed band in neighborhood window is calculated, and by neighborhood window
Center pixel gray scale g (x, y) is compared with it, off-limits to be judged as doubtful noise；The pixel grey scale being calculated
Being distributed allowed band is
m(x,y)-e(x,y)≤g(x,y)≤m(x,y)+e(x,y)；
Wherein,
E (x, y)=max (s × d (x, y), T), s >=0；
N represents number of pixels in neighborhood window in above formula, and (u, v) represents the coordinate in neighborhood window, and s is the weights of standard deviation, T
It is minimum variance.
Further, the video image noise reducing method also includes：The mask image is complementary masking image, will be described
The pixel of the correspondence position of the noise location of pixels centre two field picture of complementary masking image is filled on first two field picture, with
Obtain noise-reduced image.
Further, the generation method of the complementary masking image includes：Noise region is converted to by binary conversion treatment
Real-valued image；The image of a width and the equal size of former frame of video is constructed, noise position will be belonged in correspondence original frame of video
Pixel filling real number 1.0, is not belonging to the pixel filling 0.0 of noise position, i.e.,
In formula, F and B represents noise region and non-noise region respectively, and I is Real-valued binaryzation noise area image；Will
The Real-valued image being made up of 1.0 and 0.0 for obtaining obtains mask image M through gauss low frequency filter is smooth；Use equal size
The mask image that difference negated is sought by 1.0 images for constituting and mask image M
Further, the noise location of pixels of the complementary masking image is existed with the pixel of the correspondence position of middle two field picture
The method being filled on first two field picture includes：After complementary masking image is obtained, F will be usednTwo field picture is multiplied by covering of negating
Film imageFn+1Two field picture is multiplied by mask image M, and result of calculation image summation twice i.e. is obtained into noise-reduced image.
The beneficial effects of the invention are as follows the present invention can effectively extract the nearly all noise of video image in video monitoring system
Target, and movable information is not lost in well keeping original video, with reliability and real-time higher, Yi Jiben
Invention image after noise reduction will not introduce new noise and will not reduce original picture sharpness, can be widely used in video monitoring neck
Domain, is especially suitable for catching dynamic object.
Brief description of the drawings
The present invention is further described with reference to the accompanying drawings and examples.
The theory diagram of Fig. 1 video monitoring systems of the invention；
Fig. 2 shows the theory diagram of noise reduction system；
Fig. 3 shows the algorithm block diagram of video image noise reducing method of the invention；
Fig. 4 shows noise extracting method schematic diagram；
Fig. 5 shows construction noise filling mask image flow；
Fig. 6 (a) shows the schematic diagram of before processing；
Fig. 6 (b) shows the schematic diagram after noise reduction.
Specific embodiment
In conjunction with the accompanying drawings, the present invention is further explained in detail.These accompanying drawings are simplified schematic diagram, only with
Illustration illustrates basic structure of the invention, therefore it only shows the composition relevant with the present invention.
Embodiment 1
Conventionally, as video continuously plays generation by some two field pictures, therefore be decomposed into for video by the present invention
Some two field pictures carry out respective handling.
As shown in figure 1, the video monitoring system based on Optical Fiber Transmission of the invention, including：Video acquisition module is regarded with this
The first connected optical fiber communication modules of frequency acquisition module, first optical fiber communication modules pass through optical fiber and the second optical-fibre communications mould
Block is connected, and second optical fiber communication modules are connected with vision signal modular converter, and the vision signal modular converter is by regarding
Frequency image noise reduction module is connected with display module.
The vision signal modular converter is suitable to for the vision signal of Optical Fiber Transmission to be converted to video image.
The video image denoising module is adapted to filter out video noise.
Wherein, the video acquisition module can use camera.The display module is such as, but not limited to use liquid crystal
Display, CRT monitor.
As shown in Fig. 2 further, the video image denoising module includes：Difference image acquiring unit, obtains interval frame
The difference image of image；Mask generation unit, is connected with difference image acquiring unit, with what is included in the difference image of extraction
Noise region, and generate corresponding mask image；And the pixel filling unit being connected with mask generation unit, it is suitable to this
The pixel of the correspondence position of the noise location of pixels centre two field picture of mask image is filled on first two field picture, to obtain
Noise-reduced image.
As the present embodiment it is a kind of preferred embodiment, in order that filling nature, the border in the region of filling is realized
Seamless transition, need to carry out LPF to mask image, make edge-smoothing transition.The video image denoising module also includes：
The complementary masking image configuration unit being connected with noise area acquisition unit, the complementary masking image configuration unit is suitable to construction
Complementary masking image, will noise region Real-valued image is converted to by binary conversion treatment, and by Real-valued image warp
Gauss low frequency filter mask image, and seek the poor mask image for being negated；The pixel filling unit and complementary masking
Image configuration unit is connected, and is suitable for being carried out with the pixel of the correspondence position of middle two field picture according to the pixel of complementary masking image
Filling.
The video image denoising module, and difference image acquiring unit, mask generation unit and pixel filling unit
Specific embodiment detailed in Example 2 particular content.
The video image denoising module includes but is not limited to be realized using processor modules such as DSP, FPGA.
Embodiment 2
As shown in Figures 3 to 5, on the basis of embodiment 1, present invention also offers a kind of work of the video monitoring system
Make method, it includes：The video image noise reducing method of the video image denoising module, that is, obtain the difference diagram of interval two field picture
Picture, extracts the noise region included in the difference image, and generate corresponding mask image；And making an uproar the mask image
The pixel of the correspondence position of point location of pixels centre two field picture is filled on first two field picture, to obtain noise-reduced image.
After continuous processing of the above-mentioned original video image by the video image noise reducing method to picture frame, that is, it is reduced into
Video image after making an uproar.
Further, the preparation method of the difference image includes：
The absolute value of neighbor frame difference is obtained using frame difference method, to obtain the difference image；I.e.
G (x, y)=| gn(x,y)-gn+2(x,y)|；
In formula, gn(x, y) and gn+2(x, y) is respectively continuous FnFrame and Fn+2The gray scale at two field picture position (x, y) place
Value, n is positive integer；Wherein, FnFrame is set to first frame, Fn+2Frame is set to last frame, Fn+1Frame is set to intermediate frame.Wherein, it is described
Difference image FdifIt is suitable to be calculated by absolute value and represents two field pictures difference.In figure 3, it is specific in order to better illustrate
Implementation process, have chosen the first two field picture (first frame), the second frame (intermediate frame) and the 3rd frame (last frame) image respectively use F1, F2
Specific embodiment is illustrated with F3 expressions.N is natural number, 1,2,3,4 ..., the continuity of video image is represented with n.
Difference image FdifAs shown in Fig. 5 the first rows, there are many bright spots, these bright spots are the noise for needing removal.Together
When, the moving target profile caused comprising motion in difference image, these profiles are the important informations of video, it is necessary to retain
Come.
Further, the extracting method in the noise region includes：
First, doubtful noise is extracted, i.e., the difference image is filtered, the off-limits picture of Dynamic Extraction dispersion
Element, as doubtful noise.
Secondly, obtain noise region, i.e., doubtful noise after morphology obtains doubtful noise region, then by grader
False noise is filtered, to obtain the noise region.
Further, the specific method for extracting doubtful noise includes：
The difference image is filtered, the method for the off-limits pixel of Dynamic Extraction dispersion includes：Construction is adjacent
Domain window W, calculates pixel average m (x, y) in neighborhood window and (the pixel position of image where neighborhood window center point wherein, is represented with x and y
Put, it is also possible to be considered neighborhood window position relationship in the picture) and standard deviation d (x, y), the neighbour is calculated after being weighted
The discrete tolerance threshold e (x, y) of pixel in the window of domain；And pixel grey scale distribution allowed band in neighborhood window is calculated, and by neighborhood window
Interior center pixel gray scale g (x, y) is compared with it, off-limits to be judged as doubtful noise；I.e.
The pixel grey scale being calculated is distributed allowed band
m(x,y)-e(x,y)≤g(x,y)≤m(x,y)+e(x,y)；
Wherein,
E (x, y)=max (s × d (x, y), T), s >=0；
N represents number of pixels in neighborhood window in above formula, and (u, v) represents the coordinate in neighborhood window, and s is the weights of standard deviation, T
It is minimum variance.U is abscissa, and v is ordinate.
