CN104851086A - Image detection method for cable rope surface defect - Google Patents

Abstract

The present invention discloses an image detection method for cable rope surface defects, and aims to detect if cable rope surface images have defects. The method comprises the steps of performing graying processing on a cable rope surface image, carrying out enhancement processing on the grayed image by using an improved local grey-scale contrast method, carrying out a cutting processing on the enhanced image by using an improved maximum-correlation method, and, ultimately, performing defect determination according to the size of cut regions of the cut image. With adoption of the method, intelligent detection and judgment of cable rope surface defects can be realized, the workload and fatigue of manual visual inspection are reduced, the detection efficiency may be raised, and the method is safe and reliable.

Description

A kind of image detecting method for cable surface imperfection

Technical field

The present invention relates to a kind of image detecting method for cable surface imperfection, belong to the technical field of image procossing.

Background technology

As one of the main bearing member of bridge---cable, because exposing for a long time in atmosphere, its surface there will be destruction in various degree.In the past is mostly that manual observation detects cable surface, but manual detection very easily causes fatigue, causes and produces the problems such as undetected.

Along with scientific technological advance, image processing techniques has become one of multi-field research direction.It relies on the advantages such as speed is fast, precision is high, never tired, already bears fruit at multirow.Utilize image processing techniques to detect cable surface imperfection and not only can improve detection efficiency, reduction human cost, and lay the first stone for cable quality evaluation system objective standard.

The cable surface image obtained by cable-climbing robot all comprises noise contribution usually.Because gathering and be unavoidably subject in transport process the interference of noise.Therefore be necessary to carry out noise removal process to cable image.Denoising can use image filtering noise-removed technology; Concrete grammar has image enhaucament, image smoothing and image sharpening etc.Image enhaucament refers to the partial information of to give prominence to according to specific demand in piece image, cuts down the disposal route of other information simultaneously.Image enhaucament is not the information strengthening original image, but strengthens the ability to see things in their true light of partial information, and image enhaucament can lose the partial information of image.At present, the evaluation criterion of also ununified image enhancement effects.Certain image enchancing method is very many, can be divided into two classes by processing domain: spatial domain strengthens and frequency field strengthens.Image enhancement technique also can be divided into directly enhancing roughly and indirectly strengthen.Indirect enhancing is by improving image histogram, thus makes the contrast of image obtain enhancing.The adaptive histogram equalization of contrast-limited, color histogram equalization are methods conventional in enhancing technology indirectly.Direct Enhancement Method is improved the contrast of image.Between the brightness that picture contrast refers to light and shade region in image, various level is measured.Its in image directly strengthens in occupation of extremely important status.

In the defects detection of cable surface image, Iamge Segmentation is absolutely necessary step.Interested for user in image target area can extract by Iamge Segmentation.Iamge Segmentation is generally process image based on the uncontinuity of gradation of image and similarity.Generally Iamge Segmentation is divided into Threshold segmentation and rim detection.Rim detection is the uncontinuity based on gradation of image; Threshold segmentation is the similarity based on gradation of image.Target, according to certain automatic generation optimum segmentation threshold value, splits by Threshold segmentation from background.Threshold segmentation is the basic technology in image Segmentation Technology.Threshold segmentation method can be divided into local segmentation method and global segmentation method roughly.In local segmentation method, utilize the segmentation threshold of each regional area of neighborhood dynamic calculation at pixel place.And in global segmentation method, be use whole image-region to obtain a fixing threshold value.Due to simplicity and the validity of global segmentation method, it is used in various fields.

Summary of the invention

Technical matters: the present invention is directed to the accuracy in cable surface defects detection and practicality problem, there is provided a kind of based on gray level image, adopt local contrast Enhancement Method and maximal correlation threshold segmentation method, achieve the detection to cable surface imperfection, improve the accuracy of cable surface defects detection and the image detecting method for cable surface imperfection of practicality.

Technical scheme: the image detecting method for cable surface imperfection of the present invention, comprises the steps:

Step 1: cable surface image to be detected is carried out gray processing process, obtaining the image after gray processing is f (x, y):

f(x,y)＝0.212671×c _r(x,y)+0.71516×c _g(x,y)+0.072169×c _b(x,y)

Wherein, cable surface image to be detected is coloured image c (x, y), and it comprises 3 components: red component c _r(x, y), green component c _g(x, y), blue component c _b(x, y), and picture altitude be H pixel, width is W pixel, (x, y) represents the two-dimensional coordinate of image, and x ∈ [1, W], y ∈ [1, H];

Step 2: adopt the local gray level contrast method improved to carry out enhancing process, image z (x, y) after being enhanced to the image f (x, y) after gray processing;

Step 3: adopt the maximal correlation method improved to carry out dividing processing to the image z (x, y) after strengthening, obtain splitting rear image d (x, y).

In preferred version of the present invention, described step 2), detailed process is as follows:

Step 2.1: the image f (x, y) after gray processing is divided into some nonoverlapping wicket regions subimage wherein i represents the numbering of some wicket images; Disposal route is: be constant depending on y, the subimage that the image f (x, y) that is highly W pixel for H pixel, width to be divided into be highly H pixel, width is 1 pixel and i ∈ [1, W];

Step 2.2: calculate each wicket region subimage the mean value of grey scale pixel value when subimage is and time i ∈ [1, W]:

$u_w^i=\frac{\underset{y=1}{\overset{H}{\mathrm{\Σ}}}{f}_w^i(x,y)}{H};$

Step 2.3: the local gray level correlative value v (x, y) calculating each pixel in the subimage of each wicket region, when subimage is and time i ∈ (1, W):

$v(x,y)=\left\{\begin{array}{cc}\frac{f_w^i(x,y)-u_w^i}{f_w^i(x,y)+u_w^i}& ,f_w^i(x,y)<u_w^i\\ 0& ,f_w^x(x,y)>u_w^x\end{array}\right.;$

Step 2.4: described local gray level correlative value is mapped to image z (x, y) after the enhancing of 256 gray levels according to following formula:

$z(x,y)=\frac{v(x,y)-\min \left(v(x,y)\right)}{\max \left(v(x,y)\right)}\&times;255.$

In preferred version of the present invention, described step 3) detailed process as follows:

Step 3.1: calculate the frequency p that gray-scale value j occurs in image z (x, y) after enhancing _j, wherein j ∈ [0,255]:

$p_j=\frac{f_j}{W\&times;H}$

Wherein f _jrepresent the frequency that gray-scale value j occurs in image z (x, y);

Step 3.2: calculating probability distribution A:

$A=\left\{\begin{array}{cccc}\frac{p_0}{P_T},& \frac{p_1}{P_T},& ...,& \frac{p_T}{P_T}\end{array}\right\}$

Wherein, T is optimal segmenting threshold, and T ∈ [0,255], P ₀, P ₁implication is respectively frequency that gray-scale value 0 occurs in the picture and the frequency that gray-scale value 1 occurs in the picture, P _tfor the general probability of gray-scale value j on interval [0, T], that is:

$P_T=\underset{j=0}{\overset{T}{\mathrm{\&Sigma;}}}{p}_j,$

Work as P _twhen=0, definition A={0,0 ..., 0};

Step 3.3: the relevant C of calculating probability distribution A _a(T):

$C_A\left(T\right)=-\ln \underset{j=0}{\overset{T}{\mathrm{\&Sigma;}}}{\left\{\frac{p_j}{P_T}\right\}}^2$

Step 3.4: the relevant T of computed improved _cw(T):

T _Cw(T)＝(1-P _T) ^α·C _A(T)

Wherein, α is weight parameter;

Step 3.5: traversal T ∈ [0,255], gets and makes T _cw(T) T time maximum, as the optimal threshold obtained, then obtains image d (x, y) after segmentation:

