CN101655973A - Image enhancing method based on visual characteristics of human eyes - Google Patents

Abstract

The invention discloses an image enhancing method based on the visual characteristics of human eyes, for example, the human eyes has poor resolution ratio on grayscale when the grayscale of an image is very high or very low, the human eyes has high resolution ratio when the grayscale is moderate. The image enhancing method considers the visual characteristics of the human eyes in the image processing process and improves the defects of the prior image enhancing method. Neighborhood information plays an important role in the image smoothing and denoising process, thus adjacent pixel points canbe used for enhancing processing efficiency in the image processing process. The image enhancing method can smooth and denoise the image and enables the enhanced image to more accord with the visual characteristics of the human eyes.

Description

A kind of image enchancing method based on visual characteristics of human eyes

Technical field

The present invention relates to the graph and image processing field, be specifically related to a kind of image enchancing method based on visual characteristics of human eyes.

Background technology

In some theory, the definition of figure image intensifying is: " the figure image intensifying is the visual effect that is used for improving image, or image transitions is become to be suitable for a special kind of skill of the form of human eye, machine analysis." the figure image intensifying is the basic means of Flame Image Process, it is toward the preprocessing process that firmly is various graphical analyses when handling.The figure image intensifying is a kind of technology that improves image displaying quality, and it reaches the effect that strengthens image by some information of optionally emphasizing and suppress in the image.

In the figure image intensifying, adopted many technology to improve the visual effect of image.These technology mainly comprise based on histogrammic enhancing with based on the enhancing of spatial frequency domain.Histogram equalization is a kind of technology commonly used, and it is more effective for narrow histogrammic enhancing.But its shortcoming is responsive to noise ratio, and after handling, the brightness of image is compared with former figure bigger discrepancy.

When gradation of image was very high or very low, human eye was lower to the resolution of gray scale, and when gray scale was moderate, human eye was higher to the resolution of gray scale; And human eye is to more responsive at the noise of detail section in the noise ratio in the mild zone of image.In the design of the method for figure image intensifying, consider that the final recipient of the information after the figure image intensifying is the people, if introduce these visual characteristics of people, image processing effect will significantly improve.

In the traditional theory, to the existing a lot of methods in figure image intensifying aspect, but all the neighborhood information between the image slices vegetarian refreshments is not made full use of, and when smoothed image and elimination noise, neighborhood information has very vital role.And if the visual signature of human eye added consideration together, the visual effect of image will be improved to a great extent.

Summary of the invention

The objective of the invention is in, add the visual signature of human eye, and improve the deficiency that present method is not taken the image neighborhood information into account, propose a kind of improved image enchancing method based on visual characteristics of human eyes to the treatment of picture process.

The present invention proposes a kind of image enchancing method, comprise step based on visual characteristics of human eyes:

1. original image is handled and obtained the image that the edge strengthens;

2. the prewitt operator is orthogonal to the gray-scale value of the above-mentioned picture point that obtains, obtains gradient image;

3. the gradient image after the normalization is mapped to fuzzy field;

4. utilize inverse transformation, image is transformed to spatial domain by fuzzy field, the image that can be enhanced.

Original image handled obtaining the image that the edge strengthens, the method for realization is to utilize gradient and differential principle to combine to produce the sharpening operator to act on original image to obtain.

The prewitt operator is orthogonal to the gray-scale value of the picture point that 1. step obtain, obtains gradient image, the method for realization is that the prewitt operator is acted on through above-mentioned processed images point (i, gray-scale value X j) _Ij, obtain gradient image I (x, y).

Gradient image after the normalization is mapped to fuzzy field, and the method for realization is a new subordinate function according to definition:

$U_{\mathrm{ij}}=G(x,y)=\frac{[\mathrm{\α}{X_{\mathrm{ij}}}^L+\mathrm{\β}(X_{\mathrm{ij}}+{X_{\mathrm{ij}}}^L-{2X}_{\min })+\mathrm{\η}]}{2(X_{\max }-X_{\min })}$

Ambiguity in definition strengthens operator T then ₁, a variable μ that crosses _c, image is strengthened.

Utilize inverse transformation, image is transformed to spatial domain by fuzzy field, the image that can be enhanced, the method for realization is based on following inverse transformation:

$X_{\mathrm{ij}}=G^{-1}\left({\mathrm{\μ}}_{\mathrm{ij}}\right)=\frac{[2(X_{\max }-X_{\min }){\mathrm{\μ}}_{\mathrm{ij}}-\mathrm{\α}{X}_{\mathrm{ij}}^L-\mathrm{\η}]}{\mathrm{\β}-X_{\mathrm{ij}}^L+{2X}_{\min }}.$

A kind of image enchancing method based on visual characteristics of human eyes proposed by the invention has the characteristics of the following aspects:

(1) at gray scale when moderate, the visual characteristic that resolving power of the eye is stronger.

(2) on effect, abundant zone of details and the mild regional marginal portion of grey scale change in the image have been promoted.

(3) combine Fuzzy Set Theory and different zones is carried out the grey level stretching that do not wait.

As mentioned above, this method has strengthened the details and the contrast of image when effectively improving the gradation of image dynamic range, and the image after the enhancing more meets people's visual signature.

Description of drawings

Fig. 1 is the process flow diagram of the inventive method;

Fig. 2 is 3 * 3 image template window figure;

Fig. 3 is μ _cAdaptively selected figure.

Embodiment

Below in conjunction with accompanying drawing the image enchancing method based on visual characteristics of human eyes of the present invention is described in detail.

(1) is defined image template window figure as shown in Figure 2, need 4 template a of definition altogether, b, c, d, each template all is 3 * 3 calcspar, comes scan image with template, the difference of calculation template center pixel and surrounding pixel, if difference is greater than certain threshold value, then the gray-scale value with central pixel point multiply by certain coefficient, makes the gray-scale value of this pixel strengthen, otherwise keeps this gray values of pixel points constant.

As figure, the pixel of each square module is numbered, and supposes that its grey scale pixel value is respectively: m ₀, m ₁... m ₈

For template a,

$A_{11}=\frac{{\mathrm{<figure-callout\; id="1"\; label="m"\; filenames="A2009101924140009H1.png"\; state="\{\{state\}\}">m</figure-callout>}}_1+m_2+m_5}{3}$

$A_{12}=\frac{m_3+m_6+m_7}{3}$

M ₁＝|A ₁₂-A ₁₁|

For template b,

$A_{21}=\frac{|m_0+m_2+m_6|}{3}$

$A_{22}=\frac{|m_2+m_5+m_8|}{3}$

M ₂＝|A ₂₂-A ₂₁|

For template c,

$A_{31}=\frac{|m_0+{\mathrm{<figure-callout\; id="1"\; label="m"\; filenames="A2009101924140009H1.png"\; state="\{\{state\}\}">m</figure-callout>}}_1+m_3|}{3}$

$A_{32}=\frac{|m_5+m_7+m_8|}{3}$

M ₃＝|A ₃₂-A ₃₁|

For template d,

$A_{41}=\frac{|m_0+{\mathrm{<figure-callout\; id="1"\; label="m"\; filenames="A2009101924140009H1.png"\; state="\{\{state\}\}">m</figure-callout>}}_1+m_2|}{3}$

$A_{42}=\frac{|m_6+m_7+m_8|}{3}$

M ₄＝|A ₄₂-A ₄₁|

Ask M ₁, M ₂, M ₃, M ₄In maximal value:

M＝max(M ₁，M ₂，M ₃，M ₄)

Whether judge M greater than certain threshold value, if greater than, then with the gray-scale value m of center pixel ₄Multiply by certain coefficient (generally greater than 1).

(2) selected smoothing effect for use, the prewitt operator is orthogonal to above-mentioned processed images point (i, gray-scale value X j) simply fast _Ij, obtain gray level image I (x, y), $I(x,y)=\max \{\mathrm{\&Delta;}\frac{x}{3},\mathrm{\&Delta;}\frac{y}{3}\},$ Wherein, Δ x=f (x, y+1)+f (x, y)+f (x, y-1)-f (x-1, y+1)-f (x-1, y)-f (x-1, y-1) Δ y=f (x-1, y-1)+f (x, y-1)+f (x+1, y-1)-f (x-1, y)-f (x, y)-(x+1 y) establishes I to f _MaxBe that (the normalization gradient image gets I for x, maximal value y) $Z(x,y)=\frac{\left|I(x,y)\right|}{I_{\max }}.$ The enhancing process of this section can be represented with this section: and log (f ' (x, y))=log (T (α (x, y)))+β [log (f (x, y))-log (α (x, y))], wherein, β _s=β ₀| Ig (x, y) | maxIg, β ₀Be a real number, T () is a transforming function transformation function.

(3) according to the subordinate function of redetermination: $U_{\mathrm{ij}}=G(x,y)=\frac{[\mathrm{\&alpha;}{X_{\mathrm{ij}}}^L+\mathrm{\&beta;}(X_{\mathrm{ij}}+{X_{\mathrm{ij}}}^L-{2X}_{\min })+\mathrm{\&eta;}]}{2(X_{\max }-X_{\min })},$ Gradient image after the normalization is mapped to fuzzy field.Ambiguity in definition strengthens operator T ₁, image is carried out enhancement process.Be expressed as: $\mathrm{\&mu;}(i,j)=T_1\left({\mathrm{\&mu;}}_{\mathrm{ij}}\right)=1-\frac{1-{\mathrm{\&mu;}}_{\mathrm{ij}}^2}{1-{\mathrm{\&mu;}}_c},$ μ wherein _cBe variablely to get over a little.As Fig. 2 is μ _cAdaptively selected mode.

(4) utilize inverse transformation: $X_{\mathrm{ij}}=G^{-1}\left({\mathrm{\&mu;}}_{\mathrm{ij}}\right)=\frac{[2(X_{\max }-X_{\min }){\mathrm{\&mu;}}_{\mathrm{ij}}-\mathrm{\&alpha;}{X}_{\mathrm{ij}}^L-\mathrm{\&eta;}]}{\mathrm{\&beta;}-X_{\mathrm{ij}}^L+{2X}_{\min }}$ Image is returned spatial domain by the fuzzy field conversion, the image that can be enhanced.

Claims (1)

1, a kind of image enchancing method based on visual characteristics of human eyes is characterized in that may further comprise the steps:

1. original image is handled and obtained the image that the edge strengthens;

2. the prewitt operator is orthogonal to the gray-scale value of the point of the above-mentioned image that obtains, obtains gradient image;

3. the gradient image after the normalization is mapped to fuzzy field;

4. utilize inverse transformation, image is transformed to spatial domain by fuzzy field, the image that can be enhanced;

Wherein, original image handled obtaining the image that the edge strengthens, the method for realization is to utilize gradient and differential principle to combine to produce the sharpening operator to act on original image to obtain;

The prewitt operator is orthogonal to the gray-scale value of the picture point that 1. step obtain, obtains gradient image, the method for realization is that the prewitt operator is acted on through above-mentioned processed images point (i, gray-scale value X j) _Ij, obtain gradient image I (x, y);

Gradient image after the normalization is mapped to fuzzy field, and the method for realization is a new subordinate function according to definition:

$U_{\mathrm{ij}}=G(x,y)=\frac{[\mathrm{\&alpha;}{X_{\mathrm{ij}}}^L+\mathrm{\&beta;}(X_{\mathrm{ij}}+{X_{\mathrm{ij}}}^L-2X_{\min })+\mathrm{\&eta;}]}{2(X_{\max }-X_{\min })}$

Ambiguity in definition strengthens operator T then ₁, a variable μ that crosses _c, image is strengthened;

Utilize inverse transformation, image is transformed to spatial domain by fuzzy field, the image that can be enhanced, the method for realization is based on following inverse transformation:

$X_{\mathrm{ij}}=G^{-1}\left({\mathrm{\&mu;}}_{\mathrm{ij}}\right)=\frac{[2(X_{\max }-X_{\min }){\mathrm{\&mu;}}_{\mathrm{ij}}-\mathrm{\&alpha;}{X}_{\mathrm{ij}}^L-\mathrm{\&eta;}]}{\mathrm{\&beta;}-X_{\mathrm{ij}}^L+2X_{\min }}.$

Images