CN106611405B - Image interpolation method and device - Google Patents
Image interpolation method and device Download PDFInfo
- Publication number
- CN106611405B CN106611405B CN201510698275.1A CN201510698275A CN106611405B CN 106611405 B CN106611405 B CN 106611405B CN 201510698275 A CN201510698275 A CN 201510698275A CN 106611405 B CN106611405 B CN 106611405B
- Authority
- CN
- China
- Prior art keywords
- interpolated
- interpolation
- gray value
- edge
- pixel
- 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.)
- Active
Links
Images
Landscapes
- Image Processing (AREA)
Abstract
An image interpolation method and device, the method comprises the following steps: carrying out bicubic interpolation on pixel points to be interpolated to obtain an initial gray value; when the pixel point to be interpolated is positioned at the edge of the image, carrying out edge-oriented interpolation on the pixel point to be interpolated; when the pixel point to be interpolated is positioned at the non-image edge, taking the initial gray value as a final gray value; calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated; judging whether the satisfaction degree of the geometric duality hypothesis is smaller than a first threshold value; when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold, fusing the initial gray value and a result obtained by the edge-guided interpolation to obtain the final gray value; and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, taking the initial gray value as the final gray value. The scheme can improve the quality of the high-resolution image after interpolation.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to an image interpolation method and an image interpolation device.
Background
Image interpolation is a widely used image processing method that can increase or decrease the number of pixels for a digital image, such as interpolating a low resolution image into a high resolution image for clear display on a high resolution display.
A natural image is generally composed of basic elements such as flat areas, edges, corners, textures, and the like, and the existing interpolation technology often has a good interpolation effect only for a certain image element. For example, the bicubic interpolation method is a commonly used image interpolation method, and has the advantages that an interpolation coefficient is irrelevant to image content, the algorithm complexity is low, the interpolation effect on a flat area and an image corner is good, and phenomena of confusion, blurring and the like can occur when edges are interpolated. The other edge-oriented interpolation method for describing the geometric characteristics of the edges by using a parameter equation can obtain a smooth and clear interpolation result when the general edges with moderate scale and stable direction are interpolated, but the edge-oriented interpolation method needs to meet the geometric duality assumption, namely the geometric characteristics of the high-resolution image and the low-resolution image at corresponding positions are kept consistent, and when the geometric duality assumption is not met, the edge-oriented interpolation method can generate artifacts such as false texture, speckle noise and the like.
Therefore, the prior art has the problem of poor image quality after interpolation, and the artifacts of confusion, speckle noise and false texture in the interpolated image cannot be avoided simultaneously.
Disclosure of Invention
The invention solves the technical problem of improving the image quality after image interpolation, so that the interpolated image simultaneously avoids the artifacts of confusion, speckle noise and false texture in the interpolated image.
To solve the above problem, the present invention provides a method for image interpolation, the method comprising:
carrying out bicubic interpolation on pixel points to be interpolated to obtain an initial gray value;
judging whether the pixel point to be interpolated is at the edge of the image or not;
when the pixel point to be interpolated is positioned at the edge of the image, carrying out edge-oriented interpolation on the pixel point to be interpolated;
when the pixel point to be interpolated is not positioned at the edge of the image, taking the initial gray value as a final gray value;
calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated;
judging whether the satisfaction degree of the geometric duality hypothesis is smaller than a first threshold value;
when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold, fusing the initial gray value and a result obtained by the edge-oriented interpolation to obtain the final gray value of the pixel point to be interpolated;
and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, taking the initial gray value as the final gray value.
Optionally, the calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated includes:
satisfying M as the geometric duality assumptionDegree, M by formula
Calculating; wherein e isi＝(xi–hbic)2，xiIs the gray value h of the ith pixel point in 4 pixel points adjacent to the pixel point to be interpolatedbicRepresenting that bicubic interpolation is carried out on pixel points to be interpolated to obtain an initial gray value;
nijand expressing the difference of the gray values of the adjacent pixels in the same direction in the 4 multiplied by 4 neighborhood of the pixel to be interpolated.