$d(x,y)=\left\{\begin{array}{cc}z(x,y)=0& ,z(x,y)<T\\ z(x,y)=1& ,z(x,y)>T\end{array}\right.;$

In preferred version of the present invention, also comprise step 4: according to image d (x, y) after described segmentation, the number N of statistics d (x, y)=0, judge whether c (x, y) has defect:

The present invention is based on gray level image, adopt the maximal correlation threshold segmentation method of local contrast Enhancement Method and the improvement improved, whether there is defect according to cut zone size determination image.Current cable detection method of surface flaw is artificial visually examine, or after using cable-climbing robot to record cable video surface, manual observation video, searches defect.The present invention can free people, uses computer searching defect.With regard to image processing method, the present invention is directed to cable surface defects detection, after proposing to use local gray level to strengthen first, use maximal correlation threshold segmentation method.Namely being first combined of local gray level Enhancement Method and maximal correlation threshold segmentation method is proposed.With regard to cable surface defects detection, relative to use other Enhancement Method (as histogram equalization strengthen, the adaptive histogram equalization of contrast-limited strengthens, local normalization strengthens) and the combination of other dividing methods (maximum between-cluster variance segmentation, ultimate range segmentation, maximum entropy segmentation), the degree of accuracy that this kind combines and recall ratio all higher.

Beneficial effect: compared with prior art, the present invention has the following advantages:

With regard to cable surface defects detection, method of the present invention refers to and uses the method for image procossing to carry out cable surface defects detection, different from use manual observation in the past; Different from the image processing method that the people such as University Of Chongqing climax propose, their method carries out effective information intercepting to the image gathered, then be utilize Fast Median Filtering noise reduction, then that curved surface projection corrects, being use sobel rim detection, dynamic threshold segmentation and mathematical morphology to carry out Target Segmentation again, is then carry out defect recognition and storage by boundary scan.And method of the present invention first carries out gray processing to the image gathered, then carry out local gray level and strengthen again in conjunction with maximal correlation Threshold segmentation, then carry out defect dipoles by cut zone size.Compared with manual observation method, method cost of the present invention is low, efficiency is high, security good; Compared with the method proposed with University Of Chongqing, the advantage that the present invention brings is that travelling speed is fast, step is simple, accuracy is high, reliability is high.

With regard to image processing method, it is propose first that the local gray level that the present invention proposes strengthens in conjunction with maximal correlation segmentation, compared with existing method [Enhancement Method (adaptive histogram equalization, local normalization as color histogram equalization, contrast-limited) is in conjunction with dividing method (as maximum between-cluster variance, ultimate range, maximum entropy)], method of the present invention is easy, speed is fast, accuracy is high.

With regard to concrete steps of the present invention, adopt gray processing process original image, by originally needing the data volume of process three passages to be reduced to process passage, effectively can improve treatment effeciency, and can not accuracy rate be reduced.

Adopt local contrast Enhancement Method, and for the feature of cable image, because cable is cylindric, there is notable difference the image zone line then gathered and both sides, so consider the difference of entire image, if select overall Enhancement Method, enhancing effect will be reduced to a great extent, then the present invention selects local enhancement methods, and considers the feature of cable image: defect area gray-scale value is less than other district's gray scales usually, comparative selection degree Enhancement Method of the present invention.And the different impact of illumination and cable reflection behavior of surface can be weakened, effectively reject most of noise.

Adopting maximal correlation dividing method, is the feature according to cable image: cable image grey level histogram shows as single-peak response, and does not select method between maximum kind, and in order to improve the efficiency of method, does not select maximum entropy method equally.

And utilize this method to realize detecting to cable surface imperfection, there is the features such as noncontact, speed is fast, simple to operate, reproducible; And processing speed is fast, precision is high, the work of energy long-time stable, reliability is high.Solve the shortcoming that manual detection method wastes time and energy, efficiency is not high.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of whole process of the present invention

Fig. 2 is the simple flow chart of whole process of the present invention, and first row is the design sketch after the corresponding technical finesse of second row

Embodiment

Below in conjunction with accompanying drawing, specific embodiment of the invention is described further.Under windows operating system, select VC++2010 as programming tool, defects detection is carried out to the cable image that equipment obtains.The cable image that this example adopts a width to obtain from equipment, as detected object, finally judges whether it has defect.

Fig. 1 is the process flow diagram of whole process of the present invention.

Fig. 2 is the simple flow chart of whole process of the present invention, and the image of first row corresponds to the result after the technical finesse of second row.Result as " Threshold segmentation " technical finesse is corresponding image above it, and concrete " Threshold segmentation " technology is the maximal correlation threshold segmentation method of the improvement that the present invention adopts.

For the accuracy of cable surface defects detection and the problem of practicality, first the present invention uses gray processing technology to carry out gray processing process to the coloured image gathered, and this process can effectively reduce the data volume of algorithm process.Then adopt the local contrast method of improvement to remove most of noise of image, then adopt the maximal correlation dividing method of improvement to carry out binary segmentation to image, finally determine whether detected image has defect according to the size of cut zone.

When the cable surface defect image that will detect carries out statistical study, conveniently can carry out unified mass management to cable, set up corresponding quality evaluation system, wide accommodation, safe and reliable.

The present invention utilizes image processing techniques to realize detecting to cable surface imperfection, has the features such as noncontact, speed is fast, simple to operate, reproducible; System can use natural light as light source, processing speed is fast, precision is high, and the work of energy long-time stable, algorithm speed is fast, and reliability is high.

Concrete steps are as follows:

Step 1: cable surface image to be detected is carried out gray processing process, obtaining the image after gray processing is f (x, y):

f(x,y)＝0.212671×c _r(x,y)+0.71516×c _g(x,y)+0.072169×c _b(x,y)

Wherein, cable surface image to be detected is coloured image c (x, y), and it comprises 3 components: red component c _r(x, y), green component c _g(x, y), blue component c _b(x, y), and picture altitude be H pixel, width is W pixel, (x, y) represents the two-dimensional coordinate of image, and x ∈ [1, W], y ∈ [1, H];

The secondary figure of Fig. 2 first row the 2nd is the image after gray processing.

Concrete gray processing method also important method, maximum value process, mean value method.The method that the present invention adopts is method of weighted mean, because the sensitivity of human eye to green is the highest, minimum to the sensitivity of blueness, therefore, is weighted by above formula and on average can obtains more rational gray level image red, green, blue component.In the present invention, the fundamental purpose of gray processing process reduces the data volume of process, finally raises the efficiency.

Step 2: adopt the local gray level contrast method improved to carry out enhancing process, image z (x, y) after being enhanced to the image f (x, y) after gray processing, detailed process is as follows:

Step 2.1: the image f (x, y) after gray processing is divided into some nonoverlapping wicket regions subimage wherein i represents the numbering of some wicket images; Disposal route is: be constant depending on y, the subimage that the image f (x, y) that is highly W pixel for H pixel, width to be divided into be highly H pixel, width is 1 pixel and i ∈ [1, W].

Not overlapping wicket region in this step is strip, in fact also can choose the wicket region of other shapes according to the specific features of image, as square, circular and other shapes.The strip that the present invention selects is cylindric based on cable, and gather camera with cable-climbing robot along length of warping winch direction collection cable image, therefore select the strip wicket along length of warping winch direction, also consider the feature that cable surface optical characteristics is along its length similar; And do not select other shapes such as square, circular, consider that the surface optical characteristics along cable circumferencial direction changes greatly.

Step 2.2: calculate each wicket region subimage the mean value of grey scale pixel value when subimage is and time i ∈ [1, W]:

$u_w^i=\frac{\underset{y=1}{\overset{H}{\mathrm{\&Sigma;}}}{f}_w^i(x,y)}{H};$

Step 2.3: the local gray level correlative value v (x, y) calculating each pixel in the subimage of each wicket region, when subimage is and time i ∈ [1, W]:

$v(x,y)=\left\{\begin{array}{cc}\frac{f_w^i(x,y)-u_w^i}{f_w^i(x,y)+u_w^i}& ,f_w^i(x,y)<u_w^i\\ 0& ,f_w^x(x,y)>u_w^x\end{array}\right.;$

According to cable feature of image, in wicket region, the gray-scale value of defect area is less than the gray-scale value of background area; In zonule, extraneous illumination can think identical, and in zonule, the optical characteristics on cable surface is similar; Then in wicket, the region that gray-scale value is less than mean value is that the possibility of defect is large; And to be the possibility of noise large in the gray-scale value region larger than mean value; When ensureing accuracy of detection, region gray-scale value being greater than mean value is set to 0, and such process not only can reduce calculated amount, and can reach the effect eliminating partial noise.

Step 2.4: described local gray level correlative value is mapped to image z (x, y) after the enhancing of 256 gray levels according to following formula:

$z(x,y)=\frac{v(x,y)-\min \left(v(x,y)\right)}{\max \left(v(x,y)\right)}\&times;255;$

The secondary figure of Fig. 2 first row the 3rd is the image after Enhancement Method of the present invention strengthens.

Step 3: adopt the maximal correlation method improved to carry out dividing processing to the image z (x, y) after strengthening, obtain splitting rear image d (x, y), detailed process is as follows:

Step 3.1: calculate the frequency p that gray-scale value j occurs in image z (x, y) after enhancing _j, wherein j ∈ [0,255]:

$p_j=\frac{f_j}{W\&times;H}$

Wherein f _jrepresent the frequency that gray-scale value j occurs in image z (x, y),

Step 3.2: calculating probability distribution A:

$A=\left\{\begin{array}{cccc}\frac{p_0}{P_T},& \frac{p_1}{P_T},& ...,& \frac{p_T}{P_T}\end{array}\right\}$

Wherein, T is optimal segmenting threshold, and T ∈ [0,255], P ₀, P ₁implication is respectively gray-scale value when being 0,1, the frequency occurred in the picture.P _tfor the general probability of gray-scale value j on interval [0, T], that is:

$P_T=\underset{j=0}{\overset{T}{\mathrm{\&Sigma;}}}{p}_j,$

Work as P _twhen=0, definition A={0,0 ..., 0};

Step 3.3: the relevant C of calculating probability distribution A _a(T):

$C_A\left(T\right)=-\ln \underset{j=0}{\overset{T}{\mathrm{\&Sigma;}}}{\left\{\frac{p_j}{P_T}\right\}}^2$

Step 3.4: the relevant T of computed improved _cw(T):

T _Cw(T)＝(1-P _T) ^α·C _A(T)

Wherein, α is weight parameter;

The Maximum correlation method improved is mainly according to the feature of cable image: in defect image, defect area is less, and optimal segmenting threshold is by relevant C _a(T) impact is larger.

In this step, the value of weight parameter α is important, and because it can change choosing of optimal threshold, in this example, weight parameter α value is 1.Research shows, the degree of accuracy of detection rises along with the increase of weight parameter α value, and the recall ratio detected declines along with the increase of weight parameter α value.This is because when weight parameter alpha is larger, the optimal threshold of acquisition is less, make defect area detect reduction, and the defect correctness detected improves, then degree of accuracy rises, and the region of defect has omission, then recall ratio declines.In real image process, then show as undetected defect more, but in the defect detected, occur that the situation of false defect is less.

Step 3.5: traversal T ∈ [0,255], gets and makes T _cw(T) T time maximum, as the optimal threshold obtained, then obtains image d (x, y) after segmentation:

$d(x,y)=\left\{\begin{array}{cc}z(x,y)=0& ,z(x,y)<T\\ z(x,y)=1& ,z(x,y)>T\end{array}\right.;$

The secondary figure of Fig. 2 first row the 4th is the image after automatic Segmentation of the present invention.

Step 4: according to segmentation image d (x, y), the number N of statistics d (x, y)=0, judges whether c (x, y) has defect:

It is pointed out that the method for defect dipoles is more, the present invention considers that said method is simple and practical, and when ensureing accuracy of detection, the threshold value of judgement can change.The decision threshold chosen in this step is the feature according to cable image: the area that defect area occupies is less.

Above-described embodiment is only the preferred embodiment of the present invention; be noted that for those skilled in the art; under the premise without departing from the principles of the invention; some improvement and equivalent replacement can also be made; these improve the claims in the present invention and are equal to the technical scheme after replacing, and all fall into protection scope of the present invention.

Claims (4)

1. for an image detecting method for cable surface imperfection, it is characterized in that, the method step is as follows:

Step 1: cable surface image to be detected is carried out gray processing process, obtaining the image after gray processing is f (x, y):

f(x,y)＝0.212671×c _r(x,y)+0.71516×c _g(x,y)+0.072169×c _b(x,y)

Wherein, cable surface image to be detected is coloured image c (x, y), and it comprises 3 components: red component c _r(x, y), green component c _g(x, y), blue component c _b(x, y), and picture altitude be H pixel, width is W pixel, (x, y) represents the two-dimensional coordinate of image, and x ∈ [1, W], y ∈ [1, H];

Step 2: adopt the local gray level contrast method improved to carry out enhancing process, image z (x, y) after being enhanced to the image f (x, y) after gray processing;

Step 3: adopt the maximal correlation method improved to carry out dividing processing to the image z (x, y) after strengthening, obtain splitting rear image d (x, y).

2. the image detecting method for cable surface imperfection according to claim 1, is characterized in that, described step 2), detailed process is as follows:

Step 2.1: the image f (x, y) after gray processing is divided into some nonoverlapping wicket regions subimage wherein i represents the numbering of some wicket images; Disposal route is: be constant depending on y, the subimage that the image f (x, y) that is highly W pixel for H pixel, width to be divided into be highly H pixel, width is 1 pixel and i ∈ [1, W];

Step 2.2: calculate each wicket region subimage the mean value of grey scale pixel value when subimage is and time i ∈ [1, W]:

Step 2.3: the local gray level correlative value v (x, y) calculating each pixel in the subimage of each wicket region, when subimage is and time i ∈ (1, W):

Step 2.4: described local gray level correlative value is mapped to image z (x, y) after the enhancing of 256 gray levels according to following formula:

3. the image detecting method for cable surface imperfection according to claim 1, is characterized in that, described step 3) detailed process as follows:

Step 3.1: calculate the frequency p that gray-scale value j occurs in image z (x, y) after enhancing _j, wherein j ∈ [0,255]:

Wherein f _jrepresent the frequency that gray-scale value j occurs in image z (x, y);

Step 3.2: calculating probability distribution A:

Wherein, T is optimal segmenting threshold, and T ∈ [0,255], P ₀, P ₁implication is respectively frequency that gray-scale value 0 occurs in the picture and the frequency that gray-scale value 1 occurs in the picture, P _tfor the general probability of gray-scale value j on interval [0, T], that is:

Work as P _twhen=0, definition A={0,0 ..., 0};

Step 3.3: the relevant C of calculating probability distribution A _a(T):

Step 3.4: the relevant T of computed improved _cw(T):

T _Cw(T)＝(1-P _T) ^α·C _A(T)

Wherein, α is weight parameter;

Step 3.5: traversal T ∈ [0,255], gets and makes T _cw(T) T time maximum, as the optimal threshold obtained, then obtains image d (x, y) after segmentation:

。

4. the image detecting method for cable surface imperfection according to claim 1,2 or 3, it is characterized in that, the method also comprises step 4: according to image d (x after described segmentation, y), statistics d (x, the number N of y)=0, judges whether c (x, y) has defect: