CN1722176A - Converting the resolution of an image using interpolation and displaying the converted image - Google Patents
Converting the resolution of an image using interpolation and displaying the converted image Download PDFInfo
- Publication number
- CN1722176A CN1722176A CN 200510083358 CN200510083358A CN1722176A CN 1722176 A CN1722176 A CN 1722176A CN 200510083358 CN200510083358 CN 200510083358 CN 200510083358 A CN200510083358 A CN 200510083358A CN 1722176 A CN1722176 A CN 1722176A
- Authority
- CN
- China
- Prior art keywords
- reference pixels
- value
- image
- pixel
- pixel value
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4007—Interpolation-based scaling, e.g. bilinear interpolation
Abstract
The invention relates to a resolution transform method, an image treatment device using the resolution transform method and an image display device. An interpolation method can be used for transforming the image resolution. The interpolation method generates a weighted factor according to the pixel value and the distance value of a reference pixel. The pixel value of an interpolated pixel is determined according to the weighted factor. According to the invention, smooth and sharp images without unsmooth edges can be generated.
Description
Technical field
The present invention relates to use interpolation (interpolation) method changing image resolution (resolution) and show this changing image.
Background technology
Existing display device provides the image resolution ratio mapping function usually.For example,, then can use any known interpolating method for example near interpolation if the resolution of image is lower than the output resolution ratio of display device, the linear interpolation method, three convolution methods etc. increase image resolution ratio.In addition, in No. 2796900 communique of Jap.P. (following brief note is " patent documentation 1 ") for example, various interpolating methods are disclosed.
Described can carrying out at a high speed near interpolation still, can produce as step-like uneven edge (jagginess) on image, has the problem of picture quality variation.Described linear interpolation method is more effective near interpolation than described, can generate level and smooth image, and still, it can reduce the bright acutance (sharpness) of image, causes image blurring.Three convolution methods can provide than described near interpolation or the high-quality image of linear interpolation method, and still, its reference pixels number is many, the computing complexity.Have, these three convolution methods can increase the noise composition of image again.In the method for patent documentation 1, can obtain than described near the high-quality image of interpolation with reference to scope with relatively little, still, it is difficult to avoid uneven edge.
As mentioned above, in known method, do not have to generate do not have uneven edge, the method for level and smooth and bright sharp image.Have again, in known method, when suppressing operand, can not generate the method for high quality graphic.
Summary of the invention
The present invention puts forward for solving the existing problem of above-mentioned prior art, the objective of the invention is to, the interpolating method of the resolution conversion that can obtain the high-grade image is provided, and the image processing apparatus of carrying out this interpolating method, image display device, can generate does not have uneven edge, level and smooth and bright sharp image.
A further object of the present invention is, the interpolating method of the resolution conversion that can obtain the high-grade image is provided, and the image processing apparatus of carrying out this interpolating method, and image display device can suppress operand, generates high quality graphic.
Another purpose of the present invention is, the interpolating method of the resolution conversion that can obtain the high-grade image is provided, and the image processing apparatus of carrying out this interpolating method, image display device, even in order in the pixel value of inter polated pixel, not comprise unnecessary information, use few reference pixels to count occasion, also can be verily with in the image of original image message reflection after interpolation.
To achieve these goals, one embodiment of the present of invention comprise device, method, and system, computer program and goods use the 1st interpolating method, changing image resolution, the 1st interpolating method may further comprise the steps:
Specific inter polated pixel is so that be appended to it on image;
Around this inter polated pixel, select some reference pixels;
Ask for the distance value of each reference pixels;
Extract the pixel value of each reference pixels;
Choose the target reference pixels from some reference pixels, use the distance value and the pixel value of this target reference pixels, generate the weighting factor of described object pixel;
Weighting factor by described object pixel is determined pixel value, and the inter polated pixel that will have this pixel value is appended on the image.
An alternative embodiment of the invention comprises device, method, system, computer program and goods use an interpolating method, changing image resolution, select this interpolating method according to characteristics of image from some interpolating methods, described some interpolating methods comprise described the 1st interpolating method.
Another embodiment of the present invention comprises device, method, and system, computer program and goods, by using the 1st interpolating method or selected interpolating method, changing image resolution shows this image after conversion.
According to resolution converting method of the present invention, use the image processing apparatus and the image display device of this resolution converting method, can generate does not have uneven edge, level and smooth and bright sharp image.
According to resolution converting method of the present invention, use the image processing apparatus and the image display device of this resolution converting method, can suppress operand, generate high quality graphic.
According to resolution converting method of the present invention, use the image processing apparatus and the image display device of this resolution converting method, even in order in the pixel value of inter polated pixel, not comprise unnecessary information, use few reference pixels to count occasion, also can be verily with in the image of original image message reflection after interpolation.
Description of drawings
When considering relevant accompanying drawing,, understanding more completely the present invention and the plurality of advantages followed thereof will be obtained easily with reference to following detailed description.Wherein:
Fig. 1 represents the block scheme according to the image display device of one embodiment of the invention;
Fig. 2 represents to be added in the view data shown in Figure 1 according to the inter polated pixel of one embodiment of the invention;
Fig. 3 represents to determine according to use the 1st of one embodiment of the invention or the 2nd method the processing action flow chart of the pixel value of inter polated pixel;
Fig. 4 represents to determine according to use the 3rd method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Fig. 5 represents to determine according to use the 4th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Fig. 6 represents to determine according to use the 5th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Fig. 7 represents to determine according to use the 6th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Fig. 8 A represents that original image is input in the display device shown in Figure 1;
Fig. 8 B represents to use the linear interpolation method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 C represents to use the present invention's the 1st method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 D represents to use the present invention's the 2nd method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 E represents to use the present invention's the 3rd method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 F represents to use the present invention's the 4th method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 G represents to use the present invention's the 4th method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 H represents to use the present invention's the 5th method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 I represents to use the present invention's the 6th method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 8 J represents to use the present invention's the 6th method that Fig. 8 A original image is carried out the image that interpolation generates;
Fig. 9 A represents to be added in the view data according to the inter polated pixel of one embodiment of the invention;
Fig. 9 B represents that immediate reference pixels is to the influence of the pixel value of inter polated pixel shown in Fig. 9 A when using the linear interpolation method to determine pixel value;
Fig. 9 C represents that immediate reference pixels is to the influence of the pixel value of inter polated pixel shown in Fig. 9 A when using the present invention's the 2nd method to determine pixel value;
Fig. 9 D represents that immediate reference pixels is to the influence of the pixel value of inter polated pixel shown in Fig. 9 A when using the present invention's the 4th method to determine pixel value;
Fig. 9 E represents that immediate reference pixels is to the influence of the pixel value of inter polated pixel shown in Fig. 9 A when using the present invention's the 6th method to determine pixel value;
Figure 10 A represents according to use the present invention the 2nd method of one embodiment of the invention original image to be carried out the image that interpolation generates;
Figure 10 B represents to use the linear interpolation method that original image is carried out the image that interpolation generates;
Figure 11 represents to determine according to use the present invention the 7th of one embodiment of the invention or the 8th method the processing action flow chart of the pixel value of inter polated pixel;
Figure 12 represents to determine according to use the present invention the 9th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 13 represents to determine according to use the present invention the 10th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 14 A represents to import the original image of display device shown in Figure 1;
Figure 14 B represents to use the linear interpolation method that original image shown in Figure 14 A is carried out the image that interpolation generates;
Figure 14 C represents to use three convolution methods that original image shown in Figure 14 A is carried out the image that interpolation generates;
Figure 14 D represents to use the 7th method of the present invention that original image shown in Figure 14 A is carried out the image that interpolation generates;
Figure 14 E represents to use the 8th method of the present invention that original image shown in Figure 14 A is carried out the image that interpolation generates;
Figure 14 F represents to use the 9th method of the present invention that original image shown in Figure 14 A is carried out the image that interpolation generates;
Figure 14 G represents to use the 10th method of the present invention that original image shown in Figure 14 A is carried out the image that interpolation generates;
Figure 15 represents to determine according to use the present invention the 11st, 12,13 methods of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 16 A represents to import the original image of display device shown in Figure 1;
Figure 16 B represents to use the image that any method is carried out the interpolation generation in 1-the 10th method of the present invention to original image shown in Figure 16 A;
Figure 16 C represents to use the present invention's the 11st method that original image shown in Figure 16 A is carried out the image that interpolation generates;
Figure 16 D represents to use the 12nd method of the present invention that original image shown in Figure 16 A is carried out the image that interpolation generates;
Figure 16 E represents to use the 13rd method of the present invention that original image shown in Figure 16 A is carried out the image that interpolation generates;
Figure 16 F represents to use the linear interpolation method that original image shown in Figure 16 A is carried out the image that interpolation generates;
Figure 16 G represents to use three convolution methods that original image shown in Figure 16 A is carried out the image that interpolation generates;
Figure 16 H represents to use the image that disclosed method is carried out the interpolation generation to original image shown in Figure 16 A in patent documentation 1;
Figure 17 represents the block scheme according to the image display device of one embodiment of the invention;
Figure 18 represents to determine according to use the present invention the 14th of one embodiment of the invention or the 15th method the processing action flow chart of the pixel value of inter polated pixel;
Figure 19 represents to determine according to use the present invention the 16th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 20 represents to determine according to use the present invention the 17th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 21 A, Figure 21 B represent to determine according to use the present invention the 18th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 22 represents to determine according to use the present invention the 19th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 23 represents to determine according to use the present invention the 20th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 24 A, Figure 24 B represent to determine according to use the present invention the 21st method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 25 A represents to import the original image of display device shown in Figure 1;
Figure 25 B represents to use the linear interpolation method that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 C represents to use the present invention's the 12nd method that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 D represents to use the 16th method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 E represents to use the 17th method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 F represents to use the 18th method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 G represents to use the 2nd method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 H represents to use the 19th method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 I represents to use the 20th method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 J represents to use the 21st method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 25 K represents to use the image that disclosed method is carried out the interpolation generation to original image shown in Figure 25 A in patent documentation 1;
Figure 25 L represents to use three convolution methods that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 26 represents the air line distance from the inter polated pixel of Fig. 2 to reference pixels;
Figure 27 A, Figure 27 B represent to determine according to use the present invention the 22nd method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 28 A, Figure 28 B represent to determine according to use the present invention the 23rd method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 29 A, Figure 29 B represent to determine according to use the present invention the 24th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 30 A, Figure 30 B represent to determine according to use the present invention the 25th method of one embodiment of the invention the processing action flow chart of the pixel value of inter polated pixel;
Figure 31 A represents to use and near interpolation original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 31 B represents to use the 2nd method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 31 C represents to use the 22nd method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 31 D represents to use the 23rd method of the present invention that original image shown in Figure 25 A is carried out the image that interpolation generates;
Figure 32 A represents to have the original image of deep or light grade;
Figure 32 B represents to use the 22nd method of the present invention that original image shown in Figure 32 A is carried out the image that interpolation generates;
Figure 32 C represents to use the 23rd method of the present invention that original image shown in Figure 32 A is carried out the image that interpolation generates.
Embodiment
Below, present invention will be described in detail with reference to the accompanying embodiment.
Fig. 1 represents the block scheme according to the image display device 10 of one embodiment of the invention.
This image display device 10 can be any display device, cathode-ray tube (CRT) (cathode ray tube for example, below brief note is " CRT ") display, LCD (liquid crystaldisplay, below brief note is " LCD "), plasma display panel (plasma display panel, below brief note is " PDP "), projectors etc., it can show the image according to view data 1.
As shown in Figure 1, image display device 10 comprises image processing apparatus 9 and display part 8.If view data 1 input picture display device 10, image processing apparatus 9 detects the resolution (following brief note is " an input resolution ") of these view data 1, and the resolution (following brief note is " output resolution ratio ") of this input resolution and described display part 8 is compared.According to comparative result, this view data 1 of image processing apparatus 9 conversion becomes output resolution ratio with it from the input resolution conversion.View data 1 through above-mentioned conversion outputs to described display part 8, is illustrated in described display part 8.
As shown in Figure 1, image processing apparatus 9 comprises input data store 4, resolution test section 2, crossbar selector 3, resolution conversion portion 5, transform data storage part 6, output data storage part 7.
In one embodiment of the invention, if output resolution ratio is lower than input resolution, then some pixels (following brief note is " removing pixel ") can be removed to view data 1 by resolution conversion portion 5.Resolution conversion portion 5 selects the described pixel of removing according to input and output coordinate from view data 1.
In another embodiment of the present invention, if output resolution ratio ratio input resolution height, then resolution conversion portion 5 can increase some pixels (following brief note is " inter polated pixel ") to view data 1.Resolution conversion portion 5 determines to increase in the view data 1 part of inter polated pixel according to input and output coordinate.Have, resolution conversion portion 5 uses following described the whole bag of tricks, according to the information that is included in the view data 1, determines the pixel value of each inter polated pixel again.
6 storages of transform data storage part comprise the various data that are used for resolution conversion.
Output data storage part 7 can preferably be set, and the view data with output resolution ratio 1 that its storage has been handled is exported this view data of having handled 1 with certain unit form.
With reference to Fig. 2 and Fig. 3, the processing action of the pixel value of determining inter polated pixel according to the use the 1st or the 2nd method of one embodiment of the invention is described.
According to the 1st method of the present invention, in step S100, one of specific inter polated pixel.For example, shown in Figure 2, resolution conversion portion 5 can a specific inter polated pixel B.
In step S101, around this specific inter polated pixel, select one or more reference pixels, this reference pixels is the data of original digital image data 1.In step S101, also obtain the distance value of each reference pixels.
In order to select reference pixels, distance between inter polated pixel and its neighbor can be calculated to each inter polated pixel according to input and output coordinate by resolution conversion portion 5.Described distance can be used X, and the Y coordinate figure is represented.For example, (X1, Y1), its adjacent pixels is positioned at coordinate, and (X2, Y2), the distance between so described inter polated pixel and the described adjacent pixels can be used X, Y coordinate figure (X1-X2), (Y1-Y2) expression if inter polated pixel is positioned at coordinate.The distance value that calculates and get is stored in the transform data storage part 6 LUT.Use this LUT, one or more reference pixels can be selected to each inter polated pixel of view data 1 by resolution conversion portion 5.Have, resolution conversion portion 5 can obtain the distance value of selected each reference pixels from this LUT again.
In Fig. 2 embodiment, the 4th reference pixels A00 of 1-, A01, A10, A11 select in resolution conversion portion 5 around inter polated pixel B.To each reference pixels A00, A01, A10, A11, resolution conversion portion 5 obtains with X, the distance value that the Y coordinate figure is represented.In this embodiment, the 1st reference pixels A00 have distance value (x1, y1), the 2nd reference pixels A01 have distance value (x1, y2), the 3rd reference pixels A10 have distance value (x2, y1), the 4th reference pixels A11 have distance value (x2, y2).In Fig. 2 embodiment, selected four reference pixels, still, be not limited thereto, any amount of reference pixels can be selected by resolution conversion portion 5.
At step S102, from for example importing the pixel value that data store 4 obtains each reference pixels of obtaining at described step S101.In Fig. 2 embodiment, the 1st reference pixels A00 has pixel value a00, and the 2nd reference pixels A01 has pixel value a01, and the 3rd reference pixels A10 has pixel value a10, and the 4th reference pixels A11 has pixel value a11.
At step S103, obtain the maximal value MAX of reference pixels and the difference M between the minimum value MIN.Maximal value MAX with from the pixel value of reference pixels, choose to have a peaked pixel value corresponding, minimum value MIN is corresponding with the pixel value of choosing from the pixel value of reference pixels with minimum value.This difference M can be expressed as follows formula:
M＝MAX-MIN
On the other hand, by the pixel value of immediate reference pixels and the pixel value of each other reference pixels are compared, can obtain this difference M.Described immediate reference pixels has the reference pixels of lowest distance value.For example, in Fig. 2 embodiment, if the 4th reference pixels is corresponding to immediate reference pixels, then can obtain the difference between pixel value a11 and the pixel value a10 | a11-a10|, difference between pixel value a11 and the pixel value a00 | a11-a00|, the difference between pixel value a11 and the pixel value a01 | a11-a01|.Maximal value in the resulting difference is as difference M.
At step S104, judge whether difference M equals 0, if difference M equals 0 ("Yes" of step S104), just the pixel value of all reference pixels is identical, then handles entering step S108.If difference M is not equal to 0 ("No" of step S104), then handles and enter step S105.
In step S108, use one of reference pixels pixel value as inter polated pixel pixel value.In Fig. 2 embodiment, pixel value a00 can be as the pixel value b of inter polated pixel B.But, since a00, a01, a10, a11 have identical pixel value, and wherein any can be as the pixel value b of inter polated pixel B.
In step S105, ask for the mean value AVE of the pixel value of reference pixels.In Fig. 2 embodiment, can ask for the mean value AVE of the pixel value of reference pixels by following formula:
AVE＝(a00+a01+a10+a11)/4
Then, enter step S106, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, the max pixel value of view data 1 is 255, as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/255)
Z10＝x1*y2*(1-|a10-AVE|/255)
Z01＝x2*y1*(1-|a01-AVE|/255)
Z11＝x1*y1*(1-|a11-AVE|/255)
Then, enter step S107, use the pixel value of reference pixels, calculate the pixel value of inter polated pixel.In the present embodiment, the weighting factor with the correspondence that obtains in step S106 is weighted each pixel value.
In Fig. 2 embodiment, with weighting factor Z00, Z10, Z01, Z11 be respectively to pixel value a00, a10, and a01, a11 is weighted.So the pixel value b of inter polated pixel B can obtain from following formula:
b＝Z00*a00/(Z00+Z10+Z01+Z11)+
Z10*a10/(Z00+Z10+Z01+Z11)+
Z01*a01/(Z00+Z10+Z01+Z11)+
Z11*a11)/(Z00+Z10+Z01+Z11)
Following formula can be reduced to:
b＝(Z00*a00+Z10*a10+Z01*a01+Z11*a11)/(Z00+Z10+
Z01+Z11)
At step S109, judge whether that inter polated pixels all in the view data 1 has obtained handling, if all inter polated pixels have obtained handling ("Yes" of step S109), end process then, treated view data 1 is stored in output data storage part 7, so that show by display part 8.If also have inter polated pixel not obtain handling ("No" of step S109), then return step S100, with specific another inter polated pixel.
Fig. 8 B represents to use the linear interpolation method that Fig. 8 A original image is carried out the image that interpolation generates, and Fig. 8 C represents to use the present invention's the 1st method that Fig. 8 A original image is carried out the image that interpolation generates.Compare with this Fig. 8 B, image shown in Fig. 8 C has improved the smoothness (smoothness) of image.
The calculating of in step S106, carrying out, use the action of the 2nd method identical with the action of the 1st method of use haply.
According to the 2nd method, in step S106, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, the difference M that obtains in step S103 is as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/M)
Z10＝x1*y2*(1-|a10-AVE|/M)
Z01＝x2*y1*(1-|a01-AVE|/M)
Z11＝x1*y1*(1-|a11-AVE|/M)
Fig. 8 D represents to use the 2nd method that Fig. 8 A original image is carried out the image that interpolation generates, and compares with Fig. 8 B, and image has improved the smoothness of image shown in Fig. 8 D.
Fig. 9 A represents to be added in the view data according to the inter polated pixel of one embodiment of the invention, Fig. 9 B represents when using the linear interpolation method to determine pixel value, immediate reference pixels is to the influence of the pixel value of inter polated pixel shown in Fig. 9 A, Fig. 9 C represents that immediate reference pixels is to the influence of the pixel value of inter polated pixel shown in Fig. 9 A when using the present invention's the 2nd method to determine pixel value.
Comparison diagram 9C and Fig. 9 B use the image of the 2nd method generation to compare with using the linear interpolation method as can be known, can keep the information relevant with the pixel value of original image more.With reference to Fig. 9 B, the pixel value of inter polated pixel B is subjected to the influence of immediate reference pixels (representing with white) in Fig. 9 A.But with reference to Fig. 9 C, the influence of the pixel value of immediate reference pixels has obtained inhibition.
By to using the 2nd method Figure 10 A that generates and Figure 10 B that uses the linear interpolation method to generate to compare, use the 2nd method to compare as can be known with using the linear interpolation method, can more effectively suppress cornerwise unevenness.
In the present embodiment, difference M is as normalization factor.But, be not limited thereto, as long as the pixel value of its reflection reference pixels can use other any numerical value.For example, than the big value of (MAX-AVE) value,, can use than the value that difference M is little or big than the big value of (AVE-MIN) value.
With reference to Fig. 2 and Fig. 4, the processing action of determining the pixel value of inter polated pixel according to use the 3rd method of one embodiment of the invention is described.The processing action of the 3rd method that use Fig. 4 represents is roughly identical with the processing action of the 1st method of using Fig. 3 to represent.Difference is to eliminate step S103, S104, and S108, and, replace step S105 with step S205, replace step S106 with step S206.
In step S205, calculating mean value AVE1, it is relative to each other the mean value of pixel value of a pair of reference pixels on the diagonal line.
