CN107481197A - Image Fuzzy Complementary characterizing method - Google Patents

Abstract

The invention discloses a kind of image Fuzzy Complementary characterizing method, for solving the ropy technical problem of conventional images complementarity characterizing method image restoration.Technical scheme is to extract the complementary feature of blurred picture first, by carrying out High frequency filter to multiframe blurred picture, obtains detail of the high frequency corresponding to image；Gradient convolution operation is carried out for image high-frequency information, extracts the gradient information of image；Binarization operation is carried out to the high frequency gradient map of blurred picture, obtains the high frequency gradient binary map of blurred picture.The complementary feature of blurred picture extracted is then based on, complementary quantitative calculating is carried out using complementary calculation formula, obtains the complementary numerical values recited between blurred picture.This method is weighed to the complementarity of different blurred pictures, the image provided definitely, which is restored, for blurred picture multiframe selects frame foundation, for the sequence multiple image that picture quality is uneven, quality that the image that is characterized based on complementarity selects frame mode to improve image restoration.

Description

Image Fuzzy Complementarity Representation Method

technical field

The invention relates to an image complementarity characterization method, in particular to an image fuzzy complementarity characterization method.

Background technique

Due to the different parameters of the motion blur model, degraded images with different blur degrees will be formed, and different blurred images reflect different information of the target, which is called the complementarity of blurred images. The complementarity between images can provide more information for the sharp reconstruction of blurred images.

The paper "Li W, Zhang J, Dai Q. Exploring aligned complementary image pair for blind motion deblurring[C], IEEE Conference on Computer Vision and Pattern Recognition, 2011: 273-280" proposed a blurred image restoration algorithm based on two images , which utilizes the complementary information between images to perform multi-frame restoration, but does not clearly specify the size of the complementarity between images, and, in the frame selection process of multi-frame image restoration, this method does not mention the frame selection For a sequence of multi-frame images with uneven image quality, different frame selection methods seriously affect the quality of image restoration.

Contents of the invention

In order to overcome the shortcomings of poor image restoration quality in existing image complementarity characterization methods, the present invention provides an image fuzzy complementarity characterization method. This method first uses the method based on image filtering to extract the complementary features of the fuzzy image, and obtains the high-frequency detail information corresponding to the image by performing high-frequency filtering on multiple frames of fuzzy images; performs gradient convolution operations on the high-frequency information of the image to extract Image gradient information in the horizontal direction and vertical direction; binarize the high-frequency gradient map of the blurred image to obtain a high-frequency gradient binary map of the blurred image, which is the complementary feature of the blurred image. Then, based on the extracted complementarity features of the fuzzy images, the complementarity calculation formula is used to carry out quantitative calculation of complementarity, and finally the value of complementarity between fuzzy images is obtained. This method measures the complementarity of different blurred images, and provides a clearer basis for image frame selection for multi-frame restoration of blurred images. quality of image restoration.

The technical scheme adopted by the present invention to solve the technical problem is: a kind of image fuzzy complementarity characterization method, which is characterized in that it includes the following steps:

Step 1. For any two blurred images g ₁ and g ₂ in the sequence image with a size of 256×256, use the Gaussian filter G to perform image filtering and decomposition, and calculate the high-frequency components h ₁ and h ₂ of the blurred image:

in, is the convolution operator, and the two-dimensional Gaussian filter convolution operator G is shown in formula (2):

Perform first-order gradient differentiation on each high-frequency component image h _i in the horizontal direction and vertical direction, that is, the horizontal and vertical gradient operator convolution operation, to obtain the horizontal gradient image hd _ix and the vertical gradient image hd _iy :

Among them, d _x is the horizontal gradient operator [1,-1], d _y is the vertical gradient operator [1,-1] ^T . On the basis of calculating the gradient images in the horizontal direction and the vertical direction, further calculate the global gradient image hd _i of the extracted image:

Perform image binarization on the calculated gradient image hd _i , and use formula (6) to extract the complementary feature T _i (x, y) of the blurred image.