Optionally, the fusing the initial gray value and the result obtained by the edge-guided interpolation includes:
taking h as the final gray value of the pixel point to be interpolated, wherein h passes through a formula
Calculating to obtain; wherein h isedRepresenting the result obtained by performing edge-oriented interpolation on the pixel point to be interpolated, hbicRepresenting that bicubic interpolation is carried out on pixel points to be interpolated to obtain an initial gray value; th denotes the first threshold, M being the satisfaction of the geometric duality assumption.
Optionally, when the pixel to be interpolated is located at an image edge, performing edge-oriented interpolation on the pixel to be interpolated, including:
when edge-oriented interpolation is carried out on two adjacent pixel points to be interpolated, the same difference items which are generated in the grey value square estimation of the four adjacent pixel points of the pixel points to be interpolated and the grey value square estimation of the pixel points to be interpolated are repeatedly used and calculated, and the old difference items are replaced by the new difference items.
An embodiment of the present invention further provides an image interpolation apparatus, where the apparatus includes: the device comprises a bicubic interpolation calculation unit, a first judgment unit, an edge guide interpolation unit, a satisfaction degree calculation unit, a second judgment unit, a fusion unit and an output unit;
the bicubic interpolation calculation unit is suitable for performing bicubic interpolation on the pixel points to be interpolated to obtain an initial gray value;
the first judging unit is suitable for judging whether the pixel point to be interpolated is positioned at the edge of the image or not;
the edge-oriented interpolation unit is suitable for carrying out edge-oriented interpolation on the pixel points to be interpolated when the pixel points to be interpolated are positioned at the edge of the image;
the satisfaction degree calculation unit is suitable for calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated;
the second judging unit is adapted to judge whether the satisfaction degree of the geometric duality assumption is smaller than a first threshold value;
the fusion unit is suitable for fusing the initial gray value and the result obtained by the edge-oriented interpolation when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold;
the output unit is suitable for outputting a result obtained by fusing the initial gray value and the edge-oriented interpolation as a final gray value of the pixel point to be interpolated when the satisfaction degree of the geometric duality assumption is smaller than the first threshold; and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, or when the pixel point to be interpolated is not positioned at the edge of the image, outputting the initial gray value as the final gray value.
Optionally, the satisfaction degree calculation unit is adapted to: taking M as the satisfaction degree of the geometric duality hypothesis, wherein M passes through a formula
Calculating; wherein e isi＝(xi–hbic)，xiFor the gray value of the ith pixel point in 4 pixel points adjacent to the pixel point to be interpolated, hbic represents that the pixel point to be interpolated is subjected to bicubic interpolation to obtain an initial gray value;
nijand expressing the difference of the gray values of the adjacent pixels in the same direction in the 4 multiplied by 4 neighborhood of the pixel to be interpolated.
Optionally, the fusion unit is adapted to: taking h as the final gray value of the pixel point to be interpolated, wherein h passes through a formula
Calculating to obtain; wherein h isedRepresenting the result obtained by performing edge-oriented interpolation on the pixel point to be interpolated, hbicRepresenting that bicubic interpolation is carried out on pixel points to be interpolated to obtain an initial gray value; th denotes the first threshold, M being the satisfaction of the geometric duality assumption.
Optionally, the edge-directed interpolation unit is further adapted to:
when edge-oriented interpolation is carried out on two adjacent pixel points to be interpolated, the same difference items which are generated in the grey value square estimation of the four adjacent pixel points of the pixel points to be interpolated and the grey value square estimation of the pixel points to be interpolated are repeatedly used and calculated, and the old difference items are replaced by the new difference items.
Compared with the prior art, the technical scheme of the embodiment of the invention has the following beneficial effects:
the technical scheme of the embodiment of the invention includes that an initial gray value is obtained by performing bicubic interpolation on a pixel point to be interpolated, when the pixel point to be interpolated is positioned at the edge of an image, edge-oriented interpolation is performed on the pixel point to be interpolated, when the pixel point to be interpolated is not positioned at the edge of the image, the initial gray value is taken as the final gray value, the satisfaction degree of geometric duality hypothesis at the pixel point to be interpolated is calculated, whether the satisfaction degree of the geometric duality hypothesis is smaller than a first threshold value or not is judged, when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold value, the final gray value of the pixel point to be interpolated is obtained by fusing the initial gray value and a result obtained by the edge-oriented interpolation, and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, the initial gray value is taken as the final, the technical scheme fuses a result of bicubic interpolation and a result of edge-oriented interpolation, particularly provides a method for fusing the results, and distinguishes the condition whether to use the fusion result, namely quantifies the satisfaction degree of geometric duality hypothesis and compares the satisfaction degree with a first threshold value to determine whether to use the fusion result, thereby simultaneously avoiding the artifacts of confusion, speckle noise and false texture in an interpolated image and improving the quality of the interpolated high-resolution image.