In Fig. 2 embodiment, the 1st reference pixels A00 and the 4th reference pixels A11 form on a pair of diagonal line reference pixels relative to each other.Therefore, the mean value AVE11 of this reference pixels A00 and A11 can be calculated as follows:
AVE11＝(a00+a11)/2
Equally, the 2nd reference pixels A01 and the 3rd reference pixels A10 form on a pair of diagonal line reference pixels relative to each other.Therefore, the mean value AVE12 of this reference pixels A01 and A10 can be calculated as follows:
AVE12＝(a01+a10)/2
In step S206, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE1 that obtains at step S205, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, predefined value is bigger than the max pixel value 255 of view data 1, as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE12|/255)
Z10＝x1*y2*(1-|a10-AVE11|/255)
Z01＝x2*y1*(1-|a01-AVE11|/255)
Z11＝x1*y1*(1-|a11-AVE12|/255)
Compare with Fig. 8 B image that uses the linear interpolation method to be generated, shown in Fig. 8 E, the bright acutance of the image that uses the 3rd method to be generated is improved.
With reference to Fig. 2 and Fig. 5, the processing action of determining the pixel value of inter polated pixel according to use the 4th method of one embodiment of the invention is described.
The processing action of the 4th method that use Fig. 5 represents is roughly identical with the processing action of the 3rd method of using Fig. 4 to represent.Difference is to have increased step S303, and, replace step S206 with step S306.
In step S303, maximal value MAX and minimum value MIN according to reference pixels obtain difference M1.This difference M1 is the value of any kind of bigger than the difference M that obtains in the step S103 of Fig. 3.For example, difference M1 can be the value of representing with following formula:
M1＝MAX-MIN+α
Wherein, α is any value greater than 0, and in the present embodiment, α is set to 1.
In step S306, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE1 that obtains at step S205, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, the difference M1 that obtains in step S303 is as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE12|/M1)
Z10＝x1*y2*(1-|a10-AVE11|/M1)
Z01＝x2*y1*(1-|a01-AVE11|/M1)
Z11＝x1*y1*(1-|a11-AVE12|/M1)
Compare with Fig. 8 B image that uses the linear interpolation method to be generated, shown in Fig. 8 E, the bright acutance of the image that uses the 3rd method to be generated can be improved.
Compare with Fig. 8 B image that uses the linear interpolation method to be generated, shown in Fig. 8 F, the bright acutance of the image that uses the 4th method to be generated can be improved.
Have, compare with the linear interpolation method, shown in Fig. 9 D, the image that uses the 4th method to generate can keep the information relevant with the pixel value of original image more.With reference to Fig. 9 B, its image for using the linear interpolation method to be generated, inter polated pixel B is subjected to the influence of (representing with white) of immediate reference pixels in Fig. 9 A.But, with reference to Fig. 9 D, the image that it is generated for described the 4th method of use, the influence of immediate reference pixels has obtained inhibition.Therefore, use the 4th method to compare, can more effectively suppress cornerwise unevenness with using the linear interpolation method.
In the present embodiment, difference M1 is as normalization factor.But, as long as the pixel value of its reflection reference pixels can use any value.For example, shown in Fig. 8 G, can increase the normalization factor value, so that improve Image Smoothness.The image of image shown in Fig. 8 G for using (M1*3) to be generated as the normalization factor value, the image smoothing that this image is generated as the normalization factor value than the use M1 shown in Fig. 8 F.
With reference to Fig. 2 and Fig. 6, the processing action of determining the pixel value of inter polated pixel according to use the 5th method of one embodiment of the invention is described.
The processing action of the 5th method that use Fig. 6 represents is roughly identical with the processing action of the 3rd method of using Fig. 4 to represent.Difference is to have eliminated step S205 and S303, and, replace step S206 with step S406.
In step S406, use the pixel value that obtains at step S102, the distance value that obtains at step S101, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, use the value of predefined max pixel value 255 greater than image 1 as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-a11|/256)
Z10＝x1*y2*(1-|a10-a01|/256)
Z01＝x2*y1*(1-|a01-a10|/256)
Z11＝x1*y1*(1-|a11-a00|/256)
Shown in above-mentioned equation, replace the employed mean value AVE1 of Fig. 4 step S206, the 3rd method is used the pixel value of the reference pixels relative with target reference pixels diagonal angle, calculates the weighting factor of target reference pixels.
Compare with Fig. 8 B image that uses the linear interpolation method to be generated, shown in Fig. 8 H, the bright acutance of the image that uses the 5th method to be generated can be improved.
With reference to Fig. 2 and Fig. 7, the processing action of determining the pixel value of inter polated pixel according to use the 6th method of one embodiment of the invention is described.
The processing action of the 6th method that use Fig. 7 represents is roughly identical with the processing action of the 4th method of using Fig. 5 to represent.Difference is to have eliminated step S205, and, replace step S306 with step S506.
In step S506, use the pixel value that obtains at step S102, the distance value that obtains at step S101, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, the difference M1 that obtains at step S303 is as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-a11|/M1)
Z10＝x1*y2*(1-|a10-a01|/M1)
Z01＝x2*y1*(1-|a01-a10|/M1)
Z11＝x1*y1*(1-|a11-a00|/M1)
Shown in above-mentioned equation, replace the employed mean value AVE1 of Fig. 5 step S306, the 3rd method is used the pixel value of the reference pixels relative with target reference pixels diagonal angle, calculates the weighting factor of target reference pixels.
Compare with Fig. 8 B image that uses the linear interpolation method to be generated, shown in Fig. 8 I, the bright acutance of the image that uses the 6th method to be generated can be improved.
Have, compare with the linear interpolation method, shown in Fig. 9 E, the image that uses the 6th method to generate can keep the information relevant with the pixel value of original image more.With reference to Fig. 9 B, its image for using the linear interpolation method to be generated, inter polated pixel B is subjected to the influence of (representing with white) of immediate reference pixels in Fig. 9 A.But, with reference to Fig. 9 E, the image that it is generated for described the 6th method of use, the influence of immediate reference pixels has obtained inhibition.Therefore, use the 6th method to compare, can more effectively suppress cornerwise unevenness with using the linear interpolation method.
In the present embodiment, difference M1 is as normalization factor.But, as long as the pixel value of its reflection reference pixels can use any value.For example, shown in Fig. 8 J, can increase the normalization factor value, so that improve Image Smoothness.The image of image shown in Fig. 8 J for using (M1*1.3) to be generated as the normalization factor value, the image smoothing that this image is generated as the normalization factor value than the use M1 shown in Fig. 8 I.
With reference to Fig. 2 and Figure 11, the processing action of determining the pixel value of inter polated pixel according to use the 7th, the 8 method of one embodiment of the invention is described respectively.
The processing action of the 7th method that use Figure 11 represents is roughly identical with the processing action of the 1st method of using Fig. 3 to represent.Difference is to have increased step S605, and, replace step S106 with step S606.
In step S605, from the reference pixels that step S101 obtains, to select near reference pixels A, it is the reference pixels with lowest distance value.In the present embodiment, distance value can be with X, and the Y coordinate figure is represented.
According to the 7th method, in step S606, the pixel value that use obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor, obtain except that described near the weighting factor of each reference pixels the reference pixels A, roughly the step S106 with Fig. 3 is identical.In the present embodiment, use the max pixel value 255 of view data 1 as normalization factor.Have again, step S606 use step S101 obtain near the distance value of reference pixels A, obtain weighting factor near reference pixels A.
In Fig. 2 embodiment, if reference pixels A00 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2
Z10＝x1*y2*(1-|a10-AVE|/255)
Z01＝x2*y1*(1-|a01-AVE|/255)
Z11＝x1*y1*(1-|a11-AVE|/255)
In Fig. 2 embodiment, if reference pixels A10 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/255)
Z10＝x1*y2
Z01＝x2*y1*(1-|a01-AVE|/255)
Z11＝x1*y1*(1-|a11-AVE|/255)
In Fig. 2 embodiment, if reference pixels A01 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/255)
Z10＝x1*y2*(1-|a10-AVE|/255)
Z01＝x2*y1
Z11＝x1*y1*(1-|a11-AVE|/255)
In Fig. 2 embodiment, if reference pixels A11 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/255)
Z10＝x1*y2*(1-|a10-AVE|/255)
Z01＝x2*y1*(1-|a01-AVE|/255)
Z11＝x1*y1
Compare with Figure 14 B image that uses the linear interpolation method to be generated, shown in Figure 14 D, the bright acutance of the image that uses the 7th method to be generated can be improved.Have, the 7th method is compared with three convolution methods, can more effectively suppress the noise composition of original image again.The image of Figure 14 C is to use three convolution method generations, compares with the linear interpolation method, and it can keep the information relevant with the pixel value of original image more.But the image of Figure 14 C may comprise unnecessary information concerning the pixel value of determining inter polated pixel, so, cause existing in the image noise.
Except the practiced calculating of step S606, use the processing action of the 8th method roughly the same with the processing action of the 7th method of use.
According to the 8th method, in step S606, the pixel value that use obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor, obtain except that described near the weighting factor of each reference pixels the reference pixels A, roughly the step S106 with Fig. 3 is identical.In the present embodiment, use difference M that step S103 obtains as normalization factor.Have again, step S606 use step S101 obtain near the distance value of reference pixels A, obtain weighting factor near reference pixels A.
In Fig. 2 embodiment, if reference pixels A00 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2
Z10＝x1*y2*(1-|a10-AVE|/M)
Z01＝x2*y1*(1-|a01-AVE|/M)
Z11＝x1*y1*(1-|a11-AVE|/M)
In Fig. 2 embodiment, if reference pixels A10 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/M)
Z10＝x1*y2
Z01＝x2*y1*(1-|a01-AVE|/M)
Z11＝x1*y1*(1-|a11-AVE|/M)
In Fig. 2 embodiment, if reference pixels A01 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/M)
Z10＝x1*y2*(1-|a10-AVE|/M)
Z01＝x2*y1
Z11＝x1*y1*(1-|a11-AVE|/M)
In Fig. 2 embodiment, if reference pixels A11 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/M)
Z10＝x1*y2*(1-|a10-AVE|/M)
Z01＝x2*y1*(1-|a01-AVE|/M)
Z11＝x1*y1
Compare with Figure 14 B image that uses the linear interpolation method to be generated, shown in Figure 14 E, use image that the 8th method generated when improving bright acutance, can also keep the information relevant with the pixel value of original image.Have, shown in Figure 14 C and Figure 14 E, the 8th method is compared with three convolution methods, can more effectively suppress the noise composition of original image again.
In the present embodiment, difference M is as weighting factor.But, be not limited thereto, as long as the pixel value of its reflection reference pixels can use any value.Can use for example than the big value of (MAX-AVE) value, than the big value of (AVE-MIN) value, than the little or big value of difference M.
With reference to Fig. 2 and Figure 12, the processing action of determining the pixel value of inter polated pixel according to use the 9th method of one embodiment of the invention is described.
The processing action of the 9th method that use Figure 12 represents is roughly identical with the processing action of the 3rd method of using Fig. 4 to represent.Difference is to have increased step S605, and, replace step S206 with step S706.
In step S706, the pixel value that use obtains at step S102, the distance value that obtains at step S101, the mean value AVE1 that obtains at step S205, and normalization factor, obtain except that described near the weighting factor of each reference pixels the reference pixels A, roughly the step S206 with Fig. 4 is identical.In the present embodiment, use the big value of predefined max pixel value 255 than view data 1 as normalization factor.Have again, step S706 use step S101 obtain near the distance value of reference pixels A, obtain weighting factor near reference pixels A.
In Fig. 2 embodiment, if reference pixels A00 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2
Z10＝x1*y2*(1-|a10-AVE11|/256)
Z01＝x2*y1*(1-|a01-AVE11|/256)
Z11＝x1*y1*(1-|a11-AVE12|/256)
In Fig. 2 embodiment, if reference pixels A10 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE12|/256)
Z10＝x1*y2
Z01＝x2*y1*(1-|a01-AVE11|/256)
Z11＝x1*y1*(1-|a11-AVE12|/256)
In Fig. 2 embodiment, if reference pixels A01 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE12|/256)
Z10＝x1*y2*(1-|a10-AVE11|/256)
Z01＝x2*y1
Z11＝x1*y1*(1-|a11-AVE12|/256)
In Fig. 2 embodiment, if reference pixels A11 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE12|/256)
Z10＝x1*y2*(1-|a10-AVE11|/256)
Z01＝x2*y1*(1-|a01-AVE11|/256)
Z11＝x1*y1
Compare with Figure 14 B image that uses the linear interpolation method to be generated, shown in Figure 14 F, the image that uses the 9th method to be generated can improve bright acutance.Have, shown in Figure 14 C and Figure 14 F, the 9th method is compared with three convolution methods, can more effectively suppress the noise composition of original image again.
With reference to Fig. 2 and Figure 13, the processing action of determining the pixel value of inter polated pixel according to use the 10th method of one embodiment of the invention is described.
The processing action of the 10th method that use Figure 13 represents is roughly identical with the processing action of the 5th method of using Fig. 6 to represent.Difference is to have increased step S605, and, replace step S406 with step S806.
According to the 10th method, in step S806, use the pixel value that obtains at step S102, the distance value that obtains at step S101, and normalization factor, obtain except that described near the weighting factor of each reference pixels the reference pixels A, roughly the step S406 with Fig. 6 is identical.In the present embodiment, use the big value of predefined max pixel value 255 than view data 1 as normalization factor.Have again, step S806 use step S101 obtain near the distance value of reference pixels A, obtain weighting factor near reference pixels A.
In Fig. 2 embodiment, if reference pixels A00 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2
Z10＝x1*y2*(1-|a10-a01|/256)
Z01＝x2*y1*(1-|a01-a10|/256)
Z11＝x1*y1*(1-|a11-a00|/256)
In Fig. 2 embodiment, if reference pixels A10 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-a11|/256)
Z10＝x1*y2
Z01＝x2*y1*(1-|a01-a10|/256)
Z11＝x1*y1*(1-|a11-a00|/256)
In Fig. 2 embodiment, if reference pixels A01 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-a11|/256)
Z10＝x1*y2*(1-|a10-a01|/256)
Z01＝x2*y1
Z11＝x1*y1*(1-|a11-a00|/256)
In Fig. 2 embodiment, if reference pixels A11 is equivalent to then to use following formula to calculate reference pixels A00 respectively near reference pixels A, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-a11|/256)
Z10＝x1*y2*(1-|a10-a01|/256)
Z01＝x2*y1*(1-|a01-a10|/256)
Z11＝x1*y1
Compare with Figure 14 B image that uses the linear interpolation method to be generated, shown in Figure 14 G, the image that uses the 10th method to be generated can improve bright acutance.Have, shown in Figure 14 C and Figure 14 G, the 10th method is compared with three convolution methods, can more effectively suppress the noise composition of original image again.
With reference to Fig. 2 and Figure 15, the processing action of determining the pixel value of inter polated pixel according to use the 11st, 12,13 methods of one embodiment of the invention is described respectively.
The processing action of the 11st, 12,13 methods that use Figure 15 represents is roughly identical with the processing action of the 2nd method of using Fig. 3 to represent.Difference is to replace step S106 with step S906.
According to the 11st method, in step S906, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, use the difference M in step S103, obtain as normalization factor.Have again, in the present embodiment, become the n power of distance value.This value n is any number greater than 1, is preferably greater than 2.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝(x2*y2)
n*(1-|a00-AVE|/M)
Z10＝(x1*y2)
n*(1-|a10-AVE|/M)
Z01＝(x2*y1)
n*(1-|a01-AVE|/M)
Z11＝(x1*y1)
n*(1-|a11-AVE|/M)
With use the present invention 1-the 10th method in one of Figure 16 B image of being generated compare, shown in Figure 16 C, the bright acutance of the image that uses the 11st method to be generated can be improved.Have, the 11st method is compared with known method again, can more effectively improve the bright acutance and the smoothness of original image.Described known method comprises for example linear interpolation method (with reference to Figure 16 F), three convolution methods (with reference to Figure 16 G), the method for patent documentation 1 (with reference to Figure 16 H).
Except the practiced computing of step S906, use the processing action of the 12nd method roughly identical with the processing action of the 11st method of use.
According to the 12nd method, in step S906, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, use the difference M in step S103, obtain as normalization factor.Have, in the present embodiment, pixel value becomes the n power again.This value n is any number greater than 1, is preferably greater than 2.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝(x2*y2)*(1-|a00-AVE|/M)
n
Z10＝(x1*y2)*(1-|a10-AVE|/M)
n
Z01＝(x2*y1)*(1-|a01-AVE|/M)
n
Z11＝(x1*y1)*(1-|a11-AVE|/M)
n
With use the present invention 1-the 10th method in one of Figure 16 B image of being generated compare, shown in Figure 16 D, use image that the 12nd method generated when improving bright acutance, can also keep the smoothness of image.Have, the 12nd method is compared with known method again, can more effectively improve the bright acutance and the smoothness of original image.Described known method comprises for example linear interpolation method (with reference to Figure 16 F), three convolution methods (with reference to Figure 16 G), the method for patent documentation 1 (with reference to Figure 16 H).
Except the practiced computing of step S906, use the processing action of the 13rd method roughly identical with the processing action of the 11st method of use.
According to the 13rd method, in step S906, use the pixel value that obtains at step S102, the distance value that obtains at step S101, the mean value AVE that obtains at step S105, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, use the difference M in step S103, obtain as normalization factor.Have, in the present embodiment, distance value becomes the n power again, and pixel value becomes the p power.This value n and value p are any numbers greater than 1, are preferably greater than 2.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝(x2*y2)
n*(1-|a00-AVE|/M)
p
Z10＝(x1*y2)
n*(1-|a10-AVE|/M)
p
Z01＝(x2*y1)
n*(1-|a01-AVE|/M)
p
Z11＝(x1*y1)
n*(1-|a11-AVE|/M)
p
If n equals p, then following formula can be reduced to following formula:
Z00＝[(x2*y2)*(1-|a00-AVE|/M)]
n
Z10＝[(x1*y2)*(1-|a10-AVE|/M)]
n
Z01＝[(x2*y1)*(1-|a01-AVE|/M)]
n
Z11＝[(x1*y1)*(1-|a11-AVE|/M)]
n
With use the present invention 1-the 10th method in one of Figure 16 B image of being generated compare, shown in Figure 16 E, use image that the 13rd method generated when improving bright acutance, can also keep the smoothness of image.Have, the 13rd method is compared with known method again, can more effectively improve the bright acutance and the smoothness of original image.Described known method comprises for example linear interpolation method (with reference to Figure 16 F), three convolution methods (with reference to Figure 16 G), the method for patent documentation 1 (with reference to Figure 16 H).
In the present embodiment, by changing described power value n, p can adjust the bright acutance and the smoothness of image.For example, by added value n, the pixel value of inter polated pixel can more be subjected to influencing near the pixel value of reference pixels of it.Therefore, the bright acutance of image is improved.By added value p, the pixel value of inter polated pixel can be influenced by the average pixel value of general image more.Therefore, the smoothness of image is improved.
According to any method in said method of the present invention and other methods, resolution conversion portion 5 can be stored in the result of calculation of transform data storage part 6.Perhaps the image processing apparatus 9 of Fig. 1 can append data store is provided, and storage comprises the various data of the result of calculation that is generated by resolution conversion portion 5.Especially when as any method in the above-mentioned 11-13 method, when carrying out the power of pixel value or distance value, this supplemental data storage part can improve the processing speed of image processing apparatus 9.
For example, as shown in figure 17, append Value Data storage part 11 and can be added and be set to image processing apparatus 9.This appends Value Data storage part 11 and stores some weighting factors before the power and some values of appending as LUT in corresponding mode.Resolution conversion portion 5 can use from this and append the value of appending that Value Data storage part 11 obtains, and is multiplied by described weighting factor.
With reference to Fig. 2 and Figure 18, the processing action of determining the pixel value of inter polated pixel according to use the 14th, 15 method of one embodiment of the invention is described respectively.
The processing action of the 14th method that use Figure 18 represents is roughly identical with the processing action of the 1st method of using Fig. 3 to represent.Difference is to have increased step S1007 and S1008.
According to the 14th method, in step S1007,, obtain the value of appending of each weighting factor that obtains at step S106 from being stored in the described LUT that appends Value Data storage part 11.
In step S1008, on duty with appending of correspondence with weighting factor, so that the weighting factor that has been appended.
Except the calculating of in step S1008, carrying out, use the processing action of the 15th method roughly the same with the processing action of the 14th method of use.
Step S1008 increases by the value of appending on weighting factor, so that obtain the weighting factor that multiplies each other.
Use the described the 14th or the 15th method, the processing speed of resolution conversion portion 5 can be improved.
Except above-mentioned 1-15 method, resolution conversion portion 5 also can carry out any interpolating method according to instructions of the present invention and claims scope.For example, in instructions of the present invention and claims scope, the key element of said method, feature, function can make up mutually, or replaces mutually.For example, at least a in said method of the present invention and other method can be selected according to user's situation by resolution conversion portion 5.What perhaps resolution conversion portion 5 can be according in the feature selecting of integral body or topography said method of the present invention and other method is at least a.In order to carry out above-mentioned selection, can append selection portion in the image processing apparatus 9, be used for selecting at least a of said method of the present invention and other method according to user's situation or characteristics of image.
Have, except the method and other method of the invention described above, resolution conversion portion 5 also can carry out any known interpolating method again, for example, near interpolation, linear interpolation method, three convolution methods etc.For example, at least a in said method of the present invention and other method and above-mentioned known interpolating method can be selected according to user's situation by resolution conversion portion 5.What perhaps resolution conversion portion 5 can be according in the feature selecting of integral body or topography said method of the present invention and other method is at least a, and above-mentioned known interpolating method.In order to carry out above-mentioned selection, can append selection portion in the image processing apparatus 9, be used for selecting at least a of said method of the present invention and other method according to user's situation or characteristics of image, and above-mentioned known interpolating method.
With reference to Fig. 2 and Figure 19, the processing action of determining the pixel value of inter polated pixel according to use the 16th method of one embodiment of the invention is described.
The processing action of the 16th method that use Figure 19 represents is roughly identical with the processing action of the 1st method of using Fig. 3 to represent.Difference is to have increased step S1105 and step S1107, and, replace step S106 with step S906.
In step S1105, judge whether the difference M that obtains at step S103 is equal to or greater than a predefined selective value.If difference M, then handles action less than this predefined selective value ("No" of step S1105) and enters step S1107.If difference M is equal to or greater than this predefined selective value ("Yes" of step S1105), then handles action and enter step S105.
As above-mentioned description for Fig. 3, difference M represents poor between the maximal value MAX of reference pixels and the minimum value MIN.In the present embodiment, selective value is equivalent to any pixel value among the 0-255, can be used to select for handling the suitable interpolating method of specific inter polated pixel.For example, this selective value can be chosen as pixel value 128.Have again, in the present embodiment, suppose resolution conversion portion 5 selection the 12nd method of the present invention or linear interpolation methods.But resolution conversion portion 5 can select a kind of interpolating method at least, and it comprises any method of method of the present invention and known method.
As mentioned above, if the difference M of described reference pixels less than selective value, then the pixel value of resolution conversion portion 5 hypothesis reference pixels changes smaller.According to this feature, select linear interpolation, it is fit to improve the smoothness of image.Pixel value has the image example of less variation can enumerate the image with literal or line.
If the difference M of reference pixels is equal to or greater than selective value, then the pixel value of resolution conversion portion 5 hypothesis reference pixels changes greatly.According to this feature, select the 12nd method, it is fit to improve the bright acutance of image.Pixel value has the image example of bigger variation can enumerate the image with photo.
In step S107, use the linear interpolation method, obtain the weighting factor of each reference pixels.In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2
Z10＝x1*y2
Z01＝x2*y1
Z11＝x1*y1
In step S906, use the 12nd method of the present invention, obtain the weighting factor of each reference pixels.In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/M)
n
Z10＝x1*y2*(1-|a10-AVE|/M)
n
Z01＝x2*y1*(1-|a01-AVE|/M)
n
Z11＝x1*y1*(1-|a11-AVE|/M)
n
In the present embodiment, power value n is set at 3, still, is not limited thereto, and can use any number greater than 1, is preferably greater than 2.
In step S107, use the pixel value of reference pixels, calculate the pixel value of inter polated pixel.In the present embodiment, each pixel value is used in the weighting factor that obtains among step S906 or the step S1107 and is weighted.
In Fig. 2 embodiment, if select linear interpolation, pixel value a00 then, a10, a01, a11 use weighting factor Z00 respectively, Z10, Z01, Z11 is weighted.So the pixel value b of inter polated pixel B can use following formula to try to achieve:
b＝(Z00*a00+Z10*a10+Z01*a01+Z11*a11)/(Z00+Z10+
Z01+Z11)
Because the pixel value sum (Z00+Z10+Z01+Z11) of reference pixels is 1, following formula can further be reduced to following formula:
b＝Z00*a00+Z10*a10+Z01*a01+Z11*a11
In Fig. 2 embodiment, if select the 12nd method, pixel value a00 then, a10, a01, a11 use weighting factor Z00 respectively, Z10, Z01, Z11 is weighted.So the pixel value b of inter polated pixel B can use following formula to try to achieve:
b＝(Z00*a00+Z10*a10+Z01*a01+Z11*a11)/(Z00+Z10+
Z01+Z11)
Use the 16th method, shown in Figure 25 A-25D, can be according to topography's characteristic, the smoothness of control chart picture and bright acutance.The top of the original image shown in Figure 25 A has photograph image, and the bottom has oblique line and literal.If shown in Figure 25 B, use linear interpolation method is carried out interpolation to the original image of Figure 25 A, and then the bottom graph picture is because of fuzzy and uneven sustaining damage.If shown in Figure 25 C, use the present invention the 12nd method is carried out interpolation to the original image of Figure 25 A, and then upper image sustains damage because of uneven.By making up above-mentioned linear interpolation method and the 12nd method, shown in Figure 25 D, the 16th method can generate the upper image with good smoothness, and the bottom graph picture with the sharp property of good aquatic foods.
Have, in this embodiment, step shown in Figure 19 is shared linear interpolation and the 12nd method again, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 20, the processing action of determining the pixel value of inter polated pixel according to use the 17th method of one embodiment of the invention is described.
The processing action of the 17th method that use Figure 20 represents is roughly identical with the processing action of the 16th method of using Figure 19 to represent.Difference is to replace step S1105 with step S1205.
In step S1205, judge whether the pixel value of each reference pixels equals the maximal value MAX or the minimum value MIN of reference pixels.If the pixel value of reference pixels equals the maximal value MAX or the minimum value MIN ("Yes" of step S1205) of reference pixels, then handle action and enter step S105.If the pixel value of reference pixels is not equal to the maximal value MAX or the minimum value MIN ("No" of step S1205) of reference pixels, then handles action and enter step S1107.
In the present embodiment, if the pixel value of each reference pixels equals the maximal value MAX or the minimum value MIN of reference pixels, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are binary pictures.According to this feature, select the 12nd method, this method is suitable for improving the bright acutance of image.
In the present embodiment, if the pixel value of each reference pixels is not equal to the maximal value MAX or the minimum value MIN of reference pixels, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are multivalue images.According to this feature, select linear interpolation, this method is suitable for improving the smoothness of image.
As Figure 25 A, 25B, 25C shown in the 25E, uses the 17th method, can be according to the smoothness and the bright acutance of topography's feature control chart picture.By combination linear interpolation (Figure 25 B) and the 12nd method (Figure 25 C), shown in Figure 25 E, the bottom graph picture that the 17th method can generate the upper image with smoothness and have bright acutance.
Have, in this embodiment, step shown in Figure 20 is shared linear interpolation and the 12nd method again, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 21, the processing action of determining the pixel value of inter polated pixel according to use the 18th method of one embodiment of the invention is described.
The processing action of the 18th method that use Figure 21 represents is roughly identical with the processing action of the 16th method of using Figure 19 to represent.Difference is to have increased step S1304, and replaces step S1105 with step S1305.
In step S1304, stipulated range of choice, can use this scope, be used to select to be suitable for handling the interpolating method of specific inter polated pixel.This range of choice can be stipulated according to a predefined constant M2.This preestablishes constant M2 can be arbitrary value, and still, in the present embodiment, this preestablishes constant M2 and stipulates by following formula according to difference M:
M2＝M/S
Wherein, S is the arbitrary value greater than 2.
According to this predefined constant M2, below range of choice can be stipulated:
Greater than (MAX-M2) or less than (MIN+M2).
Above-mentioned value MAX is equivalent to the maximal value of reference pixels, and above-mentioned value MIN is equivalent to the minimum value of reference pixels.
In step S1305, whether the pixel value of judging each reference pixels is within the range of choice of described step S1304 regulation.If the pixel value of reference pixels is ("Yes" of step S1305) within described range of choice, then handle action and enter step S105.If the pixel value of reference pixels is ("No" of step S1305) not within described range of choice, then handle action and enter step S1107.
In the present embodiment, if the pixel value of each reference pixels within described range of choice, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are that pixel value changes little image, for example grey scale image.According to this feature, select the 12nd method, this method is suitable for improving the bright acutance of image.
In the present embodiment, if the pixel value of each reference pixels not within described range of choice, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are multivalue images.According to this feature, select linear interpolation, this method is suitable for improving the smoothness of image.
As Figure 25 A, 25B, 25C shown in the 25F, uses the 18th method, can be according to the smoothness and the bright acutance of topography's feature control chart picture.By combination linear interpolation (Figure 25 B) and the 12nd method (Figure 25 C), shown in Figure 25 F, the bottom graph picture that the 18th method can generate the upper image with smoothness and have bright acutance.In order to generate image shown in Figure 25 F, setting value S is 8, and value n is 3.
Have, in this embodiment, step shown in Figure 21 is shared linear interpolation and the 12nd method again, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 22, the processing action of determining the pixel value of inter polated pixel according to use the 19th method of one embodiment of the invention is described.
The processing action of the 19th method that use Figure 22 represents is roughly identical with the processing action of the 16th method of using Figure 19 to represent.Difference is to replace step S1105 with step S1115, replaces step S1107 with step S106.Have, sequence of steps shown in Figure 22 is different with sequence of steps shown in Figure 19 again.
In step S1115, judge whether the difference M that obtains at step S103 is equal to or greater than a predefined selective value.If difference M, then handles action less than this selective value ("No" of step S1115) and enters step S106.If difference M is equal to or greater than this predefined selective value ("Yes" of step S1115), then handles action and enter step S906.
As above-mentioned description for Fig. 3, difference M represents poor between the maximal value MAX of reference pixels and the minimum value MIN.In the present embodiment, selective value is equivalent to any pixel value among the 0-255, can be used to select for handling the suitable interpolating method of specific inter polated pixel.For example, this selective value can be chosen as pixel value 128.Have again, in the present embodiment, suppose resolution conversion portion 5 selection the 12nd method of the present invention or the 2nd methods.But the present invention is not limited thereto, and at least a interpolating method can be selected by resolution conversion portion 5 from method of the present invention and known method.
As mentioned above, if the difference M of described reference pixels less than selective value, then the pixel value of resolution conversion portion 5 hypothesis reference pixels changes smaller.According to this feature, select the 2nd method, it is fit to improve the smoothness of image.
If the difference M of reference pixels is equal to or greater than selective value, then the pixel value of resolution conversion portion 5 hypothesis reference pixels changes greatly.According to this feature, select the 12nd method, it is fit to improve the bright acutance of image.
In step S106, use the 2nd method, obtain the weighting factor of each reference pixels.In Fig. 2 embodiment, use following formula to calculate reference pixels A00 respectively, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝x2*y2*(1-|a00-AVE|/M)
Z10＝x1*y2*(1-|a10-AVE|/M)
Z01＝x2*y1*(1-|a01-AVE|/M)
Z11＝x1*y1*(1-|a11-AVE|/M)
In step S107, use the pixel value of reference pixels, calculate the pixel value of inter polated pixel.In the present embodiment, each pixel value is used in the weighting factor that obtains among step S906 or the step S106 and is weighted.
In Fig. 2 embodiment, pixel value a00, a10, a01, a11 use weighting factor Z00 respectively, Z10, Z01, Z11 is weighted.So the pixel value b of inter polated pixel B can use following formula to try to achieve:
b＝(Z00*a00+Z10*a10+Z01*a01+Z11*a11)/(Z00+Z10+
Z01+Z11)
Use the 19th method, as Figure 25 A, 25C, 25G, 25H, 25K, shown in the 25L, can be according to topography's characteristic, the smoothness of control chart picture and bright acutance.The top of the original image shown in Figure 25 A has photograph image, and the bottom has oblique line and literal.If shown in Figure 25 G, use the 2nd method that the original image of Figure 25 A is carried out interpolation, compare with the image that method generated in the use patent documentation 1 shown in Figure 25 K, or compare, can improve Image Smoothness and bright acutance with the image that three times convolution method generated of the use shown in Figure 25 L.But the bottom graph picture of this Figure 25 G image sustains damage because of fuzzy and unevenness.
If shown in Figure 25 C, use the 12nd method that the original image of Figure 25 A is carried out interpolation, then the upper image of this image sustains damage because of unevenness.By making up above-mentioned the 2nd method (Figure 25 G) and the 12nd method (Figure 12 C), shown in Figure 25 H, the 19th method can generate the upper image with good smoothness, and the bottom graph picture with the sharp property of good aquatic foods.
Have again, in this embodiment, shared the 2nd method of step shown in Figure 22 and the 12nd method, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 23, the processing action of determining the pixel value of inter polated pixel according to use the 20th method of one embodiment of the invention is described.
The processing action of the 20th method that use Figure 23 represents is roughly identical with the processing action of the 17th method of using Figure 20 to represent.Difference is to replace step S1205 with step S1215, replaces step S1107 with step S106.Have, sequence of steps shown in Figure 23 is different with sequence of steps shown in Figure 20 again.
In step S1215, judge whether the pixel value of each reference pixels equals the maximal value MAX or the minimum value MIN of reference pixels.If the pixel value of reference pixels equals the maximal value MAX or the minimum value MIN ("Yes" of step S1215) of reference pixels, then handle action and enter step S906.If the pixel value of reference pixels is not equal to the maximal value MAX or the minimum value MIN ("No" of step S1215) of reference pixels, then handles action and enter step S106.
In the present embodiment, if the pixel value of each reference pixels equals the maximal value MAX or the minimum value MIN of reference pixels, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are binary pictures.According to this feature, select the 12nd method, this method is suitable for improving the bright acutance of image.
In the present embodiment, if the pixel value of each reference pixels is not equal to the maximal value MAX or the minimum value MIN of reference pixels, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are multivalue images.According to this feature, select the 2nd method, this method is suitable for improving the smoothness of image.
As Figure 25 A, 25C, 25G shown in the 25I, uses the 20th method, can be according to the smoothness and the bright acutance of topography's feature control chart picture.By making up the 2nd method (Figure 25 G) and the 12nd method (Figure 25 C), shown in Figure 25 I, the bottom graph picture that the 20th method can generate the upper image with smoothness and have bright acutance.
Have again, in this embodiment, shared the 2nd method of step shown in Figure 23 and the 12nd method, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 24, the processing action of determining the pixel value of inter polated pixel according to use the 21st method of one embodiment of the invention is described.
The processing action of the 21st method that use Figure 24 represents is roughly identical with the processing action of the 18th method of using Figure 21 to represent.Difference is to replace step S1305 with step S1315, replaces step S1107 with step S106.Have, sequence of steps shown in Figure 24 is different with sequence of steps shown in Figure 21 again.
In step S1315, whether the pixel value of judging each reference pixels is within the range of choice of described step S1304 regulation.If the pixel value of reference pixels is ("Yes" of step S1315) within described range of choice, then handle action and enter step S906.If the pixel value of reference pixels is ("No" of step S1315) not within described range of choice, then handle action and enter step S106.
In the present embodiment, if the pixel value of each reference pixels within described range of choice, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are that pixel value changes little image, for example grey scale image.According to this feature, select the 12nd method, this method is suitable for improving the bright acutance of image.
In the present embodiment, if the pixel value of each reference pixels not within described range of choice, then resolution conversion portion 5 supposition original images or the part that has this reference pixels at least are multivalue images.According to this feature, select the 2nd method, this method is suitable for improving the smoothness of image.
As Figure 25 A, 25C, 25G shown in the 25J, uses the 21st method, can be according to the smoothness and the bright acutance of topography's feature control chart picture.By making up the 2nd method (Figure 25 G) and the 12nd method (Figure 25 C), shown in Figure 25 J, the bottom graph picture that the 21st method can generate the upper image with smoothness and have bright acutance.In order to generate image shown in Figure 25 J, setting value S is 8, and value n is 3.
Have, in this embodiment, step shown in Figure 24 is shared the 2nd method and the 12nd method again, so, improved bulk treatment speed.
As described in any method in the above-mentioned 16-21 method, smoothness and bright acutance that resolution conversion portion 5 can the control chart picture.In another embodiment, resolution conversion portion 5 can control the information relevant with the pixel value of original image, and this information can be used for determining the pixel value of inter polated pixel.
With reference to Fig. 2 and Figure 27, the processing action of determining the pixel value of inter polated pixel according to use the 22nd method of one embodiment of the invention is described.
In step S100, a specific inter polated pixel.For example shown in Figure 2, the 5 specific inter polated pixel B of resolution conversion portion.
In step S101, around this specific inter polated pixel, select one or more reference pixels, this reference pixels is the data of original digital image data 1.In step S101, also obtain the distance value of each reference pixels.So in Fig. 2 embodiment, the 4th reference pixels A00 of 1-, A01, A10, A11 select in resolution conversion portion 5 around inter polated pixel B.To each reference pixels A00, A01, A10, A11, resolution conversion portion 5 obtains with X, the distance value that the Y coordinate figure is represented.In this embodiment, the 1st reference pixels A00 have distance value (x1, y1), the 2nd reference pixels A01 have distance value (x1, y2), the 3rd reference pixels A10 have distance value (x2, y1), the 4th reference pixels A11 have distance value (x2, y2).In Fig. 2 embodiment, selected four reference pixels, still, be not limited thereto, any amount of reference pixels can be selected by resolution conversion portion 5.
At step S102, obtain the pixel value of each reference pixels of obtaining at described step S101.In Fig. 2 embodiment, the 1st reference pixels A00 has pixel value a00, and the 2nd reference pixels A01 has pixel value a01, and the 3rd reference pixels A10 has pixel value a10, and the 4th reference pixels A11 has pixel value a11.
In step S1401, obtain the distance value of each reference pixels, they are different with resulting distance value among the described step S101.In the present embodiment, resolution conversion portion 5 is according to using X, the distance value of Y coordinate representation, and calculated line distance value L, it is the air line distance between each reference pixels and the inter polated pixel.As shown in figure 26, the 1st reference pixels A00 has air line distance value L00=(x1*y1), the 2nd reference pixels A01 has air line distance value L01=(x2*y1), the 3rd reference pixels A10 has air line distance value L10=(x1*y2), and the 4th reference pixels A11 has air line distance value L11=(x2*y2).
In step S1404, judge whether the pixel value of immediate reference pixels equals in abutting connection with the pixel value of at least one of reference pixels.If the pixel value of immediate reference pixels equals any one pixel value ("Yes" of step S1404) in abutting connection with reference pixels, then handle action and enter step S1408.Otherwise ("No" of step S1404) then handled action and entered step S1402.
In the present embodiment, described immediate reference pixels is equivalent to have the pixel of short lines distance value L.Describedly be equivalent to along X or Y direction and specific inter polated pixel pixel adjacent in abutting connection with reference pixels.
If the pixel value of immediate reference pixels equals any one pixel value in abutting connection with reference pixels, then resolution conversion portion 5 supposition reference pixels have identical or approaching value, and suppose that further the part with this reference pixels is equivalent to for example literal or the symbolic component of view data 1.According to this feature, to select near interpolation, it is suitable for keeping the Pixel Information of original image.
If the pixel value of immediate reference pixels is not equal to any one pixel value in abutting connection with reference pixels, then resolution conversion portion 5 supposition reference pixels have different pixel values, and suppose that further the part with this reference pixels is equivalent to for example photograph image or oblique line.According to this feature, select the 2nd method of the present invention or other method, it is suitable for improving the smoothness of image.
In step S1408, use near the pixel value of reference pixels pixel value as inter polated pixel.In Fig. 2 embodiment,, then can use the pixel value a11 of the 4th reference pixels A11 if pixel value a11 is not equal in the pixel value a01 of reference pixels A01 and A10 and the a10 any one.
In step S1402, obtain the max pixel value MAX and the minimum pixel value MIN of reference pixels.This max pixel value MAX is equivalent to have the pixel value of the reference pixels of max pixel value, and this minimum pixel value MIN is equivalent to have the pixel value of the reference pixels of minimum pixel value.
In step S1403,, obtain the difference M1 of reference pixels according to described max pixel value MAX and minimum pixel value MIN.In the present embodiment, this difference M1 can be expressed as follows formula:
M1＝MAX-MIN+α
Wherein, α is the arbitrary value greater than 0.
In step S105, ask for the mean value AVE of the pixel value of reference pixels.In Fig. 2 embodiment, each reference pixels A00, A01, A10, the mean value AVE of A11 can ask for by following formula:
AVE＝(a00+a01+a10+a11)/4
Then,, use the pixel value that obtains at step S102 at step S1406, the air line distance value L that obtains at step S1401, the mean value AVE that obtains at step S105, and normalization factor obtain the weighting factor of each reference pixels.In the present embodiment, the difference M1 that will obtain in step S1403 is as normalization factor.
In Fig. 2 embodiment, use following formula to calculate reference pixels A00, A10, A01, the weighting factor Z00 of A11, Z10, Z01, the value of Z11:
Z00＝L11*(1-|a00-AVE|/M1)
Z10＝L01*(1-|a10-AVE|/M1)
Z01＝L10*(1-|a01-AVE|/M1)
Z11＝L00*(1-|a11-AVE|/M1)
Shown in above-mentioned equation, in the present embodiment, step S1406 uses the 2nd method shown in Figure 3.But, be not limited thereto, can use any method in above-mentioned or other method of the present invention, as long as it is suitable for improving Image Smoothness.
In step S107, use the pixel value of reference pixels, calculate the pixel value of inter polated pixel.In the present embodiment, the weighting factor with the correspondence that obtains in step S1406 is weighted each pixel value.
In Fig. 2 embodiment, with weighting factor Z00, Z10, Z01, Z11 be respectively to pixel value a00, a10, and a01, a11 is weighted.So the pixel value b of inter polated pixel B can obtain from following formula:
b＝a00*Z00/(Z00+Z10+Z01+Z11)+
a10*Z10/(Z00+Z10+Z01+Z11)+
a01*Z01/(Z00+Z10+Z01+Z11)+
a11*Z11/(Z00+Z10+Z01+Z11)
Following formula can be reduced to:
b＝(Z00*a00+Z10*a10+Z01*a01+Z11*a11)/(Z00+Z10+
Z01+Z11)
At step S109, judge whether that inter polated pixels all in the view data 1 has obtained handling, if all inter polated pixels have obtained handling ("Yes" of step S109), end process then, treated view data 1 is stored in output data storage part 7, so that show by display part 8.If also have inter polated pixel not obtain handling ("No" of step S109), then return step S100, with specific another inter polated pixel.
As Figure 31 A, 31B shown in the 31C, uses the 22nd method, according to topography's feature, can control smoothness and the information relevant with the original image pixel.Figure 31 A, 31B, any one image shown in the 31C generate according to the original image of Figure 25 A.Use very the most effective to the maintenance information relevant with the original image pixel value near image shown in Figure 31 A of interpolation generation.But, owing to increased noise composition, the image border unevenness of this Figure 31 A.Image is very effective to improving Image Smoothness shown in Figure 31 B that use the present invention the 2nd method generates.Shown in Figure 31 C, described by making up near interpolation (Figure 31 A) and the 2nd method (Figure 31 B), the bottom graph picture that the 22nd method can generate the upper image with smoothness and have bright acutance.
In the present embodiment, step S1404 judges the most whether pixel value near reference pixels equals any one pixel value in the reference pixels.Perhaps can judge and whether equal any one pixel value in the reference pixels near the pixel value of the relative reference pixels in reference pixels diagonal angle at step S1404.
Have, in this embodiment, shared the 2nd method of step shown in Figure 27 reaches near interpolation again, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 28, the processing action of determining the pixel value of inter polated pixel according to use the 23rd method of one embodiment of the invention is described.
The processing action of the 23rd method that use Figure 28 represents is roughly identical with the processing action of the 22nd method of using Figure 27 to represent.Difference is to have increased step S1503, and replaces step S1404 with step S1504.
Step S1503 stipulates range of choice, and it can be used to select to be suitable for handling the interpolating method of specific inter polated pixel.This range of choice can be according to stipulating near the pixel value of reference pixels.For example, in Fig. 2 embodiment, if the 4th pixel value a11 is equivalent to the nearest pixel value, then range of choice may be prescribed as the scope between (a11+ β) and (a11-γ).Above-mentioned β value is the arbitrary value greater than pixel value a11, and above-mentioned γ value is the arbitrary value less than pixel value a11.Be preferably, above-mentioned β value and γ value sum are less than peaked 50% of reference pixels.
In step S1504, judge any in abutting connection with the pixel value of reference pixels whether within the range of choice of described step S1503 regulation.In Fig. 2 embodiment, the 2nd and the 3rd reference pixels A01, A10 be equivalent to inter polated pixel B in abutting connection with reference pixels.Each pixel value a01 judges in resolution conversion portion 5, and whether a10 is within described range of choice.If the pixel value of adjacent pixels is ("Yes" of step S1504) within described range of choice, then handle action and enter step S1408.If the pixel value of adjacent pixels is ("No" of step S1504) not within described range of choice, then handle action and enter step S1402.
As Figure 31 A, 31B shown in the 31D, uses the 23rd method, can be according to the smoothness of topography's feature control chart picture and the information relevant with the original image pixel value.By making up near interpolation (Figure 31 A) and the 2nd method (Figure 31 B), shown in Figure 31 D, the bottom graph picture that the 23rd method can generate the upper image with smoothness and have bright acutance.
As Figure 32 A, 32B, shown in the 32C, especially when image had deep or light grade, the 23rd method can improve the bright acutance of image.Shown in Figure 32 B, if use the 22nd method when former deep or light grade image carries out interpolation shown in Figure 32 A, the bright acutance of image can not be greatly improved.Shown in Figure 32 C, if use the 23rd method when former deep or light grade image carries out interpolation shown in Figure 32 A, to compare with Figure 32 B image, the bright acutance of image can be improved.
Have, in this embodiment, step shown in Figure 28 is that shared the 2nd method reaches near interpolation again, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 29, the processing action of determining the pixel value of inter polated pixel according to use the 24th method of one embodiment of the invention is described.
The processing action of the 24th method that use Figure 29 represents is roughly identical with the processing action of the 23rd method of using Figure 28 to represent.Difference is to have increased step S1602, and replaces step S1503 with step S1603, replaces step S1504 with step S1604.Have, sequence of steps shown in Figure 29 can be different with sequence of steps shown in Figure 28 again.
In step S1602, use the maximal value MAX and the minimum value MIN of reference pixels, obtain difference M.Perhaps in step S1602, can by near the pixel value of reference pixels and except that near the pixel value of the reference pixels the reference pixels, obtain difference M.For example, in Fig. 2 embodiment, can obtain the poor of pixel value a11 and a10 respectively | a11-a10|, pixel value a11 and a01's is poor | a11-a01|, pixel value a11 and a00's is poor | a11-a00|.The maximal value of gained difference is as difference M.
In step S1603, the regulation range of choice, this range of choice can be by the 1st constant M3 and the 2nd constant M4 regulation.The 1st constant M3 can be an arbitrary value, and in the present embodiment, described the 1st constant M3 presses following formula and determines according to difference M:
M3＝M/E
Wherein, E is equal to or greater than 2 arbitrary value.
Described the 2nd constant M4 can be an arbitrary value, and in the present embodiment, described the 2nd constant M4 presses following formula and determines according to difference M:
M3＝M/F
Wherein, F is equal to or greater than 2 arbitrary value.
According to described the 1st constant M3 and the 2nd constant M4, can be by following definite range of choice: be equal to or greater than (a1+M3), perhaps be equal to or less than (a1-M4).Wherein, the a1 value is equivalent to the pixel value near reference pixels.
In step S1604, judge any in abutting connection with the pixel value of reference pixels whether within the range of choice of described step S1603 regulation.If the pixel value of adjacent pixels is ("Yes" of step S1604) within described range of choice, then handle action and enter step S1408.If the pixel value of adjacent pixels is ("No" of step S1604) not within described range of choice, then handle action and enter step S1403.
In the present embodiment, if any in abutting connection with the pixel value of reference pixels within described range of choice, then resolution conversion portion 5 hypothesis original images or the part that has reference pixels at least are that for example the such pixel value of grey scale image changes little image.According to this feature, to select near interpolation, it is suitable for the information that keeps relevant with the original image pixel value.
In the present embodiment, if any in abutting connection with the pixel value of reference pixels not within described range of choice, then resolution conversion portion 5 hypothesis original images or the part that has reference pixels at least are multivalue images for example, or pixel value has the image of great changes.According to this feature, select the 2nd method or other method, it is suitable for improving Image Smoothness.
Use the 24th method, according to topography's feature, smoothness that can the control chart picture, and the information relevant with the original image pixel value.
Have, in the present embodiment, step shown in Figure 29 is that shared the 2nd method reaches near interpolation again, so, improved bulk treatment speed.
With reference to Fig. 2 and Figure 30, the processing action of determining the pixel value of inter polated pixel according to use the 25th method of one embodiment of the invention is described.
The processing action of the 25th method that use Figure 30 represents is roughly identical with the processing action of the 24th method of using Figure 29 to represent.Difference is to have increased step S1404.
Use the 25th method, according to topography's feature, smoothness that can the control chart picture, and the information relevant with the original image pixel value.
Have, in the present embodiment, step shown in Figure 30 is that shared the 2nd method reaches near interpolation again, so, improved bulk treatment speed.
Have, except the method for the invention described above 16-25, resolution conversion portion 5 can be according to instructions of the present invention and the described scope of claims again, carries out any other interpolating method.For example, within instructions of the present invention and the described scope of claims, can be to the key element of said method, feature, function make up mutually and/or replace mutually.
Above with reference to description of drawings embodiments of the invention, but the present invention is not limited to the foregoing description.Can do all changes in the technology of the present invention thought range, they all belong to protection scope of the present invention.
For example, above-mentioned and other method of the present invention can be specialized with computer program.In one embodiment, image processing apparatus 9 can use the microprocessor and/or the signal processor of one or more traditional common purposes, it can carry out at least a said method of the present invention or other method according to the instruction that obtains from any medium.Described medium can be enumerated floppy disk, hard disk, and CD, disk, tape, Nonvolatile memory card, ROM etc. still, are not limited to said storing medium.
Perhaps, the present invention can be by the suitable networks of interconnection traditional structure circuit, or with its microprocessor and/or signal processor combinations with one or more traditional common purposes, through suitably programming, realizes with special IC.
Presents is based on the Japanese patent application No.2004-206408 that submits on July 13rd, 2004, and the Japanese patent application 2005-039603 that submits on February 16th, 2005, and as right of priority, its full content is added into, as a reference.
Claims (40)
1. image resolution ratio transform method comprises:
Specific inter polated pixel is so that be appended to it on image;
Around this inter polated pixel, select some reference pixels;
Ask for the distance value of each reference pixels;
Extract the pixel value of each reference pixels;
Choose the target reference pixels from some reference pixels, use the distance value and the pixel value of this target reference pixels, generate the weighting factor of described object pixel;
Weighting factor by described object pixel is determined pixel value, and the inter polated pixel that will have this pixel value is appended on the image.
2. according to the resolution converting method described in the claim 1, it is characterized in that, further comprise:
Show that this appends the image of inter polated pixel.
3. according to the resolution converting method described in the claim 1, it is characterized in that the step of described generation weighting factor further comprises:
To the target reference pixels, the difference of the pixel value between the pixel value of this target reference pixels and the 1st preset value is carried out computing;
4. according to the resolution converting method described in the claim 3, it is characterized in that the step of described generation weighting factor further comprises:
Make described margin of image element standardization with normalization factor.
5. according to the resolution converting method described in the claim 3, it is characterized in that described the 1st preset value is the average pixel value of reference pixels.
6. according to the resolution converting method described in the claim 3, it is characterized in that the mean value of the pixel value that described the 1st preset value is the target reference pixels and the pixel value of the reference pixels relative with this target reference pixels.
7. according to the resolution converting method described in the claim 3, it is characterized in that described the 1st preset value is the pixel value of the reference pixels relative with target reference pixels diagonal angle.
8. according to the resolution converting method described in the claim 4, it is characterized in that described normalization factor is equal to or greater than the max pixel value of image.
9. according to the resolution converting method described in the claim 4, it is characterized in that described normalization factor is determined by the pixel value of reference pixels.
10. according to the resolution converting method described in the claim 4, it is characterized in that described normalization factor determined by the difference of reference pixels, described difference is represented the poor of the maximal value of reference pixels and minimum value.
11. the resolution converting method according to described in the claim 4 is characterized in that, described normalization factor is determined by the maximal value and/or the minimum value of reference pixels.
12. the resolution converting method according to described in the claim 1 is characterized in that, the step of described generation weighting factor further comprises:
At least one side by improving distance value and pixel value is to the power of the value that multiplies each other, the weighting factor after obtaining multiplying each other;
Append in the step described, the weighting factor after resulting the multiplying each other is as weighting factor.
13. the resolution converting method according to described in the claim 1 is characterized in that, described target reference pixels is equivalent to whole reference pixels of selecting in described selection step.
14. the resolution converting method according to described in the claim 1 is characterized in that, described target reference pixels be equivalent in described selection step, select near the whole reference pixels the reference pixels.
15. the resolution converting method according to described in the claim 14 is characterized in that:
Further comprise use near the distance value of reference pixels, generate weighting factor near reference pixels;
Further by near the weighting factor of reference pixels, determine the pixel value of inter polated pixel.
16. an image processing apparatus comprises:
Pick-up unit, the input resolution and the output resolution ratio of detected image;
Comparison means compares described input resolution and output resolution ratio, generates comparative result;
Converting means according to described comparative result, uses at least a in some interpolating methods, will import resolution conversion and become output resolution ratio;
Wherein, described some interpolating methods comprise the 1st interpolating method, and it may further comprise the steps:
Specific inter polated pixel is so that be appended to it on image;
Around this inter polated pixel, select some reference pixels;
Ask for the distance value of each reference pixels;
Extract the pixel value of each reference pixels;
Choose the target reference pixels from some reference pixels, use the distance value and the pixel value of this target reference pixels, generate the weighting factor of described object pixel;
Weighting factor by described object pixel is determined pixel value, and the inter polated pixel that will have this pixel value is appended on the image.
17. the image processing apparatus according to described in the claim 16 is characterized in that, the generation step of described the 1st interpolating method comprises:
To the target reference pixels, the difference of the pixel value between the pixel value of this target reference pixels and the 1st preset value is carried out computing;
Make described margin of image element standardization with normalization factor.
18. the image processing apparatus according to described in the claim 17 is characterized in that, the generation step of described the 1st interpolating method further comprises:
By bringing up to the power of the value that multiplies each other, the weighting factor after obtaining multiplying each other with distance value with by at least one side of the difference of standardized pixel value;
Appending in the step of described the 1st interpolating method, the weighting factor after resulting the multiplying each other is as weighting factor.
19. the image processing apparatus according to described in the claim 17 is characterized in that, according to characteristics of image, determines in the 1st preset value and the normalization factor at least one.
20. the image processing apparatus according to described in the claim 17 is characterized in that, according to user's situation, determines in the 1st preset value and the normalization factor at least one.
21. the image processing apparatus according to described in the claim 18 is characterized in that, according to characteristics of image, determines the 1st preset value, normalization factor, and in the value that multiplies each other at least one.
22. the image processing apparatus according to described in the claim 18 is characterized in that, according to user's situation, determines the 1st preset value, normalization factor, and in the value that multiplies each other at least one.
23. the image processing apparatus according to described in claim 19 or 21 is characterized in that, determines described characteristics of image according to the pixel value of reference pixels.
24. the image processing apparatus according to described in claim 19 or 21 is characterized in that, determines described characteristics of image according to the difference of reference pixels.
25. the image processing apparatus according to described in claim 19 or 21 is characterized in that, determines described characteristics of image according to the maximal value and the minimum value of reference pixels.
26. the image processing apparatus according to described in the claim 16 is characterized in that, further comprises:
Acquisition device is used to obtain respectively and input resolution and the corresponding coordinate system of output resolution ratio, uses for converting means.
27. the image processing apparatus according to described in the claim 16 is characterized in that, further comprises:
Memory storage is used for storing at least a value of distance value and pixel value.
28. the image processing apparatus according to described in the claim 18 is characterized in that, further comprises:
Memory storage is used for the data that the memory mapping device uses, so that the weighting factor that has been multiplied each other.
29. the image processing apparatus according to described in the claim 16 is characterized in that, further comprises:
Selecting arrangement is used for selecting a kind of interpolating method according to characteristics of image from some interpolating methods.
30. the image processing apparatus according to described in the claim 29 is characterized in that, determines described characteristics of image according to the pixel value of reference pixels.
31. the image processing apparatus according to described in the claim 29 is characterized in that, according at least one determines described characteristics of image in the maximal value of reference pixels and the minimum value.
32. the image processing apparatus according to described in the claim 29 is characterized in that, determines described characteristics of image according to the difference of reference pixels.
33. the image processing apparatus according to described in the claim 29 is characterized in that, determines described characteristics of image near the pixel value of reference pixels according to what select from reference pixels.
34. the image processing apparatus according to described in the claim 29 is characterized in that, according to select from reference pixels, with determine described characteristics of image near the pixel value of the relative reference pixels in reference pixels diagonal angle.
35. the image processing apparatus according to described in the claim 32 is characterized in that, when difference equaled 0, selected interpolating method may further comprise the steps:
Increase inter polated pixel, the pixel value of this inter polated pixel equals in the reference pixels pixel value of any.
36. the image processing apparatus according to described in the claim 29 is characterized in that, described some interpolating methods further comprise the linear interpolation method.
37. the image processing apparatus according to described in the claim 29 is characterized in that, described some interpolating methods further comprise near interpolation.
38. an image display device comprises:
Image processing apparatus, the input resolution of detected image, described input resolution and output resolution ratio are compared, generate comparative result, according to described comparative result, use at least a in some interpolating methods, will import resolution conversion and become output resolution ratio, described some interpolating methods comprise the 1st interpolating method;
Display is used to show the image through conversion;
Described the 1st interpolating method may further comprise the steps:
Specific inter polated pixel is so that be appended to it on image;
Around this inter polated pixel, select some reference pixels;
Ask for the distance value of each reference pixels;
Extract the pixel value of each reference pixels;
Choose the target reference pixels from some reference pixels, use the distance value and the pixel value of this target reference pixels, generate the weighting factor of described object pixel;
Weighting factor by described object pixel is determined pixel value, and the inter polated pixel that will have this pixel value is appended on the image.
39. a computer program product is stored in and calculates on the medium that function reads, and when equipment operation, carries out a method, this method may further comprise the steps:
Specific inter polated pixel is so that be appended to it on image;
Around this inter polated pixel, select some reference pixels;
Ask for the distance value of each reference pixels;
Extract the pixel value of each reference pixels;
Choose the target reference pixels from some reference pixels, use the distance value and the pixel value of this target reference pixels, generate the weighting factor of described object pixel;
Weighting factor by described object pixel is determined pixel value, and the inter polated pixel that will have this pixel value is appended on the image.
40. one kind is calculated the medium that function reads, the instruction of storage computation machine is used to carry out a method, and this method may further comprise the steps:
Specific inter polated pixel is so that be appended to it on image;
Around this inter polated pixel, select some reference pixels;
Ask for the distance value of each reference pixels;
Extract the pixel value of each reference pixels;
Choose the target reference pixels from some reference pixels, use the distance value and the pixel value of this target reference pixels, generate the weighting factor of described object pixel;
Weighting factor by described object pixel is determined pixel value, and the inter polated pixel that will have this pixel value is appended on the image.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP206408/04 | 2004-07-13 | ||
| JP2004206408 | 2004-07-13 | ||
| JP039603/05 | 2005-02-16 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1722176A true CN1722176A (en) | 2006-01-18 |
| CN100552714C CN100552714C (en) | 2009-10-21 |
Family
ID=35912466
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005100833586A Expired - Fee Related CN100552714C (en) | 2004-07-13 | 2005-07-12 | Image resolution ratio transform method, image processing apparatus and image display device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100552714C (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102158622A (en) * | 2010-01-22 | 2011-08-17 | 佳能株式会社 | Image reading apparatus and control method thereof |
| CN101682686B (en) * | 2008-01-25 | 2011-11-30 | 索尼株式会社 | Apparatus and method for image processing |
| CN104169970A (en) * | 2012-04-18 | 2014-11-26 | 索尼公司 | Method and optical system for determining a depth map of an image |
-
2005
- 2005-07-12 CN CNB2005100833586A patent/CN100552714C/en not_active Expired - Fee Related
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101682686B (en) * | 2008-01-25 | 2011-11-30 | 索尼株式会社 | Apparatus and method for image processing |
| CN102158622A (en) * | 2010-01-22 | 2011-08-17 | 佳能株式会社 | Image reading apparatus and control method thereof |
| CN102158622B (en) * | 2010-01-22 | 2013-03-20 | 佳能株式会社 | Image reading apparatus and control method thereof |
| CN104169970A (en) * | 2012-04-18 | 2014-11-26 | 索尼公司 | Method and optical system for determining a depth map of an image |
| US9639947B2 (en) | 2012-04-18 | 2017-05-02 | Sony Corporation | Method and optical system for determining a depth map of an image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100552714C (en) | 2009-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1203662C (en) | Automatic grey-degree grade compensation apparatus, method and recording medium thereof | |
| CN101079949A (en) | Image processing apparatus and method, recording medium, and program | |
| CN1324873C (en) | Boundary detection method between areas having different features in image data | |
| CN1922630A (en) | Image processing device, image processing system, image processing method, image processing program, and integrated circuit device | |
| CN1805531A (en) | Information processing apparatus, information processing method, recording medium, and program | |
| CN1774031A (en) | Image processing apparatus and image processing method as well as computer program | |
| CN1645241A (en) | Imaging apparatus, method and device for processing images | |
| CN1200571C (en) | Orthogonal transformation, inverse orthogonal transformation method and device, and encoding and decoding method and device | |
| CN1640142A (en) | Method and apparatus for coding of wavelet transformed coefficients | |
| CN1399243A (en) | Display device, display method and record medium with display control program | |
| CN1207896C (en) | Image data compressing method and restoring method | |
| CN1396767A (en) | Image display device | |
| CN1262986C (en) | Image processing method and image processing apparatus | |
| CN1713256A (en) | Signal processing circuit and method for self-luminous type display | |
| CN1178490C (en) | Data processing method, data processor, and program recorded medium | |
| CN1229758C (en) | Resolution ratio transforming device and method of orthogonal transformation image | |
| CN1271783C (en) | Distortion compensator | |
| CN1647524A (en) | Image conversion device and image conversion method | |
| CN1304617A (en) | Interpolation processor and recording medium recording interpolation processing program | |
| CN1722176A (en) | Converting the resolution of an image using interpolation and displaying the converted image | |
| CN1879126A (en) | Average maintenance interpolation operating circuit, pixel interpolation circuit, average maintenance interpolation operating method, pixel interpolatin method | |
| CN1703726A (en) | Image processing device, image display device, image processing method, and image display method | |
| CN1302395C (en) | Image processing method and image processing apparatus | |
| CN1684515A (en) | Information processing apparatus and method, recording medium, and program | |
| CN1647523A (en) | Image conversion device and image conversion 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20091021 Termination date: 20190712 |