Among them, the binarization threshold k _i is:

In the formula, max(hd _i ) represents the maximum pixel gray value of the extracted gradient image hd _i , and min(hd _i ) represents the minimum pixel gray value of the extracted gradient image hd _i .

Step 2. For the calculated binary images T ₁ and T ₂ , use them as the complementary features of the fuzzy image, based on the method of mathematical set theory, use the formula (8), the complementarity of the fuzzy images g ₁ and g ₂ To characterize.

Formula ( ₈ ) expresses the magnitude of complementarity between blurred images _g1 and g2. Among them, T ₁ ∪ T ₂ means to calculate the union of binary images T ₁ and T _2. The specific method is to perform point-to-point OR operation on the corresponding position pixels of the image. The calculation result is a 256×256 binary image, T ₁ ∩ T ₂ Indicates to calculate the intersection of binary images T ₁ and T ₂ , the specific method is to perform point-to-point AND operation on the corresponding position pixels, and the calculation result is also a 256×256 binary image. ||T|| ₁ means to calculate its l ₁ norm for the binary image T:

The beneficial effect of the present invention is that: the method first adopts the method based on image filtering to extract the complementary features of the fuzzy image, and obtains the high-frequency detail information corresponding to the image by performing high-frequency filtering on multi-frame fuzzy images; The gradient convolution operation extracts the image gradient information in the horizontal direction and the vertical direction respectively; performs a binarization operation on the high-frequency gradient map of the blurred image to obtain a high-frequency gradient binary map of the blurred image, which is the blurred image. Complementary features of images. Then, based on the extracted complementarity features of the fuzzy images, the complementarity calculation formula is used to carry out quantitative calculation of complementarity, and finally the value of complementarity between fuzzy images is obtained. This method measures the complementarity of different blurred images, and provides a clearer basis for image frame selection for multi-frame restoration of blurred images. quality of image restoration.

The present invention will be described in detail below in combination with specific embodiments.

detailed description

The specific steps of the image fuzzy complementarity characterization method of the present invention are as follows:

Step 1, extracting the complementary features of the blurred image.

For any two blurred images g ₁ and g ₂ in the sequence image with a size of 256×256, the Gaussian filter G is used for image filtering and decomposition, and the high frequency components h ₁ and h ₂ of the blurred image are calculated:

in, is a convolution operator, and the two-dimensional Gaussian filter convolution operator G is as follows:

Perform first-order gradient differentiation on each high-frequency component image h _i in the horizontal direction and vertical direction, that is, the horizontal and vertical gradient operator convolution operation, to obtain the horizontal gradient image hd _ix and the vertical gradient image hd _iy :

Among them, d _x is the horizontal gradient operator [1,-1], d _y is the vertical gradient operator [1,-1] ^T . On the basis of calculating the gradient images in the horizontal direction and the vertical direction, further calculate the global gradient image hd _i of the extracted image:

Perform image binarization on the calculated gradient image hd _i , and use formula (6) to extract the complementary feature T _i (x, y) of the blurred image.

Among them, the binarization threshold k _i is:

In the formula, max(hd _i ) represents the maximum pixel gray value of the extracted gradient image hd _i , and min(hd _i ) represents the minimum pixel gray value of the extracted gradient image hd _i .

Step 2: Fuzzy image complementarity characterization.

For the calculated binary images T ₁ and T ₂ , as the complementarity features of blurred images, based on the idea of mathematical set theory, formula (8) is used to characterize the complementarity of blurred images g ₁ and g ₂ .