Further, in the technical scheme of the embodiment of the invention, when edge-directed interpolation is performed on two adjacent pixel points to be interpolated, the same difference item which is used for calculating the square estimation of the difference between the gray values of the four adjacent pixel points of the pixel point to be interpolated and the pixel point to be interpolated is repeatedly used, and the old difference item is replaced by the new difference item, so that the calculation amount of the edge-directed interpolation can be reduced.
Drawings
FIG. 1 is a flow chart of an image interpolation method in an embodiment of the present invention;
FIG. 2 is a schematic diagram of interpolation when obtaining a high resolution image of twice the original image size according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a positional relationship between a center pixel and adjacent pixels according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an embodiment of the present invention for replacing the same difference term when calculating an estimate of the gray-scale difference squared between a pixel to be interpolated and 4 adjacent pixels;
fig. 5 is a schematic structural diagram of an image interpolation apparatus in an embodiment of the present invention.
Detailed Description
As mentioned in the background art, the quality of an image obtained by interpolating an image in the prior art is poor, and artifacts such as blurring, speckle noise, and false texture in the interpolated image cannot be avoided at the same time.
The embodiment of the invention obtains an initial gray value by performing bicubic interpolation on a pixel point to be interpolated, performs edge-oriented interpolation on the pixel point to be interpolated when the pixel point to be interpolated is positioned at the edge of an image, takes the initial gray value as the final gray value when the pixel point to be interpolated is not positioned at the edge of the image, calculates the satisfaction degree of the geometric duality hypothesis at the pixel point to be interpolated, judges whether the satisfaction degree of the geometric duality hypothesis is smaller than a first threshold value, fuses the initial gray value and the result obtained by the edge-oriented interpolation when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold value to obtain the final gray value of the pixel point to be interpolated, and takes the initial gray value as the final gray value when the satisfaction degree of the geometric duality hypothesis is greater than or equal to the first threshold value, the technical scheme integrates the result of bicubic interpolation and the result of edge-oriented interpolation, particularly provides a method for integrating the result, and distinguishes the condition whether to use the integration result, thereby simultaneously avoiding the artifacts of confusion, speckle noise and false texture in the interpolated image and improving the quality of the interpolated high-resolution image.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
Fig. 1 is a flowchart of an image interpolation method in an embodiment of the present invention. The image interpolation method as shown in fig. 1 may include:
step S101: and carrying out bicubic interpolation on the pixel points to be interpolated to obtain an initial gray value.
In a specific implementation, the pixel to be interpolated may be a central pixel or an adjacent pixel. Referring to FIG. 3, a black square h0As the central pixel, a black triangular block h1Is the adjacent pixel.
In a specific implementation, since bicubic interpolation is a relatively stable calculation method, the result of bicubic interpolation is used as the initial gray value of the pixel to be interpolated. Furthermore, the initial use of the results of bicubic interpolation may simplify the operation.
Step S102: judging whether the pixel point to be interpolated is at the edge of the image or not; and executing step S103 when the pixel point to be interpolated is at the edge of the image, otherwise executing step S107.
In specific implementation, whether the pixel point to be interpolated is at the edge of the image can be judged according to factors such as gradient change and the like.
Step S103: and carrying out edge-oriented interpolation on the pixel points to be interpolated.
In a specific implementation, when the pixel to be interpolated is at an edge of an image, edge-oriented interpolation is performed on the pixel to be interpolated, and when the pixel to be interpolated is at a non-edge of the image, bicubic interpolation is performed on the pixel to be interpolated, that is, the initial gray value calculated in step S101 is used as the gray value of the pixel to be interpolated in the high-resolution image.
In particular implementations, the result of computing the edge-directed interpolation may employ a formula
Wherein x isiIs the gray value, w, of the ith pixel point in 4 pixel points adjacent to the pixel point to be interpolatediDetermining the weight according to the similarity between the ith pixel point of 4 pixel points adjacent to the pixel point to be interpolated and the pixel point to be interpolated, wherein the greater the similarity is, the greater the w isiThe larger. w is aiCan be according to the formula wi＝(E(ni 2)+c)-2Where c is a constant, E (n)i 2) For estimating the square of the gray difference value of the ith pixel point and the pixel point to be interpolated,
nijand expressing the difference of the gray values of the adjacent pixels in the same direction in the 4 multiplied by 4 neighborhood of the pixel to be interpolated.
In a specific implementation, in the process of performing edge-oriented interpolation, when performing edge-oriented interpolation on two adjacent pixel points to be interpolated, the above E (n) may be specifically calculatedi 2) In the process, the adjacent pixel point to be interpolated can be repeatedly used to calculate E (n)i 2) Phase of passingThe homodyne term and the old difference term is replaced with the new difference term.
Step S104: and calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated.
Since the important basis of the edge-oriented interpolation method is the assumption of geometric duality, the assumption is that the geometric characteristics of the image content are consistent in the high-resolution image and the low-resolution image, such as the image edges. Edge-guided interpolation methods can produce artifacts if the geometric duality assumption cannot be satisfied. For example, the edges of the grass image are small in scale and weak in intensity, and the interpolated pixels differ significantly from the pixels mapped directly into the high resolution image, making the speckle effect very noticeable. As another example, using edge-guided interpolation methods for corners in an image creates burrs at the corners.
Therefore, the satisfaction of the geometric duality assumption is related to the usage effect of the edge-oriented interpolation method.
In a specific implementation, calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated may include:
taking M as the satisfaction degree of the geometric duality hypothesis, wherein M passes through a formula
Calculating; wherein e isi＝xi-hbic，xiRepresenting the gray value h of one of four pixel points adjacent to the pixel point to be interpolatedbicRepresenting an initial gray value obtained by performing bicubic interpolation on a pixel point to be interpolated;
nijand expressing the difference of the gray values of the adjacent pixels in the same direction in the 4 multiplied by 4 neighborhood of the pixel to be interpolated.
The embodiment of the invention reflects the satisfaction degree of the geometric duality assumption by using the calculated M.
Step S105: judging whether the satisfaction degree of the geometric duality hypothesis is smaller than a first threshold value; when the satisfaction degree of the geometric duality assumption is smaller than the first threshold, step S106 is performed, otherwise step S107 is performed.
In a specific implementation, the setting of the first threshold may be set by performing different adjustments, and is not limited herein.
Step S106: and fusing the initial gray value and the result obtained by the edge-oriented interpolation to obtain the final gray value of the pixel point to be interpolated.
In specific implementation, when the satisfaction degree of the geometric duality assumption is smaller than the first threshold, h may be used as the final gray value of the pixel to be interpolated, where h passes through a formula
Calculating to obtain; wherein h isedRepresenting the result obtained by performing edge-oriented interpolation on the pixel point to be interpolated, hbicCarrying out bicubic interpolation on pixel points to be interpolated to obtain an initial gray value; th denotes the first threshold value, and M denotes the degree of satisfaction of the geometric duality assumption.
The embodiment of the invention fuses a bicubic interpolation result and an edge-oriented interpolation result, and one of the conditions required for fusing the two results is that the satisfaction degree of the geometric duality assumption is smaller than the first threshold value. Therefore, according to the technical scheme of the embodiment of the invention, by actually providing a method for evaluating the satisfaction degree of the geometric duality hypothesis and adding the size judgment between the geometric duality hypothesis and the first threshold, whether results obtained by bicubic interpolation and edge-guided interpolation are used simultaneously can be accurately measured, and then interpolation is performed on pixels to be interpolated in different areas by adopting a proper method.
Step S107: and taking the initial gray value as the final gray value.
In specific implementation, when a pixel point to be interpolated is in a non-edge region, bicubic interpolation can be directly adopted, because the interpolation effect of the edge-oriented interpolation method in the non-edge region is similar to that of the bicubic interpolation, and in this case, compared with the edge-oriented interpolation method, the bicubic interpolation is used, the complexity of operation can be simplified, and the operation speed is increased.
In specific implementation, when the adjacent pixel is used as the pixel point to be interpolated, E (n) needs to be calculated by using the interpolated central pixel2 i)。
For further explaining the process of calculating the final gray value of the pixel to be interpolated according to the technical solution of the present invention, please refer to fig. 2.
Fig. 2 is a schematic diagram of interpolation for obtaining a high-resolution image with a size twice that of an original image in the embodiment of the present invention. The description is made with reference to the flowchart of fig. 1.
As shown in the figure, assume that the pixel point to be interpolated is h shown by a solid square0The adjacent pixel point in 4 × 4 neighborhood is P5、P6、P9And P10. Point P represented by black dot0-P15Are pixels in the low resolution image.
Firstly, calculating the initial gray value h of the point to be interpolated by utilizing 16 pixels nearest to the pixel point to be interpolated in bicubic interpolationbic. It should be noted that those skilled in the art can understand how the bicubic interpolation can calculate the initial gray value hbicAnd will not be described herein.
Next, step 102 is executed, and if the determination result is yes, step 103 is executed. In step S103, an edge-oriented interpolation of the pixel to be interpolated is calculated, and the edge-oriented interpolation uses the nearest 4 pixels to find the pixel to be interpolated. Here using the formula
Calculating to obtain the edge guide interpolation hedWherein w isiFor weighting, the low resolution can be determined according to the adjacent degree on the premise of satisfying the assumption of geometric dualityEstimating adjacent pixel point as P by gray difference between rate pixel points5、P6、P9And P10The similarity between the ith pixel point and the pixel point to be interpolated is determined, so as to determine wiI.e. using the formula wi＝(E(ni 2)+c)-2To calculate wiWherein
nijIndicating the difference between the gray values of the neighboring pixels in the same direction in the 4 × 4 neighborhood of the pixel to be interpolated, and the arrow in fig. 2 indicates that E (n) is obtained2 1) Two pixels with time difference in gray, E (n)2 1) The calculation formula of (A) is as follows: e (n)1 2)＝((p0-p5)2+(p1-p6)2+...+(p9-p14)2) (iii) p in the formula0-p5To express P0Pixel point and P5The gray level difference of the pixel points and the meanings of other terms in the calculation formula are similar, and are not described in detail.
It is noted that only the calculation E (n) is given in fig. 22 1) I.e. the calculation E (n) is indicated only by arrows2 1) And (4) solving the pixel points with the gray difference. Those skilled in the art will understand that other three adjacent pixel points P can be obtained6、P9And P10W ofiCalculation E (n) to be listed2 i) The formula (2) is calculated.
Calculating the edge guide interpolation hedAfter that, step S104 is executed. The embodiment of the invention uses the formula
M is calculated to measure the satisfaction of the geometric duality assumption. When the step S105 is executed to determine whether the satisfaction degree M of the geometric duality assumption is smaller than the first threshold th, the step S106 is executed to obtain the initial gray value h by bicubic interpolationbicAnd edge guided interpolation hedPerforming fusion, using in particular the formula
And calculating the final gray value h of the pixel point to be interpolated in the high-resolution image.
In one implementation, as shown in FIG. 3, block h0As a central pixel, a triangular block h1And h2Is a neighboring pixel when the central pixel h0Adjacent pixel h1And h2When the central pixel h is used as a pixel point to be interpolated, the central pixel h which is already interpolated needs to be utilized0To calculate E (n)2 i)。
In the specific implementation, as shown in fig. 4, two adjacent pixel points h to be interpolated are processed0And h'0When the edge-oriented interpolation is carried out, the square estimation of the difference between the gray values of four adjacent pixel points of the pixel point to be interpolated and the pixel point to be interpolated is repeatedly used, namely, the calculation E (n) is repeatedly used2 i) The same difference terms that have occurred and the old difference terms are replaced with the new difference terms. It should be noted that h0And h'0Is any two adjacent center pixels, the next one and h'0The neighboring center pixel is in the calculation of E (n)2 i) Just, h 'can be calculated repeatedly'0The same difference terms in edge-directed interpolation, and so on. Meanwhile, it should be noted that fig. 4 only shows the case where the pixel to be interpolated is the central pixel, and the same difference terms that have appeared can be multiplexed when edge-directed interpolation is performed between two adjacent pixels, so that the computation amount of the algorithm can be further reduced.
Fig. 5 is a schematic structural diagram of an image interpolation apparatus in an embodiment of the present invention. The image interpolation apparatus 50 as shown in the figure may include: a bicubic interpolation calculation unit 501, a first determination unit 502, an edge guide interpolation unit 503, a satisfaction degree calculation unit 504, a second determination unit 505, a fusion unit 506, and an output unit 507;
the bicubic interpolation calculation unit 501 is adapted to perform bicubic interpolation on the pixel points to be interpolated to obtain an initial gray value;
the first determining unit 502 is adapted to determine whether the pixel to be interpolated is located at an image edge;
the edge-oriented interpolation unit 503 is adapted to perform edge-oriented interpolation on the pixel to be interpolated when the pixel to be interpolated is at the edge of the image;
the satisfaction degree calculating unit 504 is adapted to calculate the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated;
the second determining unit 505 is adapted to determine whether the satisfaction degree of the geometric duality assumption is smaller than a first threshold;
the fusion unit 506 is adapted to fuse the initial gray value and the result of the edge-oriented interpolation when the satisfaction degree of the geometric duality assumption is smaller than the first threshold;
the output unit 507 is adapted to output, when the satisfaction degree of the geometric duality assumption is smaller than the first threshold, the result obtained by fusing the initial gray value and the edge-oriented interpolation as a final gray value of the pixel to be interpolated; and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, or when the pixel point to be interpolated is not positioned at the edge of the image, outputting the initial gray value as the final gray value.
In a specific implementation, the satisfaction degree calculation unit 503 is adapted to: taking M as the satisfaction degree of the geometric duality hypothesis, wherein M passes through a formula
Calculating; wherein e isi＝(xi–hbic)2，xiRepresenting the gray values, h, of four pixels adjacent to the pixel to be interpolatedbicRepresenting that bicubic interpolation is carried out on pixel points to be interpolated to obtain an initial gray value;
nijrepresenting the gray value of the adjacent pixel points in the same direction in the 4 multiplied by 4 neighborhood of the pixel point to be interpolatedThe difference between them.
In a specific implementation, the fusion unit 505 is adapted to: taking h as the final gray value of the pixel point to be interpolated, wherein h passes through a formula
Calculating to obtain; wherein h isedAnd the result obtained by performing edge-oriented interpolation on the pixel points to be interpolated is represented, and th represents the first threshold.
In a specific implementation, the edge-directed interpolation unit 503 is further adapted to: when edge-oriented interpolation is carried out on two adjacent pixel points to be interpolated, the same difference items which are generated in the grey value square estimation of the four adjacent pixel points of the pixel points to be interpolated and the grey value square estimation of the pixel points to be interpolated are repeatedly used and calculated, and the old difference items are replaced by the new difference items.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer-readable storage medium, and the storage medium may include: ROM, RAM, magnetic or optical disks, and the like.
Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (6)
1. An image interpolation method, comprising:
performing bicubic interpolation on the pixel points to be interpolated to obtain an initial gray value before judging whether the pixel points to be interpolated are positioned at the edge of the image;
after carrying out bicubic interpolation on pixel points to be interpolated to obtain an initial gray value, judging whether the pixel points to be interpolated are positioned at the edge of an image or not;
when the pixel point to be interpolated is at the edge of the image, performing edge-oriented interpolation on the pixel point to be interpolated, including: calculating the edge-directed interpolation of the pixel points to be interpolated only by 4 adjacent pixel points of the pixel points to be interpolated, and when performing the edge-directed interpolation on two adjacent pixel points to be interpolated, repeatedly using the same difference terms which are used for calculating the gray values of four adjacent pixel points of the pixel points to be interpolated and the gray value square estimation of the pixel points to be interpolated, and replacing the old difference terms by new difference terms;
when the pixel point to be interpolated is not positioned at the edge of the image, taking the initial gray value as a final gray value;
calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated;
judging whether the satisfaction degree of the geometric duality hypothesis is smaller than a first threshold value;
when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold, fusing the initial gray value and a result obtained by the edge-oriented interpolation to obtain the final gray value of the pixel point to be interpolated;
and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, taking the initial gray value as the final gray value.
2. The image interpolation method according to claim 1, wherein the calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated comprises:
taking M as the satisfaction degree of the geometric duality hypothesis, wherein M passes through a formula
Calculating; wherein e isi＝(xi–hbic)2，xiIs the gray value h of the ith pixel point in 4 pixel points adjacent to the pixel point to be interpolatedbicRepresenting an initial gray value obtained by performing bicubic interpolation on a pixel point to be interpolated;
nijto indicate a waitThe difference of the gray values of the adjacent pixels in the same direction in the 4 × 4 neighborhood of the interpolated pixel.
3. The image interpolation method according to claim 2, wherein the fusing the initial gray-scale value and the result of the edge-guided interpolation comprises:
taking h as the final gray value of the pixel point to be interpolated, wherein h passes through a formula
Calculating to obtain; wherein h isedRepresenting the result obtained by performing edge-oriented interpolation on the pixel point to be interpolated, hbicRepresenting an initial gray value obtained by performing bicubic interpolation on a pixel point to be interpolated; th denotes the first threshold, M being the satisfaction of the geometric duality assumption.
4. An image interpolation apparatus, characterized by comprising: the device comprises a bicubic interpolation calculation unit, a first judgment unit, an edge guide interpolation unit, a satisfaction degree calculation unit, a second judgment unit, a fusion unit and an output unit;
the bicubic interpolation calculation unit is suitable for performing bicubic interpolation on the pixel points to be interpolated to obtain an initial gray value before judging whether the pixel points to be interpolated are positioned at the edge of the image;
the first judging unit is suitable for judging whether the pixel point to be interpolated is positioned at the edge of the image or not after carrying out bicubic interpolation on the pixel point to be interpolated to obtain an initial gray value;
the edge-oriented interpolation unit is suitable for performing edge-oriented interpolation on the pixel points to be interpolated when the pixel points to be interpolated are positioned at the edge of the image; the edge-oriented interpolation of the pixel points to be interpolated is calculated only by 4 adjacent pixel points of the pixel points to be interpolated, when the edge-oriented interpolation is carried out on two adjacent pixel points to be interpolated, the same difference items which are calculated and appear in the gray value square estimation of four adjacent pixel points of the pixel points to be interpolated and the gray value square estimation of the pixel points to be interpolated are repeatedly used, and the old difference items are replaced by the new difference items;
the satisfaction degree calculation unit is suitable for calculating the satisfaction degree of the geometric duality assumption at the pixel point to be interpolated;
the second judging unit is adapted to judge whether the satisfaction degree of the geometric duality assumption is smaller than a first threshold value;
the fusion unit is suitable for fusing the initial gray value and the result obtained by the edge-oriented interpolation when the satisfaction degree of the geometric duality hypothesis is smaller than the first threshold;
the output unit is suitable for outputting a result obtained by fusing the initial gray value and the edge-oriented interpolation as a final gray value of the pixel point to be interpolated when the satisfaction degree of the geometric duality assumption is smaller than the first threshold; and when the satisfaction degree of the geometric duality hypothesis is larger than or equal to the first threshold value, or when the pixel point to be interpolated is not positioned at the edge of the image, outputting the initial gray value as the final gray value.
5. The image interpolation apparatus according to claim 4, wherein the satisfaction degree calculation unit is adapted to: taking M as the satisfaction degree of the geometric duality hypothesis, wherein M passes through a formula
Calculating; wherein e isi＝(xi–hbic)2，xiIs the gray value h of the ith pixel point in 4 pixel points adjacent to the pixel point to be interpolatedbicRepresenting that bicubic interpolation is carried out on pixel points to be interpolated to obtain an initial gray value;
nijand expressing the difference of the gray values of the adjacent pixels in the same direction in the 4 multiplied by 4 neighborhood of the pixel to be interpolated.
6. The image interpolation apparatus according to claim 5, wherein the fusion unit is adapted to: taking h as the final gray value of the pixel point to be interpolated, wherein h passes through a formula
Calculating to obtain; wherein h isedRepresenting the result obtained by performing edge-oriented interpolation on the pixel point to be interpolated, hbicRepresenting that bicubic interpolation is carried out on pixel points to be interpolated to obtain an initial gray value; th denotes the first threshold, M being the satisfaction of the geometric duality assumption.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510698275.1A CN106611405B (en) | 2015-10-23 | 2015-10-23 | Image interpolation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510698275.1A CN106611405B (en) | 2015-10-23 | 2015-10-23 | Image interpolation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106611405A CN106611405A (en) | 2017-05-03 |
| CN106611405B true CN106611405B (en) | 2020-04-07 |
Family
ID=58613425
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510698275.1A Active CN106611405B (en) | 2015-10-23 | 2015-10-23 | Image interpolation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106611405B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107734281A (en) * | 2017-09-22 | 2018-02-23 | 维沃移动通信有限公司 | A kind of image magnification method and mobile terminal |
| CN112565887B (en) * | 2020-11-27 | 2023-06-20 | 紫光展锐(重庆)科技有限公司 | Video processing method, device, terminal and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101815157A (en) * | 2009-02-24 | 2010-08-25 | 虹软(杭州)科技有限公司 | Image and video amplification method and relevant image processing device |
| CN102800069A (en) * | 2012-05-22 | 2012-11-28 | 湖南大学 | Image super-resolution method for combining soft decision self-adaptation interpolation and bicubic interpolation |
| CN103500437A (en) * | 2013-09-27 | 2014-01-08 | 中国人民解放军空军工程大学 | Fast NEDI (new edge-directed interpolation) image interpolation implementation method |
-
2015
- 2015-10-23 CN CN201510698275.1A patent/CN106611405B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101815157A (en) * | 2009-02-24 | 2010-08-25 | 虹软(杭州)科技有限公司 | Image and video amplification method and relevant image processing device |
| CN102800069A (en) * | 2012-05-22 | 2012-11-28 | 湖南大学 | Image super-resolution method for combining soft decision self-adaptation interpolation and bicubic interpolation |
| CN103500437A (en) * | 2013-09-27 | 2014-01-08 | 中国人民解放军空军工程大学 | Fast NEDI (new edge-directed interpolation) image interpolation implementation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106611405A (en) | 2017-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Feichtenhofer et al. | A perceptual image sharpness metric based on local edge gradient analysis | |
| Zhou et al. | Image zooming using directional cubic convolution interpolation | |
| Rehman et al. | SSIM-based non-local means image denoising | |
| US20160112659A1 (en) | Image processing apparatus and image processing method, and program | |
| KR20060124502A (en) | False contour correction method and display apparatus to be applied to the same | |
| CN102479380B (en) | Image interpolation method capable of improving image resolution and device utilizing image interpolation method | |
| CN106169174B (en) | Image amplification method | |
| JP5107342B2 (en) | Image improvement to increase accuracy smoothing characteristics | |
| US20090226097A1 (en) | Image processing apparatus | |
| US11145032B2 (en) | Image processing apparatus, method and storage medium for reducing color noise and false color | |
| CN111062884B (en) | Image enhancement method and device, storage medium and terminal equipment | |
| US20150187051A1 (en) | Method and apparatus for estimating image noise | |
| Mahajan et al. | Adaptive and non-adaptive image interpolation techniques | |
| US20090022402A1 (en) | Image-resolution-improvement apparatus and method | |
| JP2010134700A (en) | Image evaluation device and image evaluation method | |
| CN110866882A (en) | Layered joint bilateral filtering depth map restoration algorithm based on depth confidence | |
| US8213736B2 (en) | Image processing device and image processing method | |
| CN106611405B (en) | Image interpolation method and device | |
| Bouma et al. | Precise local blur estimation based on the first-order derivative | |
| KR20130036430A (en) | Method for generating high resolution depth image from low resolution depth image, and medium recording the same | |
| CN108416770B (en) | Image quality evaluation method based on visual saliency | |
| CN111445435A (en) | No-reference image quality evaluation method based on multi-block wavelet transform | |
| JP2014130427A (en) | Image quality estimation device, image quality estimation method, and image quality estimation program | |
| JP2012227791A (en) | Image processor, image processing method, program and recording medium | |
| Nasonov et al. | Adaptive image deringing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |