CN100479497C - Method and device of vedio noise reduction - Google Patents
Method and device of vedio noise reduction Download PDFInfo
- Publication number
- CN100479497C CN100479497C CNB2007100024519A CN200710002451A CN100479497C CN 100479497 C CN100479497 C CN 100479497C CN B2007100024519 A CNB2007100024519 A CN B2007100024519A CN 200710002451 A CN200710002451 A CN 200710002451A CN 100479497 C CN100479497 C CN 100479497C
- Authority
- CN
- China
- Prior art keywords
- current point
- absolute difference
- brightness value
- point
- current
- 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
Images
Landscapes
- Picture Signal Circuits (AREA)
- Image Analysis (AREA)
Abstract
This invention provides a method and device for video noise reduction including: receiving video signals and carrying out low pass filtration to the signals, determining if the current points of the signals are flat region or static region points, when they the flat ones, the values of the current points are the result of low pass filtration, when they are static region points, they are processed evenly in multiple images. This invention also discloses a video noise-reduction device including a field storage unit, a low pass operation unit, an edge determining unit, a noise system evaluation unit and an iteration operation unit.
Description
Technical field
The present invention relates to a kind of noise reduction technology, relate in particular to a kind of method and apparatus of vedio noise reduction.
Background technology
In the video signal transmission process, can introduce white noise, the frequency distribution of described white noise is on entire spectrum, and energy is very little." noise " that show as continuous flicker of white noise in picture, therefore, white noise can reduce the display effect of picture.Picture display effect is the key factor that the consumer considers when shopping goods, so provide the function that reduces noise in the image to improve picture display effect, is one of each display production firm problem that institute must consideration when producing display screen.
The reason that white noise produces is a lot, thus very difficult from the source with its elimination.The existing method of eliminating white noise has low pass filtering method and averaging of multiple image.Below the simple principle of describing low pass filtering method and averaging of multiple image.
Low pass filtering method is a kind of method of on frequency domain picture signal being handled.When the frequency characteristic of analysis image signal, the high fdrequency component of the edge of one sub-picture, jump part and white noise representative image signal, this high fdrequency component The corresponding area is called as borderline region, large-area background area is the low frequency component of representative image signal then, and this low frequency component The corresponding area is called as flat site.But with low pass filtering method filtering HFS, thereby remove noise, image is obtained smoothly.The normal low pass filter that uses carries out filtering in low pass filtering method, and low pass filter can be realized by the multiply accumulating unit.
The principle of averaging of multiple image is, white noise often shows as additive noise, and promptly the power of white noise and primary signal is uncorrelated mutually, and its mean value is zero, just can adopt averaging of multiple image to remove noise in this case.This method also is known as the interative computation method.
In conjunction with above-mentioned noise-reduction method, in existing noise reduction technology, generally use denoising device to carry out noise reduction, as shown in Figure 1, described denoising device mainly comprises memory cell, low pass arithmetic element and interative computation unit.
Memory cell is used to store current field data and preceding field data.Described memory cell comprises: when the front court memory cell, be used to store current field data; The previous field memory cell is used to store the previous field data; The first two memory cell is used to store the first two field data.
The low pass arithmetic element is responsible for picture point is done the low-pass filtering computing.The result that the low pass arithmetic element calculates this imports the interative computation unit.The interative computation unit carries out weighted average computing between the field then to memory cell application front court corresponding points data.Its result imports memory cell, is provided with the back and uses.
In above-mentioned low pass arithmetic element, owing to there is not Boundary Detection mechanism, whole image is all carried out the calculating of low-pass filtering, cause the obscurity boundary in the picture.
In above-mentioned interative computation unit, use the precondition of interative computation method to be, image is static, and noise type is additive noise, and the rest image The corresponding area is called as stagnant zone, and the moving image The corresponding area is called as the moving region.For static image, this processing method has good effect, and for the image that moves, owing to there is not motion detection mechanism, whole image is all carried out average weighted calculating between the field, and " hangover " phenomenon appears in the picture that causes moving violent.In addition, because the interative computation method is at additive white noise, and the energy of the energy of white noise and signal itself is irrelevant.So under the situation of same energy noise, for the energy higher signal, signal to noise ratio is bigger.And for the lower signal of energy, signal to noise ratio can be less.When signal to noise ratio hour if also image is weighted on average with same standard, then can cause deterioration to the original details of picture, influence image effect.
Summary of the invention
Embodiments of the invention provide a kind of method and apparatus of vedio noise reduction, can solve the relatively poor problem of prior art noise reduction.
Embodiments of the invention provide a kind of vedio noise reduction method, comprising:
Receiving video signals, and described vision signal carried out low-pass filtering treatment;
Whether the current point of determining vision signal is flat site point or stagnant zone point;
When described current point was flat site point, the brightness value of current point was got the result of low-pass filtering treatment;
When described current point is stagnant zone point, current point is carried out averaging of multiple image handle;
When described current point is borderline region point or motion region, then get when the brightness value of the current point of front court brightness value as current point.
The embodiment of the invention also discloses a kind of vedio noise reduction device, comprising:
The field memory cell, the result of field data and current field data before being used to store, described front court data comprise current previous field and/or the video data of the first two;
The video reception unit is used for receiving video signals;
The low pass arithmetic element is used for the current point of the vision signal that receives is carried out low-pass filtering treatment;
The border determining unit is used for determining current some boundary marker according to the preceding field data and the current point data of field cell stores;
The noise factor assessment unit, be used to receive the current point data that preceding field data, the border determining unit of current point data, a memory cell determine current some boundary marker and low pass arithmetic element are handled, according to current some boundary marker, current point data, the current point data of low pass arithmetic element processing and the amount of exercise that described preceding field data is determined current point that the border determining unit is determined, determine noise factor according to the amount of exercise of current point then;
The interative computation unit, be used for when the border determining unit determines that according to the current some boundary marker of determining current point is flat site point, the value of current point is got the current point data that the low pass arithmetic element is handled, when the border determining unit determined that current point is borderline region point, the value of current point was got current point data; And according to the value of noise factor, current point and preceding field data current point is carried out averaging of multiple image and handle.
Embodiments of the invention are by dividing into picture flat site and borderline region, only the flat site point is carried out low-pass filtering treatment, and borderline region point is not carried out low-pass filtering treatment, thus make the entire image picture obtain smoothly, do not influence the effect of borderline region.In interative computation, adopted motion detection mechanism, only the stagnant zone point is carried out average weighted calculating between the field, and motion region is not handled, thereby the phenomenon of " hangover " appears in the picture of having avoided moving violent.
Description of drawings
Fig. 1 shows the denoising device of prior art;
Vedio noise reduction processing method when Fig. 2 shows the interlacing receiving video signals of the embodiment of the invention;
Fig. 3 shows the flow chart of finding the solution noise factor K of the embodiment of the invention;
Vedio noise reduction processing method when Fig. 4 shows the receiving video signals line by line of the embodiment of the invention;
Fig. 5 shows the vedio noise reduction device of the embodiment of the invention;
Fig. 6 is the differentiate schematic diagram of the embodiment of the invention in determining the border process.
Embodiment
Understand and realization the present invention the existing embodiments of the invention of describing in conjunction with the accompanying drawings for the ease of persons skilled in the art.
Embodiments of the invention provide a kind of vedio noise reduction method, and the vedio noise reduction method of the embodiment of the invention had both been used low pass filtering method, also used averaging of multiple image.Image frame is distinguished flat site and borderline region, only the flat site point is carried out low-pass filtering treatment, and borderline region point is not carried out low-pass filtering treatment, thereby make the entire image picture obtain smoothly, do not influence the effect of borderline region.Image frame is divided into stagnant zone and moving region, only the stagnant zone point is carried out average weighted calculating between the field, and motion region is not handled, thereby the phenomenon of " hangover " appears in the picture of having avoided moving violent.The vedio noise reduction method of the embodiment of the invention is described in conjunction with Fig. 2 and Fig. 3 below by two embodiment.Vedio noise reduction processing method when wherein, embodiment 1 has described the interlacing receiving video signals of the embodiment of the invention; Vedio noise reduction processing method when embodiment 2 has described the receiving video signals line by line of the embodiment of the invention.In the following embodiments, the value of each point is a brightness value.
Embodiment 1
Vedio noise reduction processing method when as shown in Figure 2, present embodiment is described the interlacing receiving video signals.
The method that obtains the absolute difference of horizontal direction specifically comprises:
The absolute difference of any before in the front court, obtaining current point and current point;
In previous field, obtain the absolute difference of the lastrow of current line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points; For convenience of description, in present specification, lastrow is meant the lastrow of current line, and next line is meant the next line of current line, and the first two is meant the previous field of previous field.
In previous field, obtain the absolute difference of next line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
In the first two, obtain the absolute difference of current line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
From above-mentioned absolute difference, determine a maximum absolute difference, the maximum absolute difference of determining is defined as the horizontal direction absolute difference.
The method that obtains the absolute difference of vertical direction specifically comprises:
Obtain to work as current point in front court and previous field lastrow when the absolute difference between the corresponding points of prostatitis;
Obtain to work as current point in front court and previous field next line when the absolute difference between the corresponding points of prostatitis;
Obtain the previous field lastrow and work as absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field next line;
Obtain the first two current line and work as absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field lastrow;
Obtain the first two current line and work as absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field next line;
From above-mentioned absolute difference, determine a maximum absolute difference, the maximum absolute difference of determining is defined as the vertical direction absolute difference.
The absolute difference of the horizontal direction that step 24, basis are determined and the absolute difference of vertical direction determine whether the current point of vision signal is the flat site point.
The current point of determining vision signal is that the flat site point methods is as follows: the absolute difference of determined level direction whether less than the absolute difference of first predetermined value or vertical direction whether less than second predetermined value, if determine that then current point is the flat site point.Wherein first predetermined value can be identical with second predetermined value, also can be different.
Obtain to work as when the current point in front court and the first two current line the absolute difference of prostatitis corresponding points, wherein, when current point was flat site point, current point was got the result of low pass computing; When current point was borderline region point, current point was got the original input when the front court;
Whether judge described absolute difference less than the 3rd predetermined value, if determine that then working as the current point in front court is the stagnant zone point.
If current point when being fringe region point or motion region, is then got when the brightness value of the current point of the front court brightness value as current point, otherwise calculate the brightness value of current point by following formula:
Y_out＝Y_cur[temp]*K/32+Y_pp[temp]*(32-K)/32
Wherein, Y_cur is the array of the first two input brightness for the array when front court input brightness, Y_pp, and temp is the sequence number of current point, and K is noise factor (computational methods of noise is referring to the description of back); When current point was flat site, Y_cur got the result of low pass computing; When current point was borderline region point, Y_cur got the original input when the front court.The benefit of doing like this is that borderline region is protected, and output at last is the 3D noise reduction result after two-dimensional space filtering and one dimension time filtering.
The computational methods of noise factor K are described below as shown in Figure 3.
The amount of exercise of step 32, the current point of calculating, momental computational methods are: brightness value and the previous field current line of getting current point are worked as the absolute difference of prostatitis corresponding points brightness value.
According to the embodiment of the invention, because unlike signal intensity signal to noise ratio difference also needs according to the signal strength signal intensity of current some the K value to be proofreaied and correct.Signal strength signal intensity can be distinguished different intervals, each interval is provided with a coefficient.This coefficient can be arranged between 0 to 1, and this coefficient has reacted the noise reduction dynamics to unlike signal intensity, but should satisfy the big more rule of the big more correction coefficient of signal strength signal intensity.For example, Power table can be divided into 8 parts, as when signal strength values between 0-255 the time, can 32 for portion is divided into 8 parts, at each part a coefficient is set.After having determined the K value of current point, can determine the signal strength signal intensity interval at current some place according to the signal strength signal intensity of current point according to the Para table.Obtain correction coefficient according to this signal strength signal intensity interval and Power table then, and multiply by the K value, obtain corrected coefficient k with this correction coefficient.
Embodiment 2
Vedio noise reduction processing method when as shown in Figure 4, present embodiment is described line by line receiving video signals.
The method of the absolute difference of acquisition horizontal direction is as follows:
The absolute difference of any before in the front court, obtaining current point and current point;
In previous field, obtain the absolute difference of current line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
From above-mentioned absolute difference, determine a maximum absolute difference, the maximum absolute difference of determining is defined as the horizontal direction absolute difference.
The method that obtains the absolute difference of vertical direction specifically comprises:
Obtain to work as current point in front court and previous field current line when the absolute difference between the corresponding points of prostatitis;
Obtain the first two current line and work as absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field current line;
From above-mentioned absolute difference, determine a maximum absolute difference, the maximum absolute difference of determining is defined as the vertical direction absolute difference.
The absolute difference of the horizontal direction that step 44, basis are determined and the absolute difference of vertical direction determine whether the current point of vision signal is the flat site point.
The current point of determining vision signal is that the flat site point methods is as follows: the absolute difference of determined level direction whether less than the absolute difference of first predetermined value or vertical direction whether less than second predetermined value, if determine that then current point is the flat site point.Wherein, first predetermined value can be identical with second predetermined value, also can be different.
Obtain to work as when current point in front court and previous field current line the absolute difference of prostatitis corresponding points, wherein, when current point was flat site point, current point was got the result of low pass computing; When current point was borderline region point, current point was got the original input when the front court;
Whether judge described absolute difference less than the 3rd predetermined value, if determine that then working as the current point in front court is the stagnant zone point.
Embodiment 3
As shown in Figure 5, embodiments of the invention also disclose a kind of vedio noise reduction device, and the vedio noise reduction device comprises: video reception unit (not shown), a memory cell 1, low pass arithmetic element 2, border determining unit 3, noise factor assessment unit 4 and interative computation unit 5.For convenience of description, be the vedio noise reduction device that example is described the embodiment of the invention with the interlacing receiving video data below.
The video reception unit is used for receiving video signals.
Field memory cell 1 is used to store previous field and the brightness data of the first two, when other unit (border determining unit 3, noise factor assessment unit 4 and interative computation unit 5) sends when request to field memory cell 1, can obtain the data of previous field lastrow and next line, and the first two current line is when the data in prostatitis.
Low pass arithmetic element 2 is used for carrying out the low pass computing when the current point in front court, and for increasing the flexibility and the accuracy of computing, it is adjustable that low pass arithmetic element 2 is set to adjustable, the every rank of exponent number coefficient.Consider to realize cost and effect, can use maximum 9 rank filtering as low pass arithmetic element 2.
Border determining unit 3 is used for the high band part in level and vertical both direction detected image, to determine whether current point is the borderline region point.Thereby provide the foundation of calculating noise coefficient for noise factor assessment unit 4.
Border determining unit 3 can have multiple implementation, and the embodiment of the invention adopts adjacent point-to-point transmission to ask the mode of first derivative to determine the border.Determine adjacent first derivative earlier at 2: try to achieve this 2 differences, ask its absolute value then, promptly ask absolute difference.If absolute difference greater than predetermined value, thinks that then current point is the HFS of image (being the borderline region point), otherwise, think that current point is the flat site point.For on the image more arbitrarily, border determining unit 3 can be between adjacent 3 (promptly when front court, previous field and the first two), carry out differentiate to 4 pairs of points of horizontal direction and 5 pairs of points of vertical direction, to judge the horizontal boundary and the vertical boundary of image.Its concrete grammar is as shown in Figure 6:
Horizontal derivative 1 calculates when current point in front court and current point more preceding carries out differentiate;
Horizontal derivative 2, previous field lastrow a bit carry out differentiate and calculate before prostatitis corresponding points and this corresponding points;
Horizontal derivative 3, previous field next line a bit carry out differentiate and calculate before prostatitis corresponding points and this corresponding points;
Horizontal derivative 4, the first two current line a bit carry out differentiate and calculate before prostatitis corresponding points and this corresponding points;
Vertical derivatives 1 is calculated when the current point in front court and previous field lastrow carry out differentiate when the prostatitis corresponding points;
Vertical derivatives 2 is calculated when the current point in front court and previous field next line carry out differentiate when the prostatitis corresponding points;
Vertical derivatives 3, previous field lastrow are carried out differentiate calculating when the prostatitis corresponding points when prostatitis corresponding points and previous field next line;
Vertical derivatives 4, the first two current line are carried out differentiate calculating when the prostatitis corresponding points when prostatitis corresponding points and previous field lastrow;
Vertical derivatives 5, the first two current line are carried out differentiate calculating when the prostatitis corresponding points when prostatitis corresponding points and previous field next line.
In Fig. 6, real circle represents that when the prostatitis corresponding points empty circle is for working as the more preceding of prostatitis corresponding points, and each square frame is other pixel.To the derivative maximizing of four sub-level directions, if this maximum greater than horizontal boundary thresholding (also claiming first predetermined value), then being judged to be current point is the horizontal direction border; In like manner, to the derivative maximizing of five vertical direction, if this maximum, judges then that current point is the vertical direction border greater than vertical boundary thresholding (also claiming second predetermined value).When if current point is horizontal direction border or vertical direction border, send boundary marker to noise factor assessment unit 4.
Noise factor assessment unit 4 is used to receive video data, the video data when the current point in front court after 2 processing of low pass arithmetic element, the boundary marker of border determining unit 3 outputs and front court (previous field and/or the first two field) video data of a memory cell 1 output when the current point in front court, determine the amount of exercise of current point according to the current point data of low pass arithmetic element processing, boundary marker, preceding field data and the current point data of border determining unit 3 outputs, and determine noise factor K, and to interative computation unit output noise COEFFICIENT K according to the amount of exercise of current point.Described noise factor assessment unit comprises: receiving element is used to receive the current point data that preceding field data, the border determining unit of current point data, a memory cell determine current some boundary marker and low pass arithmetic element are handled; Motion amount determination unit is according to the current point data of current definite boundary marker of border determining unit, current point data, the processing of low pass arithmetic element and the amount of exercise that preceding field data is determined current point; The noise factor determining unit is used for determining noise factor according to the amount of exercise of current point.
Interative computation unit 5 be used for spatially to colleague and same column correspondence mutually in abutting connection with between pixel carry out iteration filtering noise reduction.The noise factor K that interative computation unit 5 provides according to the noise factor estimation unit carries out iteration filtering and calculates.Computational methods are as follows:
Y_out＝Y_cur[temp]*K/32+Y_pp[temp]*(32-K)/32
In this formula, if during current some flat site, Y_cur gets the result of low pass computing.If when current point was borderline region, Y_cur got the original input when the front court.The benefit of doing like this is that borderline region is protected.Output at last is the 3D noise reduction result after two-dimensional space filtering and one dimension time filtering.
It should be noted, above only to being described when the zone line point of front court, for when the borderline region point of front court (as, the point of first row, the point of last column, the point of first row, the point of last row), can directly adopt initial data, also can adopt the result after the low-pass filtering when the front court.
In addition, in the above embodiments, only the processing to the regional current point in centre is described, in actual applications, can handle equally other point (removing the borderline region point) according to receive mode, like this, utilize the description of the preceding paragraph again, just can obtain whole noise reduction result borderline region point.
According to embodiments of the invention,, can avoid because the phenomenon of obscurity boundary that noise reduction causes and motion hangover has promoted the picture display quality effectively by vision signal is carried out the 3D noise reduction process.In addition, because the interative computation method has been considered the influence of signal to noise ratio to signal energy, can further improve image quality.
Though described the present invention by embodiment, those of ordinary skills know, without departing from the spirit and substance in the present invention, just can make the present invention that many distortion and variation are arranged, and scope of the present invention is limited to the appended claims.
Claims (14)
1, a kind of vedio noise reduction method is characterized in that, comprising:
Receiving video signals, and described vision signal carried out low-pass filtering treatment;
Whether the current point of determining vision signal is flat site point or stagnant zone point;
When described current point was flat site point, the brightness value of current point was got the result of low-pass filtering treatment;
When described current point is stagnant zone point, current point is carried out averaging of multiple image handle;
When described current point is borderline region point or motion region, then get when the brightness value of the current point of front court brightness value as current point.
2, method according to claim 1 is characterized in that, the mode of described receiving video signals comprises that reception line by line and interlacing receive.
3, method according to claim 2 is characterized in that, whether the current point of described definite vision signal is that flat site point specifically comprises:
Obtain the brightness value absolute difference of horizontal direction;
Obtain the brightness value absolute difference of vertical direction;
The brightness value absolute difference of determined level direction whether less than the brightness value absolute difference of first predetermined value or vertical direction whether less than second predetermined value, if determine that then current point is the flat site point.
4, method according to claim 3 is characterized in that, when receiving video signals line by line, the step of the brightness value absolute difference of described acquisition horizontal direction specifically comprises:
Brightness value absolute difference before in the front court, obtaining current point and current point between any;
In previous field, obtain the brightness value absolute difference of current line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
From above-mentioned brightness value absolute difference, determine a maximum brightness value absolute difference, and with the brightness value absolute difference of this maximum brightness value absolute difference as horizontal direction.
5, method according to claim 3 is characterized in that, when the interlacing receiving video signals, the step of the brightness value absolute difference of described acquisition horizontal direction specifically comprises:
Brightness value absolute difference before in the front court, obtaining current point and current point between any;
In previous field, obtain the brightness value absolute difference of lastrow when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
In previous field, obtain the brightness value absolute difference of next line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
In the first two, obtain the brightness value absolute difference of current line when prostatitis corresponding points and the horizontal consecutive points of these corresponding points;
From above-mentioned brightness value absolute difference, determine a maximum brightness value absolute difference, and with the brightness value absolute difference of this maximum brightness value absolute difference as horizontal direction.
6, method according to claim 3 is characterized in that, when receiving video signals line by line, the step of the brightness value absolute difference of described acquisition vertical direction specifically comprises:
Obtain to work as current point in front court and previous field current line when the brightness value absolute difference between the corresponding points of prostatitis;
Obtain the first two current line and work as brightness value absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field current line;
From above-mentioned brightness value absolute difference, determine a maximum brightness value absolute difference, and with the brightness value absolute difference of this maximum brightness value absolute difference as vertical direction.
7, method according to claim 3 is characterized in that, when the interlacing receiving video signals, the step of the brightness value absolute difference of described acquisition vertical direction specifically comprises:
Obtain to work as current point in front court and previous field lastrow when the brightness value absolute difference between the corresponding points of prostatitis;
Obtain to work as current point in front court and previous field next line when the brightness value absolute difference between the corresponding points of prostatitis;
Obtain the previous field lastrow and work as brightness value absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field next line;
Obtain the first two current line and work as brightness value absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field lastrow;
Obtain the first two current line and work as brightness value absolute difference between the corresponding points of prostatitis when prostatitis corresponding points and previous field next line;
From above-mentioned brightness value absolute difference, determine a maximum brightness value absolute difference, and with the brightness value absolute difference of this maximum brightness value absolute difference as vertical direction.
8, method according to claim 3 is characterized in that, when receiving video signals line by line, the current point of described definite vision signal is that stagnant zone point specifically comprises:
Obtain to work as when current point in front court and previous field current line the brightness value absolute difference of prostatitis corresponding points, wherein, when current point was flat site point, current point was got the result of low pass computing; When current point was borderline region point, current point was got the original input when the front court;
Whether judge described brightness value absolute difference less than the 3rd predetermined value, if determine that then working as the current point in front court is the stagnant zone point.
9, method according to claim 3 is characterized in that, when the interlacing receiving video signals, the current point of described definite vision signal is that stagnant zone point specifically comprises:
Obtain to work as when the current point in front court and the first two current line the brightness value absolute difference of prostatitis corresponding points, wherein, when current point was flat site point, current point was got the result of low pass computing; When current point was borderline region point, current point was got the original input when the front court;
Whether judge described brightness value absolute difference less than the 3rd predetermined value, if determine that then working as the current point in front court is the stagnant zone point.
10, method according to claim 1 is characterized in that, describedly current point is carried out averaging of multiple image handles and specifically to comprise:
Calculate the brightness value of current point by following formula:
Y_out=Y_cur[temp] * K/ the 4th predetermined value+Y_pp[temp] * (predetermined value of the 4th predetermined value-K)/the 4th;
Wherein, Y_out is the brightness value of current point, and Y_cur is the array when front court input brightness, Y_pp is the array of the first two input brightness, and temp is the sequence number of current point, and K is a noise factor, when current point was borderline region point or motion region, described noise factor was got the 4th predetermined value.
11, method according to claim 10 is characterized in that, described method also comprises: the amount of exercise according to current point is determined noise factor.
12, method according to claim 11 is characterized in that, described method also comprises: the signal strength signal intensity according to current point is proofreaied and correct noise factor.
13, a kind of vedio noise reduction device is characterized in that, comprising:
The field memory cell, the result of field data and current field data before being used to store, described front court data comprise current previous field and/or the video data of the first two;
The video reception unit is used for receiving video signals;
The low pass arithmetic element is used for the current point of the vision signal that receives is carried out low-pass filtering treatment;
The border determining unit is used for determining current some boundary marker according to the preceding field data and the current point data of field cell stores;
The noise factor assessment unit, be used to receive the current point data that preceding field data, the border determining unit of current point data, a memory cell determine current some boundary marker and low pass arithmetic element are handled, according to current some boundary marker, current point data, the current point data of low pass arithmetic element processing and the amount of exercise that described preceding field data is determined current point that the border determining unit is determined, determine noise factor according to the amount of exercise of current point then;
The interative computation unit, be used for when the border determining unit determines that according to boundary marker current point is flat site point, the value of current point is got the current point data that the low pass arithmetic element is handled, and when the border determining unit determined that current point is borderline region point, the value of current point was got current point data; And according to the value of noise factor, current point and preceding field data current point is carried out averaging of multiple image and handle.
14, vedio noise reduction device according to claim 13 is characterized in that, described noise factor assessment unit comprises:
Receiving element is used to receive the current point data that preceding field data, the border determining unit of current point data, a memory cell determine current some boundary marker and low pass arithmetic element are handled;
Motion amount determination unit is according to the current point data of current definite boundary marker of border determining unit, current point data, the processing of low pass arithmetic element and the amount of exercise that preceding field data is determined current point;
The noise factor determining unit is used for determining noise factor according to the amount of exercise of current point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100024519A CN100479497C (en) | 2007-01-22 | 2007-01-22 | Method and device of vedio noise reduction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100024519A CN100479497C (en) | 2007-01-22 | 2007-01-22 | Method and device of vedio noise reduction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101001317A CN101001317A (en) | 2007-07-18 |
| CN100479497C true CN100479497C (en) | 2009-04-15 |
Family
ID=38693116
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2007100024519A Expired - Fee Related CN100479497C (en) | 2007-01-22 | 2007-01-22 | Method and device of vedio noise reduction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100479497C (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104702558B (en) * | 2013-12-05 | 2018-03-09 | 上海数字电视国家工程研究中心有限公司 | The phase noise elimination method of ofdm system |
| CN103702016B (en) * | 2013-12-20 | 2017-06-09 | 广东威创视讯科技股份有限公司 | Vedio noise reduction method and device |
-
2007
- 2007-01-22 CN CNB2007100024519A patent/CN100479497C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN101001317A (en) | 2007-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1223181C (en) | Noise reduction | |
| CN102281386B (en) | Method and device for performing adaptive denoising on video image | |
| JP5708916B2 (en) | Image evaluation method, image evaluation system, and program | |
| WO2016017450A1 (en) | Image-processing device, image processing method, program, and image sensor | |
| CN104023166B (en) | A kind of environment self-adaption video image noise reducing method and device | |
| CN102356631A (en) | Image processing device, signal processing device, and program | |
| CN101478637A (en) | History-based spatio-temporal noise reduction | |
| CN1248160C (en) | Method and apparatus for edge detection | |
| CN102281388A (en) | Method and apparatus for adaptively filtering image noise | |
| CN101915901A (en) | Magnetic resonance imaging method and device | |
| CN104077746A (en) | Gray level image processing method and device | |
| CN103778611A (en) | Switch weighting vector median filter method utilizing edge detection | |
| CN102946505A (en) | Self-adaptive motion detection method based on image block statistics | |
| CN104021533A (en) | Real-time image denoising method and device | |
| CN102509311A (en) | Motion detection method and device | |
| CN100479497C (en) | Method and device of vedio noise reduction | |
| CN101237523B (en) | Main edge detection method and noise reduction method and device based on this method | |
| CN104202555A (en) | Method and device for deinterlacing | |
| CN101354783B (en) | Method and apparatus for detecting edge | |
| JP2009044594A5 (en) | ||
| CN101726829A (en) | Method for automatically focusing zoom lens | |
| CN103096009B (en) | Image processing apparatus and method and image display device and method | |
| CN1636393A (en) | Method and apparatus for improving conversion from SD to HDTV | |
| CN102013100A (en) | Image quality discrimination method based on remote sensing image phase correlation | |
| US8189874B2 (en) | Method for motion detection of horizontal line |
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: 20090415 Termination date: 20180122 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |