CN105608686A - Long and short exposure wide dynamic feather blending method based on mean filtering - Google Patents

Abstract

The invention discloses a long and short exposure wide dynamic feather blending method based on mean filtering, and the method comprises the steps: adjusting an operator through a brightness region, and enabling a brightness region in a long exposure image to be extracted; solving the mean value of a result obtained at step S1, carrying out the binarization processing of the mean value, and obtaining a binary mask composed of 0 and 1; carrying out the mean filtering of the binary mask, enabling the binary mask to generate a feather effect, and obtaining a feather mask; extracting short exposure information and long exposure information: (1) carrying out the multiply operation of the feather mask obtained at step S3 and a short exposure image; (2), solving the negation of a feather mask image obtained at step S3, and carrying out the multiply operation of the negation result and the long exposure image; carrying out the adding operation of the result obtained at step S4; carrying out the data correction of the result obtained step S5, and outputting the data correction result. The method is suitable for FPGA implementation, is simple in implementation, occupies fewer hardware resources, directly achieves the integration of original pixels, and is very high in instantaneity.

Description

The wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter

Technical field

The present invention relates to a kind of wide dynamic fusion algorithm, particularly relate to the wide dynamic emergence of a kind of exposure of the length based on mean filterFusion method.

Background technology

In various wide dynamic techniques, the wide dynamic technique merging based on multiframe is comparatively general one, and blending algorithm is also manyKind is various, and as the method such as fusion, segment fusion of pyramid, comparatively speaking, these algorithms are realized complicated, take hardwareResource is more, and real-time is poor, is unfavorable for Realization of Product.

Summary of the invention

The object of the invention is to overcome the deficiencies in the prior art, a kind of prerequisite do not change original image pixels value is as far as possible providedUnder, length exposure image is directly merged, is applicable to the hard-wired exposure of the length based on mean filter wide dynamic emergence fusion sideMethod.

The object of the invention is to be achieved through the following technical solutions: the wide dynamic emergence of a kind of exposure of the length based on mean filterFusion method, comprises the following steps:

S1: extract highlight regions, adjust operator by luminance area the highlight regions in long exposure image is extracted;

S2: set up binaryzation and cover plate, the result that step S1 is obtained is averaged, and utilizes average to carry out binary conversion treatment, obtains oneIndividual by the 0 and 1 binaryzation illiteracy plate forming;

S3: cover plate and sprout wings, binaryzation is covered to plate and carry out mean filter, make it produce feather effect, obtain emergence and cover plate L " (x, y);

S4: extract short exposure information and long exposure information:

(1) plate is covered in emergence step S3 being obtained and short exposure image does multiply operation;

(2) negate of plate image is covered in emergence step S3 being obtained, and does multiply operation with long exposure image;

S5: length exposure information merges, and the result that step S4 is obtained is done add operation;

S6: the result of step S5 is carried out to data correction, then output.

Described extraction highlight regions method is: first each pixel F (x, y) of long exposure image F is normalized, soAfter bring formula L (x, y)=1/ (1+F (x, y) into^-10) obtain high luminance area area image L.

The described method of setting up binaryzation illiteracy plate is: first ask for the average M of area image L in high luminance area in step S1, average M'sAcquiring method is that all pixels are added, and divided by the total pixel number of image;

Traversal high luminance area area image L, proceeds as follows each pixel L (x, y): when L (x, y) value is more than or equal to MTime, L ' (x, y)=1, in the time that L (x, y) value is less than M, L ' (x, y)=0, that is:

$L^{\&prime;}(x,y)=\left\{\begin{array}{cc}1& L(x,y)\&GreaterEqual;M\\ 0& L(x,y)<M\end{array}\right.$

Wherein, M is the average of high luminance area area image L, the binaryzation illiteracy plate of L ' for obtaining.

The method of described extraction short exposure information and long exposure information is:

(1) plate L is covered in emergence step S3 being obtained, and " (x, y) does multiply operation with short exposure image B, obtains a short exposure letterBreath image B ', that is:

B′(x,y)＝L″(x,y)·B(x,y)

Wherein, B ' is short exposure frame, and B is short exposure image;

(2) negate of plate image is covered in emergence step S3 being obtained, and does multiply operation with long exposure image F, obtains long exposing to the sunOptical information image F ', that is:

F′(x,y)＝F(x,y)·[1-L″(x,y)]

Wherein, F ' is long exposure information image, and F is long exposure image.

The method that described length exposure information merges is: the long exposure information image F ' that S4 is obtained and short exposure frame B ' enterRow is added and merges, and obtains the image O after merging, and in order to regulate syncretizing effect, is multiplied by a factor sigma and does effect adjusting, that is:

O(x,y)＝σ·B′(x,y)+(1-σ)·F′(x,y)

Wherein, σ span is [0,1], and O is the image after merging.

The method of described data correction is: traversing graph, as each pixel O (x, y) of O, judges the value size of O (x, y), whenO (x, y) is greater than at 255 o'clock, and the value of data is made as to 255; In the time that the value of O (x, y) is less than 0, the value of data is made as to 0, that is:

$O(x,y)=\left\{\begin{array}{cc}255& O(x,y)>255\\ 0& O(x,y)<0\end{array}\right..$

The invention has the beneficial effects as follows: compare the blending algorithm such as fusion, segment fusion of traditional pyramid, this fusion is calculatedMethod realizes simple, takies hardware resource less, and this fusion method directly merges image pixel, and real-time is very high.

(1) this fusion method need to adopt convolution realization except covering plate emergence, and other steps are all the operations based on pixel,Be applicable to FPGA and realize, realize simple;

(2) this fusion method is the fusion based on pixel, length exposure information is directly merged to the image maximum after therefore mergingThe maintenance of degree former figure information, can not produce unpredictable problem.

Brief description of the drawings

Fig. 1 is fusion method flow chart of the present invention.

Detailed description of the invention

Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail, but protection scope of the present invention is not limited to following instituteState.

As shown in Figure 1, the wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter, comprises the following steps:

S1: extract highlight regions, adjust operator by luminance area the highlight regions in long exposure image is extracted;

Described extraction highlight regions method is: first each pixel F (x, y) of long exposure image F is normalized, soAfter bring formula L (x, y)=1/ (1+F (x, y) into^-10) obtain high luminance area area image L.

S2: set up binaryzation and cover plate, the result that step S1 is obtained is averaged, and utilizes average to carry out binary conversion treatment, obtains oneIndividual by the 0 and 1 binaryzation illiteracy plate forming;

The described method of setting up binaryzation illiteracy plate is: first ask for the average M of area image L in high luminance area in step S1, average M'sAcquiring method is that all pixels are added, and divided by the total pixel number of image;

Traversal high luminance area area image L, proceeds as follows each pixel L (x, y): when L (x, y) value is more than or equal to MTime, L ' (x, y)=1, in the time that L (x, y) value is less than M, L ' (x, y)=0, that is:

$L^{\&prime;}(x,y)=\left\{\begin{array}{cc}1& L(x,y)\&GreaterEqual;M\\ 0& L(x,y)<M\end{array}\right.$

Wherein, M is the average of high luminance area area image L, the binaryzation illiteracy plate of L ' for obtaining.

S3: cover plate and sprout wings, binaryzation is covered to plate and carry out mean filter, make it produce feather effect, obtain emergence and cover plate L " (x, y);

The described method of setting up the emergence of binaryzation illiteracy plate is: the binaryzation obtaining in S2 is covered to plate and adopt known mean filter sideMethod, makes it produce feather effect, and obtains the illiteracy plate L " (x, y) that sprouts wings.

S4: extract short exposure information and long exposure information:

(1) plate L is covered in emergence step S3 being obtained, and " (x, y) does multiply operation with short exposure image B, obtains a short exposure letterBreath image B ', that is:

B′(x,y)＝L″(x,y)·B(x,y)

Wherein, B ' is short exposure frame, and B is short exposure image;

(2) negate of plate image is covered in emergence step S3 being obtained, and does multiply operation with long exposure image F, obtains long exposing to the sunOptical information image F ', that is:

F′(x,y)＝F(x,y)·[1-L″(x,y)]

Wherein, F ' is long exposure information image, and F is long exposure image.

S5: length exposure information merges, and the result that step S4 is obtained is done add operation;

The method that described length exposure information merges is: the long exposure information image F ' that S4 is obtained and short exposure frame B 'Be added fusion, obtain the image O after merging, in order to regulate syncretizing effect, be multiplied by a factor sigma and do effect adjusting, that is:

O(x,y)＝σ·B′(x,y)+(1-σ)·F′(x,y)

Wherein, σ span is [0,1], and O is the image after merging.

S6: the result of step S5 is carried out to data correction, then output;

The method of described data correction is: traversing graph, as each pixel O (x, y) of O, judges the value size of O (x, y), whenO (x, y) is greater than at 255 o'clock, and the value of data is made as to 255; In the time that the value of O (x, y) is less than 0, the value of data is made as to 0, that is:

$O(x,y)=\left\{\begin{array}{cc}255& O(x,y)>255\\ 0& O(x,y)<0\end{array}\right..$

The technical issues that need to address of the present invention have:

(1) by long exposure image and short exposure image separately useful information split;

(2) how to set up the illiteracy plate (employing average) of sprouting wings;

(3) how to utilize the illiteracy plate of sprouting wings that two width image corresponding regions are split;

(4) how by a synthetic two parts useful information image without splicing vestige.

The above is only the preferred embodiment of the present invention, is to be understood that the present invention is not limited to disclosed form herein,Should not regard the eliminating to other embodiment as, and can be used for various other combinations, amendment and environment, and can be described hereinIn contemplated scope, change by technology or the knowledge of above-mentioned instruction or association area. And the change that those skilled in the art carry outDo not depart from the spirit and scope of the present invention with changing, all should be in the protection domain of claims of the present invention.

Claims (6)

1. the wide dynamic emergence fusion method of the exposure of the length based on mean filter, is characterized in that, comprises the following steps:

S1: extract highlight regions, adjust operator by luminance area the highlight regions in long exposure image is extracted;

S2: set up binaryzation and cover plate, the result that step S1 is obtained is averaged, and utilizes average to carry out binary conversion treatment, obtains oneIndividual by the 0 and 1 binaryzation illiteracy plate forming;

S3: cover plate and sprout wings, binaryzation is covered to plate and carry out mean filter, make it produce feather effect, obtain emergence and cover plate;

S4: extract short exposure information and long exposure information:

(1) plate is covered in emergence step S3 being obtained and short exposure image does multiply operation;

(2) negate of plate image is covered in emergence step S3 being obtained, and does multiply operation with long exposure image;

S5: length exposure information merges, and the result that step S4 is obtained is done add operation;

S6: the result of step S5 is carried out to data correction, then output.

2. the wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter according to claim 1, its feature existsIn: described extraction highlight regions method is: first each pixel F (x, y) of long exposure image F is normalized, soAfter bring formula L (x, y)=1/ (1+F (x, y) into^-10) obtain high luminance area area image L.

3. the wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter according to claim 1, its feature existsIn: the described method of setting up binaryzation illiteracy plate is: first ask for the average M of area image L in high luminance area in step S1, average M'sAcquiring method is that all pixels are added, and divided by the total pixel number of image;

Traversal high luminance area area image L, proceeds as follows each pixel L (x, y): when L (x, y) value is more than or equal to MTime, L ' (x, y)=1, in the time that L (x, y) value is less than M, L ' (x, y)=0, that is:

$L^{\&prime;}(x,y)=\left\{\begin{array}{cc}1& L(x,y)\&GreaterEqual;M\\ 0& L(x,y)<M\end{array}\right.$

Wherein, M is the average of high luminance area area image L, the binaryzation illiteracy plate of L ' for obtaining.

4. the wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter according to claim 1, its feature existsIn: the method for described extraction short exposure information and long exposure information is:

(1) plate L is covered in emergence step S3 being obtained, and " (x, y) does multiply operation with short exposure image B, obtains a short exposure letterBreath image B ', that is:

B′(x,y)＝L″(x,y)·B(x,y)

Wherein, B ' is short exposure frame, and B is short exposure image;

(2) negate of plate image is covered in emergence step S3 being obtained, and does multiply operation with long exposure image F, obtains long exposing to the sunOptical information image F ', that is:

F′(x,y)＝F(x,y)·[1-L″(x,y)]

Wherein, F ' is long exposure information image, and F is long exposure image.

5. the wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter according to claim 1, its feature existsIn: the method that described length exposure information merges is: the long exposure information image F ' that S4 is obtained and short exposure frame B ' enterRow is added and merges, and obtains the image O after merging, and in order to regulate syncretizing effect, is multiplied by a factor sigma and does effect adjusting, that is:

O(x,y)＝σ·B′(x,y)+(1-σ)·F′(x,y)

Wherein, σ span is [0,1], and O is the image after merging.

6. the wide dynamic emergence fusion method of a kind of exposure of the length based on mean filter according to claim 1, its feature existsIn: the method for described data correction is: traversing graph, as each pixel O (x, y) of O, judges the value size of O (x, y), whenO (x, y) is greater than at 255 o'clock, and the value of data is made as to 255; In the time that the value of O (x, y) is less than 0, the value of data is made as to 0, that is: $O(x,y)=\left\{\begin{array}{cc}255& O(x,y)>255\\ 0& O(x,y)<0\end{array}\right..$