The above _formula expresses the magnitude of complementarity between blurred images _g1 and g2. Among them, T ₁ ∪ T ₂ means to calculate the union of binary images T ₁ and T _2. The specific method is to perform a point-to-point "OR" operation on the corresponding pixels of the image. The calculation result is a 256×256 binary image, T ₁ ∩ T ₂ means to calculate the intersection of binary images T ₁ and T _2. The specific method is to perform point-to-point "AND" operation on the corresponding position pixels, and the calculation result is also a 256×256 binary image. ||T|| ₁ means to calculate its l ₁ norm for the binary image T:

Claims (1)

1. An image fuzzy complementarity characterization method is characterized in that comprising the following steps: Step 1. For any two blurred images g ₁ and g ₂ in the sequence image with a size of 256×256, use the Gaussian filter G to perform image filtering and decomposition, and calculate the high-frequency components h ₁ and h ₂ of the blurred image: <mrow><msub><mi>h</mi><mi>i</mi></msub><mo>=</mo><msub><mi>g</mi><mi>i</mi></msub><mo>-</mo><mi>G</mi><mo>&amp;CircleTimes;</mo><msub><mi>g</mi><mi>i</mi></msub><mo>,</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></mrow> in, is the convolution operator, and the two-dimensional Gaussian filter convolution operator G is shown in formula (2): <mrow><mi>G</mi><mo>=</mo><mfrac><mn>1</mn><mn>16</mn></mfrac><mfenced open = "[" close = "]"><mtable><mtr><mtd><mn>1</mn></mtd><mtd><mn>2</mn></mtd><mtd><mn>1</mn></mtd></mtr><mtr><mtd><mn>2</mn></mtd><mtd><mn>4</mn></mtd><mtd><mn>2</mn></mtd></mtr><mtr><mtd><mn>1</mn></mtd><mtd><mn>2</mn></mtd><mtd><mn>1</mn></mtd></mtr></mtable></mfenced><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></mrow> Perform first-order gradient differentiation on each high-frequency component image h _i in the horizontal direction and vertical direction, that is, the horizontal and vertical gradient operator convolution operation, to obtain the horizontal gradient image hd _ix and the vertical gradient image hd _iy : <mrow><msub><mi>hd</mi><mrow><mi>i</mi><mi>x</mi></mrow></msub><mo>=</mo><msub><mi>h</mi><mi>i</mi></msub><mo>&amp;CircleTimes;</mo><msub><mi>d</mi><mi>x</mi></msub><mo>,</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow></mrow> <mrow><msub><mi>hd</mi><mrow><mi>i</mi><mi>y</mi></mrow></msub><mo>=</mo><msub><mi>h</mi><mi>i</mi></msub><mo>&amp;CircleTimes;</mo><msub><mi>d</mi><mi>y</mi></msub><mo>,</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>4</mn><mo>)</mo></mrow></mrow> Among them, d _x is the horizontal gradient operator [1,-1], d _y is the vertical gradient operator [1,-1] ^T ; on the basis of calculating the gradient image in the horizontal direction and the vertical direction, further calculate and extract the image The global gradient image hd _i : <mrow><msub><mi>hd</mi><mi>i</mi></msub><mo>=</mo><msqrt><mrow><msup><msub><mi>hd</mi><mrow><mi>i</mi><mi>x</mi></mrow></msub><mn>2</mn></msup><mo>+</mo><msup><msub><mi>hd</mi><mrow><mi>i</mi><mi>y</mi></mrow></msub><mn>2</mn></msup></mrow></msqrt><mo>,</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>5</mn><mo>)</mo></mrow></mrow> Carry out image binarization operation on the calculated gradient image hd _i , and use formula (6) to extract the complementary feature T _i (x, y) of the blurred image; <mrow><msub><mi>T</mi><mi>i</mi></msub><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow><mo>=</mo><mfenced open = "{" close = ""><mtable><mtr><mtd><mrow><mn>1</mn><mo>,</mo></mrow></mtd><mtd><mrow><msub><mi>hd</mi><mi>i</mi></msub><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow><mo>&amp;GreaterEqual;</mo><msub><mi>k</mi><mi>i</mi></msub></mrow></mtd></mtr><mtr><mtd><mrow><mn>0</mn><mo>,</mo></mrow></mtd><mtd><mrow><msub><mi>hd</mi><mi>i</mi></msub><mrow><mo>(</mo><mi>x</mi><mo>,</mo><mi>y</mi><mo>)</mo></mrow><mo>&lt;</mo><msub><mi>k</mi><mi>i</mi></msub></mrow></mtd></mtr></mtable></mfenced><mo>,</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>6</mn><mo>)</mo></mrow></mrow> Among them, the binarization threshold k _i is: <mrow><msub><mi>k</mi><mi>i</mi></msub><mo>=</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>(</mo><mi>m</mi><mi>a</mi><mi>x</mi><mo>(</mo><mrow><msub><mi>hd</mi><mi>i</mi></msub></mrow><mo>)</mo><mo>-</mo><mi>m</mi><mi>i</mi><mi>n</mi><mo>(</mo><mrow><msub><mi>hd</mi><mi>i</mi></msub></mrow><mo>)</mo><mo>)</mo></mrow><mo>,</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mn>2</mn><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>7</mn><mo>)</mo></mrow></mrow> In the formula, max(hd _i ) represents the maximum pixel gray value of the extracted gradient image hd _i , and min(hd _i ) represents the minimum pixel gray value of the extracted gradient image hd _i ; Step 2. For the calculated binary images T ₁ and T ₂ , use them as the complementary features of the fuzzy image, based on the method of mathematical set theory, use the formula (8), the complementarity of the fuzzy images g ₁ and g ₂ To characterize; <mrow><mi>c</mi><mi>o</mi><mi>m</mi><mrow><mo>(</mo><msub><mi>g</mi><mn>1</mn></msub><mo>,</mo><msub><mi>g</mi><mn>2</mn></msub><mo>)</mo></mrow><mo>=</mo><mfrac><mrow><mo>|</mo><mo>|</mo><msub><mi>T</mi><mn>1</mn></msub><mo>&amp;cup;</mo><msub><mi>T</mi><mn>2</mn></msub><mo>-</mo><msub><mi>T</mi><mn>1</mn></msub><mo>&amp;cap;</mo><msub><mi>T</mi><mn>2</mn></msub><mo>|</mo><msub><mo>|</mo><mn>1</mn></msub></mrow><mrow><mo>|</mo>mo><mo>|</mo><msub><mi>T</mi><mn>1</mn></msub><mo>&amp;cup;</mo><msub><mi>T</mi><mn>2</mn></msub><mo>|</mo><msub><mo>|</mo><mn>1</mn></msub></mrow></mfrac><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>8</mn><mo>)</mo></mrow></mrow> Formula (8) represents the complementarity between blurred images g ₁ and g ₂ ; among them, T ₁ ∪ T ₂ means to calculate the union of binary images T ₁ and T ₂ , the specific method is to perform point-to-point or Operation, the calculation result is a 256×256 binary image, T ₁ ∩ T ₂ means to calculate the intersection of the binary image T ₁ and T ₂ , the specific method is to perform point-to-point AND operation on the corresponding position pixels, and the calculation result is also a 256×256 binary image; ||T|| ₁ means to calculate its l ₁ norm for the binary image T: <mrow><mo>|</mo><mo>|</mo><mi>T</mi><mo>|</mo><msub><mo>|</mo><mn>1</mn></msub><mo>=</mo><munderover><mo>&amp;Sigma;</mo><mrow><mi>x</mi><mo>=</mo><mn>1</mn></mrow><mi>M</mi></munderover><munderover><mo>&amp;Sigma;</mo><mrow><mi>y</mi><mo>=</mo><mn>1</mn></mrow><mi>N</mi></munderover><mo>|</mo><msub><mi>t</mi><mrow><mi>x</mi><mi>y</mi></mrow></msub><mo>|</mo><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>9</mn><mo>)</mo></mrow><mo>.</mo></mrow> 1