CN102332154A - Method and system for enhancing color images of cotton pseudo foreign fibers under non-uniform illumination - Google Patents

Abstract

The invention discloses a method for enhancing color images of cotton pseudo foreign fibers under non-uniform illumination. The method comprises the following steps of: 1, reading original color images of the cotton pseudo foreign fibers in; 2, converting the original color images of the cotton pseudo foreign fibers from an RGB (red, green and blue) space to an HSI (hue, saturation and intensity) space; 3, enhancing components I of the converted HSI color images; 4, performing enhancement post-processing on the images enhanced by the step 3; and 5, converting the HSI images processed by the step 4 back to the RGB space. In addition, the invention also discloses a system for implementing the method. The system comprises an image reading module, a first image conversion module, an image enhancement module, an image enhancement post-processing module and a second image conversion module which are used for realizing the five steps respectively. By utilizing the method and system disclosed by the invention, the color images of the cotton pseudo foreign fibers can be effectively enhanced, so that the subsequent segmentation and identification of foreign fibers and the pseudo foreign fibers can be conveniently realized.

Description

Pseudo-foreign fiber colour-image reinforcing method of cotton under the uneven illumination and system

Technical field

The present invention relates to technical field of image processing, relate in particular to pseudo-foreign fiber colour-image reinforcing method of cotton and system under a kind of uneven illumination.

Background technology

Pseudo-foreign fiber in the cotton; Be meant at cotton planting, pluck, ted sneak in the process in the cotton with foreign fiber in size, color, be difficult to some non-fiber soft or hard impurity of distinguishing in shape, mainly comprise: cotton leaf, cottonseed bits, cotton stalk, blade of grass etc.The content of pseudo-foreign fiber in gined cotton is more than foreign fiber; And its imaging is from aspect and the very difficult differentiations of foreign fiber such as color, shapes; Cotton foreign fiber detection and accurate measurement influence to based on machine vision are very big; Therefore need pseudo-foreign fiber and foreign fiber to be distinguished through some Flame Image Process and machine vision technique.The figure image intensifying is a link the most basic in Flame Image Process and the machine vision, is the prerequisite of image segmentation, graphical analysis and pattern-recognition.

The method of figure image intensifying is through certain means original image to be added some information or transform data, and some unwanted characteristic in interested characteristic or inhibition (covering) image is complementary image and eye response characteristic in the outstanding selectively image.In figure image intensifying process, do not analyze the reason of image deterioration, the image after the processing not necessarily approaches original image.Because the characteristics of image are had nothing in common with each other, and are difficult to find the enhancement algorithms of a suitable all images.

Retinex is the abbreviation of retina " Retina " and cerebral cortex " Cortex "; Be people such as EdwinLand through a large amount of visions and psychology experiment, sum up propose how to regulate the color of the object that perceives and the model of brightness about human visual system (Human Visual System).The traditional image enhancement algorithms; Can only strengthen a certain category feature of image like linearity, nonlinear transformation, image sharpening etc.; Like the dynamic range of compressed image, or strengthen edge of image etc., the Retinex model can be in the gray scale dynamic range compression; The edge strengthens and three aspects of colored constancy reach balance, thereby is suitable for the enhancing of the pseudo-foreign fiber image of cotton.

Because mostly the pseudo-foreign fiber of cotton is smoits, color is generally cream colour, shallow brown or dark-brown, therefore strengthens through coloured image more to help pseudo-foreign fiber identification of targets.Color space is because the computation model difference can be divided into rgb space and HSI space.Wherein the rgb space use is commonplace, but its not too is fit to human visual characteristic, has therefore produced other various colors spatial representations.

The HSI color space is from the mankind's vision system angle; With tone (Hue), color saturation (Saturation) and brightness (Intensity) color is described; Because people's vision far is better than the sensitivity to color to the sensitivity of brightness; Handle and identification for the ease of color, some systems often adopt the HSI color space, and it more meets human vision property than rgb color space.And when using the Retinex model that the HSI coloured image is strengthened, only need luminance component I is asked its reflected image R, so its processing speed obviously is superior to the rgb space Flame Image Process.

Summary of the invention

The technical matters that (one) will solve

The technical matters that the present invention will solve is: pseudo-foreign fiber colour-image reinforcing method of cotton and system under a kind of uneven illumination are provided; The pseudo-foreign fiber coloured image of cotton that the machine vision platform is collected strengthens; To solve cotton foreign fiber based on the machine vision pseudo-foreign fiber identification problem in detecting, be convenient to follow-up cutting apart and work such as identification.

(2) technical scheme

For addressing the above problem, one aspect of the present invention provides the cotton under a kind of uneven illumination pseudo-foreign fiber colour-image reinforcing method, may further comprise the steps:

S1: read in the pseudo-foreign fiber coloured image of original cotton;

S2: the pseudo-foreign fiber image of original color cotton is transformed into the HSI space from rgb space;

S3: the I component to the HSI coloured image that obtains after the conversion strengthens;

S4: the image to after strengthening through step S3 carries out enhancement post-processing;

S5: will return rgb space through the HSI image transitions after step S4 handles.

Preferably, said step S3 specifically comprises:

S31: gray scale maximal value and the minimum value of finding out said HSI coloured image I component;

S32: according to the gray scale maximal value of said I component and the tonal range of minimum value adjustment image I component is [0,255];

S33: with the gradation conversion of image I component is to the number space;

S34: the best luminosity image that obtains the image I component;

S35: with best luminosity image from the logarithm space conversion to the index space;

S36: with original cotton HSI coloured image I component divided by best luminosity image, the I component image that is enhanced.

Preferably, the step of the best luminosity image of acquisition image I component comprises among the said step S34:

S341: initialization luminosity image;

S342: initialization is apart from array;

S343: loop initialization number of times;

S344: initialization step-length;

S345: begin circulation;

S346: the gradient of calculating the luminosity image;

S347: through gradient and the new luminosity image of step size computation;

S348: the distance of calculating the luminosity image;

S349: compare with last round-robin distance,, calculate and finish if new distance greater than the distance of last time, then withdraws from circulation; Otherwise, if circulation has surpassed the cycle index of setting next time, then withdraw from circulation, calculate and finish, if circulation next time also in the cycle index of setting, is then changeed step S345, continue circulation next time.

Preferably, the computing method of luminosity image gradient are following among the said step S346:

T＝-Δl+α(l-s)

Wherein T is the gradient of luminosity image, and l is the luminosity image, and s is the original I component image, and Δ l is for to carry out the Laplace operator computing to original luminosity image, and α is the similarity parameter of luminosity image and original image.

Preferably, the computing method of new luminosity image are following among the said step S347:

l＝max(l-dt*T，s)

Wherein dt is a step-length.

Preferably, the computing method of said step S348 luminosity image distance are following:

D _j＝||l _j-l _j-1||

Wherein Dj is the j time round-robin distance, l _J-1Be the j-1 time round-robin luminosity image, l _jBe the j time round-robin luminosity image.

Preferably, said step S4 specifically comprises:

S41: the I component image to after strengthening is proofreaied and correct;

S42: will proofread and correct the I component of image and the I component of original image and merge.

Preferably, the correction of among the said step S41 I component image after strengthening being carried out is that Gamma proofreaies and correct, and its method is following:

R＝s/((l _b/w)∧(1-(1/gamma)))

Wherein R is the image after strengthening, and s is the original I component image, l _bBe best luminosity image, w is and complete 255 matrixes of s same scale that gamma is a correction coefficient.

Preferably, said step S42 will proofread and correct the method that the I component of I component and the original image of image merges and be: the I component of original image is replaced with the I component after proofreading and correct.

Preferably, among the said step S5:

1)、

When 0≤H＜120, to the formula of rgb space be with the HSI image transitions after handling:

B＝I(1-S)

$R=I[1+\frac{S\cos H}{\cos (60-H)}]$

G＝3I-(R+B)

2)、

When 120≤H＜240, to the formula of rgb space be with the HSI image transitions after handling:

H＝H-120

R＝I(1-S)

$G=I[1+\frac{S\cos H}{\cos (60-H)}]$

B＝3I-(R+G)

3)、

When 240≤H≤360, to the formula of rgb space be with the HSI image transitions after handling:

H＝H-240

G＝I(1-S)

$B=I[1+\frac{S\cos H}{\cos (60-H)}]$

R＝3I-(G+B)

Wherein, H, S, I are respectively the value of H component in the HSI image, S component and I component; R, G, B are respectively the value of R component in the RGB image, G component and B component.

On the other hand, the present invention also provides a kind of system that is used to realize the pseudo-foreign fiber colour-image reinforcing method of cotton under the above-mentioned uneven illumination, comprising:

The image read module is used to read in the pseudo-foreign fiber coloured image of original cotton;

First image conversion module, the pseudo-foreign fiber image of original color cotton that is used for said image read module is read is transformed into the HSI space from rgb space;

The Image Enhancement Based piece is used for the I component of the HSI coloured image that obtains after the conversion of said first image conversion module is strengthened;

Image enhancement post-processing module is used for the image after the said Image Enhancement Based piece enhancing is carried out enhancement post-processing;

Second image conversion module is used for the HSI image transitions after the said image enhancement post-processing resume module is returned rgb space.

(3) beneficial effect

The pseudo-foreign fiber image intensifying system of cotton of the present invention is handled the image of machine vision collection; I component to coloured image carries out enhancement process; The color characteristic that has farthest kept image; And effectively improved the execution speed of algorithm, helped the online detection of the pseudo-foreign fiber of cotton.

Description of drawings

Fig. 1 is according to the pseudo-foreign fiber image Enhancement Method of embodiment of the invention cotton process flow diagram;

Fig. 2 for according to embodiment of the invention step S3 to conversion after the process flow diagram that strengthens of the I component of the HSI coloured image that obtains;

Fig. 3 is the process flow diagram according to embodiment of the invention calculating optimum luminosity image;

Fig. 4 is the process flow diagram according to embodiment of the invention enhancement post-processing;

Fig. 5 is the structural representation block diagram according to the pseudo-foreign fiber image intensifying system of embodiment of the invention cotton.

Embodiment

Below in conjunction with accompanying drawing and embodiment the present invention is elaborated as follows.

Embodiment one:

As shown in Figure 1, the pseudo-foreign fiber colour-image reinforcing method of the cotton under a kind of uneven illumination may further comprise the steps:

S1: read in the pseudo-foreign fiber coloured image of original cotton.

S2: the pseudo-foreign fiber image of original color cotton is transformed into the HSI space from rgb space.

Because of the HSI model separates the chromatic information of carrying in luminance component and the width of cloth coloured image, so can keep realizing the quick enhancing of image under the constant situation of the pseudo-foreign fiber color of cotton.

The method that the RGB image is converted into the HSI image has multiple, and in the present embodiment, wherein the conversion formula of each component is following:

$\begin{array}{c}\mathrm{\θ}=\arccos \left\{\frac{\frac{1}{2}[(R-G)+(R-B)]}{{[{(R-G)}^2+(R-B)(G-B)]}^{1/2}}\right\}\\ S=1-\frac{3}{(R+G+B)}\left[\min (R,G,B)\right]\end{array}$

$I=\frac{1}{3}(R+G+B)$

Certainly, in other embodiments of the invention, also can adopt other conversion method carry out shown in RGB to the conversion of HIS image.

S3: the I component to the HSI coloured image that obtains after the conversion strengthens, and its computing formula is following:

$\mathrm{Minimize}:F\left[l\right]={\∫}_{\mathrm{\Ω}}({\left(\right|\▿l\left|\right)}^2+\mathrm{\α}{(l-s)}^2)\mathrm{dxdy}$

$\begin{array}{cccc}\mathrm{Subject}& \mathrm{to}:l\&GreaterEqual;s,\mathrm{and}<\&dtri;l,\stackrel{\&RightArrow;}{n}>=0& \mathrm{on}& \&PartialD;\mathrm{\&Omega;}\end{array}$

Wherein Ω is an image collection; L is a luminance picture; S is an original image; presentation video edge, α is the similarity parameter of luminosity image and original image.

As shown in Figure 2, said step S3 specifically comprises:

S31: gray scale maximal value and the minimum value of finding out said HSI coloured image I component;

S32: according to the gray scale maximal value of said I component and the tonal range of minimum value adjustment image I component is [0,255];

S33:, be to the number space with the gradation conversion of image I component according to variable Retinex algorithm;

S34: obtain the best luminosity image of image I component, as shown in Figure 3, specifically may further comprise the steps:

S341: initialization luminosity image in the present embodiment, makes initial luminosity image l=s;

S342: initialization is apart from array, in the present embodiment, makes initial distance array D=(0...0);

S343: the loop initialization number of times in the present embodiment, makes cycle index n=10 (in other embodiments of the invention, this cycle index can also be other natural number);

S344: the initialization step-length in the present embodiment, makes initial step length dt=0.1;

S345: begin circulation;

S346: calculate the gradient of luminosity image, its computing method are following:

T＝-Δl+α(l-s)

Wherein T is the gradient of luminosity image, and l is the luminosity image, and s is the original I component image, and Δ l is for to carry out the Laplace operator computing to original luminosity image, and α is the similarity parameter of luminosity image and original image.The results showed that in the present embodiment, the reinforced effects of the pseudo-foreign fiber coloured image of cotton is relatively good when α is 0.1-0.2;

S347: through gradient and the new luminosity image of step size computation, its computing method are following:

l＝max(l-dt*T，s)

Be that new luminosity image is got the maximal value among l-dt*T and the s, wherein dt is a step-length.Through evidence, in the present embodiment, the better and rapid speed of step S3 reinforced effects when dt is 0.1;

S348: calculate the distance of luminosity image, its computing method are following:

D _j＝||l _j-l _j-1||

Wherein Dj is the j time round-robin distance, l _J-1Be the j-1 time round-robin luminosity image, l _jBe the j time round-robin luminosity image;

S349: compare with last round-robin distance,, calculate and finish if new distance greater than the distance of last time, then withdraws from circulation; Otherwise, if circulation has surpassed the cycle index of setting next time, then withdraw from circulation, calculate and finish, if circulation next time also in the cycle index of setting, is then changeed step S345, continue circulation next time.

S35: with best luminosity image from the logarithm space conversion to the index space;

S36: with original cotton HSI coloured image I component divided by best luminosity image, the I component image that is enhanced.

S4: the image to after strengthening through step S3 carries out enhancement post-processing, and to solve the mistake Enhancement problem that variable Retinex algorithm brings, the pseudo-foreign fiber of the cotton that the color that is guaranteed is constant strengthens image;

As shown in Figure 4, said step S4 specifically comprises:

S41: the I component image to after strengthening is proofreaied and correct;

The correction of among the said step S41 I component image after strengthening being carried out is that Gamma proofreaies and correct, and its method is following:

R＝s/((l _b/w)∧(1-(1/gamma)))

Wherein R is the image after strengthening, and s is the original I component image, l _bBe best luminosity image, w is and complete 255 matrixes of s same scale that gamma is a correction coefficient.Through evidence, in the present embodiment, the reinforced effects of the pseudo-foreign fiber coloured image of cotton is relatively good when gamma gets 2-2.2.

S42: will proofread and correct the I component of image and the I component of original image and merge, the I component that is about to original image replaces with the I component after proofreading and correct, the pseudo-foreign fiber coloured image of the cotton based on the HSI space that can be enhanced.

S5: will return rgb space through the HSI image transitions after step S4 handles, so that the follow-up pseudo-foreign fiber Target Recognition of image segmentation and cotton of further carrying out.

Among the said step S5:

1)、

When 0≤H＜120, to the formula of rgb space be with the HSI image transitions after handling:

B＝I(1-S)

$R=I[1+\frac{S\cos H}{\cos (60-H)}]$

G＝3I-(R+B)

2)、

When 120≤H＜240, to the formula of rgb space be with the HSI image transitions after handling:

H＝H-120

R＝I(1-S)

$G=I[1+\frac{S\cos H}{\cos (60-H)}]$

B＝3I-(R+G)

3)、

When 240≤H≤360, to the formula of rgb space be with the HSI image transitions after handling:

H＝H-240

G＝I(1-S)

$B=I[1+\frac{S\cos H}{\cos (60-H)}]$

R＝3I-(G+B)

Wherein, H, S, I are respectively the value of H component in the HSI image, S component and I component; R, G, B are respectively the value of R component in the RGB image, G component and B component.

Embodiment two:

As shown in Figure 5, a kind of system that is used to realize the pseudo-foreign fiber colour-image reinforcing method of cotton under the above-mentioned uneven illumination comprises:

The image read module is used to read in the pseudo-foreign fiber coloured image of original cotton;

First image conversion module, the pseudo-foreign fiber image of original color cotton that is used for said image read module is read is transformed into the HSI space from rgb space;

The Image Enhancement Based piece is used for the I component of the HSI coloured image that obtains after the conversion of said first image conversion module is strengthened;

Image enhancement post-processing module is used for the image after the said Image Enhancement Based piece enhancing is carried out enhancement post-processing;

Second image conversion module is used for the HSI image transitions after the said image enhancement post-processing resume module is returned rgb space.

Adopt pseudo-foreign fiber colour-image reinforcing method of cotton and system under the uneven illumination of the present invention; Can be according to the parameter of actual conditions adjustment model, can be effectively receive the even or cotton layer of uneven illumination to block the pseudo-foreign fiber coloured image of the cotton not too clearly of causing to strengthen to various.

Above embodiment only is used to explain the present invention; And be not limitation of the present invention; The those of ordinary skill in relevant technologies field under the situation that does not break away from the spirit and scope of the present invention, can also be made various variations and modification; Therefore all technical schemes that are equal to also belong to category of the present invention, and scope of patent protection of the present invention should be defined by the claims.

Claims (11)

1. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination is characterized in that, may further comprise the steps:

S1: read in the pseudo-foreign fiber coloured image of original cotton;

S2: the pseudo-foreign fiber image of original color cotton is transformed into the HSI space from rgb space;

S3: the I component to the HSI coloured image that obtains after the conversion strengthens;

S4: the image to after strengthening through step S3 carries out enhancement post-processing;

S5: will return rgb space through the HSI image transitions after step S4 handles.

2. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 1 is characterized in that said step S3 specifically comprises:

S31: gray scale maximal value and the minimum value of finding out said HSI coloured image I component;

S32: according to the gray scale maximal value of said I component and the tonal range of minimum value adjustment image I component is [0,255];

S33: with the gradation conversion of image I component is to the number space;

S34: the best luminosity image that obtains the image I component;

S35: with best luminosity image from the logarithm space conversion to the index space;

S36: with original cotton HSI coloured image I component divided by best luminosity image, the I component image that is enhanced.

3. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination under the uneven illumination as claimed in claim 2 is characterized in that, the step that obtains the best luminosity image of image I component among the said step S34 comprises:

S341: initialization luminosity image;

S342: initialization is apart from array;

S343: loop initialization number of times;

S344: initialization step-length;

S345: begin circulation;

S346: the gradient of calculating the luminosity image;

S347: through gradient and the new luminosity image of step size computation;

S348: the distance of calculating the luminosity image;

S349: compare with last round-robin distance,, calculate and finish if new distance greater than the distance of last time, then withdraws from circulation; Otherwise, if circulation has surpassed the cycle index of setting next time, then withdraw from circulation, calculate and finish, if circulation next time also in the cycle index of setting, is then changeed step S345, continue circulation next time.

4. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 3 is characterized in that the computing method of luminosity image gradient are following among the said step S346:

T＝-Δl+α(l-s)

Wherein T is the gradient of luminosity image, and l is the luminosity image, and s is the original I component image, and Δ l is for to carry out the Laplace operator computing to original luminosity image, and α is the similarity parameter of luminosity image and original image.

5. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 4 is characterized in that the computing method of new luminosity image are following among the said step S347:

l＝max(l-dt*T，s)

Wherein dt is a step-length.

6. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 3 is characterized in that the computing method of said step S348 luminosity image distance are following:

D _j＝||l _j-l _j-1||

Wherein Dj is the j time round-robin distance, l _J-1Be the j-1 time round-robin luminosity image, l _jBe the j time round-robin luminosity image.

7. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 1 is characterized in that said step S4 specifically comprises:

S41: the I component image to after strengthening is proofreaied and correct;

S42: will proofread and correct the I component of image and the I component of original image and merge.

8. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 7 is characterized in that, the correction of among the said step S41 I component image after strengthening being carried out is that Gamma proofreaies and correct, and its method is following:

R＝s/((l _b/w)∧(1-(1/gamma)))

Wherein R is the image after strengthening, and s is the original I component image, l _bBe best luminosity image, w is and complete 255 matrixes of s same scale that gamma is a correction coefficient.

9. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 7; It is characterized in that said step S42 will proofread and correct the method that the I component of I component and the original image of image merges and be: the I component of original image is replaced with the I component after proofreading and correct.

10. the pseudo-foreign fiber colour-image reinforcing method of the cotton under the uneven illumination as claimed in claim 1 is characterized in that, among the said step S5:

1)、

When 0≤H＜120, to the formula of rgb space be with the HSI image transitions after handling:

B＝I(1-S)

$R=I[1+\frac{S\cos H}{\cos (60-H)}]$

G＝3I-(R+B)

2)、

When 120≤H＜240, to the formula of rgb space be with the HSI image transitions after handling:

H＝H-120

R＝I(1-S)

$G=I[1+\frac{S\cos H}{\cos (60-H)}]$

B＝3I-(R+G)

3)、

When 240≤H≤360, to the formula of rgb space be with the HSI image transitions after handling:

H＝H-240

G＝I(1-S)

$B=I[1+\frac{S\cos H}{\cos (60-H)}]$

R＝3I-(G+B)

Wherein, H, S, I are respectively the value of H component in the HSI image, S component and I component; R, G, B are respectively the value of R component in the RGB image, G component and B component.

11. a system that is used for realizing the pseudo-foreign fiber colour-image reinforcing method of cotton under each uneven illumination of claim 1-10 is characterized in that, comprising:

The image read module is used to read in the pseudo-foreign fiber coloured image of original cotton;

First image conversion module, the pseudo-foreign fiber image of original color cotton that is used for said image read module is read is transformed into the HSI space from rgb space;

The Image Enhancement Based piece is used for the I component of the HSI coloured image that obtains after the conversion of said first image conversion module is strengthened;

Image enhancement post-processing module is used for the image after the said Image Enhancement Based piece enhancing is carried out enhancement post-processing;

Second image conversion module is used for the HSI image transitions after the said image enhancement post-processing resume module is returned rgb space.

Images