Wherein, neighborhood window W can select 10*10, or 12*12, or 15*15, and preferably 15*15, such as Fig. 5 selection are 15*
15 neighborhood window.
It is substantially high according to the dispersion degree segmentation dispersion of gray scale in neighborhood window to carry out image segmentation using the present invention
Pixel, as shown in Figure 2.Neighborhood window scans full images by " Z " font, according to equal in neighborhood window after each displacement in scanning process
Value and variance only can realize rapid computations in the upper of a preceding result of calculation by changing data boundary.At the step
As shown in the second row in Fig. 5, all of doubtful noise is all separated image after reason.
The specific method for obtaining noise region includes：
Image segmentation is carried out according to above-mentioned standard, is 1.5 such as but not limited to 1.5 by the image actionradius after segmentation,
Or 2.0, or the circular mask of 2.5 pixels carries out morphological dilation (first step morphology), is asked by connecting computing afterwards
Take connected component (the second morphology).By after above-mentioned two steps Morphological scale-space, it is possible to obtain doubtful noise region, such as Fig. 5
Shown in middle the third line.
By the image of above-mentioned Morphological scale-space false noise is filtered by grader；Grader specifically is used, that is, is propped up
Holding vector machine (SVM classifier) carries out category filter, is classified as feature using size, brightness, three parameters of circularity.
By after grader screening, the disturbance caused from moving target in video can be filtered effectively, be regarded from without influence
The effect of frequency.Two kinds of results are only exported due to grader：Noise and pseudo- noise, and have because feature compares simple classification utensil
High real-time, the classification time of single image is in millisecond rank.
As the present embodiment it is a kind of preferred embodiment, be can see from the third line of Fig. 5, by SVMs
The region of moving target has been removed after classification, and remaining is all noise region.Be can see from local 3D rendering, to this
Region carries out the image-region after binaryzation with the pixel of surrounding without excessive, only 0 and 1 two value.Therefore, in order that image
Pixel filling pixel is carried out as mask and tend to nature, it is to avoid the pixel of filling has aberration with adjacent pixels.Covered by complementation
Film image is realized optimizing pixel filling treatment.
Specific embodiment includes：In order that filling nature, seamless transition is realized on the border in the region of filling, need to be to covering
Film image carries out LPF, makes edge-smoothing transition.The mask image is complementary masking image, will the complementary masking
The pixel of the correspondence position of the noise location of pixels centre two field picture of image is filled on first two field picture, to obtain noise reduction
Image.
Further, the generation method of the complementary masking image includes：
Noise region is converted into Real-valued image by binary conversion treatment；I.e.
The image of a width and the equal size of former frame of video is constructed, the pixel that noise position is belonged in correspondence original frame of video is filled out
Number 1.0 is enriched, the pixel filling 0.0 of noise position is not belonging to, i.e.,
In formula, F and B represents noise region and non-noise region respectively, and I is Real-valued binaryzation noise area image；Will
The Real-valued image being made up of 1.0 and 0.0 for obtaining obtains mask image M through gauss low frequency filter is smooth；Use equal size
The mask image that difference negated is sought by 1.0 images for constituting and mask image M
By the noise location of pixels of the complementary masking image with the pixel of the correspondence position of middle two field picture in first frame figure
The method being filled on picture includes：After complementary masking image is obtained, F will be usednTwo field picture is multiplied by the mask image for negatingFn+1Two field picture is multiplied by mask image M, and result of calculation image summation twice i.e. is obtained into noise-reduced image.Mask image pair
Use the filling of pixel is converted into the multiplication and addition of image, by after pixel filling, the noise of intermediate frame video is obtained
Effective removal, as shown in Figure 5.
Fig. 6 (a) and Fig. 6 (b) show noise reduction comparison diagram, and treatment of the invention is should be apparent that from the figure
Effect.
Because optical fiber monitoring is that this noise like is had in analog signal, therefore transmitting procedure, therefore present invention is particularly suitable for
Overcome distinctive noise produced by the video of fiber form transmission.
With above-mentioned according to desirable embodiment of the invention as enlightenment, by above-mentioned description, relevant staff is complete
Various changes and amendments can be carried out without departing from the scope of the technological thought of the present invention' entirely.The technology of this invention
Property scope is not limited to the content on specification, it is necessary to its technical scope is determined according to right.
Claims (4)
1. a kind of method of work of video monitoring system, it is characterised in that
The video monitoring system, including：The first optical-fibre communications mould that video acquisition module is connected with the video acquisition module
Block, first optical fiber communication modules are connected by optical fiber with the second optical fiber communication modules, second optical fiber communication modules with
Vision signal modular converter is connected, and the vision signal modular converter is connected by video image denoising module with display module；
And
The method of work of the video monitoring system includes the video image noise reducing method of the video image denoising module, i.e.,
The difference image of interval two field picture is obtained, the noise region included in the difference image is extracted, and generate corresponding mask
Image；And
The noise location of pixels of the mask image is carried out with the pixel of the correspondence position of middle two field picture on first two field picture
Filling, to obtain noise-reduced image；
The preparation method of the difference image includes：
The absolute value of neighbor frame difference is obtained using frame difference method, to obtain the difference image；I.e.
G (x, y)=| gn(x,y)-gn+2(x,y)|；
In formula, gn(x, y) and gn+2(x, y) is respectively continuous FnFrame and Fn+2The gray value at two field picture position (x, y) place, n
It is positive integer；Wherein, FnFrame is set to first frame, Fn+1Frame is set to intermediate frame, Fn+2Frame is set to last frame；
The extracting method in the noise region includes：
The difference image is filtered, the off-limits pixel of Dynamic Extraction dispersion, as doubtful noise；And
By the doubtful noise after morphology obtains doubtful noise region, then false noise is filtered by grader, to obtain
Obtain the noise region；
The difference image is filtered, the method for the off-limits pixel of Dynamic Extraction dispersion includes：
Construction neighborhood window W, calculates pixel average m (x, y) and standard deviation d (x, y) in neighborhood window, and this is calculated after being weighted
The discrete tolerance threshold e (x, y) of pixel in neighborhood window；And
Pixel grey scale distribution allowed band in neighborhood window is calculated, and center pixel gray scale g (x, y) in neighborhood window is compared with it
Compared with off-limits to be judged as doubtful noise；I.e.
The pixel grey scale being calculated is distributed allowed band
m(x,y)-e(x,y)≤g(x,y)≤m(x,y)+e(x,y)；
Wherein,
E (x, y)=max (s × d (x, y), T), s >=0；
N represents number of pixels in neighborhood window in above formula, and (u, v) represents the coordinate in neighborhood window, and s is the weights of standard deviation, and T is for most
Small variance.
2. the method for work of video monitoring system according to claim 1, it is characterised in that
The mask image is complementary masking image, i.e.,
By the pixel of the correspondence position of the middle two field picture of noise location of pixels use of the complementary masking image on first two field picture
It is filled, to obtain noise-reduced image.
3. the method for work of video monitoring system according to claim 2, it is characterised in that
The generation method of the complementary masking image includes：
Noise region is converted into Real-valued image by binary conversion treatment；I.e.
The image of a width and the equal size of former frame of video is constructed, the pixel filling reality of noise position will be belonged in correspondence original frame of video
Number 1.0, is not belonging to the pixel filling 0.0 of noise position, i.e.,
In formula, F and B represents noise region and non-noise region respectively, and I is Real-valued binaryzation noise area image；
The Real-valued image being made up of 1.0 and 0.0 that will be obtained obtains mask image M through gauss low frequency filter is smooth；With same
The poor mask image for being negated is sought by 1.0 images for constituting and mask image M etc. size
4. the method for work of video monitoring system according to claim 3, it is characterised in that by the complementary masking image
The method that is filled on first two field picture of pixel of correspondence position of noise location of pixels centre two field picture include：
After complementary masking image is obtained, F will be usednTwo field picture is multiplied by the mask image for negatingFn+1Two field picture is multiplied by be covered
Film image M, and result of calculation image summation twice is obtained into noise-reduced image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510056556.7A CN104601961B (en) | 2015-02-03 | 2015-02-03 | A kind of method of work of the video monitoring system based on Optical Fiber Transmission |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510056556.7A CN104601961B (en) | 2015-02-03 | 2015-02-03 | A kind of method of work of the video monitoring system based on Optical Fiber Transmission |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104601961A CN104601961A (en) | 2015-05-06 |
| CN104601961B true CN104601961B (en) | 2017-06-23 |
Family
ID=53127421
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510056556.7A Expired - Fee Related CN104601961B (en) | 2015-02-03 | 2015-02-03 | A kind of method of work of the video monitoring system based on Optical Fiber Transmission |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104601961B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014240A (en) * | 2010-12-01 | 2011-04-13 | 深圳市蓝韵实业有限公司 | Real-time medical video image denoising method |
| CN202050503U (en) * | 2011-03-11 | 2011-11-23 | 孙飞 | Multichannel uncompressed digital video optical fiber transmission device |
| CN102629970A (en) * | 2012-03-31 | 2012-08-08 | 广东威创视讯科技股份有限公司 | Denoising method and system for video images |
| CN102636736A (en) * | 2012-05-10 | 2012-08-15 | 山东电力集团公司济南供电公司 | Online monitoring system for partial discharge of middle-high voltage power cable |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5450668B2 (en) * | 2010-02-15 | 2014-03-26 | シャープ株式会社 | Signal processing apparatus, control program, and integrated circuit |
-
2015
- 2015-02-03 CN CN201510056556.7A patent/CN104601961B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014240A (en) * | 2010-12-01 | 2011-04-13 | 深圳市蓝韵实业有限公司 | Real-time medical video image denoising method |
| CN202050503U (en) * | 2011-03-11 | 2011-11-23 | 孙飞 | Multichannel uncompressed digital video optical fiber transmission device |
| CN102629970A (en) * | 2012-03-31 | 2012-08-08 | 广东威创视讯科技股份有限公司 | Denoising method and system for video images |
| CN102636736A (en) * | 2012-05-10 | 2012-08-15 | 山东电力集团公司济南供电公司 | Online monitoring system for partial discharge of middle-high voltage power cable |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104601961A (en) | 2015-05-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110246108B (en) | Image processing method, device and computer readable storage medium | |
| Zhu et al. | Haze removal method for natural restoration of images with sky | |
| CN104616274B (en) | A kind of multi-focus image fusing method based on salient region extraction | |
| CN106846339A (en) | A kind of image detecting method and device | |
| CN105654091A (en) | Detection method and apparatus for sea-surface target | |
| CN105427257A (en) | Image enhancement method and apparatus | |
| CN105303561A (en) | Image preprocessing grayscale space division method | |
| CN102194443A (en) | Display method and system for window of video picture in picture and video processing equipment | |
| CN103778434A (en) | Face recognition method based on multi-resolution multi-threshold local binary pattern | |
| Chen et al. | Saliency-directed image interpolation using particle swarm optimization | |
| CN105225238A (en) | A kind of gray space division methods of the Image semantic classification based on mean filter | |
| CN109377464A (en) | A kind of Double plateaus histogram equalization method and its application system of infrared image | |
| DE112016005482T5 (en) | Object detection with adaptive channel features | |
| CN110111347A (en) | Logos extracting method, device and storage medium | |
| CN111223110A (en) | Microscopic image enhancement method and device and computer equipment | |
| CN106529535A (en) | License plate positioning method in complex lighting environment based on wavelet transform | |
| CN101853500A (en) | Colored multi-focus image fusing method | |
| CN109255799A (en) | A kind of method for tracking target and system based on spatially adaptive correlation filter | |
| CN115131229A (en) | Image noise reduction and filtering data processing method and device and computer equipment | |
| Maity et al. | Background modeling and foreground extraction in video data using spatio-temporal region persistence features | |
| CN117456376A (en) | Remote sensing satellite image target detection method based on deep learning | |
| CN104601861B (en) | A kind of noise-reduction method and system for optical fiber monitoring video sequence | |
| CN104601961B (en) | A kind of method of work of the video monitoring system based on Optical Fiber Transmission | |
| CN110633705A (en) | Low-illumination imaging license plate recognition method and device | |
| CN105069764A (en) | Image denoising method and system based on edge tracking |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170623 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |