CN1668066A - Method for enhancing visual effect of image edge - Google Patents

Abstract

This invention relates to a method for strengthening image edge visual result including the following steps: a, carrying out a second order filter computation to a target pixel, b, judging the position of the target pixel at the image edge, c, computing the edge increased gain according to the filter compute result and its edge position, d, computing the strengthened pixel value and protecting the result.

Description

Strengthen the method for visual effect of image edge

Technical field

The invention belongs to technical field of image processing, particularly a kind of method that strengthens visual effect of image edge.

Background technology

Because human eye is responsive especially to the image border, the visual effect that therefore strengthens the image border can improve the definition of image, makes image clear more beautiful.Moreover in the transmission course of picture signal, image edge information belongs to HFS, is easy to decay, so just makes the image border that shows on display device become fuzzyyer, and visual effect is poorer.Therefore, the visual effect of enhancing image border becomes and improves the very important link of image picture quality.

The enhancing visual effect of image edge is also referred to as luminance transient and improves (Luminance TransientImprovement), is called for short LTI.Traditional LTI method has horizontal filtering, two-dimensional filtering etc., and what have in these methods causes distorted signals easily, as the horizontal filtering algorithm; The algorithm that has is too complicated, realizes that hardware costs is too big, for example the two-dimensional filtering algorithm.Application number is the edge enhancing method that the patent of invention of US6580835 discloses a kind of statistical nature based on the image local zone, this method is not adjusted the gain that strengthens in the position at edge according to pixel, will inevitably cause and strengthen back edge of image random fluctuation, influence visual effect, and this method desired data amount is very big, is not suitable for hardware and realizes.

Summary of the invention

Purpose of the present invention is exactly defective or the deficiency that exists at prior art, and a kind of method that strengthens visual effect of image edge is provided.This method can obviously strengthen the visual effect of image border on the one hand, and amount of calculation is smaller on the other hand, and hardware is realized simple efficient.

This method may further comprise the steps: (a) object pixel is carried out second-order filter and calculate; (b) judge the position of object pixel in the image border; (c) according to the second-order filter result of calculation of object pixel and its in the position at edge, edge calculation strengthens gain; (d) pixel value after calculate strengthening and result of calculation protected.

Step (a) is utilized equation (1) that object pixel is carried out second-order filter and is calculated:

$F_m^{\′}=[-\frac{1}{8},-\frac{1}{4},-\frac{1}{8},\frac{1}{4},\frac{1}{2},\frac{1}{4},-\frac{1}{8},-\frac{1}{4},-\frac{1}{8}]\×---\left(1\right)$

${[F_{m-4},F_{m-3},F_{m-2},F_{m-1},F_m,F_{m+1},F_{m+2},F_{m+3},F_{m+4}]}^T$

F wherein _mPixel value before the filtering of expression object pixel, F ' _mRepresent filtered pixel value.

Step (b) judges that object pixel needs following steps in the position of image border: (1) calculates the first-order difference of each pixel of original image; (2) according to the first-order difference of object pixel adjacent with its level about the relation of first-order difference of each seven pixel, find two pixels that are positioned at this edge termination; (3) according to the pixel value of object pixel relation, judge the position of object pixel in the image border with the pixel value of two pixels that are positioned at edge termination.

Step (1) is calculated the first-order difference of each pixel with formula (2):

D _m＝F _m-F _m-1 (2)

Step (2) utilizes formula (3-1) to judge the relation of object pixel and seven pixels in its left side, judges the relation of object pixel and its seven pixels in the right with formula (3-2):

T _m+i＝D _m×D _m+i＞0，i＝[-7，-1] (3-1)

T _m+j＝D _m×D _m+j＞0，j＝[1，7] (3-2)

T wherein _M+iBe to represent F at 1 o'clock _M+iPoint and F _mPoint all is in the edge (rising edge or trailing edge) of same nature.

Use formula (4-1) to find the pixel that is positioned at this edge left end in the step (2) again, (4-2) finds the pixel that is positioned at this edge right-hand member with formula:

l＝max(m+i)，P _m+i＝1 (4-1)

r＝min(m+j)，P _m+j＝1 (4-2)

Wherein l is the rightest point in all pixels that satisfy formula (3-1), and r is for satisfying the most left point in all pixels of formula (3-2), that is l is for to be in the most left pixel in same edge with object pixel, and r is the rightest pixel in same edge.

Step (3) utilizes formula (5) to calculate the position of object pixel at the edge:

P _m＝|F _m-(F _l+F _r)/2| (5)

P in the formula _mCan characterize the position of object pixel at the edge, this value is big more, and expression m point is the closer to the port at this edge.

Step (c) is utilized be enhanced edge gain of formula (6):

F″ _m＝F′ _m×(P _m×C ₁+C ₂) (6)

C wherein ₁, C ₂Be constant, and all greater than zero.

Step (d) is utilized the pixel value after equation (7) calculates enhancement process:

${\stackrel{\&OverBar;}{F}}_m=\left\{\begin{array}{cc}{F}_m& F_m^{\&prime;\&prime;}\&le;C_3\\ F_m+F_m^{\&prime;\&prime;}& C_3<F_m^{\&prime;\&prime;}<\\ F_m+C_4& F_m^{\&prime;\&prime;}\&GreaterEqual;C_4\end{array}{C}_4\right.---\left(7\right)$

C wherein ₃, C ₄Be constant, and all greater than zero.

This invention can strengthen various edge of image visions, and especially, this invention is applicable to the convergent-divergent computing of television image signal.

Because human eye is responsive especially to the image border, so the visual effect of marginal portion has determined the visual effect of entire image to a great extent.But edge of image is ever-changing, is not enough with traditional edge enhancing method only.At this point, this invention utilizes pixel at marginal position, and the point that is in diverse location is adopted different gains, can improve reinforced effects and accuracy greatly like this.Simultaneously, this is invented, and amount of calculation is smaller on the other hand, and hardware is realized simple efficient.

Description of drawings

Fig. 1 is a block diagram of the present invention;

Fig. 2 is the second-order filter schematic diagram;

Fig. 3 seeks edge end points schematic diagram;

Fig. 4 is the edge calculation position view;

Fig. 5 strengthens back pixel value schematic diagram for calculating.

Embodiment

As shown in Figure 1, the image of input at first calculates the second-order filter value of object pixel according to equation (1), and concrete preparation method can be with reference to figure 2.Then the residing marginal position of object pixel is judged.At first utilize the first-order difference of formula (2) each pixel of calculating input image during judgement, then according to the object pixel level adjacent about seven pixel difference values, utilize formula (3-1), (3-2) judge whether object pixel and these pixels are positioned at the edge of same nature (this step can with reference to figure 3), take out the neighbor pixel that all and object pixel are positioned at the edge of same nature then, utilize formula (4-1), (4-2) find with object pixel and be in the most left of same edge, the rightest pixel, utilize at last formula (5) promptly can be regarded as object pixel in the position at this edge, computational methods can be with reference to figure 4.The edge that then utilizes formula (6) to calculate object pixel strengthens gain.Utilize formula (7) to obtain the pixel value of final edge enhancement process at last, can be when calculating this value with reference to figure 5.

Fig. 2 has illustrated that employing equation (1) calculates the method for object pixel second-order filter.F among the figure _mPixel value before the filtering of expression object pixel, F ' _mRepresent filtered pixel value.Solid line is represented the original image imported, and dotted line is represented the second-order filter value.

Fig. 3 has illustrated that employing equation (2), (3-1), (3-2), (4-1), (4-2) seek the method for edge end points.M represents the object pixel horizontal coordinate among the figure.As shown in the figure, the first-order difference of some m is being for just, and the first-order difference of some m-1, m-2, m-3, m-4 is also for just, and all the other points are to bear.Therefore learn by equation (3-1) that some m-1, m-2, m-3, m-4 and target are in the same nature edge, i.e. rising edge.Learn by formula (4-1) that again some m-4 is the end points at this edge.Fig. 3 has only illustrated the method for seeking the edge left end point, seeks right endpoint and can analogize.

Fig. 4 has illustrated that employing equation (5) calculates the method for object pixel marginal position of living in.M is the object pixel horizontal coordinate among the figure, and l is the most left pixel horizontal coordinate in this edge, and r is the rightest pixel horizontal coordinate.F _m, F _l, F _rBe respectively m, l, 3 corresponding pixel value of r.P only has been described among the figure when object pixel edge of living in is trailing edge _mComputational methods, other situations can be analogized.

Fig. 5 has illustrated the pixel value method after employing equation (7) calculates the object pixel enhancement process.Pixel value before dotted line is represented to strengthen among the figure is realized the pixel value after expression strengthens, and dotted line is represented the F in the equation (7) " _m, C ₃, C ₄Be the C in the equation (7) ₃, C ₄, during actual the realization, C ₃Can be made as 60, C ₄Can be made as 10.

Claims (5)

1, strengthens the method for visual effect of image edge, it is characterized in that this method may further comprise the steps: (a) object pixel is carried out second-order filter and calculate; (b) judge the position of object pixel in the image border; (c) according to the second-order filter result of calculation of object pixel and its in the position at edge, edge calculation strengthens gain; (d) pixel value after calculate strengthening and result of calculation protected.

2, the method for enhancing visual effect of image edge as claimed in claim 1 is characterized in that step (a) utilizes equation (1) that object pixel is carried out second-order filter and calculate

$F_m^{\&prime;}=[-\frac{1}{8},-\frac{1}{4},-\frac{1}{8},\frac{1}{4},\frac{1}{2},\frac{1}{4}-\frac{1}{8},-\frac{1}{4},-\frac{1}{8}]\&times;---\left(1\right)$

${[F_{m-4},F_{m-3},F_{m-2},F_{m-1},F_m,F_{m+1},F_{m+2},F_{m+3},F_{m+4}]}^T$

F wherein _mPixel value before the filtering of expression object pixel, F _m' represent filtered pixel value.

3, the method for enhancing visual effect of image edge as claimed in claim 1, it is characterized in that the method for step (b) judgement object pixel in the position of image border may further comprise the steps: (1) calculates the first-order difference of each pixel of original image; (2) according to the first-order difference of object pixel adjacent with its level about the relation of first-order difference of each seven pixel, find two pixels that are positioned at this edge termination; (3) according to the pixel value of object pixel relation, judge the position of object pixel in the image border with the pixel value of two pixels that are positioned at edge termination;

Step (1) is calculated the first-order difference of each pixel with formula (2)

D _m＝F _m-F _m-1 (2)

Step (2) utilizes formula (3-1) to judge the relation of object pixel and seven pixels in its left side, judges the relation of object pixel and its seven pixels in the right with formula (3-2)

T _m+i＝D _m×D _m+i＞0，i＝[-7，-1] (3-1)

T _m+j＝D _m×D _m+j＞0，j＝[1，7] (3-2)

T wherein _M+iBe to represent F at 1 o'clock _M+iPoint and F _mPoint all is in the edge (rising edge or trailing edge) of same nature,

Use formula (4-1) to find the pixel that is positioned at this edge left end in the step (2) again, (4-2) finds the pixel that is positioned at this edge right-hand member with formula

l＝max(m+i)，P _m+i＝1 (4-1)

r＝min(m+j)，P _m+j＝1 (4-2)

Wherein l is the rightest point in all pixels that satisfy formula (3-1), and r is for satisfying the most left point in all pixels of formula (3-2), that is l is for to be in the most left pixel in same edge with object pixel, and r is the rightest pixel in same edge;

Step (3) utilizes formula (5) to calculate the position of object pixel at the edge

P _m＝|F _m-(F _l+F _r)/2| (5)

P in the formula _mCan characterize the position of object pixel at the edge, this value is big more, and expression m point is the closer to the port at this edge.

4, the method for enhancing visual effect of image edge as claimed in claim 1 is characterized in that step (c) utilizes be enhanced edge gain of formula (6)

F _m″＝F _m′×(P _m×C ₁+C ₂) (6)

C wherein ₁, C ₂Be constant, and all greater than zero.

5, the method for enhancing visual effect of image edge as claimed in claim 1 is characterized in that step (d) utilizes the pixel value after equation (7) calculates enhancement process

${\stackrel{\&OverBar;}{F}}_m=\left\{\begin{array}{cc}{F}_m& F_m^{\&prime;\&prime;}\&le;C_3\\ F_m+F_m^{\&prime;\&prime;}& C_3<F_m^{\&prime;\&prime;}<\\ F_m+C_4& F_m^{\&prime;\&prime;}\&GreaterEqual;C_4\end{array},C_4\right.---\left(7\right)$

C wherein ₃, C ₄Be constant, and all greater than zero.

Images