CN102395019A - Deinterlacing method - Google Patents
Deinterlacing method Download PDFInfo
- Publication number
- CN102395019A CN102395019A CN201110375109XA CN201110375109A CN102395019A CN 102395019 A CN102395019 A CN 102395019A CN 201110375109X A CN201110375109X A CN 201110375109XA CN 201110375109 A CN201110375109 A CN 201110375109A CN 102395019 A CN102395019 A CN 102395019A
- Authority
- CN
- China
- Prior art keywords
- pixel
- interpolation
- boundary
- information
- pixels
- 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
Images
Landscapes
- Television Systems (AREA)
Abstract
A deinterlacing method is characterized by: through reliably detecting each direction of a pixel point to be interpolated, determining possibility (reliability) of regarding the each direction as a boundary direction of the pixel point to be interpolated; selecting the direction with the highest reliability as the boundary direction of the pixel point to be interpolated; performing vertical direction interpolation and the interpolation based on the boundary direction to the pixel point to be interpolated, wherein an interpolation result is related to the reliability. Therefore, the interpolation result is accurate and a better deinterlacing effect can be obtained.
Description
Technical field
The present invention relates to technical field of image processing, particularly interlace-removing method.
Background technology
In order to reduce the bandwidth of vision signal, and make circuit be easier to design, the traditional tv system, for example PAL/NTSC/SECAM mostly adopts horizontal-interlace technique.In simple terms, horizontal-interlace technique is come display reproduction with interleaved mode exactly.Under horizontal-interlace technique, the vertical resolution of image is lower, and will produce such as visible interference such as edge flares inevitably.And when screen was big more, this interference was more for remarkable.Along with the popularization of development of science and technology and DTV, modern video equipment changes the high definition epoch that show line by line gradually over to.Play old video content in modern times on the display device, just need the vision signal of interlacing be converted into vision signal line by line, promptly go interlacing to handle.Need between two row of interlace signal, estimate the one-row pixels point, or between adjacent two pixels of vertical direction of interlace signal, estimate a new pixel, i.e. interpolation pixel.
Deinterlacing technique can be divided into two big types of 2D and 3D, and 2D technology requisite part in the 3D technology often.When having comprised motion in the image sequence, introduce reliable 2D technology and can prevent distortion and guarantee certain picture quality.The main challenge of 2D technology is to recover the border texture details of image how as far as possible, and the lowland reduces the loss of image resolution ratio as far as possible.
In existing interlaced field in the method for reconstructing of disappearance row, the most directly method be with the mean value of neighbouring 2 points (or multiple spot) of current calculation level or weighted average serve as the value that will calculate, comprise bilinear interpolation, bicubic interpolation method etc.The subject matter that this type of interpolation method faced is to have serious distortion phenomenon in the image of being rebuild usually, causes in the result images object boundary place sawtooth to occur and flicker appears in motion sequence when playing.In addition, prior art has also proposed the interpolation method based on the border, though the border interpolation method can provide higher resolution, the border judgement also erroneous judgement occurs through regular meeting, and wrong boundary information will greatly damage quality of reconstructed images.
Summary of the invention
The problem that the present invention solves provides a kind of interlace-removing method, goes the interlacing effect preferably to provide, and avoids image fault.
For addressing the above problem, the present invention provides a kind of interlace-removing method, comprising:
For the interpolation pixel, obtain the existing pixel information in its neighborhood;
According to the existing pixel information in the said neighborhood, obtain boundary direction detection information through all directions of said interpolation pixel;
Said boundary direction detection information is carried out reliability detect, obtain the corresponding reliable degree, said reliability reflects that current boundary direction detects the possibility of the corresponding direction of information as the boundary direction of interpolation pixel;
Choose the highest boundary direction of reliability and detect the boundary direction of the corresponding direction of information as the interpolation pixel;
Said interpolation pixel is carried out vertical direction interpolation and based on the boundary direction interpolation of said interpolation pixel, said vertical direction interpolation result and said boundary direction interpolation result carried out the pixel value that weighting summation obtains said interpolation pixel; The weight of said boundary direction interpolation result is relevant with the reliability of the boundary direction of said interpolation pixel, and the weight sum of said vertical direction interpolation result and said boundary direction interpolation result is 1; Said vertical direction is the vertical direction through said interpolation pixel.
Compared with prior art; Above-mentioned interlace-removing method has the following advantages: through the reliability through all directions of interpolation pixel is detected; Confirm the possibility (reliability) of all directions, and therefrom select the boundary direction of the highest direction of reliability, the interpolation of said interpolation pixel is combined the interpolation of vertical direction and based on the interpolation of boundary direction as the interpolation pixel as the boundary direction of said interpolation pixel; And it is all relevant with said reliability; Therefore, interpolation result is more accurate, has obtained to go preferably the interlacing effect.
Description of drawings
Fig. 1 is a kind of execution mode schematic flow sheet of interlace-removing method of the present invention;
Fig. 2 obtains a kind of mode sketch map that boundary direction detects information in the interlace-removing method of the present invention;
Fig. 3 obtains the another kind of mode sketch map that boundary direction detects information in the interlace-removing method of the present invention;
Fig. 4 is a kind of mode sketch map of consistency detection in the interlace-removing method of the present invention;
Fig. 5 is the another kind of mode sketch map of consistency detection in the interlace-removing method of the present invention;
Fig. 6 is a kind of mode sketch map of correlation detection in the interlace-removing method of the present invention;
Fig. 7 is the another kind of mode sketch map of correlation detection in the interlace-removing method of the present invention.
Embodiment
As previously mentioned, prior art when going interlacing to handle, when the interpolation method that adopts based on the border, can be because the border error in judgement obtain the boundary information of mistake.Therefore, the interpolation result that is obtained is also not accurate enough, influences picture quality.Present invention is directed at this, proposed a kind of new interlace-removing method, through the reliability of boundary direction is related with interpolation result, to obtain to go preferably the interlacing effect.
Fig. 1 shows a kind of execution mode of interlace-removing method of the present invention, and it comprises:
Execution in step s1 for the interpolation pixel, obtains the existing pixel information in its neighborhood;
Execution in step s2 according to the existing pixel information in the said neighborhood, obtains the boundary direction detection information through all directions of said interpolation pixel;
Execution in step s3 carries out reliability to said boundary direction detection information and detects, and obtains the corresponding reliable degree, and said reliability reflects that current boundary direction detects the possibility of the corresponding direction of information as the boundary direction of interpolation pixel;
Execution in step s4 chooses the highest boundary direction of reliability and detects the boundary direction of the corresponding direction of information as the interpolation pixel;
Execution in step s5; Said interpolation pixel is carried out vertical direction interpolation and based on the boundary direction interpolation of said interpolation pixel, said vertical direction interpolation result and said boundary direction interpolation result carried out the pixel value that weighting summation obtains said interpolation pixel; The weight of said boundary direction interpolation result is relevant with the reliability of the boundary direction of said interpolation pixel, and the weight sum of said vertical direction interpolation result and said boundary direction interpolation result is 1; Said vertical direction is the vertical direction through said interpolation pixel.
In the above-mentioned execution mode, through to the Boundary Detection through the direction of interpolation pixel, obtain its corresponding reliable degree information, said reliability reflects that current boundary direction detects the possibility of the corresponding direction of information as the boundary direction of interpolation pixel.After obtaining said reliability information, therefrom choose the boundary direction of the direction of the highest process interpolation pixel of reliability as the interpolation pixel.And, merging with the interpolation of interpolation pixel vertical direction with based on the interpolation of the boundary direction of said interpolation pixel, the reliability with said boundary direction during fusion is confirmed the interpolation proportion in fusion based on boundary direction.Thereby the result after the interpolation can reflect the border condition of interpolation pixel more realistically, and interpolation result is more accurate, has obtained to go preferably the interlacing effect.
Below be elaborated through concrete instance each step to above-mentioned execution mode.
With reference to shown in Figure 2, go the interlaced video signal before interlacing is handled to comprise multirow pixel (representing) information with round dot, for example i-1 is capable, i is capable and i+1 is capable etc.Suppose; Interpolation pixel when going interlacing to handle (representing) with asterisk point at i-1 capable and i capable between; Then i-1 is capable and i the is capable pixel can be thought and is in the interpolation pixel neighborhood of a point, and the pixel that certain i+1 is capable also can be thought and is in the interpolation pixel neighborhood of a point.Continue supposition, said interpolation pixel is positioned at its j that is expert at row, is positioned at then that (i-1 is j) with (i, j) the existing pixel of two positions is two adjacent in vertical direction pixels of said interpolation pixel.And i-1 is capable and i other pixels on capable should be through on a plurality of incline directions of said interpolation pixel with said interpolation pixel adjacent pixels point.Definition is that benchmark is tilted to the left and is a kind of dominant direction with said vertical direction, and is that benchmark is tilted to the right and is another kind of dominant direction with said vertical direction.Then, for a certain dominant direction, wherein comprised the incline direction of the said interpolation pixel of a plurality of processes.
For example; Line through asterisk point among Fig. 2 has been represented wherein a kind of incline direction that is tilted to the right; Existing pixel (i-1 on the i-1 of said incline direction process is capable; J+4) and the existing pixel of i on capable (i j-4) is said interpolation pixel adjacent two pixels on said incline direction.The offset identical (horizontal offset is 4) of two pixels that said i-1 two pixels capable and on i is capable are adjacent in vertical direction with said interpolation pixel, and offset (skew left, a skew to the right) in the opposite direction.
Then, for the boundary direction that obtains the said incline direction that is tilted to the right detects information, need to confirm earlier the corresponding calculated window.For example, can with (i-1, j+4) with (i, j-4) two existing pixel horizontal-shifts are less than or equal to 2 existing pixel and include in the calculation window, then the width of said calculation window is 2, highly is 1.Particularly, in said calculation window, also comprise: (i-1 is j+6) with (i, j-2) two locational existing pixels; (i-1 is j+5) with (i, j-3) two locational existing pixels; (i-1 is j+3) with (i, j-5) two locational existing pixels; (i-1 is j+2) with (i, j-6) two locational existing pixels.Can find out that the incline direction that the existing pixel of other in the said calculation window is confirmed is also with (i-1 is j+4) with (i, j-4) incline direction confirmed of two existing pixels is identical.Need to prove that the confirming of the calculation window here is merely for example, is not in order to limit.In general, the size of calculation window is relevant with incline direction, and is big more with respect to the inclination angle of vertical direction, and calculation window also can be corresponding big more.For example, by (i-1 is j+1) with (i, the calculation window that j-1) incline direction confirmed of two locational existing pixels is corresponding is just less relatively.Certainly, the size of the calculation window of each incline direction also can be identical, can regulate according to actual needs.
Calculate the difference between a plurality of existing pixels of said calculation window respectively, i.e. (i-1 is j+6) with (i in calculating respectively; J-2) difference between two locational existing pixels, (i-1 is j+5) with (i, j-3) the difference, (i-1 between two locational existing pixels; J+3) and (i; J-5) difference between two locational existing pixels, (i-1 is j+2) with (i, j-6) difference and the (i-1 between two locational existing pixels; J+4) and (i, j-4) difference between two locational existing pixels.
The absolute value of the above-mentioned difference that obtains is added up, obtain the difference degree between existing pixel on the corresponding said incline direction of said calculation window, with the boundary direction detection information of said difference degree as the corresponding incline direction of said calculation window.The said computational process that adds up is by following formulate:
Wherein, n is the mark of said incline direction, C
nFor existing pixel adjacent on the said incline direction with the interpolation pixel with respect to the horizontal offset of neighbor pixel on the interpolation pixel vertical direction (big more with respect to the inclination angle of vertical direction usually, C
nBig more), k is the vertical position in the corresponding calculation window of said incline direction, and l is the interior horizontal level of calculation window of corresponding said incline direction, and kdifw is the relevant weight coefficient of the corresponding calculation window of said incline direction, f
w(i-1+k, j+C
n+ l) and f
w(i-1+k, j+C
n+ l) representative needs the pixel value of the existing pixel of calculated difference respectively.For by (i-1 is j+4) with (i, the j-4) incline directions confirmed of two locational existing pixels, n=4, C
n=3, height_n=1, width_n=2.
In other implementations; Also can be with adding up after the above-mentioned squared difference that obtains; Obtain the difference degree between existing pixel on the corresponding incline direction of said calculation window, with the boundary direction detection information of said difference degree as the corresponding incline direction of said calculation window.
The account form that still adds up with absolute value is an example, and the acquisition process that detects information for the boundary direction of the incline direction that is tilted to the left also can carry out with reference to aforesaid way, and the computing formula that respective absolute values adds up is following:
Wherein each parameter can be with reference to the corresponding parameter in the formula (1).
The scope of choosing when more than confirming calculation window is the existing pixel of the capable and i of i-1 on capable, in other implementations, also can be with the pixel in the i-1 is capable and i the is capable adjacent lines reference information during also as the said difference degree of acquisition.
Continuation is with reference to Fig. 2, because (i is j+4) with (i+1; J-4) incline direction confirmed of two locational existing pixels with (i-1, j+4) with (i, j-4) incline direction confirmed of two existing pixels is identical; At definite (i-1; J+4) and (when i, calculation window that j-4) incline direction confirmed of two existing pixels is corresponding, also can confirm (an i according to appearance; J+4) and (i+1, the calculation window that j-4) incline direction confirmed of two locational existing pixels is corresponding (for example above-mentioned width is 2, highly be 1 calculation window).Subsequently, the difference between each the existing pixel in two calculation windows is carried out that difference absolute value adds up (can with reference to formula (1) and formula (2)) or difference square adds up, detect information with final acquisition boundary direction.
With reference to Fig. 3, because (i-1 is j+4) with (i-2; J+12) incline direction confirmed of two locational existing pixels be in that (i-1 is j+4) with (i is j-4) on the extended line of the incline direction confirmed of two existing pixels; At definite (i-1; J+4) and (when i, calculation window that j-4) incline direction confirmed of two existing pixels is corresponding, also can confirm (an i-1 according to appearance; J+4) and (i-2, the calculation window that j+12) incline direction confirmed of two locational existing pixels is corresponding (for example above-mentioned width is 2, highly be 1 calculation window).Subsequently, the difference between each the existing pixel in two calculation windows is carried out difference absolute value add up, said cumulative process can be by following formulate:
Wherein, shift comprises central point on the incline direction of said extended line with respect to the horizontal offset of interpolation pixel, and other parameters can be with reference to formula (1).For (i, j-4), (i-1, j+4), (i-2; J+12) on the incline direction that comprises said extended line that three locational existing pixels are confirmed; (i-1, j+4) this locational pixel is said central point, its horizontal offset with respect to the interpolation pixel is 4.
With reference to formula (2) and formula (3) also can push away the absolute value accumulation calculating formula of the incline direction that is tilted to the left.
Certainly; With preceding similar; Also can obtain the difference degree between existing pixel on the corresponding incline direction of said calculation window, with the boundary direction detection information of said difference degree as the corresponding incline direction of said calculation window with adding up after the above-mentioned squared difference that obtains.
After obtaining boundary direction detection information, detect to obtain the corresponding reliable degree through reliability, said reliability reflects that current boundary direction detects the possibility of the corresponding direction of information as the boundary direction of interpolation pixel.
In order to ensure the accuracy of the boundary direction of the interpolation pixel that finally obtains, reduce the influence of disturbing factors such as noise, after obtaining boundary direction detection information, can carry out consistency detection.The purpose of consistency detection is in order to obtain the difference degree at pixel on current direction.And in interpolation pixel neighborhood of a point, the closer to current direction to be detected (through the incline direction of interpolation pixel), correlation is good more on the locus, but noise resisting ability is relatively poor, and institute's amount of information that provides is few more; On the contrary, distance consistency check correlation far away more is poor, but has antinoise preferably with information provides value, should do the balance selection in the practical application.
With reference to shown in Figure 4, continue that the capable and i j between capable classifies example as and describes at i-1 with the interpolation pixel, the direction that need carry out consistency detection is served as reasons, and (i-1 is j+4) with (i, j-4) two incline directions that existing pixel is confirmed.Then, detect on the identical direction of the corresponding direction of information the difference between pixel (consistency detection value) in twos through obtaining in the said neighborhood with said boundary direction, and with said boundary direction detection information comparison, obtain the consistency detection result.
Said pixel in twos can be the existing pixel in the said neighborhood; For example among Fig. 4 by (i, j) with (i+1, j-8) incline direction confirmed of two locational existing pixels with by (i-1; J+4) and (i, j-4) incline direction confirmed of two existing pixels is identical.Then, can obtain that (i is j) with (i+1, the j-8) difference between two locational existing pixels, and with (i-1 is j+4) with (i, the boundary direction detection information that j-4) incline direction confirmed of two existing pixels is corresponding relatively obtain the consistency detection result.That is the difference that, said difference and said boundary direction is detected information is as the consistency detection result.
Perhaps, said pixel in twos also can be other interpolation pixels in existing pixel and the said neighborhood in the said neighborhood, and the pixel value of said other interpolation pixels obtains through existing pixel interpolation around it.For example among Fig. 4 by (i-1, the incline direction that existing pixel on j-4) and the interpolation pixel that i-1 is capable and the j-8 of i between capable lists are confirmed with by (i-1 is j+4) with (i, j-4) incline direction confirmed of two existing pixels is identical.Wherein, the interpolation pixel that the j-8 between the capable and i of i-1 is capable lists can be according to existing pixel information on every side, and through bilinear interpolation, existing interpolation method such as bicubic interpolation method obtains.For example, can be according to (i-1 is j-8) with (i, j-8) two existing pixel interpolation obtain, and promptly obtain through its vertical direction interpolation.Then; Can obtain (i-1; Difference between existing pixel j-4) and i-1 is capable and the j-8 of i between capable lists interpolation pixel, and with (i-1 is j+4) with (i; J-4) two definite corresponding boundary direction detection information of incline direction of existing pixel compare, and obtain the consistency detection result.That is the difference that, said difference and said boundary direction is detected information is as the consistency detection result.Difference between said existing pixel and said interpolation pixel can be calculated by following formula and obtain:
chkdifL
n＝abs(f
w(i-1，j)-(f
w(i-1，j+c
n)+f
w(i，j+c
n))/2) (4)
Wherein each parameter can be with reference to the corresponding parameter in the formula (1).
Similarly; Also can choose (i+1; Difference between existing pixel j+4) and i is capable and the j+8 of i+1 between capable lists interpolation pixel and said boundary direction detection information relatively obtain the consistency detection result, and its detailed calculated is repeated no more.
Perhaps, said pixel in twos can also be other interpolation pixels in the said neighborhood, and the pixel value of said other interpolation pixels obtains through existing pixel interpolation around it.For example; The incline direction that interpolation pixel that the capable and i j+8 between capable lists by i-1 and the interpolation pixel that i is capable and the j of i+1 between capable lists are confirmed with by (i-1; J+4) and (i, j-4) incline direction confirmed of two existing pixels is identical.Wherein, said two interpolation pixels can be according to having pixel information on every side, and through bilinear interpolation, existing interpolation method such as bicubic interpolation method obtains.For example, the interpolation pixel that the j between the capable and i+1 of i is capable lists can be according to (i is j) with (i+1, j) two existing pixel interpolation obtain, and promptly obtain through its vertical direction interpolation.And the interpolation pixel that i-1 is capable and the j+8 of i between capable lists can be according to (i-1 is j+8) with (i, j+8) two existing pixel interpolation obtain, and promptly obtain through its vertical direction interpolation.Subsequently, with the difference between said two interpolation pixels with (i-1 is j+4) with (i, the boundary direction detection information that j-4) incline direction confirmed of two existing pixels is corresponding relatively obtain the consistency detection result.That is the difference that, said difference and said boundary direction is detected information is as the consistency detection result.
Above-mentioned difference between pixel in twos all is in interpolation pixel neighborhood of a point, to choose one group of difference to compare; In other implementations; Can also choose many group differences; Many groups difference is averaged, and mean value and said boundary direction detection information are compared, obtain the consistency detection result.For example, the difference between the difference between above-mentioned two existing pixels, existing pixel and other interpolation pixels, the difference between other interpolation pixels are asked on average, said mean value and said boundary direction detection information are compared.Certainly, also can from said many group differences, choose one group of maximum difference, with one group of difference of said maximum and said boundary direction detection information relatively, acquisition consistency detection result.The consistency detection result who is obtained is more little, representes that current direction to be detected is big more as the possibility of the boundary direction of interpolation pixel.
Fig. 5 shows the another kind of mode sketch map of consistency detection; Its consistency detection corresponding with Fig. 4 is similar; Can be through choosing one group of difference between pixel, the difference of choosing between many group pixels is averaged, and chooses one group of difference of maximum of the difference between many group pixels; Compare with said boundary direction detection information, obtain the consistency detection result.Its consistency detection corresponding with Fig. 4 difference is to choose the mode of pixel in twos.For example, pixel is chosen along (i-1 is j+4) with (i, j-4) two existing pixels on the incline direction confirmed of two existing pixels: (i-1 is j+4) with (i-2, j+12) locational two existing pixels in twos.
Because also there are situation such as intersect edges and angle in real image in handling between pixel, be the accuracy that further increases the reliability detection, also can the correlation of current direction to be detected be detected.Particularly; Choose along said boundary direction and detect the corresponding direction of information, two pixels adjacent with said interpolation pixel; Two adjacent in vertical direction with two selected pixels respectively pixels; And two pixels adjacent in vertical direction with said interpolation pixel, the line between selected a plurality of pixels and said interpolation pixel constitutes two right-angled triangles up and down of shared hypotenuse; Two triangles are up and down carried out correlation detection, obtain relevance detection results.
With reference to shown in Figure 6, with the capable pixel A of i-1
x, A
0, A
xThe pixel B capable with i
-xBetween line constitute to go up right-angled triangle, the said hypotenuse of going up right-angled triangle is through the interpolation pixel.With the capable pixel B of i
x, B
0, B
-xThe pixel A capable with i-1
xBetween line constitute right-angled triangle down, the said hypotenuse of right-angled triangle down passes through the interpolation pixel equally.Then, said correlation detection realizes through following formula:
adj
R＝max(abs(B
-x-avg)-fnt，0)+max(abs(A
x-avg)-fnt，0)
fnt＝min(abs(A
0-B
0)/2，fdelta)
fdelta＝max(0，C
0-C
1*min(delta
up，delta
down))
delta
up＝max(A
-x，A
0，A
x，B
-x)-min(A
-x，A
0，A
x，B
-x)
delta
down＝max(B
-x，B
0，B
x，A
x)-min(B
-x，B
0，B
x，A
x)
Wherein, A
-x, A
0, A
x, B
x, B
0, B
-xBe respectively the pixel value of institute's corresponding pixel points, avg is A
0, B
0The average pixel value of institute's corresponding pixel points, C
0For with the relevant threshold value of bit wide of current pending image, C
1For with delta
Up, delta
DownRelevant weight coefficient, C
1Span can be 0 to 511.
Correlation detection value adj
RMore little, show that direction to be detected is big more as the possibility of the boundary direction of interpolation pixel.
Fig. 7 shows the another kind of mode sketch map of correlation detection, and its correlation detection corresponding with Fig. 6 is similar, and difference is that up and down leg-of-mutton formation is different.With reference to shown in Figure 7, with the capable pixel A of i-1
-x, A
0, A
xThe pixel B capable with i
xBetween line constitute to go up right-angled triangle, the said hypotenuse of going up right-angled triangle is through the interpolation pixel.With the capable pixel B of i
x, B
0, B
-xThe pixel A capable with i-1
-xBetween line constitute right-angled triangle down, the said hypotenuse of right-angled triangle down passes through the interpolation pixel equally.Then, said correlation detection realizes through following formula:
adj
L＝max(abs(A
-x-avg)-fnt，0)+max(abs(B
x-avg)-fnt，0)
fnt＝min(abs(A
0-B
n)/2，fdelta)
fdelta＝max(0，C
0-C
1*min(delta
up，delta
down))
delta
up＝max(A
x，A
0，A
-x，B
x)-min(A
x，A
0，A
-x，B
x)
delta
down＝max(B
-x，B
0，B
x，A
-x)-min(B
-x，B
0，B
x，A
-x)
Wherein, A
-x, A
0, A
x, B
x, B
0, B
-xBe respectively the pixel value of institute's corresponding pixel points, avg is A
0, B
0The average pixel value of institute's corresponding pixel points, C
0For with the relevant threshold value of figure place of current pending image, C
1For with delta
Up, delta
DownRelevant weight coefficient, C
1Span can be 0 to 511.
Correlation detection value adj
LMore little, show that direction to be detected is big more as the possibility of the boundary direction of interpolation pixel
For obtaining more comprehensive reliability testing result; Make reliability can reflect the possibility of direction to be detected more realistically as the boundary direction of the pixel of interpolation; Can also obtain the gradient testing result according to said consistency detection value and said correlation detection value and the corresponding direction of said boundary direction detection information.
Said gradient detects and comprises:
Compare boundary direction detection information with respect to the incline direction that is tilted to the left and is tilted to the right of the inclination angle of vertical direction identical (being that position offset is identical);
If the boundary direction of the said incline direction that is tilted to the left detects the boundary direction detection information of information more than or equal to the said incline direction that is tilted to the right, then obtain the corresponding gradient detected value of the said incline direction that is tilted to the left according to corresponding consistency detection value and correlation detection value;
Otherwise, then obtain the corresponding gradient detected value of the said incline direction that is tilted to the right according to corresponding consistency detection value and correlation detection value.
For the corresponding gradient detected value of the incline direction that is tilted to the left, it is calculated by following formula and obtains:
grad
n＝kwd[n]*[djfL
n-kneg*difR
n]-kwc[n]*chkdifR
n-kc*adj
R
For the corresponding gradient detected value of the incline direction that is tilted to the right, it is calculated by following formula and obtains:
grad
n＝kwd[n]*[difR
n-kneg*difL
n]-kwc[n]*kchk*chkdifL
n-kc*adj
L
Wherein kwd [n] be on each incline direction with the relevant weight coefficient of boundary direction detection information, Kneg is a gradient factor, kwc [n] is the relevant weight coefficient of consistency detection on each incline direction, Kc is the relevant weight coefficient of correlation detection.Above-mentioned coefficient all can be regulated according to practical application, repeats no more here.
The grad that is calculated
nBig more this incline direction n that indicating is that the possibility of interpolation pixel boundary direction is big more, through certain searching algorithm and tactful maximizing grad
n, then its pairing direction can be thought the boundary direction of current interpolation pixel.In search procedure, can limit the scope of being searched for through certain experience knowledge, promptly can improve the speed of search, also can reject impossible boundary direction, make resulting direction more near real boundary direction.For example can through calculate current point laterally with vertical on the pixel value distributed degrees, perhaps calculate around the current point methods such as simple sobel operator gradient direction, estimate the border and possibly direction confirm the maximum value search scope.
In addition, in order further to improve the accuracy of the boundary direction that is obtained, but the continuity of the adjacent direction of the said incline direction of integrated survey also.Said continuity is meant, supposes that the border that on incline direction n, shows is the border that is tilted to the left, and then requiring on incline direction n-1 and the incline direction n+1 at least, one of which or 2 all is the border that is tilted to the left.Particularly, realize through the continuity detected value that compares each incline direction.The continuity of said incline direction detects by following formula and calculates acquisition:
positive
n＝sign(difL
n-difR
n)
At difL
n-difR
nValue greater than 0 o'clock, positive
n=1; At difL
n-difR
nValue equal at 0 o'clock, positive
n=0; At difL
n-difR
nValue less than 0 o'clock, positive
n=-1.
Then this continuity is defined as:
positive
n＝positive
n-1or/and?positive
n＝positive
n+1
Be positive
nAs long as and positive
N-1, positive
N+1In any one equate, just can think that adjacent direction meets continuity, certainly as if positive
nWith positive
N-1, positive
N+1Equate that all then adjacent direction also meets continuity.
From each incline direction to be detected, choose the Grad maximum and meet the boundary direction of successional incline direction as the interpolation pixel.Then, calculate by following formula based on the interpolation of boundary direction:
For the boundary direction of the incline direction that is tilted to the right as the interpolation pixel
value
max＝(f
w(i，j+C
n)+f
w(i+1，j-C
n)+1)/2
For the boundary direction of the incline direction that is tilted to the left as the interpolation pixel
value
max＝(f
w(i，j-C
n)+f
w(i+1，j+C
n)+1)/2
Wherein, C
nBe the horizontal offset of existing pixel adjacent on the said incline direction with respect to neighbor pixel on the interpolation pixel vertical direction, f with the interpolation pixel
w(i, j+C
n), f
w(i+1, j-C
n) or f
w(i, j-C
n), f
w(i+1, j+C
n) pixel value of used existing pixel when representing along right bank boundary direction or left bank boundary direction interpolation respectively.
For preventing that possible false border from causing the image fault phenomenon, also in final interpolation calculation, introduce protection mechanism, with the mean value value of two adjacent on the interpolation pixel vertical direction existing pixels
AvgAs the border do not exist or the undesirable situation of testing result under interpolation, and in other cases, then regulate this mean value and scale based on the interpolation of boundary direction according to the gradient of boundary direction, maybe with what avoid that distortion appears in image.The Grad that Boundary Detection result's quality can directly be passed through to be calculated obtains, and this Grad is big more, and pairing border reliability is good more, and vice versa.Size through Grad obtains integration percentage calculating as follows:
Wherein, ThC is integration percentage check threshold value to gradient when calculating, and ThA is the associated weight coefficient of integration percentage when calculating, and the span of 2d_alpha is [0,1].
Usually, ThA and ThC all can calculate through test of many times, and adjust according to the image effect after the interpolation, to obtain comparatively suitable value.
The interpolation result of the interpolation pixel that finally obtains is following:
Dresult＝value
avg*(1-2d_alpha)+value
max*2d_alpha
In sum; Interlace-removing method of the present invention passes through the reliability of all directions of process interpolation pixel is detected; Confirm the possibility (reliability) of all directions, and therefrom select the boundary direction of the highest direction of reliability, the interpolation of said interpolation pixel is combined the interpolation of vertical direction and based on the interpolation of boundary direction as the interpolation pixel as the boundary direction of said interpolation pixel; And it is all relevant with said reliability; Therefore, interpolation result is more accurate, has obtained to go preferably the interlacing effect
More than disclose many aspects of the present invention and execution mode, it will be understood by those skilled in the art that others of the present invention and execution mode.Disclosed many aspects and execution mode just are used to illustrate among the present invention, are not to be to qualification of the present invention, and real protection range of the present invention and spirit should be as the criterion with claims.
Claims (12)
1. an interlace-removing method is characterized in that, comprising:
For the interpolation pixel, obtain the existing pixel information in its neighborhood;
According to the existing pixel information in the said neighborhood, obtain boundary direction detection information through all directions of said interpolation pixel;
Said boundary direction detection information is carried out reliability detect, obtain the corresponding reliable degree, said reliability reflects that current boundary direction detects the possibility of the corresponding direction of information as the boundary direction of interpolation pixel;
Choose the highest boundary direction of reliability and detect the boundary direction of the corresponding direction of information as the interpolation pixel;
Said interpolation pixel is carried out vertical direction interpolation and based on the boundary direction interpolation of said interpolation pixel, said vertical direction interpolation result and said boundary direction interpolation result carried out the pixel value that weighting summation obtains said interpolation pixel; The weight of said boundary direction interpolation result is relevant with the reliability of the boundary direction of said interpolation pixel, and the weight sum of said vertical direction interpolation result and said boundary direction interpolation result is 1; Said vertical direction is the vertical direction through said interpolation pixel.
2. interlace-removing method as claimed in claim 1 is characterized in that, according to the existing pixel information in the said neighborhood, the corresponding boundary direction detection information of all directions of obtaining the said interpolation pixel of process comprises:
Choose a plurality of incline directions as the dominant direction that detects, said incline direction is confirmed calculation window according to each said incline direction through said interpolation pixel in said neighborhood;
In the corresponding calculation window of a certain incline direction, calculate respectively and the difference of said interpolation pixel between the adjacent existing pixel of said incline direction;
According to the said difference that obtains in the said calculation window, obtain the difference degree between existing pixel on the corresponding incline direction of said calculation window, with the boundary direction detection information of said difference degree as the corresponding incline direction of said calculation window.
3. interlace-removing method as claimed in claim 2 is characterized in that, said dominant direction comprises: with said vertical direction is that benchmark is tilted to the left, and is that benchmark is tilted to the right with said vertical direction;
Choosing said incline direction comprises: two neighbor pixels choosing respectively with the above interpolation pixel of said vertical direction are positioned at two pixels with delegation; Two selected pixels are identical with respect to the position offset of two neighbor pixels of the above interpolation pixel of said vertical direction, and offset is in the opposite direction; Confirm that along the direction of two selected pixels be said incline direction.
4. interlace-removing method as claimed in claim 2; It is characterized in that; According to the said difference that obtains in the said calculation window; Obtain that the difference degree between existing pixel comprises on the corresponding incline direction of said calculation window: the absolute value of each difference that obtains in the calculation window is carried out weighted accumulation obtain the difference degree between existing pixel on the corresponding incline direction of said calculation window, the weight coefficient that uses in the weighted accumulation is big or small relevant with calculation window.
5. interlace-removing method as claimed in claim 2; It is characterized in that; According to the said difference that obtains in the said calculation window; The difference degree that obtains to have between pixel on the corresponding incline direction of said calculation window comprises: obtain the difference degree between existing pixel on the corresponding incline direction of said calculation window with carrying out weighted accumulation after each squared difference that obtains in the calculation window, the weight coefficient that uses in the weighted accumulation is relevant with the calculation window size.
6. interlace-removing method as claimed in claim 1 is characterized in that, said boundary direction detection information is carried out the reliability detection comprise:
Obtain in the said neighborhood on the direction identical difference between pixel in twos with the corresponding direction of said boundary direction detection information; And with said boundary direction detection information relatively; Obtain the consistency detection result, confirm that according to said consistency detection result said boundary direction detects the reliability of the corresponding direction of information.
7. interlace-removing method as claimed in claim 6 is characterized in that, said pixel in twos is the existing pixel in the said neighborhood; Perhaps said pixel in twos is existing pixel and other interior interpolation pixels of said neighborhood in the said neighborhood, and the pixel value of said other interpolation pixels obtains through existing pixel interpolation around it; Perhaps said pixel in twos is other interpolation pixels in the said neighborhood, and the pixel value of said other interpolation pixels obtains through existing pixel interpolation around it.
8. interlace-removing method as claimed in claim 6 is characterized in that, obtains one group of said difference between pixel in twos, and compares with said boundary direction detection information, obtains the consistency detection result; Perhaps, obtain the said difference between pixel in twos of many groups, many groups difference is averaged, said mean value and said boundary direction detection information are compared, obtain the consistency detection result; Perhaps, obtain the said differences between pixel in twos of many groups, get one group of wherein maximum difference, with one group of difference of said maximum and said boundary direction detection information relatively, acquisition consistency detection result.
9. interlace-removing method as claimed in claim 1 is characterized in that, said boundary direction detection information is carried out the reliability detection comprise:
Choose along said boundary direction and detect the corresponding direction of information, two pixels adjacent with said interpolation pixel; Two adjacent in vertical direction with two selected pixels respectively pixels; And two pixels adjacent in vertical direction with said interpolation pixel, the line between selected a plurality of pixels and said interpolation pixel constitutes two right-angled triangles up and down of shared hypotenuse;
Two triangles are up and down carried out correlation detection, obtain relevance detection results, confirm that according to said relevance detection results said boundary direction detects the reliability of the corresponding direction of information.
10. interlace-removing method as claimed in claim 1 is characterized in that, said boundary direction detection information is carried out the reliability detection comprise:
Obtain in the said neighborhood with said boundary direction and detect on the identical direction of the corresponding direction of information the difference between pixel in twos, and with said boundary direction detection information comparison, acquisition consistency detection result;
Choose along said boundary direction and detect the corresponding direction of information, two pixels adjacent with said interpolation pixel; Two adjacent in vertical direction with two selected pixels respectively pixels; And two pixels adjacent in vertical direction with said interpolation pixel, the line between selected a plurality of pixels and said interpolation pixel constitutes two right-angled triangles up and down of shared hypotenuse; Two triangles are up and down carried out correlation detection, obtain relevance detection results;
According to said consistency detection result and said relevance detection results and said boundary direction detection information, obtain the gradient testing result;
Confirm that according to said gradient testing result said boundary direction detects the reliability of the corresponding direction of information.
11. interlace-removing method as claimed in claim 10; It is characterized in that; Said boundary direction detection information is carried out the reliability detection also to be comprised: after obtaining the gradient testing result; Detect the continuity that said boundary direction detects the adjacent direction of the corresponding direction of information, confirm that according to said gradient testing result and said continuity said boundary direction detects the reliability of the corresponding direction of information.
12. interlace-removing method as claimed in claim 10 is characterized in that, said gradient testing result obtains through following step:
The boundary direction of relatively more identical with respect to the inclination angle of the vertical direction incline direction that is tilted to the left and is tilted to the right detects information;
If the boundary direction of the said incline direction that is tilted to the left detects the boundary direction detection information of information more than or equal to the said incline direction that is tilted to the right, then obtain the corresponding gradient detected value of the said incline direction that is tilted to the left according to corresponding consistency detection result and relevance detection results;
Otherwise, then obtain the corresponding gradient detected value of the said incline direction that is tilted to the right according to corresponding consistency detection result and relevance detection results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110375109.XA CN102395019B (en) | 2011-11-22 | 2011-11-22 | Deinterlacing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110375109.XA CN102395019B (en) | 2011-11-22 | 2011-11-22 | Deinterlacing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102395019A true CN102395019A (en) | 2012-03-28 |
| CN102395019B CN102395019B (en) | 2014-05-21 |
Family
ID=45862218
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110375109.XA Active CN102395019B (en) | 2011-11-22 | 2011-11-22 | Deinterlacing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102395019B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103024332A (en) * | 2012-12-26 | 2013-04-03 | 电子科技大学 | Video de-interlacing method based on edge and motion detection |
| CN103647919B (en) * | 2013-12-24 | 2017-01-18 | 上海交通大学 | Edge-sensitive intra-field de-interlacing method |
| CN111885338A (en) * | 2020-07-31 | 2020-11-03 | 青岛信芯微电子科技股份有限公司 | Video de-interlacing processing method and device |
| WO2022057221A1 (en) * | 2020-09-17 | 2022-03-24 | 海信视像科技股份有限公司 | Receiving device and non-volatile storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11331773A (en) * | 1998-05-15 | 1999-11-30 | Victor Co Of Japan Ltd | Image interpolation method |
| CN1738404A (en) * | 2004-08-19 | 2006-02-22 | 瑞昱半导体股份有限公司 | Intra-field interpolation method and associated apparatus |
| CN101018286A (en) * | 2007-02-09 | 2007-08-15 | 天津大学 | De-interlacing method with the motive detection and self-adaptation weight filtering |
| US20080024658A1 (en) * | 2003-05-30 | 2008-01-31 | Samsung Electronics Co., Ltd. | Edge direction based image interpolation method |
| CN101197995A (en) * | 2006-12-07 | 2008-06-11 | 深圳艾科创新微电子有限公司 | Edge self-adapting de-interlacing interpolation method |
| US20080279479A1 (en) * | 2007-05-07 | 2008-11-13 | Mstar Semiconductor, Inc | Pixel interpolation apparatus and method thereof |
| CN101309376A (en) * | 2008-06-13 | 2008-11-19 | 北京中星微电子有限公司 | Method and device for eliminating alternate line |
| CN102025960A (en) * | 2010-12-07 | 2011-04-20 | 浙江大学 | Motion compensation de-interlacing method based on adaptive interpolation |
-
2011
- 2011-11-22 CN CN201110375109.XA patent/CN102395019B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11331773A (en) * | 1998-05-15 | 1999-11-30 | Victor Co Of Japan Ltd | Image interpolation method |
| US20080024658A1 (en) * | 2003-05-30 | 2008-01-31 | Samsung Electronics Co., Ltd. | Edge direction based image interpolation method |
| CN1738404A (en) * | 2004-08-19 | 2006-02-22 | 瑞昱半导体股份有限公司 | Intra-field interpolation method and associated apparatus |
| CN101197995A (en) * | 2006-12-07 | 2008-06-11 | 深圳艾科创新微电子有限公司 | Edge self-adapting de-interlacing interpolation method |
| CN101018286A (en) * | 2007-02-09 | 2007-08-15 | 天津大学 | De-interlacing method with the motive detection and self-adaptation weight filtering |
| US20080279479A1 (en) * | 2007-05-07 | 2008-11-13 | Mstar Semiconductor, Inc | Pixel interpolation apparatus and method thereof |
| CN101309376A (en) * | 2008-06-13 | 2008-11-19 | 北京中星微电子有限公司 | Method and device for eliminating alternate line |
| CN102025960A (en) * | 2010-12-07 | 2011-04-20 | 浙江大学 | Motion compensation de-interlacing method based on adaptive interpolation |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103024332A (en) * | 2012-12-26 | 2013-04-03 | 电子科技大学 | Video de-interlacing method based on edge and motion detection |
| CN103024332B (en) * | 2012-12-26 | 2015-05-27 | 电子科技大学 | Video de-interlacing method based on edge and motion detection |
| CN103647919B (en) * | 2013-12-24 | 2017-01-18 | 上海交通大学 | Edge-sensitive intra-field de-interlacing method |
| CN111885338A (en) * | 2020-07-31 | 2020-11-03 | 青岛信芯微电子科技股份有限公司 | Video de-interlacing processing method and device |
| WO2022057221A1 (en) * | 2020-09-17 | 2022-03-24 | 海信视像科技股份有限公司 | Receiving device and non-volatile storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102395019B (en) | 2014-05-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100394782C (en) | Image adaptive deinterlacing method and device based on edge | |
| JP4212657B2 (en) | Adaptive diagonal interpolation for image resolution enhancement | |
| CN101946514B (en) | Estimation of true motion vectors using an adaptive search range | |
| JP3836159B2 (en) | A system to convert interlaced video to progressive video using edge correlation | |
| JP4553481B2 (en) | Scanning line interpolation device | |
| US7292738B1 (en) | Method and system for dynamic angle interpolation in image processing | |
| US7414671B1 (en) | Systems and methods for display object edge detection and pixel data interpolation in video processing systems | |
| CN101309385B (en) | Alternate line eliminating process method based on motion detection | |
| CN101197995B (en) | Edge self-adapting de-interlacing interpolation method | |
| CN101163224A (en) | Super-resolution device and method | |
| CN102395019B (en) | Deinterlacing method | |
| CN103093458A (en) | Detecting method and detecting device for key frame | |
| US20060018564A1 (en) | Interpolation pixel generation circuit | |
| CN101309376B (en) | Method and device for eliminating alternate line | |
| US20080080790A1 (en) | Video signal processing apparatus and video signal processing method | |
| NL1030787C2 (en) | Judder detection apparatus determines whether detection pattern similar to judder pattern is actual judder based on blind pattern detection | |
| CN102170549A (en) | Edge correlation image intra-field de-interlacing algorithm of edge pre-judgment | |
| CN201222771Y (en) | High speed edge self-adapting de-interlaced interpolation device | |
| CN102404601A (en) | Method for detecting stereoscopic images | |
| TWI400688B (en) | Method and apparatus for an angular-directed spatial deinterlacer | |
| CN101001387B (en) | Method of deciding image local information change minimum direction using multiple decision | |
| KR100931110B1 (en) | Deinterlacing apparatus and method using fuzzy rule-based edge recovery algorithm | |
| CN104780295A (en) | Video denoising system based on noise correlation | |
| JP5449001B2 (en) | Video noise detector | |
| JPH11331773A (en) | Image interpolation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| ASS | Succession or assignment of patent right |
Owner name: SHANGHAI SHUNJIU ELECTRONIC TECHNOLOGY CO., LTD. Free format text: FORMER OWNER: HUAYA MICRO-ELECTRONIC (SHANGHAI) CO., LTD. Effective date: 20130307 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20130307 Address after: 201203, Shanghai 690 Zhangjiang Road, Pudong No. 5 Building No. 2 floor Applicant after: HUAYA MICROELECTRONICS (SHANGHAI) CO.,LTD. Address before: 201203, room 5, building 690, No. 201-202 blue wave road, Zhangjiang hi tech park, Shanghai, Pudong New Area Applicant before: Huaya Microelectronics (Shanghai) Co., Ltd. |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |