CN103269412B - A kind of noise-reduction method of video image and device - Google Patents
A kind of noise-reduction method of video image and device Download PDFInfo
- Publication number
- CN103269412B CN103269412B CN201310137164.4A CN201310137164A CN103269412B CN 103269412 B CN103269412 B CN 103269412B CN 201310137164 A CN201310137164 A CN 201310137164A CN 103269412 B CN103269412 B CN 103269412B
- Authority
- CN
- China
- Prior art keywords
- frame
- pixels
- block
- pixel
- currently processed
- 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
Landscapes
- Picture Signal Circuits (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the invention discloses a kind of noise-reduction method of video image and device, are related to image procossing, the noise of image are largely eliminated, and effectively prevent the motion blur phenomenon of moving region.Concrete scheme is：The all pixels point in the range of the N*N in frame under process is carried out frame in filtering centered on currently processed pixel and obtain the second block of pixels；The all pixels point in the range of the N*N in reference frame is carried out frame in filtering centered on the first pixel and obtain the 4th block of pixels；The weighted value of the time-domain filtering of currently processed pixel is obtained according to the second block of pixels and the 4th block of pixels；The time-domain filtering component of currently processed pixel is obtained according to weighted value；Frame in filtering is carried out to currently processed pixel obtains airspace filter component；The gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value.During the present invention is used for the denoising of video image.
Description
Technical field
The present invention relates to image processing field, more particularly to a kind of noise-reduction method of video image and device.
Background technology
With the continuous development of science and technology, digital camera being used widely in people's daily life, therefore such as
By Image Denoising, undistorted restore original image becomes the emphasis problem of this area research as far as possible for what.
During video acquisition, have various noises and be mingled in the image sequence of collection, wherein great majority
Noise is all random, it is to be appreciated that the correlation in the image sequence gathered using digital camera between image is very
Height, therefore can reduce the noise of image sequence by image is carried out temporal weighted average, using in this method
Can be very good the random noise of interframe is removed, but the superposition for simply simply image being carried out in time domain is averaging meeting by image
The middle region that there is motion thickens, and filters that depth is deeper, and blooming is more serious.Therefore occur in that time-space domain filters
Algorithm, the algorithm first by motion detection technique come the moving region in detection image, then carry out again time domain and spatial domain it
Between adaptive-filtering.Specifically, can by between the block of relevant position in the block in frame under process and reference frame away from
From the weight for obtaining time-domain filtering and airspace filter, then according to the weight of time-domain filtering and the weight of airspace filter to current place
Reason frame is filtered, to reduce the noise of frame under process.During being filtered using time-space domain filtering algorithm, in order to
Ensure time-domain filtering and the sky obtained according to the distance between block of relevant position in the block in frame under process and reference frame
Domain filtering weight accuracy, calculate apart from when, need consider noise impact, prior art provide two kinds of solution party
Method, a kind of method be according to the gain of equipment obtain image sequence noise, calculate apart from when remove noise impact, another
Kind of method be using the variance of smooth region in frame under process as noise, then according to the size of the noise to calculating
Distance is compensated.
During stating filtering in realization, inventor has found prior art, and at least there are the following problems：Increased according to equipment
In benefit or frame under process, the noise of the frame under process that the variance of smooth region is obtained is all inaccurate, therefore calculates
Also can be present certain error in the distance between the block in frame under process and the block of relevant position in reference frame, so will
The weight of the time-domain filtering and airspace filter obtained according to distance is caused to there is also error, so that being filtered to image
During there is the situation generation that noise eliminates the appearances smear of unclean or moving region.
Content of the invention
Embodiments of the invention provide a kind of noise-reduction method of video image and device, largely eliminate image
Noise, and effectively prevent the motion blur phenomenon of moving region.
For reaching above-mentioned purpose, embodiments of the invention are adopted the following technical scheme that：
A first aspect of the present invention, provides a kind of noise-reduction method of video image, including：
When frame under process is not the first frame of image sequence, at the current place centered on currently processed pixel
The pixel in the range of N*N is chosen in reason frame as the first block of pixels, all pixels point in first block of pixels is carried out
Frame in filtering obtains the second block of pixels；
The pixel in the range of N*N is chosen using centered on the first pixel in reference frame as the 3rd block of pixels, to institute
State all pixels point in the 3rd block of pixels carry out frame in filtering obtain the 4th block of pixels；Wherein, the reference frame is worked as described
The former frame of pre-treatment frame, location is existed first pixel with the currently processed pixel in the reference frame
In the frame under process, location is identical；
The time-domain filtering of the currently processed pixel is obtained according to second block of pixels and the 4th block of pixels
Weighted value；
The time-domain filtering component of the currently processed pixel is obtained according to the weighted value；
Frame in filtering is carried out to the currently processed pixel obtains airspace filter component；
The currently processed pixel is updated according to the time-domain filtering component, the airspace filter component and the weighted value
The gray value of point.
In conjunction with a first aspect, in a kind of possible implementation, described in institute centered on currently processed pixel
The pixel in frame under process in the range of selection N*N is stated as the first block of pixels, to all pictures in first block of pixels
Vegetarian refreshments carry out frame in filtering obtain the second block of pixels before, also include：
Judge that whether the frame under process is the first frame of described image sequence；
When the frame under process is for the first frame of described image sequence, the first two field picture is stored in frame buffer,
Or, carry out being stored in after frame in filtering process in the frame buffer the first frame figure, so as to by first two field picture
Reference frame as the second two field picture in described image sequence.
In conjunction with first aspect and above-mentioned possible implementation, in alternatively possible implementation, described according to institute
The weighted value of the time-domain filtering of the second block of pixels and the 4th block of pixels acquisition currently processed pixel is stated, including：
Difference is taken by corresponding with all pixels point in the 4th block of pixels for second block of pixels, and all differences are entered
Range information is obtained after row weighted average；Wherein, the difference including but not limited to following any one：Absolute difference, all square
Difference；
The weighted value of the currently processed pixel is obtained according to the range information.
In conjunction with first aspect and above-mentioned possible implementation, in alternatively possible implementation, described according to institute
The time-domain filtering component that weighted value obtains the currently processed pixel is stated, including：
Judge the weighted value whether less than predetermined threshold value；
When the weighted value is less than the threshold value, the first of the currently processed pixel is obtained according to the weighted value
Time-domain filtering component, and using the first time-domain filtering component as the time-domain filtering component；
When the weighted value is more than or equal to the threshold value, first time-domain filtering is obtained according to the weighted value and is divided
Amount, and in taking to the gray value of the first time-domain filtering component, the airspace filter component and the currently processed pixel
Value obtains the time-domain filtering component.
In conjunction with first aspect and above-mentioned possible implementation, in alternatively possible implementation, when having updated
After stating the gray value of all pixels point in frame under process, also include：
The frame under process according to gray value has been updated updates the frame buffer.
In conjunction with first aspect and above-mentioned possible implementation, in alternatively possible implementation, the frame in filter
Ripple mode including but not limited to following any one：Mean filter, bilateral filtering, NL filtering.
A second aspect of the present invention, provides a kind of denoising device of video image, including：
Processing unit, for when frame under process be not image sequence the first frame when, in currently processed pixel being
The heart chooses pixel in the range of N*N as the first block of pixels in the frame under process, in first block of pixels
All pixels point carries out frame in filtering and obtains the second block of pixels；
The processing unit, is additionally operable to centered on the first pixel choose the pixel in the range of N*N in reference frame
As the 3rd block of pixels, frame in filtering is carried out to all pixels point in the 3rd block of pixels and obtains the 4th block of pixels；Wherein,
The reference frame is the former frame of the frame under process, first pixel location and institute in the reference frame
State currently processed pixel location in the frame under process identical；
First acquisition unit, for second block of pixels that obtained according to the processing unit and the 4th block of pixels
Obtain the weighted value of the time-domain filtering of the currently processed pixel；
Second acquisition unit, the weighted value for being got according to the first acquisition unit obtain the current place
The time-domain filtering component of reason pixel；
Filter unit, obtains airspace filter component for carrying out frame in filtering to the currently processed pixel；
First updating block, for according to the time-domain filtering component, the airspace filter component and the weighted value more
The gray value of newly described currently processed pixel.
In conjunction with second aspect, in a kind of possible implementation, also include：
Judging unit, for being selected in the frame under process centered on currently processed pixel in the processing unit
The pixel in the range of N*N is taken as the first block of pixels, frame in filtering is carried out to all pixels point in first block of pixels
Before obtaining the second block of pixels, judge that whether the frame under process is the first frame of described image sequence；
Memory cell, for when the frame under process is for the first frame of described image sequence, the first two field picture being deposited
Storage in frame buffer, or, the first frame figure is carried out being stored in after frame in filtering process in the frame buffer, so as to by institute
The first two field picture is stated as the reference frame of the second two field picture in described image sequence.
In conjunction with second aspect and above-mentioned possible implementation, in alternatively possible implementation, described first obtains
Unit is taken, including：
First processing module, for taking difference by all pixels point is corresponding in second block of pixels and the 4th block of pixels
Value, and range information is obtained after all differences are weighted averagely；Wherein, the difference is including but not limited to following any one
Kind：Absolute difference, mean square deviation；
Acquisition module, the range information for being obtained according to the first processing module obtain the currently processed picture
The weighted value of vegetarian refreshments.
In conjunction with second aspect and above-mentioned possible implementation, in alternatively possible implementation, described second obtains
Unit is taken, including：
Whether judge module, for judging the weighted value less than predetermined threshold value；
Second processing module, for when the judge module obtains the weighted value less than the threshold value, according to described
Weighted value obtains the first time-domain filtering component of the currently processed pixel, and using the first time-domain filtering component as institute
State time-domain filtering component；
3rd processing module, for when the judge module obtain the weighted value more than or equal to the threshold value when, root
The first time-domain filtering component is obtained according to the weighted value, and the first time-domain filtering component, the airspace filter are divided
The gray value of amount and the currently processed pixel takes intermediate value and obtains the time-domain filtering component.
In conjunction with second aspect and above-mentioned possible implementation, in alternatively possible implementation, also include：
Second updating block, for after the gray value for having updated the point of all pixels in the frame under process, root
The frame buffer is updated according to the frame under process for having updated gray value.
In conjunction with second aspect and above-mentioned possible implementation, in alternatively possible implementation, the frame in filter
Ripple mode including but not limited to following any one：Mean filter, bilateral filtering, NL filtering.
The noise-reduction method of video image provided in an embodiment of the present invention and device, in frame under process, will be currently to locate
The all pixels point in the range of N*N centered on reason pixel carries out frame in filtering and obtains the second block of pixels, and in reference frame,
All pixels point in the range of N*N centered on the first pixel is carried out frame in filtering and obtains the 4th block of pixels, Ran Houzai
The weighted value of the time-domain filtering of currently processed pixel is obtained according to the second block of pixels and the 4th block of pixels, and is obtained according to weighted value
The time-domain filtering component of currently processed pixel is taken, and frame in filtering is carried out to currently processed pixel and obtains airspace filter component,
Then the gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value, is so being obtained
Before the weighted value of the time-domain filtering for taking currently processed pixel, to by currently processed pixel and centered on the first pixel
Block of pixels carry out frame in filtering, further according to filtered block of pixels obtain weighted value, it is ensured that the accuracy of weighted value, so as to
The noise for largely removing image is ensure that, and effectively prevent the motion blur phenomenon of moving region.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
Accompanying drawing to be used needed for technology description is had to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, may be used also
To obtain other accompanying drawings according to these accompanying drawings.
A kind of noise-reduction method flow chart of video image that Fig. 1 is provided for one embodiment of the invention；
A kind of noise-reduction method flow chart of video image that Fig. 2 is provided for another embodiment of the present invention；
A kind of denoising device composition schematic diagram of video image that Fig. 3 is provided for another embodiment of the present invention；
The denoising device composition schematic diagram of another kind of video image that Fig. 4 is provided for another embodiment of the present invention；
A kind of denoising device composition schematic diagram of video image that Fig. 5 is provided for another embodiment of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the present invention, rather than whole embodiments.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
The embodiment of the present invention provides a kind of noise-reduction method of video image, as shown in figure 1, the method can include：
101st, when frame under process is not the first frame of image sequence, located currently centered on currently processed pixel
The pixel in the range of N*N is chosen in reason frame as the first block of pixels, frame in is carried out to all pixels point in the first block of pixels
Filtering obtains the second block of pixels.
Specifically, when frame under process is not required to the first frame in the image sequence of the video image of process, need
Pixel in frame under process is carried out noise reduction process one by one.For currently processed pixel, noise reduction process is being carried out
Before, in order to accurately calculate the weighted value of the time-domain filtering of currently processed pixel, can be first with currently processed pixel
Centered on, the pixel in the range of N*N is chosen in frame under process as the first block of pixels, and by first block of pixels
All pixels point carries out frame in filtering and obtains the second block of pixels.
102nd, pixel N*N in the range of is chosen using centered on the first pixel in reference frame as the 3rd block of pixels,
Frame in filtering is carried out to all pixels point in the 3rd block of pixels obtains the 4th block of pixels.
Wherein, accordingly, it is also possible to which frame in filtering is carried out to corresponding region in reference frame, specifically, can be with
In reference frame, present position is with currently processed pixel centered on identical first pixel of location in frame under process,
The pixel in the range of N*N is chosen in reference frame as the 3rd block of pixels, and all of pixel in the 3rd block of pixels is clicked through
The filtering of row frame in obtains the 4th block of pixels.Wherein, reference frame is the former frame of frame under process, and the first pixel is in reference frame
Location in frame under process is identical with currently processed pixel for location.
Optionally, it is also possible to frame in filtering is not carried out to the 3rd block of pixels in reference frame centered on the first pixel.
103rd, the weighted value of the time-domain filtering of currently processed pixel is obtained according to the second block of pixels and the 4th block of pixels.
Wherein, after the second block of pixels and the 4th block of pixels is obtained through frame in filtering, just can be according to the second pixel
Block and the 4th block of pixels obtain the weighted value of the time-domain filtering of currently processed pixel, specific calculate can be by the second pixel
In block and the 4th block of pixels, all of pixel is corresponded to and takes difference, and further difference is weighted averagely, and according to adding
The result that weight average is obtained obtains the weighted value of the time-domain filtering of currently processed pixel.
104th, the time-domain filtering component of currently processed pixel is obtained according to weighted value.
Wherein, when the weight of the time-domain filtering for getting currently processed pixel according to the second block of pixels and the 4th block of pixels
After value, the time-domain filtering component of currently processed pixel just can be obtained according to the weighted value.
105th, frame in filtering is carried out to currently processed pixel obtains airspace filter component.
106th, the gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value.
Wherein, after the time-domain filtering component, airspace filter component and the weight that get currently processed pixel, just
Currently processed pixel can be filtered according to domain filtered components, airspace filter component and weighted value, and according to filtering knot
Fruit updates the gray value of currently processed pixel.
It should be noted that the filtering mode of the frame in filtering in the embodiment of the present invention can be mean filter, bilateral filter
Any one in ripple, NL (Non Local) filtering, and it is not limited to the above-mentioned frame in filtering mode that enumerates.
The noise-reduction method of video image provided in an embodiment of the present invention, in frame under process, will be with currently processed pixel
The all pixels point in the range of N*N centered on point carries out frame in filtering and obtains the second block of pixels, and in reference frame, will be with the
All pixels point in the range of N*N centered on one pixel carries out frame in filtering and obtains the 4th block of pixels, then further according to the
Two block of pixels and the 4th block of pixels obtain the weighted value of the time-domain filtering of currently processed pixel, and are obtained according to weighted value current
The time-domain filtering component of pixel is processed, and frame in filtering is carried out to currently processed pixel and obtains airspace filter component, Ran Hougen
The gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value, is so being obtained currently
Before the weighted value of the time-domain filtering for processing pixel, to the pixel by currently processed pixel and centered on the first pixel
Block carries out frame in filtering, obtains weighted value further according to filtered block of pixels, it is ensured that the accuracy of weighted value, so as to ensure that
The noise of image is largely removed, and effectively prevent the motion blur phenomenon of moving region.
The embodiment of the present invention provides a kind of noise-reduction method of video image, as shown in Fig. 2 the method can include：
In order to improve the signal to noise ratio of image/video, need to carry out denoising to video image, and in video image
Each two field picture in each pixel when being processed, in order to be able to preferably remove noise, need with frame under process
Former frame as reference frame, therefore before processing to each pixel in frame under process, due to image sequence
The first two field picture in row can first judge that whether frame under process is the first frame of image sequence, when current without reference to frame
When processing the first frame that frame is image sequence, the first two field picture can be stored in frame buffer, to make the first two field picture
Reference frame for the second two field picture in image sequence.When frame under process is not the first frame of image sequence, can be currently to locate
The former frame of reason frame carries out denoising for reference frame to all pixels point in frame under process, specifically to all pixels point
In the concrete denoising process of currently processed picture point can be carried out following steps 201 to step 209.
Optionally, when judge to obtain frame under process as image sequence the first frame when, in order to further reduce noise
Impact, a certain degree of frame in filtering can be carried out to the first two field picture, then again will be through filtered first frame of frame in
Image is stored in frame buffer, to use during the second two field picture of subsequent treatment.
201st, when frame under process is not the first frame of image sequence, located currently centered on currently processed pixel
The pixel in the range of N*N is chosen in reason frame as the first block of pixels, frame in is carried out to all pixels point in the first block of pixels
Filtering obtains the second block of pixels.
Wherein, when judging to obtain the first frame that frame under process is not image sequence, now just can be to currently processed
The currently processed pixel in all pixels point in frame is processed.Specifically, centered on currently processed pixel, working as
The pixel in the range of N*N is chosen in pre-treatment frame as the first block of pixels, wherein, N is the integer more than 1, then again to choosing
All pixels point in the first block of pixels for taking carries out frame in filtering and obtains the second block of pixels.For example, work as N=5, frame in filtering
When filtering mode is frame in mean filter, can be in frame under process, centered on currently processed pixel, choose 5*5
In the range of 25 pixels constitute the first block of pixels, and using frame in mean filter filtering mode in the first block of pixels
25 pixels carry out frame in filtering, then obtain the second block of pixels.
202nd, pixel N*N in the range of is chosen using centered on the first pixel in reference frame as the 3rd block of pixels,
Frame in filtering is carried out to all pixels point in the 3rd block of pixels obtains the 4th block of pixels.
Accordingly, in reference frame, with location in reference frame with currently processed pixel in frame under process
Centered on middle the first pixel of location identical, the pixel in the range of N*N is chosen as the 3rd block of pixels, and to the
All pixels point in three block of pixels carries out frame in filtering and obtains the 4th block of pixels.Wherein, before reference frame is frame under process
One frame, the first pixel in reference frame location with currently processed pixel location phase in frame under process
With.
Optionally, as the former frame of frame under process is likely to be the picture frame through denoising, therefore with first
After the 3rd block of pixels being chosen in reference frame centered on pixel, it is also possible to do not carry out frame in filtering to the 3rd block of pixels,
Directly the weighted value of the time-domain filtering of currently processed pixel is calculated using the second block of pixels and the 3rd block of pixels.
203rd, take difference by corresponding with all pixels point in the 4th block of pixels for the second block of pixels, and all differences are carried out plus
Range information is obtained after weight average.
Wherein, after the second block of pixels and the 4th block of pixels is got, can be by the second block of pixels for getting and
The gray value of the pixel of all of pixel correspondence position in four block of pixels takes difference, and all differences for getting are entered
The range information of currently processed pixel and the first pixel is obtained after row weighted average.Wherein, the difference includes but does not limit
In following any one：Absolute difference, mean square deviation.
204th, the weighted value of the time-domain filtering of currently processed pixel is obtained according to range information.
Wherein, after the range information for getting currently processed pixel and the first pixel, just can be according to distance
The weighted value of the time-domain filtering of the currently processed pixel of acquisition of information, specific weighted value are proportional with range information.Example
Such as, the weighted value of time-domain filtering can be obtained by the following method：
If (dif ＞ thr2) AlphaIIR=0；
Else if (dif ＜ thr1) AlphaIIR=250；
Else AlphaIIR=250* (thr2-dif)/(thr2-thr1)；
Wherein, AlphaIIR is the weighted value of the time-domain filtering of currently processed pixel, and dif is range information, thr1,
Thr2 is default threshold value, and the threshold value can be arranged according to the gain of equipment, and gain is bigger, and threshold value is bigger.
205th, judge weighted value whether less than predetermined threshold value.
Wherein, in order to further ensure the noise of removal image as big as possible, and motor area is further avoided
The motion blur phenomenon in domain, after the weighted value of time-domain filtering of currently processed pixel is got, it can be determined that the weighted value is
No less than default threshold value.Wherein the threshold value can be arranged according to the gain of equipment.
206th, when weighted value is less than threshold value, divided according to the first time-domain filtering that weighted value obtains currently processed pixel
Amount, and using the first time-domain filtering component as time-domain filtering component.
Specifically, when judging the weighted value to be obtained less than threshold value, show currently processed pixel for stagnant zone, so
Just the first time-domain filtering component of currently processed pixel directly can be obtained according to weighted value, and first time-domain filtering is divided
Amount is used as time-domain filtering component.Wherein, the first time-domain filtering component can be obtained according to below equation：
P_time1=(1-AlphaIIR) * CurrPixel+AlphaIIR*PastPixel (1)
Wherein, P_timel is the first time-domain filtering component, and AlphaIIR for the weighted value of time-domain filtering, CurrPixel is
The gray value of currently processed pixel, PastPixel are the gray value of the first pixel, and the first pixel is residing in reference frame
Location in frame under process is identical with currently processed pixel for position.
207th, when weighted value is more than or equal to threshold value, the first time-domain filtering component is obtained according to weighted value, and to first
The gray value of time-domain filtering component, airspace filter component and currently processed pixel takes intermediate value and obtains time-domain filtering component.
Specifically, when judging weighted value to be obtained more than or equal to threshold value, show currently processed pixel for moving region,
In order to the generation of motion motion blur phenomenon can be effectively avoided, then first the filter of the first time domain can be calculated according to above-mentioned formula (1)
Wave component, is then taken again from the gray value three of the first time-domain filtering component, airspace filter component and currently processed pixel
Intermediate value, used as time-domain filtering component.
208th, airspace filter component is obtained to current as process vegetarian refreshments carries out frame in filtering.
209th, the gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value.
Wherein, when the power of the time-domain filtering component, airspace filter component and time-domain filtering for getting currently processed pixel
After weight values, just currently processed pixel after filtering can be obtained according to time-domain filtering component, airspace filter component and weighted value
Gray value, then update the gray value of currently processed pixel according to result.Specifically, after currently processed pixel is filtered
The calculating of gray value can be：
P_out=(1-AlphaIIR) * P_spa+AlphaIIR*P_time (2)
Wherein, P_out be currently processed pixel filtered after the gray value that obtains, AlphaIIR is the power of time-domain filtering
Weight values, P_spa are airspace filter component, and P_time is time-domain filtering component.
After the gray value for having updated the point of all pixels in frame under process, frame buffer is updated, ash will have been updated
The frame under process of angle value is stored in frame buffer, during to process the next frame in frame under process, frame under process is made
For reference frame.
It should be noted that the filtering mode of the frame in filtering in the embodiment of the present invention can be mean filter, bilateral filter
Any one in ripple, NL filtering, and it is not limited to the above-mentioned frame in filtering mode that enumerates.
The noise-reduction method of video image provided in an embodiment of the present invention, in frame under process, will be with currently processed pixel
The all pixels point in the range of N*N centered on point carries out frame in filtering and obtains the second block of pixels, and in reference frame, will be with the
All pixels point in the range of N*N centered on one pixel carries out frame in filtering and obtains the 4th block of pixels, then further according to the
Two block of pixels and the 4th block of pixels obtain the weighted value of the time-domain filtering of currently processed pixel, and are obtained according to weighted value current
The time-domain filtering component of pixel is processed, and frame in filtering is carried out to currently processed pixel and obtains airspace filter component, Ran Hougen
The gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value, is so being obtained currently
Before the weighted value of the time-domain filtering for processing pixel, to the pixel by currently processed pixel and centered on the first pixel
Block carries out frame in filtering, obtains weighted value further according to filtered block of pixels, it is ensured that the accuracy of weighted value, so as to ensure that
The noise of image is largely removed, and effectively prevent the motion blur phenomenon of moving region.
Also, after the weighted value for getting time-domain filtering, according to the size of weighted value, in front and back content frame gap relatively
Little region, directly obtains time-domain filtering component by mathematic interpolation, further ensures making an uproar for removal image as big as possible
Sound, for the region that in front and back's content frame gap is larger, obtains time-domain filtering component by medium filtering, further avoid motion
The motion blur phenomenon in region.
Embodiments of the invention provide a kind of denoising device of video image, as shown in figure 3, the device includes：Process single
Unit 31, first acquisition unit 32, second acquisition unit 33, filter unit 34, the first updating block 35.
Processing unit 31, for when frame under process be not image sequence the first frame when, with currently processed pixel be
Pixel in the range of N*N is chosen as the first block of pixels in the frame under process in center, in first block of pixels
All pixels point carry out frame in filtering obtain the second block of pixels.
The processing unit 31, is additionally operable to centered on the first pixel choose the pixel in the range of N*N in reference frame
O'clock as the 3rd block of pixels, frame in filtering is carried out to all pixels point in the 3rd block of pixels and obtains the 4th block of pixels；Its
In, the reference frame is the former frame of the frame under process, first pixel location in the reference frame
In the frame under process, location is identical with the currently processed pixel.
First acquisition unit 32, for second block of pixels that obtained according to the processing unit 31 and the 4th picture
Plain block obtains the weighted value of the time-domain filtering of the currently processed pixel.
Second acquisition unit 33, for the weighted value that got according to the first acquisition unit 32 obtain described ought
The time-domain filtering component of pre-treatment pixel.
Filter unit 34, obtains airspace filter component for carrying out frame in filtering to the currently processed pixel.
First updating block 35, for according to the time-domain filtering component, the airspace filter component and the weighted value
Update the gray value of the currently processed pixel.
Further, as shown in figure 4, the device can also include：Judging unit 36, memory cell 37.
Judging unit 36, in the processing unit 31 in the frame under process centered on currently processed pixel
Pixel in the range of middle selection N*N carries out frame in as the first block of pixels to all pixels point in first block of pixels
Before filtering obtains the second block of pixels, judge that whether the frame under process is the first frame of described image sequence.
Memory cell 37, for when the frame under process is for the first frame of described image sequence, by the first two field picture
It is stored in frame buffer, or, the first frame figure is carried out being stored in after frame in filtering process in the frame buffer, so that will
First two field picture is used as the reference frame of the second two field picture in described image sequence.
Further, the first acquisition unit 32 can include：First processing module 321, acquisition module 322.
First processing module 321, for corresponding with all pixels point in the 4th block of pixels by second block of pixels
Difference is taken, and range information is obtained after all differences are weighted averagely；Wherein, the difference is including but not limited to following appoints
Meaning is a kind of：Absolute difference, mean square deviation.
Acquisition module 322, the range information for being obtained according to the first processing module 321 obtain described current
Process the weighted value of pixel.
Further, the second acquisition unit 33 can include：Judge module 331, Second processing module the 332, the 3rd
Processing module 333.
Whether judge module 331, for judging the weighted value less than predetermined threshold value.
Second processing module 332, for when the judge module 331 obtain the weighted value less than the threshold value when, root
Obtain the first time-domain filtering component of the currently processed pixel according to the weighted value, and by the first time-domain filtering component
As the time-domain filtering component.
3rd processing module 333, for obtaining the weighted value more than or equal to the threshold value when the judge module 331
When, the first time-domain filtering component is obtained according to the weighted value, and the first time-domain filtering component, the spatial domain are filtered
The gray value of wave component and the currently processed pixel takes intermediate value and obtains the time-domain filtering component.
Further, the device can also include：Second updating block 38.
Second updating block 38, for after the gray value for having updated the point of all pixels in the frame under process,
The frame under process according to gray value has been updated updates the frame buffer.
Further, the frame in filtering mode including but not limited to following any one：Mean filter, bilateral filtering,
NL is filtered.
The denoising device of video image provided in an embodiment of the present invention, in frame under process, will be with currently processed pixel
The all pixels point in the range of N*N centered on point carries out frame in filtering and obtains the second block of pixels, and in reference frame, will be with the
All pixels point in the range of N*N centered on one pixel carries out frame in filtering and obtains the 4th block of pixels, then further according to the
Two block of pixels and the 4th block of pixels obtain the weighted value of the time-domain filtering of currently processed pixel, and are obtained according to weighted value current
The time-domain filtering component of pixel is processed, and frame in filtering is carried out to currently processed pixel and obtains airspace filter component, Ran Hougen
The gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value, is so being obtained currently
Before the weighted value of the time-domain filtering for processing pixel, to the pixel by currently processed pixel and centered on the first pixel
Block carries out frame in filtering, obtains weighted value further according to filtered block of pixels, it is ensured that the accuracy of weighted value, so as to ensure that
The noise of image is largely removed, and effectively prevent the motion blur phenomenon of moving region.
Also, after the weighted value for getting time-domain filtering, according to the size of weighted value, in front and back content frame gap relatively
Little region, directly obtains time-domain filtering component by mathematic interpolation, further ensures making an uproar for removal image as big as possible
Sound, for the region that in front and back's content frame gap is larger, obtains time-domain filtering component by medium filtering, further avoid motion
The motion blur phenomenon in region.
Embodiments of the invention provide a kind of denoising device of video image, as shown in figure 5, including：At least one is processed
Device 41, memory 42, communication interface 43 and bus 44, at least one processor 41, memory 42 and communication interface 43 pass through
Bus 44 connects and completes mutual communication, wherein：
The bus 44 can be industry standard architecture (Industry Standard Architecture, ISA)
Bus, external equipment interconnection (Peripheral Component Interconnect, PCI) bus or extension industrial standard body
Architecture (Extended Industry Standard Architecture, EISA) bus etc..The bus 44 can be divided into ground
Location bus, data/address bus, controlling bus etc..For ease of representing, only represented with a thick line in Fig. 5, it is not intended that only one
Root bus or a type of bus.
The memory 42 is used for storing executable program code, and the program code includes computer-managed instruction.Storage
Device 42 may include high-speed RAM memory, it is also possible to also include nonvolatile memory (non-volatile memory), example
As at least one magnetic disc store.
The processor 41 is probably a central processing unit (Central Processing Unit, CPU), or
Specific integrated circuit (Application Specific Integrated Circuit, ASIC), or be arranged to implement
One or more integrated circuits of the embodiment of the present invention.
The communication interface 43, is mainly used in realizing the communication between the equipment of the present embodiment.
The processor 41, for when frame under process be not image sequence the first frame when, with currently processed pixel
Centered in the frame under process choose N*N in the range of pixel as the first block of pixels, to first block of pixels
Interior all pixels point carries out frame in filtering and obtains the second block of pixels；N*N is chosen centered on the first pixel in reference frame
In the range of pixel as the 3rd block of pixels, carry out frame in filtering to all pixels point in the 3rd block of pixels and obtain the
Four block of pixels；Wherein, the reference frame is the former frame of the frame under process, and first pixel is in the reference frame
Location in the frame under process is identical with the currently processed pixel for location；According to second picture
Plain block and the 4th block of pixels obtain the weighted value of the time-domain filtering of the currently processed pixel；Obtained according to the weighted value
Take the time-domain filtering component of the currently processed pixel；Frame in filtering is carried out to the currently processed pixel obtains spatial domain filter
Wave component；The currently processed pixel is updated according to the time-domain filtering component, the airspace filter component and the weighted value
The gray value of point.
Further, the processor 41, is additionally operable to described at the current place centered on currently processed pixel
The pixel in the range of N*N is chosen in reason frame as the first block of pixels, all pixels point in first block of pixels is carried out
Before frame in filtering obtains the second block of pixels, judge that whether the frame under process is the first frame of described image sequence.
The memory 42, is additionally operable to when the frame under process is for the first frame of described image sequence, by the first frame
Image is stored in frame buffer, or, the first frame figure is carried out being stored in after frame in filtering process in the frame buffer, with
Just using first two field picture as the second two field picture in described image sequence reference frame.
Further, the processor 41, is additionally operable to all pictures in second block of pixels and the 4th block of pixels
Vegetarian refreshments is corresponded to and takes difference, and obtains range information after all differences are weighted averagely；Institute is obtained according to the range information
State the weighted value of currently processed pixel；Wherein, the difference including but not limited to following any one：Absolute difference,
Mean square deviation.
Further, whether the processor 41, be additionally operable to judge the weighted value less than predetermined threshold value；When the weight
When value is less than the threshold value, first time-domain filtering component of the currently processed pixel is obtained according to the weighted value, and will
The first time-domain filtering component is used as the time-domain filtering component；When the weighted value is more than or equal to the threshold value, root
The first time-domain filtering component is obtained according to the weighted value, and the first time-domain filtering component, the airspace filter are divided
The gray value of amount and the currently processed pixel takes intermediate value and obtains the time-domain filtering component.
Further, the processor 41, is additionally operable to when the ash of all pixels point updated in the frame under process
After angle value, the frame buffer is updated according to the frame under process for having updated gray value.
Further, the frame in filtering mode including but not limited to following any one：Mean filter, bilateral filtering,
NL is filtered.
The denoising device of video image provided in an embodiment of the present invention, in frame under process, will be with currently processed pixel
The all pixels point in the range of N*N centered on point carries out frame in filtering and obtains the second block of pixels, and in reference frame, will be with the
All pixels point in the range of N*N centered on one pixel carries out frame in filtering and obtains the 4th block of pixels, then further according to the
Two block of pixels and the 4th block of pixels obtain the weighted value of the time-domain filtering of currently processed pixel, and are obtained according to weighted value current
The time-domain filtering component of pixel is processed, and frame in filtering is carried out to currently processed pixel and obtains airspace filter component, Ran Hougen
The gray value of currently processed pixel is updated according to time-domain filtering component, airspace filter component and weighted value, is so being obtained currently
Before the weighted value of the time-domain filtering for processing pixel, to the pixel by currently processed pixel and centered on the first pixel
Block carries out frame in filtering, obtains weighted value further according to filtered block of pixels, it is ensured that the accuracy of weighted value, so as to ensure that
The noise of image is largely removed, and effectively prevent the motion blur phenomenon of moving region.
Also, after the weighted value for getting time-domain filtering, according to the size of weighted value, in front and back content frame gap relatively
Little region, directly obtains time-domain filtering component by mathematic interpolation, further ensures making an uproar for removal image as big as possible
Sound, for the region that in front and back's content frame gap is larger, obtains time-domain filtering component by medium filtering, further avoid motion
The motion blur phenomenon in region.
Through the above description of the embodiments, those skilled in the art can be understood that the present invention can be borrowed
Help software to add the mode of required common hardware to realize, naturally it is also possible to by hardware, but in many cases the former be more preferably
Embodiment.Based on such understanding, the portion that technical scheme is substantially contributed to prior art in other words
Divide and can be embodied in the form of software product, the computer software product is stored in the storage medium that can read, such as count
The floppy disk of calculation machine, hard disk or CD etc., use including some instructions so that computer equipment (can be personal computer,
Server, or the network equipment etc.) execute method described in each embodiment of the present invention.
The above, the only specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, and any
Those familiar with the art the invention discloses technical scope in, the change or replacement that can readily occur in, all answer
It is included within the scope of the present invention.Therefore, protection scope of the present invention with the scope of the claims should be
Accurate.
Claims (10)
1. a kind of noise-reduction method of video image, it is characterised in that include：
When frame under process is not the first frame of image sequence, in the frame under process centered on currently processed pixel
Pixel in the range of middle selection N*N carries out frame in as the first block of pixels to all pixels point in first block of pixels
Filtering obtains the second block of pixels；
The pixel in the range of N*N is chosen using centered on the first pixel in reference frame as the 3rd block of pixels, to described the
All pixels point in three block of pixels carries out frame in filtering and obtains the 4th block of pixels；Wherein, the reference frame is the current place
Reason frame former frame, first pixel in the reference frame location with the currently processed pixel described
In frame under process, location is identical；
The weight of the time-domain filtering of the currently processed pixel is obtained according to second block of pixels and the 4th block of pixels
Value；
The time-domain filtering component of the currently processed pixel is obtained according to the weighted value；
Frame in filtering is carried out to the currently processed pixel obtains airspace filter component；
The currently processed pixel is updated according to the time-domain filtering component, the airspace filter component and the weighted value
Gray value；
The time-domain filtering for obtaining the currently processed pixel according to second block of pixels and the 4th block of pixels
Weighted value, including：
Take difference by corresponding with all pixels point in the 4th block of pixels for second block of pixels, and all differences are carried out plus
Range information is obtained after weight average；Wherein, the difference including but not limited to following any one：Absolute difference, mean square deviation；
The weighted value of the currently processed pixel is obtained according to the range information.
2. the noise-reduction method of video image according to claim 1, it is characterised in that described with currently processed pixel
Centered in the frame under process choose N*N in the range of pixel as the first block of pixels, to first block of pixels
Interior all pixels point carry out frame in filtering obtain the second block of pixels before, also include：
Judge that whether the frame under process is the first frame of described image sequence；
When the frame under process is for the first frame of described image sequence, the first two field picture is stored in frame buffer, or, will
The first frame figure is stored in the frame buffer after carrying out frame in filtering process, so as to using first two field picture as institute
State the reference frame of the second two field picture in image sequence.
3. the noise-reduction method of video image according to claim 1, it is characterised in that described obtained according to the weighted value
The time-domain filtering component of the currently processed pixel, including：
Judge the weighted value whether less than predetermined threshold value；
When the weighted value is less than the threshold value, the first time domain of the currently processed pixel is obtained according to the weighted value
Filtered components, and using the first time-domain filtering component as the time-domain filtering component；
When the weighted value is more than or equal to the threshold value, the first time-domain filtering component is obtained according to the weighted value,
And intermediate value is taken to the gray value of the first time-domain filtering component, the airspace filter component and the currently processed pixel obtain
Obtain the time-domain filtering component.
4. the noise-reduction method of the video image according to any claim in claim 1-3, it is characterised in that work as renewal
After the gray value of all pixels point in the complete frame under process, also include：
The frame under process according to gray value has been updated updates frame buffer.
5. the noise-reduction method of the video image according to any claim in claim 1-3, it is characterised in that the frame
Interior filtering mode including but not limited to following any one：Mean filter, bilateral filtering, NL filtering.
6. a kind of denoising device of video image, it is characterised in that include：
Processing unit, for when frame under process be not image sequence the first frame when, centered on currently processed pixel
The pixel in the range of N*N is chosen in the frame under process as the first block of pixels, to all in first block of pixels
Pixel carries out frame in filtering and obtains the second block of pixels；
The processing unit, be additionally operable to using centered on the first pixel in reference frame choose N*N in the range of pixel as
3rd block of pixels, carries out frame in filtering and obtains the 4th block of pixels to all pixels point in the 3rd block of pixels；Wherein, described
Reference frame is the former frame of the frame under process, and with described worked as in the reference frame by location for first pixel
Pre-treatment pixel location in the frame under process is identical；
First acquisition unit, second block of pixels and the 4th block of pixels for being obtained according to the processing unit are obtained
The weighted value of the time-domain filtering of the currently processed pixel；
Second acquisition unit, the weighted value for being got according to the first acquisition unit obtain the currently processed picture
The time-domain filtering component of vegetarian refreshments；
Filter unit, obtains airspace filter component for carrying out frame in filtering to the currently processed pixel；
First updating block, for updating institute according to the time-domain filtering component, the airspace filter component and the weighted value
State the gray value of currently processed pixel；
The first acquisition unit, including：
First processing module, for taking difference by all pixels point is corresponding in second block of pixels and the 4th block of pixels,
And range information is obtained after being weighted all differences averagely；Wherein, the difference including but not limited to following any one：
Absolute difference, mean square deviation；
Acquisition module, the range information for being obtained according to the first processing module obtain the currently processed pixel
The weighted value.
7. the denoising device of video image according to claim 6, it is characterised in that also include：
Judging unit, for choosing N* in the processing unit centered on currently processed pixel in the frame under process
Pixel in the range of N carries out frame in filtering and obtains as the first block of pixels to all pixels point in first block of pixels
Before second block of pixels, judge that whether the frame under process is the first frame of described image sequence；
Memory cell, for when the frame under process is for the first frame of described image sequence, the first two field picture being stored in
In frame buffer, or, the first frame figure is carried out being stored in after frame in filtering process in the frame buffer, so as to by described the
One two field picture is used as the reference frame of the second two field picture in described image sequence.
8. the denoising device of video image according to claim 6, it is characterised in that the second acquisition unit, including：
Whether judge module, for judging the weighted value less than predetermined threshold value；
Second processing module, for when the judge module obtains the weighted value less than the threshold value, according to the weight
Value obtains the first time-domain filtering component of the currently processed pixel, and using the first time-domain filtering component as described when
Domain filtered components；
3rd processing module, for when the judge module obtains the weighted value more than or equal to the threshold value, according to institute
State weighted value and obtain the first time-domain filtering component, and to the first time-domain filtering component, the airspace filter component and
The gray value of the currently processed pixel takes intermediate value and obtains the time-domain filtering component.
9. the denoising device of the video image according to any claim in claim 68, it is characterised in that also wrap
Include：
Second updating block, for after the gray value for having updated the point of all pixels in the frame under process, according to more
Newly the frame under process of complete gray value updates frame buffer.
10. the denoising device of the video image according to any claim in claim 68, it is characterised in that described
Frame in filtering mode including but not limited to following any one：Mean filter, bilateral filtering, NL filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310137164.4A CN103269412B (en) | 2013-04-19 | 2013-04-19 | A kind of noise-reduction method of video image and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310137164.4A CN103269412B (en) | 2013-04-19 | 2013-04-19 | A kind of noise-reduction method of video image and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103269412A CN103269412A (en) | 2013-08-28 |
| CN103269412B true CN103269412B (en) | 2017-03-08 |
Family
ID=49013021
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310137164.4A Expired - Fee Related CN103269412B (en) | 2013-04-19 | 2013-04-19 | A kind of noise-reduction method of video image and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103269412B (en) |
Families Citing this family (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103632352B (en) * | 2013-11-01 | 2017-04-26 | 华为技术有限公司 | Method for time domain noise reduction of noise image and related device |
| CN104010114B (en) * | 2014-05-29 | 2017-08-29 | 广东威创视讯科技股份有限公司 | Video denoising method and device |
| CN105338221B (en) * | 2014-08-15 | 2019-01-15 | 联想(北京)有限公司 | A kind of image processing method and electronic equipment |
| CN106612386B (en) * | 2015-10-27 | 2019-01-29 | 北京航空航天大学 | A kind of noise-reduction method of joint spatial-temporal correlation properties |
| US20190045224A1 (en) * | 2016-02-04 | 2019-02-07 | Mediatek Inc. | Method and apparatus of non-local adaptive in-loop filters in video coding |
| CN105872317B (en) * | 2016-03-31 | 2019-04-05 | 北京奇艺世纪科技有限公司 | A kind of video image filtering method and device based on time domain |
| US20210042885A1 (en) * | 2018-02-08 | 2021-02-11 | Kowa Company, Ltd. | Image processing device, image processing method, and image processing program |
| CN109514043B (en) * | 2018-12-13 | 2020-11-24 | 南昌市龙诚电器设备有限公司 | Effective welding control system of welding robot |
| CN109741363B (en) * | 2019-01-11 | 2023-07-14 | 湖南国科微电子股份有限公司 | Motion judgment method and device based on block difference value and electronic equipment |
| CN109639929B (en) * | 2019-01-11 | 2020-10-30 | 珠海全志科技股份有限公司 | Image noise reduction method, computer device and computer readable storage medium |
| CN109859126B (en) | 2019-01-17 | 2021-02-02 | 浙江大华技术股份有限公司 | Video noise reduction method and device, electronic equipment and storage medium |
| CN111010495B (en) * | 2019-12-09 | 2023-03-14 | 腾讯科技（深圳）有限公司 | Video denoising processing method and device |
| CN113011433B (en) * | 2019-12-20 | 2023-10-13 | 杭州海康威视数字技术股份有限公司 | Filtering parameter adjusting method and device |
| CN111369465B (en) * | 2020-03-04 | 2024-03-08 | 东软医疗系统股份有限公司 | CT dynamic image enhancement method and device |
| CN111583151B (en) * | 2020-05-09 | 2023-05-12 | 浙江大华技术股份有限公司 | Video noise reduction method and device, and computer readable storage medium |
| CN114885073B (en) * | 2021-02-05 | 2024-03-15 | 腾讯科技（深圳）有限公司 | Video denoising method and device, storage medium and electronic equipment |
| CN113612996B (en) * | 2021-07-30 | 2024-02-09 | 百果园技术(新加坡)有限公司 | Video denoising method and device based on time domain filtering |
| CN113674316A (en) * | 2021-08-04 | 2021-11-19 | 浙江大华技术股份有限公司 | Video noise reduction method, device and equipment |
| CN114666583B (en) * | 2022-03-14 | 2023-03-21 | 中山大学 | Video coding preprocessing method based on time-space domain filtering |
| CN114998158B (en) * | 2022-08-03 | 2022-10-25 | 武汉市聚芯微电子有限责任公司 | Image processing method, terminal device and storage medium |
| CN117455802B (en) * | 2023-12-25 | 2024-04-05 | 榆林金马巴巴网络科技有限公司 | Noise reduction and enhancement method for image acquisition of intrinsic safety type miner lamp |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101262559A (en) * | 2008-03-28 | 2008-09-10 | 北京中星微电子有限公司 | A method and device for eliminating sequential image noise |
| US7542622B1 (en) * | 2003-06-02 | 2009-06-02 | The Trustees Of Columbia University In The City Of New York | Spatio-temporal treatment of noisy images using brushlets |
| CN101964863A (en) * | 2010-05-07 | 2011-02-02 | 镇江唐桥微电子有限公司 | Self-adaptive time-space domain video image denoising method |
| CN102769722A (en) * | 2012-07-20 | 2012-11-07 | 上海富瀚微电子有限公司 | Time-space domain hybrid video noise reduction device and method |
| CN103024248A (en) * | 2013-01-05 | 2013-04-03 | 上海富瀚微电子有限公司 | Motion-adaptive video image denoising method and device |
-
2013
- 2013-04-19 CN CN201310137164.4A patent/CN103269412B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7542622B1 (en) * | 2003-06-02 | 2009-06-02 | The Trustees Of Columbia University In The City Of New York | Spatio-temporal treatment of noisy images using brushlets |
| CN101262559A (en) * | 2008-03-28 | 2008-09-10 | 北京中星微电子有限公司 | A method and device for eliminating sequential image noise |
| CN101964863A (en) * | 2010-05-07 | 2011-02-02 | 镇江唐桥微电子有限公司 | Self-adaptive time-space domain video image denoising method |
| CN102769722A (en) * | 2012-07-20 | 2012-11-07 | 上海富瀚微电子有限公司 | Time-space domain hybrid video noise reduction device and method |
| CN103024248A (en) * | 2013-01-05 | 2013-04-03 | 上海富瀚微电子有限公司 | Motion-adaptive video image denoising method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103269412A (en) | 2013-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103269412B (en) | A kind of noise-reduction method of video image and device | |
| CN103455984B (en) | A kind of Kinect depth image acquisition method and device | |
| EP3099044B1 (en) | Multi-frame noise reduction method and terminal | |
| TWI714397B (en) | Method, device for video processing and computer storage medium thereof | |
| JP2009070123A (en) | Image processing device and method | |
| CN105894460A (en) | Image filtering method and device | |
| CN103533214A (en) | Video real-time denoising method based on kalman filtering and bilateral filtering | |
| CN102968814B (en) | A kind of method and apparatus of image rendering | |
| CN104331863B (en) | A kind of image filtering denoising method | |
| CN101241592B (en) | High frame frequency infrared image sequence movement target real time restoration method | |
| CN107437238A (en) | A kind of adaptive recursive noise reduction method and device of image block | |
| CN109903254B (en) | Improved bilateral filtering method based on Poisson nucleus | |
| CN107341787A (en) | Method, apparatus, server and the automobile that monocular panorama is parked | |
| CN113538245A (en) | Degradation model-based super-resolution image reconstruction method and system | |
| CN106713762A (en) | Image processing method and device | |
| CN111260590B (en) | Image noise reduction method and related product | |
| CN111369451B (en) | Image restoration model, method and device based on complex task decomposition regularization | |
| TW200643820A (en) | Color interpolation method with directed weights | |
| JP5490117B2 (en) | Image processing device | |
| CN106911934B (en) | Blocking effect removing method and deblocking filter | |
| CN102542530A (en) | Motion-blurred image defining device | |
| CN107004258A (en) | A kind of video smoothing method and device | |
| Xu et al. | A new framework of normalized convolution for superresolution using robust certainty | |
| Bilcu et al. | A modified sigma filter for noise reduction in images | |
| CN100459660C (en) | A method and apparatus of removing image noise |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170308 Termination date: 20200419 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |