CN104202555A - Method and device for deinterlacing - Google Patents
Method and device for deinterlacing Download PDFInfo
- Publication number
- CN104202555A CN104202555A CN201410515279.7A CN201410515279A CN104202555A CN 104202555 A CN104202555 A CN 104202555A CN 201410515279 A CN201410515279 A CN 201410515279A CN 104202555 A CN104202555 A CN 104202555A
- Authority
- CN
- China
- Prior art keywords
- pixel
- value
- interpolation
- motion
- field
- 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.)
- Granted
Links
Landscapes
- Television Systems (AREA)
Abstract
The invention provides a method and device for deinterlacing. The method comprises the following steps: using difference values between pixels, corresponding to the pixel for interpolation, in the previous field and the subsequent field and the pixel values, corresponding to the next pixel for interpolation, in the previous field and the subsequent field, as well as the movement weight value of the previous interpolation pixel to calculate the movement weight value of the pixel for interpolation to calculate the inter-field interpolation and the in-field interpolation of the pixel for interpolation; using the inter-field interpolation, the in-field interpolation and the movement weight value of the pixel for interpolation to calculate the initial pixel value of the pixel for interpolation; calculating the relativity of the current field in multiple directions according to the difference values among the pixel values of the pixels in multiple directions in the current field; selecting multiple pixels in the direction with the maximal relativity and the pixel value of the pixel corresponding to the pixel for interpolation in the previous field for filtering; calculating the compensation pixel value, and calculating the determined pixel value of the pixel for interpolation according to the initial pixel value, the compensation pixel value and the in-field interpolation. The device is used for realizing deinterlacing by adopting the method. The device for deinterlacing is small in amount of calculation, lower in requirement on a hardware device and good in output picture quality.
Description
Technical field
The present invention relates to video signal processing field, particularly, is a kind of device that interlacing scan is converted to the interlace-removing method of lining by line scan and realizes this method.
Background technology
Digital broadcast television Technology Need is the form transmission with Frame by vision signal, and the data of transmitting often comprise multirow data.And digital broadcast television technology comprises technology and the interleaved technology of lining by line scan, as shown in Figure 1, the packet that progressive scanning technology is transmitted is containing the data of every a line, as shown in Figure 1 be three row data wherein, the first row data are for comprising three pixel a, b, c, the second row comprises three pixel x, y, z, and the third line comprises three pixel d, e, f.As each field data that digital broadcast television technique transmitted includes the data of every a line, need the data volume of transmission excessive, very high to bandwidth requirement, therefore, traditional broadcast television technique normally adopts horizontal-interlace technique to save bandwidth.
In the data that horizontal-interlace technique is transmitted, each field data only comprises the data of odd-numbered line or the data of even number line, and the odd even line number of the data transmitted of two adjacent field data is just in time complementary.As shown in Figure 2, preceding field F (n-1), when front court F (n), subsequent fields F (n+1) be respectively the three adjacent field data of transmitting in horizontal-interlace technique.Wherein, when front court F (n) is the field of carrying out deinterlacing, wherein pixel a, b, c and pixel d, e, f be place when adjacent odd-numbered line or the even number line of front court F (n), between two row data, lack even number line or odd-numbered line, the data of actual transmissions are not comprise pixel y, x, z.The row that the row that pixel k, g in preceding field F shown in Fig. 2 (n-1), h form and pixel l, i in subsequent fields F (n+1), j form is all institute's even number line on the scene or odd-numbered line, and preceding field F (n-1) and subsequent fields F (n+1) all lack the data of odd-numbered line or even number line.
But horizontal-interlace technique has but been sacrificed picture quality, the image of its transmission can not meet the requirement of spectators to picture quality.In addition, current most of display device all adopts progressive scanning technology, and the deinterlacing technique that therefore interleaved data is changed into the data of lining by line scan is the important technology in current image date transmission process field.Visible, the input field receiving in horizontal-interlace technique comprises strange and even, and strange field only comprises the odd-numbered line of source figure, and even field only comprises the even number line of source figure, goes interlacing field will be converted to frame exactly, all row that frame comprises image.Therefore, deinterlacing technique be exactly need between the adjacent two row data of each place transmission, insert the even number of lines that lacked according to or odd-numbered line data.Take in Fig. 2 when front court F (n) be example, the object of deinterlacing technique is exactly for the even number line that between pixel a, b, c place row data and pixel d, e as front court F (n), f place row data, insertion lacks or pixel interpolating value y, x, the z of odd-numbered line.
Suppose that in Fig. 2, the signal frame as front court F (n) is comprised of even number line, pixel a, b, c place row data and pixel d, e, f place row data are even number line, when front court F (n) lacks odd-numbered line data.Deinterlacing technique be exactly need to adjacent even number of lines according between insert corresponding odd-numbered line data, so deinterlacing technique generates the pixel value of the odd-numbered line pixel having lacked in front court F (n) exactly.Pixel to be generated is commonly referred to interpolation pixel, the pixel of supposing interpolation is the pixel x as front court F (n), the pixel y adjacent with the left side of pixel x in same a line is when front court F (n) is by the pixel of the deinterlacing technique generation of three Motion Adaptives, namely go up an interpolating pixel, the pixel z adjacent with the right side of pixel x in same a line is the pixel of next interpolation after pixel x interpolation completes, namely next interpolation pixel.
In the deinterlacing technique of application, what effect was best is the interlace-removing method based on motion compensation, large but the algorithm of motion compensation is realized operand at present, and high to hardware device requirement, application cost is high.But not in movement compensating algorithm, preferably the interlace-removing method based on Motion Adaptive, its step comprises that motion detection and weights calculate, field interpolation calculates, between interpolation calculation, the weights mixing field interpolate value finally drawing by motion detection and between Output rusults after interpolation.For motion detection tendency, be static image, weights bias toward interpolation between field, otherwise bias toward field interpolation.Therefore, the accuracy requirement of motion detection is very high, otherwise can cause the phenomenons such as serious sawtooth, emergence and spot.
Method for testing motion performance is relatively good and to realize being simply three detection methods.Three traditional detections are that calculating previous field and difference latter and locational corresponding pixel points interpolation pixel judge whether interpolation point exists obvious motion, if difference is greater than threshold value, thinks that interpolation is pointed out to have aobvious motion, otherwise think that interpolation pixel is static.
For the very fast image of some motions, the difference of the pixel of interpolation point position, field, front and back tends to very little, so three traditional detections are easy to sentence it as static by mistake.Therefore, people improve three traditional method for testing motion, the patent of invention that is CN101600061B and CN1207906C as notification number discloses two kinds of different interlace-removing methods based on three method for testing motion, although these methods have been improved the performance of three motion detection, but its operand is very big, to hardware require too highly, also raised the cost of realizing of interlace-removing method.
Summary of the invention
Main purpose of the present invention is to provide the interlace-removing method that a kind of operand is little and output image quality is high.
Another object of the present invention be to provide a kind of to hardware requirement the low and measured de-interlacing apparatus of synthetic image matter.
In order to realize above-mentioned main purpose, interlace-removing method provided by the invention comprise according to preceding field with in subsequent fields with the pixel value of pixel corresponding to interpolation pixel between difference judge whether interpolation pixel exists obvious motion, and, application preceding field with in subsequent fields with the pixel value of pixel corresponding to interpolation pixel between difference, preceding field with in subsequent fields with the pixel value of pixel corresponding to next interpolation pixel between difference and the motion weighted value of the upper interpolating pixel motion weighted value that calculates interpolation pixel, interpolation and field interpolation between the field of calculating interpolation pixel, and interpolation between applied field, the motion weighted value of field interpolation and interpolation pixel calculates the original pixel values of interpolation pixel, according to the difference between the pixel value of the pixel in multiple directions in front court, calculate the correlation when front court multiple directions, the pixel value of the pixel that a plurality of pixels in the direction of selection correlation maximum and preceding field interpolation pixel are corresponding carries out filtering, calculate compensation pixel value, and according to original pixel values, compensation pixel value and field interpolation calculate definite pixel value of interpolation pixel.
From such scheme, between applied field, interpolation, field interpolation and motion weighted value calculate the original pixel values of interpolation pixel, the operand of original pixel values is little, and the original pixel values that application motion weighted value calculates can calculate according to the motion conditions of pixel the pixel value of each pixel, the original pixel values of calculating is more accurate, and the picture quality of output is high.
And, calculate after the original pixel values of interpolation pixel, also compensate calculating, and the operand that compensation is calculated is little, but the motion state that can effectively treat interpolating pixel compensates, the picture quality of output is higher, avoids motion state misjudgment and the phenomenons such as sawtooth, emergence appear in the image that causes.
Further scheme is, the step of calculating the motion weighted value of interpolation pixel is: application preceding field with in subsequent fields with the pixel value of pixel corresponding to interpolation pixel between difference calculate the first motion state value, application preceding field with in subsequent fields with the pixel value of pixel corresponding to next interpolation pixel between difference calculate the second motion state value, in the first motion state value, while having two to be expressed as motion in the motion weighted value of the second motion state value and a upper interpolating pixel, the motion weighted value of interpolation pixel is made as to motion value, otherwise, the motion weighted value of interpolation pixel is made as to static value.
As can be seen here, by the motion weighted value of the first motion state value, the second motion state value and a upper interpolating pixel is carried out to the motion weighted value that simple logical operation can obtain interpolation pixel, the operand of the motion weighted value of interpolation pixel is less, to hardware device, requires lower.
An optional scheme is, the step of calculating the motion weighted value of interpolation pixel is: application preceding field with in subsequent fields with the pixel value of pixel corresponding to interpolation pixel between difference calculate the first motion state value, application preceding field with in subsequent fields with the pixel value of pixel corresponding to next interpolation pixel between difference calculate the second motion state value, the motion weighted value of interpolation pixel is half of motion weighted value sum of the first motion state value, the second motion state value and a upper interpolating pixel.
As can be seen here, the moving situation that can clearly reflect interpolation pixel by calculating the first motion state value, the second motion state value and a upper interpolating pixel motion weighted value, the numerical value of the motion weighted value calculating thus reflects the motion state of interpolation pixel better, and the raising of the quality of output image is had to obvious benefit.
For realizing another above-mentioned object, de-interlacing apparatus provided by the invention comprises motion detection block, application preceding field with in subsequent fields with the pixel value of pixel corresponding to interpolation pixel between difference, preceding field and subsequent fields in the pixel value of pixel corresponding to next interpolation pixel between difference and the motion weighted value of the motion weighted value calculating interpolation pixel of a upper interpolating pixel, also be provided with interpolation computing module between field, interpolation between the field of calculating interpolation pixel, field interpolation computing module, the field interpolation of calculating interpolation pixel, and mixing computing module, interpolation between applied field, the motion weighted value of field interpolation and interpolation pixel calculates the original pixel values of interpolation pixel, also be provided with standard of compensation computing module, according to the difference between the pixel value of the pixel in multiple directions in front court, calculate the correlation when front court multiple directions, the pixel value of the pixel that a plurality of pixels in the direction of selection correlation maximum and preceding field are corresponding with interpolation pixel carries out filtering, calculate compensation pixel value, compensation determination module, according to original pixel values, compensation pixel value and field interpolation calculate definite pixel value of interpolation pixel.
From such scheme, while calculating the pixel value of interpolation pixel, calculating in conjunction with motion detection, motion weight calculation, compensation pixel value, can make the interpolation pixel value calculating, the pixel value of the interpolation pixel of calculating is more accurate, and operand is little, to hardware device require lowly, the picture quality generating is good.
Accompanying drawing explanation
Fig. 1 is the pixel map of data that progressive scanning technology is transmitted.
Fig. 2 is preceding field in interlace-removing method, when the pixel map of front court and subsequent fields.
Fig. 3 is the block diagram of de-interlacing apparatus embodiment of the present invention.
Fig. 4 is the flow chart of interlace-removing method embodiment of the present invention.
Below in conjunction with drawings and Examples, the invention will be further described.
Embodiment
Interlace-removing method of the present invention is applied in digital broadcast television transmission technology, for the technology that interleaved data transaction is become to line by line scan, namely generates the pixel value of the one-row pixels point lacking in each.De-interlacing apparatus of the present invention is the device of the data that interleaved data transaction is become to line by line scan that realize of application interlace-removing method.
The data that horizontal-interlace technique is transmitted as shown in Figure 2, include a plurality of adjacent fields, and each only comprises the data of odd-numbered line or the data of even number line.For example, three adjacent be respectively preceding field F (n-1), as front court F (n), subsequent fields F (n+1), when front court F (n) is the field of carrying out deinterlacing, pixel a, b, c and pixel d, e, f are the pixels when the adjacent even number line of front court F (n), lack the pixel of odd-numbered line between two row pixels.Pixel y, x, z are exactly the pixel that needs the odd-numbered line of calculating, generation in interlace-removing method, wherein pixel x is the pixel of interpolation, pixel y is a upper interpolating pixel point as calculated before calculating pixel point x, and pixel z is the pixel of next interpolation after pixel x.
The row that pixel k, g in preceding field F shown in Fig. 2 (n-1), h form is the pixel of preceding field F (n-1) odd-numbered line, the row that pixel l, i in subsequent fields F (n+1), j form is the pixel of odd-numbered line in subsequent fields F (n+1), visible, preceding field F (n-1) and subsequent fields F (n+1) all lack the pixel of even number line.
Referring to Fig. 3, de-interlacing apparatus of the present invention has interpolation calculation module 10 and compensation computing module 20, wherein interpolation calculation module 10 has interpolation computing module 13 between motion detection block 11, field, field interpolation computing module 14 and mixing computing module 15, and compensation computing module 20 has standard of compensation computing module 21 and compensation determination module 22.
Motion detection block 11 is for calculating and judging whether interpolation pixel exists obvious motion, and calculates the motion weighted value of interpolation pixel.Between, interpolation computing module 13, field interpolation computing module 14 are respectively used to calculate interpolation and field interpolation between the field of interpolation pixel, mix computing module 15 application interpolation pixels motion weighted value, between interpolation, field interpolation calculate the initial pixel of interpolation pixel.
Picture quality in order to ensure output is higher, the phenomenons such as the sawtooth of minimizing image, emergence, spot, after calculating the original pixel values of interpolation image, also need to treat interpolating pixel and carry out motion compensation calculations, motion compensation calculations realizes by compensating computing module 20.Standard of compensation computing module 21 is for calculating the reference data of compensation, and compensation determination module 22 calculates the concrete numerical value of compensation according to the result of calculation of standard of compensation computing module 21, obtain the final definite pixel value of interpolation pixel.
Below in conjunction with Fig. 4, introduce the operation principle of de-interlacing apparatus modules, and the flow process of explanation interlace-removing method.
First, execution step S1, receives original view data, the data that raw image data is many, and each is odd field or even field, and odd field only comprises the pixel of odd-numbered line, and even field only comprises the demonstration of even number line.
Then, motion detection block 11 execution step S2, the motion weight of calculating interpolation pixel.For the interpolation pixel x in Fig. 2, choose pixel g corresponding with interpolation pixel x in preceding field F (n-1) and the corresponding pixel h of next interpolation pixel z, and choose pixel i corresponding with interpolation pixel x in subsequent fields F (n+1) and the corresponding pixel j of next interpolation pixel z, in conjunction with the upper interpolating pixel y as front court F (n), judge the motion conditions of interpolation pixel x.The motion weight of interpolation pixel can be 0 or 1 integer, can be also the decimal between 0 to 1, the invention provides two kinds of different computational methods.
Whether obviously first method is to judge interpolation pixel x motion, the motion weighted value that this method is calculated is only 0 or 1, motion weighted value is 0, represent not obviously motion of interpolation pixel x, static value namely, if motion weighted value is 1, represent that interpolation pixel x exists obviously motion, namely motion value.Concrete computational methods are, absolute difference and a threshold value between the pixel value of the pixel at preceding field F (n-1) and subsequent fields F (n+1) same position place are compared, if this absolute difference is less than threshold value, can judge that interpolation pixel does not exist obvious motion, if absolute difference is greater than threshold value, judge that interpolation pixel exists obviously motion.
For example, difference and the threshold value of choosing the absolute value between the pixel value of pixel g in preceding field F (n-1) and the pixel value of the middle pixel i of subsequent fields F (n+1) compare, and judge between two pixels it is not have motion or have obviously motion, and judgment value is labeled as to the first motion state value Y1.When two pixels of judgement do not exist motion, the first motion state value Y1 assignment is 0, judges that when two pixels exist obviously motion, the first motion state value Y1 assignment is 1.Given threshold is Th, can obtain Y1=(| fn-1 (g)-fn+1 (i) | >Th), wherein fn-1 (g) is the pixel value of pixel g, and fn+1 (i) is the pixel value of pixel i.
In like manner, absolute value differences and the threshold value chosen between the pixel value of pixel h in preceding field F (n-1) and the pixel value of the middle pixel j of subsequent fields F (n+1) compare, judge between two pixels and whether do not exist motion or existence obviously to move, and the second motion state value is labeled as to Y2.Therefore can obtain Y2=(| fn-1 (h)-fn+1 (j) | >Th), wherein fn-1 (h) is the pixel value of pixel h, and fn+1 (j) is the pixel value of pixel j.
And, the motion weighted value of working as a upper interpolating pixel y in front court F (n) is labeled as to motiony.If having two values in the motion weighted value motiony of the first motion state value Y1, the second motion state value Y2 and a upper interpolating pixel y is 1, judgement is moved when the interpolation pixel x of front court F (n) exists obviously, otherwise, think that interpolation pixel x does not exist motion or moves not obvious.Suppose the motion weighted value of interpolation pixel x to be labeled as motionx, can obtain logical formula motionx=Y1 & & motiony||Y1 & & Y2||Y2 & & motion.By above-mentioned logical formula, can calculate the motion weighted value of interpolation pixel x.Distinguishingly, when calculating first interpolation pixel, the motion weighted value mark motiony of a upper interpolating pixel is set to 1.
Second method is that the motion weighted value motionx of interpolation pixel x is used to fractional representation, and its span is between 0 to 1.And, two threshold values are set, be respectively Th1 and Th2, preferably, the value of two threshold values is respectively Th1=8 and Th2=16.
While calculating the motion weighted value of interpolation pixel x, absolute value differences and the threshold value chosen between the middle pixel value of pixel g of preceding field F (n-1) and the pixel value of subsequent fields F (n+1) pixel i compare, if the absolute value of two pixel interpolation is less than threshold value Th1, | fn-1 (g)-fn+1 (i) | <Th1, thinks the first motion state value motionx1=0.If the absolute value of two pixel interpolation is greater than threshold value Th1, but be less than Th2+Th1, be Th1<|fn-1 (g)-fn+1 (i) | <Th2+Th1, think 0<motionx1<1, for example, can use formula motionx1=(| fn-1 (g)-fn+1 (i) |-Th1)/Th2 to calculate the value of motionx1.If the absolute value of two pixel interpolation is greater than Th2+Th1, | fn-1 (g)-fn+1 (i) | >Th2+Th1, thinks the first motion state value motionx1=1.
In like manner, absolute value differences and the threshold value chosen between the middle pixel value of pixel h of preceding field F (n-1) and the pixel value of the pixel j of subsequent fields F (n+1) compare, if | fn-1 (h)-fn+1 (j) | <Th1, thinks the second motion state value motionx2=0; If Th1<|fn-1 (h)-fn+1 (j) | <Th2+Th1; think 0<motionx2<1, for example, can use formula motionx2=(| fn-1 (h)-fn+1 (j) |-Th1)/Th2 to calculate the value of motionx2; If | fn-1 (h)-fn+1 (j) | >Th2+Th1, thinks the second motion state value motionx2=1.
Finally, in conjunction with the first motion state value motionx1, the value of the motion weighted value motiony of the second motion state value motionx2 and a upper interpolating pixel is calculated motion state and the motion weighted value of interpolation pixel x, the motion weighted value motionx of interpolation pixel x is the first motion state value motionx1, half of the motion weighted value motiony sum of the second motion state value motionx2 and a upper interpolating pixel y, namely by following formula, calculate the motion weighted value motionx:Motionx=(motionx1+motionx2+motiony)/2 of interpolation pixel x.If result of calculation motionx>1, thinks motionx=1, if 0< is motionx<1, think motionx=motionx.
Calculate after the motion weighted value of interpolation pixel x, interpolation computing module 13 and field interpolation computing module 14 execution step S3 between, calculate respectively interpolation and field interpolation between the field of interpolation pixel x.
Between calculated field during interpolation, choose in preceding field F (n-1) and subsequent fields F (n+1) with pixel value value of averaging of the pixel of interpolation pixel x correspondence position and calculate.For example choose the pixel value of the pixel i of pixel g in preceding field F (n-1) and subsequent fields F (n+1), according to following formula, calculate interpolation ft:ft=[fn-1 (g)+fn+1 (i) between the field of interpolation pixel x]/2.
During calculated field interpolate value, adopt in same and obtain field interpolation with the pixel value mean value of the upper and lower two row respective pixel of interpolation pixel x, for example choose pixel b in front court F (n) and the pixel value of pixel e, calculate as follows field interpolation fs:fs=[fn (b)+fn (e) of interpolation pixel x]/2, wherein fn (b), fn (e) they are respectively the pixel values of pixel b and pixel e.
Then, mix computing module 15 execution step S4, calculate the original pixel values fadp of interpolation pixel x.Calculate the original pixel values fadp of interpolation pixel x, need to calculate in conjunction with interpolation ft, field interpolation fs and tri-numerical value of motion weighted value Motionx between field.For example, can calculate according to following formula original pixel values fadp:fadp=fs * Motionx+ (1-Motionx) * ft of interpolation pixel x.
Above-mentioned step S1 is less to step S4 operand, therefore lower to the hardware requirement of electronic equipment, only need to calculate the data of a line in preceding field F (n-1) and subsequent fields F (n+1), so do not need line buffer to carry out the data of buffer memory preceding field F (n-1) and subsequent fields F (n+1).
As only relied on the motion detection result of interpolation pixel x, determine the pixel value of interpolation pixel x, result of calculation often may not be very accurate, the pixel of componental movement may be mistaken for staticly, the result of 20 couples of step S4 of compensation computing module that therefore need to be by Motion Adaptive compensates calculating.
While carrying out compensation calculating, first, execution step S5, the correlation that standard of compensation computing module 21 calculates when front court F (n) multiple directions, choose pixel g and the pixel a in front court F (n) in preceding field F (n-1), b, c, d, e, f carries out computing, the diagonal of calculating the pixel of working as upper and lower two row of front court F (n) represents possible edge direction, by calculating, compare the absolute value between the pixel on different directions, the correlation of respective pixel and interpolation pixel x in judgement multiple directions, the correlation that absolute difference is less is larger, direction corresponding to minimal difference is exactly edge direction.
Preferably, a plurality of different directions comprise diagonal, adjacent inclined direction and vertical direction, and if pixel a, x, f are in diagonal, pixel b, f are in adjacent inclined direction, pixel b, x, e in the vertical direction.Therefore, can calculate respectively the correlation on different directions by six formula below.
X1=|fn(a)-fn(e)|+|fn(b)-fn(f)|；
X2=|fn(b)-fn(d)|+|fn(c)-fn(e)|;
X3=|fn(b)-fn(e)|×2;
K1=|fn(a)-fn(f)|;
K2=?X3/2;
K3=|fn(c)-fn(d)|。
Then, select the pixel of pixel corresponding to a plurality of pixels in a direction of correlation maximum and preceding field interpolation pixel to carry out medium filtering.Medium filtering is exactly the median of choosing pixel value in a plurality of pixels.First select absolute difference minimum in X1, X2, X3, carry out the medium filtering of 5.Consider if a direction of correlation maximum is 90 degree, in front court F (n), the line directional dependency of pixel b and pixel e is maximum, the medium filtering of 5 will deteriorate to pixel b, e, 3 medium filterings of g so, and the result calculating is more excellent.Therefore, in sentencing X1, X2, X3, after minimum absolute difference value, need to further do as judged:
If X1 is the least absolute value difference in X1, X2, X3, relatively the size of K1 and K2, if K1< is K2, selects pixel a, b, e, f, g to carry out medium filtering, otherwise selected pixels point b, e, g carry out medium filtering.
If X2 is the least absolute value difference in X1, X2, X3, relatively the size of K2 and K3, if K3< is K2, selects pixel b, c, d, f, g to carry out medium filtering, otherwise selected pixels point b, e, g carry out medium filtering.
If X3 is the least absolute value difference in X1, X2, X3, directly selected pixels point b, e, g carry out medium filtering.
Visible, what the present invention adopted is 5 medium filtering sides, the sawtooth compared with 3 median filter method in can removal of images, and the picture quality of output is better.Carry out after above-mentioned medium filtering, can calculate the compensation pixel value fstd of interpolation pixel x, so far, the step that is finished S6.
Then, compensation determination module 22 execution step S7 choose a pixel value as definite pixel value of interpolation pixel x between original pixel values fadp, the compensation pixel value fstd of interpolation pixel x and field interpolation fs, namely compensate to judge and calculate.Compensation judges that calculating is that above-mentioned three numerical value are carried out to the calculating of absolute value differences compensation judgment, and selecting the less value of absolute value interpolation is the de-interlacing interpolation result of calculation of interpolation pixel x, namely determines pixel value comp.For example, if | fstd-fadp|<|fstd-fs|, determine that pixel value comp is original pixel values fadp, otherwise, field interpolation fs chosen as definite pixel value comp.
Preferably, compensation judges that computing module 22 also can adopt medium filtering to replace above-mentioned calculating, by following formula calculative determination pixel value comp:comp=median (fstd, fs, fadp).Wherein median is median operation.Through above-mentioned computing, can calculate and determine pixel value comp, and finally perform step S8, by the definite pixel value comp output calculating.
The pixel value that calculates interpolation pixel by above-mentioned method can compensate the pixel of odd-numbered line or the pixel of even number line that each place lacks, and the operand of interlace-removing method of the present invention is little, low to hardware requirement, thus the application cost of reduction deinterlacing technique.And after application compensation is calculated, the phenomenons such as sawtooth that can removal of images, emergence, improve the quality of image.
Certainly, above-described embodiment is only the preferred embodiment of the present invention, during practical application, can also have more variation, for example, while calculating the motion weighted value of interpolation pixel, can use other formula; Or, while calculating the correlation of multiple directions, can adopt more or a part of direction as calculating benchmark, such change does not affect enforcement of the present invention.
Finally it is emphasized that; the invention is not restricted to above-mentioned execution mode, as judge the whether obviously change of the circular of motion of pixel, the variations such as change of calculating the specific formula for calculation of original pixel values, compensation pixel value also should be included in the protection range of the claims in the present invention.
Claims (10)
1. interlace-removing method, is characterized in that: comprising:
Application preceding field with in subsequent fields with the pixel value of pixel corresponding to interpolation pixel between difference, preceding field and subsequent fields in the pixel value of pixel corresponding to next interpolation pixel between difference and the motion weighted value of the upper interpolating pixel motion weighted value that calculates described interpolation pixel;
Calculate interpolation and field interpolation between the field of described interpolation pixel, and apply described between the motion weighted value of interpolation, described field interpolation and described interpolation pixel calculate the original pixel values of described interpolation pixel;
According to the difference between the pixel value of the pixel in multiple directions in front court, calculate the correlation when front court multiple directions, in a plurality of pixels in the direction of selection correlation maximum and preceding field, the pixel value of the pixel corresponding with described interpolation pixel carries out filtering, calculates compensation pixel value;
According to described original pixel values, described compensation pixel value and described field interpolation, calculate definite pixel value of described interpolation pixel.
2. interlace-removing method according to claim 1, is characterized in that:
The step of calculating the motion weighted value of described interpolation pixel is: application preceding field with in subsequent fields with the pixel value of pixel corresponding to described interpolation pixel between difference calculate the first motion state value, application preceding field with in subsequent fields with the pixel value of pixel corresponding to described next interpolation pixel between difference calculate the second motion state value, in described the first motion state value, while having two to be expressed as motion in the motion weighted value of described the second motion state value and a described upper interpolating pixel, the motion weighted value of described interpolation pixel is made as to motion value, otherwise, the motion weighted value of described interpolation pixel is made as to static value.
3. interlace-removing method according to claim 1, is characterized in that:
The step of calculating the motion weighted value of described interpolation pixel is: application preceding field with in subsequent fields with the pixel value of pixel corresponding to described interpolation pixel between difference calculate the first motion state value, application preceding field with in subsequent fields with the pixel value of pixel corresponding to described next interpolation pixel between difference calculate the second motion state value, the motion weighted value of described interpolation pixel is half of motion weighted value sum of described the first motion state value, described the second motion state value and a described upper interpolating pixel.
4. interlace-removing method according to claim 3, is characterized in that:
The motion weighted value of described the first motion state value, described the second motion state value and described interpolation pixel is the numerical value between 0 to 1.
5. according to the interlace-removing method described in claims 1 to 3 any one, it is characterized in that:
The step of calculating described definite pixel value is that described original pixel values, described compensation pixel value and described field interpolation are carried out to medium filtering.
6. according to the interlace-removing method described in claims 1 to 3 any one, it is characterized in that:
Calculate in the step of described compensation pixel value, a plurality of described directions at least comprise diagonal, adjacent inclined direction or vertical direction.
7. interlace-removing method device, is characterized in that: comprising:
Motion detection block, application preceding field with in subsequent fields with the pixel value of pixel corresponding to described interpolation pixel between difference, preceding field and subsequent fields in the pixel value of pixel corresponding to next interpolation pixel between difference and the motion weighted value of the upper interpolating pixel motion weighted value that calculates described interpolation pixel;
Between interpolation computing module, calculate described interpolation pixel field between interpolation;
Field interpolation computing module, calculates the field interpolation of described interpolation pixel;
Mix computing module, the motion weighted value of applying interpolation, described field interpolation and described interpolation pixel between described calculates the original pixel values of described interpolation pixel;
Standard of compensation computing module, according to the difference between the pixel value of the pixel in multiple directions in front court, calculate the correlation when front court multiple directions, the pixel value of the pixel that a plurality of pixels in the direction of selection correlation maximum and preceding field are corresponding with described interpolation pixel carries out filtering, calculates compensation pixel value;
Compensate determination module, according to described original pixel values, described compensation pixel value and described field interpolation, calculate definite pixel value of described interpolation pixel.
8. de-interlacing apparatus according to claim 7, is characterized in that:
Described motion detection block application preceding field with in subsequent fields with the pixel value of pixel corresponding to described interpolation pixel between difference calculate the first motion state value, application preceding field with in subsequent fields with the pixel value of pixel corresponding to described next interpolation pixel between difference calculate the second motion state value, in described the first motion state value, while having two to be expressed as motion in the motion weighted value of described the second motion state value and a described upper interpolating pixel, the motion weighted value of described interpolation pixel is made as to motion value, otherwise, the motion weighted value of described interpolation pixel is made as to static value.
9. de-interlacing apparatus according to claim 7, is characterized in that:
Described motion detection block application preceding field with in subsequent fields with the pixel value of pixel corresponding to described interpolation pixel between difference calculate the first motion state value, application preceding field with in subsequent fields with the pixel value of pixel corresponding to described next interpolation pixel between difference calculate the second motion state value, the motion weighted value of described interpolation pixel is half of motion weighted value sum of described the first motion state value, described the second motion state value and a described upper interpolating pixel.
10. according to the de-interlacing apparatus described in claim 7 to 9 any one, it is characterized in that:
Described compensation determination module carries out medium filtering to described original pixel values, described compensation pixel value and described field interpolation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410515279.7A CN104202555B (en) | 2014-09-29 | 2014-09-29 | Interlace-removing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410515279.7A CN104202555B (en) | 2014-09-29 | 2014-09-29 | Interlace-removing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104202555A true CN104202555A (en) | 2014-12-10 |
| CN104202555B CN104202555B (en) | 2017-10-20 |
Family
ID=52087775
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410515279.7A Active CN104202555B (en) | 2014-09-29 | 2014-09-29 | Interlace-removing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104202555B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105704415A (en) * | 2016-02-02 | 2016-06-22 | 福州瑞芯微电子股份有限公司 | Video de-interleaving method and device |
| CN105744201A (en) * | 2014-12-31 | 2016-07-06 | 展讯通信（上海）有限公司 | Television signal de-interlacing processing method and device |
| CN105898179A (en) * | 2015-12-14 | 2016-08-24 | 乐视云计算有限公司 | Deinterlace method and device for interlaced video |
| CN113261276A (en) * | 2019-01-09 | 2021-08-13 | 西安诺瓦星云科技股份有限公司 | Deinterlacing interpolation method, device and system, video processing method and storage medium |
| CN113852830A (en) * | 2021-09-23 | 2021-12-28 | 杭州国芯科技股份有限公司 | Median filtering video de-interlacing method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201222771Y (en) * | 2008-05-30 | 2009-04-15 | 深圳艾科创新微电子有限公司 | High speed edge self-adapting de-interlaced interpolation device |
| CN101510985A (en) * | 2009-02-27 | 2009-08-19 | 西安交通大学 | Self-adapting de-interleave method for movement compensation accessory movement |
| CN102148953A (en) * | 2011-04-20 | 2011-08-10 | 广东威创视讯科技股份有限公司 | Method and device for detecting three-field motion of de-interlacing processing and de-interlacing system |
| CN102186045A (en) * | 2011-04-20 | 2011-09-14 | 广东威创视讯科技股份有限公司 | Three-field motion detection method and device for deinterlacing processing and deinterlacing system |
| CN102215368A (en) * | 2011-06-02 | 2011-10-12 | 中山大学 | Motion self-adaptive de-interlacing method based on visual characteristics |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100412501B1 (en) * | 2001-11-30 | 2003-12-31 | 삼성전자주식회사 | Pixel-data selection device for motion compensation and method of the same |
| CN102497524B (en) * | 2011-12-05 | 2014-02-12 | 四川虹微技术有限公司 | Edge adaptive de-interlacing interpolation method |
| CN102497523B (en) * | 2011-12-05 | 2014-02-12 | 四川虹微技术有限公司 | Median-filtering de-interlacing method |
| CN103369208B (en) * | 2013-07-15 | 2016-12-28 | 青岛海信电器股份有限公司 | Self adaptation interlace-removing method and device |
-
2014
- 2014-09-29 CN CN201410515279.7A patent/CN104202555B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201222771Y (en) * | 2008-05-30 | 2009-04-15 | 深圳艾科创新微电子有限公司 | High speed edge self-adapting de-interlaced interpolation device |
| CN101510985A (en) * | 2009-02-27 | 2009-08-19 | 西安交通大学 | Self-adapting de-interleave method for movement compensation accessory movement |
| CN102148953A (en) * | 2011-04-20 | 2011-08-10 | 广东威创视讯科技股份有限公司 | Method and device for detecting three-field motion of de-interlacing processing and de-interlacing system |
| CN102186045A (en) * | 2011-04-20 | 2011-09-14 | 广东威创视讯科技股份有限公司 | Three-field motion detection method and device for deinterlacing processing and deinterlacing system |
| CN102215368A (en) * | 2011-06-02 | 2011-10-12 | 中山大学 | Motion self-adaptive de-interlacing method based on visual characteristics |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105744201A (en) * | 2014-12-31 | 2016-07-06 | 展讯通信（上海）有限公司 | Television signal de-interlacing processing method and device |
| CN105744201B (en) * | 2014-12-31 | 2019-01-18 | 展讯通信（上海）有限公司 | The de-interlaced treating method and apparatus of TV signal |
| CN105898179A (en) * | 2015-12-14 | 2016-08-24 | 乐视云计算有限公司 | Deinterlace method and device for interlaced video |
| CN105704415A (en) * | 2016-02-02 | 2016-06-22 | 福州瑞芯微电子股份有限公司 | Video de-interleaving method and device |
| CN105704415B (en) * | 2016-02-02 | 2018-10-02 | 福州瑞芯微电子股份有限公司 | A kind of video deinterleaves treating method and apparatus |
| CN113261276A (en) * | 2019-01-09 | 2021-08-13 | 西安诺瓦星云科技股份有限公司 | Deinterlacing interpolation method, device and system, video processing method and storage medium |
| CN113261276B (en) * | 2019-01-09 | 2023-08-22 | 西安诺瓦星云科技股份有限公司 | De-interlacing interpolation method, de-interlacing interpolation device, de-interlacing interpolation system, video processing method and storage medium |
| CN113852830A (en) * | 2021-09-23 | 2021-12-28 | 杭州国芯科技股份有限公司 | Median filtering video de-interlacing method |
| CN113852830B (en) * | 2021-09-23 | 2023-12-29 | 杭州国芯科技股份有限公司 | Median filtering video de-interlacing method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104202555B (en) | 2017-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101803363B (en) | Method and apparatus for line-based motion estimation in video image data | |
| JP4303745B2 (en) | Image display apparatus and method, image processing apparatus and method | |
| US9148622B2 (en) | Halo reduction in frame-rate-conversion using hybrid bi-directional motion vectors for occlusion/disocclusion detection | |
| CN100379260C (en) | Noise measurement apparatus for image signal and method thereof | |
| EP0830018A2 (en) | Method and system for motion detection in a video image | |
| CN101663885A (en) | Image processing device and method, and image display device and method | |
| CN104202555A (en) | Method and device for deinterlacing | |
| CN101207707A (en) | System and method for advancing frame frequency based on motion compensation | |
| JP2008158162A (en) | Image display device and method, and image processor and method | |
| CN105282475A (en) | Mobile subtitle detection and compensation method and system | |
| CN102447870A (en) | Detection method for static objects and motion compensation device | |
| JP4181614B2 (en) | Image display apparatus and method, image processing apparatus and method | |
| CN101232568B (en) | Method for image deinterlacing and related device | |
| JP2009159298A (en) | Video signal processing apparatus and video signal processing method | |
| CN101282415A (en) | Motion adaptive image processing | |
| KR100910208B1 (en) | Apparatus and method for deinterlacing | |
| US8917354B2 (en) | Motion detection in video fields | |
| CN102497492B (en) | Detection method for subtitle moving in screen | |
| KR100594780B1 (en) | Apparatus for converting image and method the same | |
| KR101069712B1 (en) | A Method and Apparatus for Intra Field Scan-Line Interpolation Using Weighted Median of Edge Direction | |
| KR100936916B1 (en) | Apparatus and method for deinterlacing | |
| US9324131B1 (en) | Method and apparatus for motion adaptive deinterlacing with reduced artifacts | |
| KR101361649B1 (en) | Video signal processor | |
| KR100252943B1 (en) | scan converter | |
| KR100284697B1 (en) | Line interpolation apparatus and method using line memory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220125 Address after: Rooms 1306-1309, 13 / F, 19 science Avenue West, Hong Kong Science Park, Shatin, New Territories, China Patentee after: BUILDWIN INTERNATIONAL (ZHUHAI) LTD. Address before: 519015 3rd Floor, Stereo Science and Technology Building, 184 Bailian Road, Jida, Zhuhai City, Guangdong Province Patentee before: BUILDWIN INTERNATIONAL (ZHUHAI) Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220317 Address after: Rooms 1306-1309, 13 / F, 19 science Avenue West, Hong Kong Science Park, Shatin, New Territories, China Patentee after: BUILDWIN INTERNATIONAL (ZHUHAI) LTD. Address before: 519015 3rd Floor, Stereo Science and Technology Building, 184 Bailian Road, Jida, Zhuhai City, Guangdong Province Patentee before: BUILDWIN INTERNATIONAL (ZHUHAI) Ltd. |