KR102133330B1 - Apparatus and method for recognizing crack in sturcutre - Google Patents

Abstract

A crack recognition apparatus for a structure according to an embodiment of the present invention includes a crack image receiving unit receiving a crack image of the structure, a histogram analyzing unit analyzing a histogram of the crack image received from the crack image receiving unit, and calculating a bimodality. Based on the bimodality calculated by the analysis unit, the presence or absence of a shadow in a crack image is determined, and a shadow determination unit is divided into a shadow image with a shadow and a non-shadow image without a shadow, and the shadow image is obtained from the shadow determination unit. A shadow removal unit that performs a shadow removal operation only for crack images classified as, Recognize a crack in a shadow removal image where the shadow removal operation is performed by the shadow removal unit, or in the non-shadow image separated by the shadow determination unit And a crack recognition unit that recognizes cracks.

Description

Structure crack recognition device and method{APPARATUS AND METHOD FOR RECOGNIZING CRACK IN STURCUTRE}

The present invention relates to an apparatus and method for crack detection of structures using bimodality.

Investigation of cracks in concrete structures is an important judgment factor in the precise safety diagnosis of concrete structures.

In general, a visual observation method is used to investigate cracks in a concrete structure. However, the visual observation method has a disadvantage of low accuracy and reliability.

Recently, research has been actively conducted on a method of measuring cracks in concrete structures by image processing.

Prior art related to the present invention is Republic of Korea Patent Registration No. 10-1772916 (August 31, 2017). The preceding document discloses a method and system for crack detection of a tunnel lining surface using an image and artificial intelligence algorithm based crack detection automation program.

A method for detecting cracks occurring on a concrete surface using a synthetic product neural network (CNN) technique and discriminating final cracks from noises using the characteristics of cracks is introduced in the prior art.

However, the method for removing the influence due to the shadow is not disclosed at all, and if there is a shadow on the surface of the structure, it is difficult to accurately detect cracks because it is affected by the shadow.

An object of the present invention is to provide a crack recognition apparatus and method for a structure capable of accurately recognizing a crack by removing a shadow signal from a crack image of the structure.

Another object of the present invention is to provide a crack recognition apparatus and method for a structure capable of quickly and accurately recognizing a crack by determining the presence or absence of a shadow based on bimodality.

Another object of the present invention is to perform a shadow removal operation only for images classified as shadow images, and omit the shadow removal process for non-shadow images, so that crack recognition of structures that can shorten the processing time required for crack recognition It provides an apparatus and method.

The objects of the present invention are not limited to the objects mentioned above, and other objects and advantages of the present invention not mentioned can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. In addition, it will be readily appreciated that the objects and advantages of the present invention can be realized by means of the appended claims and combinations thereof.

An apparatus for recognizing a crack of a structure according to an embodiment of the present invention includes a crack image receiving unit for receiving a crack image of the structure; A histogram analysis unit to calculate a bimodality by analyzing a histogram of the crack image received from the crack image reception unit; A shadow determination unit for determining the presence or absence of a shadow in a crack image based on the bimodality calculated by the histogram analysis unit, and classifying a shadow image with a shadow and a non-shadow image without a shadow; A shadow removal unit that performs a shadow removal operation only on the crack image divided into the shadow image by the shadow determination unit; And a crack recognition unit for recognizing a crack in the shadow removal image in which the shadow removal operation is performed by the shadow removal unit, or for recognizing a crack in the non-shadow image separated by the shadow determination unit.

Preferably, the histogram analysis unit includes: a threshold calculation unit configured to calculate a histogram of the received crack image and set a point at which the variance between classes becomes maximum as a threshold value; A polynomial estimator for dividing an image into two groups based on the calculated threshold and estimating a quadratic polynomial for variables and frequencies in each group; A peak calculating unit calculating a peak value including a direction and a magnitude of a peak of each group from the estimated quadratic polynomial; And a bimodality calculator configured to calculate bimodality using the calculated peak value.

In addition, in the threshold calculation unit, the variance between classes may be calculated based on Equation 1 below.

[Equation 1]

는 클래스간 분산이고, m1과 m2는 각 클래스의 평균이며, P1과 P2는 가중치로 통계적으로 픽셀 값이 나타날 확률임)(here

Is the variance between classes, m1 and m2 are averages of each class, and P1 and P2 are the probability of statistically appearing pixel values as weights)

In addition, in the threshold value calculation unit, the threshold value for maximizing the variance between classes may be calculated based on Equation 2 below.

[Equation 2]

는 임계값이고, x는 화소임)(here

Is a threshold, x is a pixel)

In addition, in the polynomial estimator, a quadratic polynomial of each group divided into two groups based on the threshold may be estimated based on Equation 3 below.

[Equation 3]

와

는 2차 다항식의 계수임)(here

Wow

Is the second-order polynomial coefficient)

In addition, in the peak calculator, the peak value may be calculated based on Equation 4 below from the estimated quadratic polynomial.

[Equation 4]

(Where z _k is the peak value of the two groups)

Also, in the bimodality calculating unit, the bimodality may be calculated by substituting the peak value into Equation 5 below.

[Equation 5]

In addition, when the bimodality is greater than the threshold, the shadow determination unit may classify the corresponding image as the shadow image.

In addition, the shadow removal unit may perform the shadow removal operation based on Equation 6 below.

[Equation 6]

는 그림자 제거 연산 이전의 그림자 영상이고,

는 그림자 영상의 메디안 필터링 결과이며,

는 그림자 제거 연산 이후의 그림자 제거 영상임)(here

Is the shadow image before the shadow removal operation,

Is the median filtering result of the shadow image,

Is a shadow removal image after the shadow removal operation)

In addition, the crack recognition unit recognizes a crack in the shadow removal image or the non-shadow image, wherein the crack recognition extracts a crack candidate through binarization based on contrast, and then extracts the size and circularity of the extracted object. It can be recognized as a crack when it satisfies the setting conditions.

On the other hand, according to another embodiment of the present invention provides a method for recognizing the crack of a structure using the crack recognition device of the structure. Structure crack recognition method according to an embodiment of the present invention (a) in the crack image receiving unit, receiving a crack image of the structure; (b) calculating a bimodality by analyzing the histogram of the crack image received from the crack image receiver in the histogram analysis unit; (c) in the shadow determination unit, determining the presence or absence of a shadow in a crack image based on the bimodality calculated from the histogram analysis unit and classifying the shadow image as a shadow image and a non-shadow image as a shadow image. ; (d) in the shadow removing unit, performing the shadow removing operation on the crack image divided into the shadow image by the shadow determining unit; And (e) in the crack recognition unit, recognizing a crack in the shadow removal image in which the shadow removal operation is performed, or recognizing a crack in the non-shadow image separated by the shadow determination unit.

In this case, the step (b) may include: (b-1) calculating, by the threshold calculating unit, a histogram of the received crack image to set a point at which the variance between classes becomes the maximum as a threshold value; (b-2) estimating a quadratic polynomial for variables and frequencies in each group by dividing the image into two groups based on the calculated threshold value in the polynomial estimator; (b-3) calculating a peak value including a direction and a magnitude of a peak of each group from the estimated second order polynomial in the peak calculator; And (b-4) calculating the bimodality using the calculated peak value in the bimodality calculating unit.

In addition, in step (b-1), in the threshold calculation unit, the inter-class variance may be calculated based on Equation 1 below.

[Equation 1]

는 클래스간 분산이고, m1과 m2는 각 클래스의 평균이며, P1과 P2는 가중치로 통계적으로 픽셀 값이 나타날 확률임)(here

Is the variance between classes, m1 and m2 are averages of each class, and P1 and P2 are the probability of statistically appearing pixel values as weights)

In addition, in the step (b-1), in the threshold value calculating unit, the threshold value for maximizing the variance between classes may be calculated based on Equation 2 below.

[Equation 2]

는 임계값이고, x는 화소임)(here

Is a threshold, x is a pixel)

In addition, in the step (b-2), in the polynomial estimator, a quadratic polynomial of each group divided into two groups based on the threshold may be estimated based on Equation 3 below.

[Equation 3]

와

는 2차 다항식의 계수임)(here

Wow

Is the second-order polynomial coefficient)

In addition, in step (b-3), in the peak calculator, the peak value may be calculated based on Equation 4 below from the estimated quadratic polynomial.

[Equation 4]

(Where z _k is the peak value of the two groups)

In addition, in the step (b-4), in the bimodality calculating unit, the bimodality may be calculated by substituting the peak value into Equation 5 below.

[Equation 5]

In addition, in step (c), when the bimodality is greater than the threshold value, the shadow determination unit may classify the corresponding image as the shadow image.

In addition, in the step (d), in the shadow removal unit, the shadow removal operation may be performed based on Equation 6 below.

[Equation 6]

는 그림자 제거 연산 이전의 그림자 영상이고,

는 그림자 영상의 메디안 필터링 결과이며,

는 그림자 제거 연산 이후의 그림자 제거 영상임)(here

Is the shadow image before the shadow removal operation,

Is the median filtering result of the shadow image,

Is a shadow removal image after the shadow removal operation)

In the step (e), the crack recognition unit recognizes the crack of the shadow removal image or the non-shadow image, wherein the crack recognition extracts a crack candidate through binarization based on contrast, and then extracts the crack. If the set condition is satisfied by using the size and circularity of the object, it can be recognized as a crack.

According to the apparatus and method for crack recognition of a structure according to the present inventors, there is an advantageous technical effect that can accurately recognize a crack by removing a shadow signal from the crack image of the structure.

For example, according to the apparatus and method for crack recognition of a structure of the present invention, the histogram analysis unit analyzes the histogram of the image received by the shadow recognition unit at the crack image reception unit. In addition, the shadow determination unit determines whether a shadow is present based on the bimodality analyzed by the histogram analysis unit. In addition, the shadow removal unit performs a shadow removal operation only for the image in which the shadow determined by the shadow determination unit exists. In addition, the crack recognition unit recognizes a crack with respect to a crack image removed by the shadow removal unit or a crack image without shadow. According to the configuration as described above, according to the structure crack recognition apparatus of the present invention, there is an advantage that the structure crack can be recognized quickly and accurately compared to the conventional one.

In addition, according to the apparatus and method for detecting cracks in a structure of the present invention, the histogram analysis unit calculates a histogram of an image and sets a point at which the variance between classes becomes maximum as a threshold. In addition, the histogram analysis unit estimates a quadratic polynomial for the frequency and variables of the two groups divided based on the threshold. After that, the histogram analysis unit has an advantage of calculating the size and direction of the peaks in each group from the estimated polynomial, and measuring the bimodality using the calculated peaks.

In addition, according to the apparatus and method for crack recognition of a structure of the present invention, it is possible to perform a shadow removal operation only for an image classified as a shadow image and omit a shadow removal process for a non-shadow image. Accordingly, according to the present invention, there is an advantage that can significantly shorten the processing time required to recognize the crack of the structure.

In addition to the above-described effects, the concrete effects of the present invention will be described together while describing the specific matters for carrying out the invention.

1 is a block diagram briefly showing a crack recognition device of a structure according to an embodiment of the present invention.
Figure 2 is a block diagram showing the detailed configuration of the histogram analysis unit in the crack recognition device of the structure according to an embodiment of the present invention.
3 is a flowchart illustrating a method for recognizing a crack in a structure according to an embodiment of the present invention.
4 is a photograph showing a shadow image (a) with a shadow and a non-shadow image (b) with no shadow among the crack images of the structure.
5 is a graph showing histograms of a shadow image and a non-shadow image.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the drawings so that those skilled in the art to which the present invention pertains can easily practice. The present invention can be implemented in many different forms and is not limited to the embodiments described herein.

In order to clearly describe the present invention, parts irrelevant to the description are omitted, and the same reference numerals are assigned to the same or similar elements throughout the specification. In addition, some embodiments of the present invention will be described in detail with reference to exemplary drawings. In adding reference numerals to the components of each drawing, the same components may have the same reference numerals as possible even though they are displayed on different drawings. In addition, in describing the present invention, when it is determined that detailed descriptions of related well-known structures or functions may obscure the subject matter of the present invention, detailed descriptions thereof may be omitted.

In describing the components of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are only for distinguishing the components from other components, and the essence, order, order, or number of the components is not limited by the terms. When a component is described as being "connected", "coupled" or "connected" to another component, the component may be directly connected to or connected to the other component, but different components between each component It should be understood that the "intervenes" may be, or each component may be "connected", "coupled" or "connected" through other components.

In addition, in implementing the present invention, components may be subdivided for convenience of description, but these components may be implemented in one device or module, or one component may be multiple devices or modules. It can be implemented by being divided into.

1 is a block diagram schematically showing a crack recognition device of a structure according to an embodiment of the present invention, and FIG. 2 is a block diagram showing a detailed configuration of a histogram analysis unit among crack recognition devices of a structure according to an embodiment of the present invention It is.

Referring to FIG. 1, a crack recognition apparatus 1000 of a structure according to an embodiment of the present invention includes a crack image receiving unit 100, a histogram analysis unit 200, a shadow determination unit 300, and a shadow removal unit 400 , And a crack recognition unit 500.

Crack image receiving unit 100 receives the crack image of the structure. Here, the structure refers to a concrete structure, etc., and the crack image of the structure refers to an image of a crack occurring on a concrete surface.

The histogram analysis unit 200 analyzes the histogram of the crack image received from the crack image reception unit 100. In addition, the histogram analysis unit 200 calculates bimodality through a histogram analysis of crack images.

The shadow determination unit 300 determines the presence or absence of a shadow in the crack image based on the bimodality calculated by the histogram analysis unit 200.

As a result of determining the presence or absence of a shadow in the crack image based on the bimodality, the shadow determination unit 300 shows an image with a shadow (hereinafter referred to as a'shadow image') and an image without a shadow (hereinafter referred to as a'non-shadow'). Video).

The shadow removal unit 400 performs the shadow removal operation only on the crack image divided into the shadow image by the shadow determination unit 300,

The crack recognition unit 500 recognizes a crack in an image in which the shadow removal operation is performed by the shadow removal unit 400 (hereinafter, referred to as a “shadow removal image”). Alternatively, the crack recognition unit 500 recognizes the crack from the image in which the shadow does not exist in the shadow determination unit 300, that is, an image classified as a non-shadow image.

Specifically, the histogram analysis unit 200 includes a threshold calculation unit 210, a polynomial estimation unit 220, a peak calculation unit 230, and a bimodality calculation unit 240 as illustrated in FIG. 2. do.

The threshold calculating unit 210 calculates a histogram of the received crack image and sets a point at which the variance between classes becomes maximum as a threshold value.

For example, in the threshold calculating unit 210, the inter-class variance may be calculated based on Equation 1 below.

[Equation 1]

는 클래스간 분산이고, m1과 m2는 각 클래스의 평균이며, P1과 P2는 가중치로 통계적으로 픽셀 값이 나타날 확률을 나타낸다.here,

Is the variance between classes, m1 and m2 are averages of each class, and P1 and P2 are the weights, which indicate the probability of statistically appearing pixel values.

In addition to this, in the threshold calculating unit 210, a threshold value for maximizing the variance between classes may be calculated based on Equation 2 below.

[Equation 2]

는 임계값이고, x는 화소를 나타낸다.here,

Is a threshold value, and x represents a pixel.

The polynomial estimator 220 divides the image into two groups based on the calculated threshold value and estimates a quadratic polynomial for variables and frequencies in each group.

For example, in the polynomial estimator 220, the second-order polynomial of each group divided into two groups based on the threshold may be estimated based on Equation 3 below.

[Equation 3]

와

는 2차 다항식의 계수를 나타낸다.here,

Wow

Denotes the coefficient of the second-order polynomial.

On the other hand, the peak calculator 230 calculates a peak value including the direction and magnitude of the peaks of each group from the estimated quadratic polynomial.

For example, in the peak calculating unit 230, the peak value is obtained based on the following equation (4) from the estimated quadratic polynomial.

[Equation 4]

Here, z _k represents the peak values of the two groups.

The bimodality calculator 240 calculates bimodality using the calculated peak value.

For example, in the bimodality calculator 240, the bimodality may be calculated by substituting the peak value into Equation 5 below.

[Equation 5]

Meanwhile, when the bimodality is greater than a threshold, the shadow determination unit 300 classifies the corresponding image as the shadow image.

The shadow removal unit 400 may perform the shadow removal operation based on Equation 6 below.

[Equation 6]

는 그림자 제거 연산 이전의 그림자 영상이고,

는 그림자 영상의 메디안 필터링 결과이며,

는 그림자 제거 연산 이후의 그림자 제거 영상을 나타낸다.here,

Is the shadow image before the shadow removal operation,

Is the median filtering result of the shadow image,

Indicates a shadow removal image after the shadow removal operation.

Meanwhile, the crack recognition unit 500 recognizes a crack in a shadow removal image or a non-shadow image, but the crack recognition extracts a crack candidate through binarization based on contrast, and then uses the size and circularity of the extracted object. Therefore, it can be recognized as a crack when the setting conditions are satisfied.

Hereinafter, a method for recognizing cracks in a structure according to an embodiment of the present invention will be described.

3 is a flowchart illustrating a method for recognizing a crack in a structure according to an embodiment of the present invention.

Referring to FIG. 3, a crack recognition method of a structure according to an embodiment of the present invention includes a crack image reception step (S100), a histogram analysis step (S200), a shadow determination step (S300), a shadow removal step (S400), and Crack recognition step (S500).

Crack image receiving step (S100) refers to the step of receiving the crack image of the structure in the crack image receiving unit (100, see FIG. 1).

The histogram analysis step S200 refers to a step in which the histogram analysis unit 200 (refer to FIG. 1) analyzes the histogram of the received crack image to calculate bimodality.

The shadow determination step S300 refers to a step of determining the presence or absence of a shadow in the crack image based on the bimodality calculated by the histogram analysis unit 200 (see FIG. 1) in the shadow determination unit 300 (see FIG. 1 ). As described above, in the shadow determining step S300, a shadow image with a shadow and a non-shadow image with no shadow are distinguished.

The shadow removal step (S400) refers to a step of performing a shadow removal operation on a crack image divided into a shadow image by the shadow removal unit 400 (refer to FIG. 1).

In the crack recognition step (S500), the crack recognition unit 500 (refer to FIG. 1) recognizes a crack in the shadow removal image in which the shadow removal operation is performed, or in a non-shadow image separated by the shadow determination unit 300 (refer to FIG. 1). It is a step to recognize cracks.

For example, the histogram analysis step S200 includes a threshold calculation step, a polynomial estimation step, a peak calculation step, and a bimodality calculation step.

Specifically, in the threshold calculation step, the threshold calculation unit 210 (refer to FIG. 2) calculates a histogram of the received crack image, and sets a point at which variance between classes becomes maximum as a threshold value.

In this case, in the threshold calculation unit 210 (see FIG. 2 ), the inter-class variance may be calculated based on Equation 1 below.

[Equation 1]

는 클래스간 분산이고, m1과 m2는 각 클래스의 평균이며, P1과 P2는 가중치로 통계적으로 픽셀 값이 나타날 확률을 나타낸다.here,

Is the variance between classes, m1 and m2 are averages of each class, and P1 and P2 are the weights, which indicate the probability of statistically appearing pixel values.

In addition, in the threshold calculation unit 210 (refer to FIG. 2 ), the threshold value for maximizing the variance between classes may be calculated based on Equation 2 below.

[Equation 2]

는 임계값이고, x는 화소를 나타낸다.here,

Is a threshold value, and x represents a pixel.

Next, in the polynomial estimation step, the polynomial estimator 220 (refer to FIG. 2) divides the image into two groups based on the calculated threshold value to estimate the quadratic polynomial for variables and frequencies in each group.

For example, in the polynomial estimator 220 (refer to FIG. 2 ), the quadratic polynomial of each group divided into two groups based on the threshold may be estimated based on Equation 3 below.

[Equation 3]

와

는 2차 다항식의 계수를 나타낸다.here,

Wow

Denotes the coefficient of the second-order polynomial.

Next, in the peak calculation step, the peak calculation unit 230 (refer to FIG. 2) calculates a peak value including the direction and magnitude of the peaks of each group from the estimated quadratic polynomial.

For example, in the peak calculating unit 230 (refer to FIG. 2), the peak value is obtained from the estimated quadratic polynomial based on Equation 4 below.

[Equation 4]

Here, z _k represents the peak values of the two groups.

Next, in the bimodality calculating step, the bimodality calculating unit 240 (refer to FIG. 2) calculates the bimodality using the calculated peak value.

For example, in the bimodality calculator 240 (see FIG. 2 ), the bimodality may be calculated by substituting the peak value into Equation 5 below.

[Equation 5]

On the other hand, in the shadow determining step S300, the shadow determining unit 300 (see FIG. 1) determines whether the bimodality satisfies the set condition. If the bimodality is greater than the threshold, the corresponding image is classified as a shadow image. . And in the shadow removal step (S400), the shadow removal unit 400 (see FIG. 1) performs a shadow removal operation. The shadow removal operation can be performed based on Equation 6 below.

[Equation 6]

는 그림자 제거 연산 이전의 그림자 영상이고,

는 그림자 영상의 메디안 필터링 결과이며,

는 그림자 제거 연산 이후의 그림자 제거 영상을 나타낸다.here,

Is the shadow image before the shadow removal operation,

Is the median filtering result of the shadow image,

Indicates a shadow removal image after the shadow removal operation.

On the other hand, in the crack recognition step (S500), the crack recognition unit 500 (refer to FIG. 1) recognizes a crack in a shadow removal image or a non-shadow image.

For example, crack recognition is performed by extracting a crack candidate through binarization based on contrast, and then recognized as a crack when a set condition is satisfied using the size and circularity of the extracted object.

4 is a photograph showing a shadow image (a) with a shadow and a non-shadow image (b) with no shadow among the crack images of the structure. And FIG. 5 is a graph showing histograms of a shadow image and a non-shadow image.

Looking at the shape of the histograms of the two images shown in FIG. 5, it can be confirmed that the histograms of the shadow images exhibit bimodal characteristics, and that there is a difference from the non-shadow images representing unimodal characteristics.

The bimodality of the shadow image and the non-shadow image measured by the bimodality measurement method presented in the present invention was found to be 0.86 and 0, respectively, and shows that the shadow image and the non-shadow image can be distinguished by calculating the bimodality.

As described above, according to the configuration and operation of the present invention, by removing the shadow signal from the crack image of the structure, it is possible to accurately recognize the crack.

As a specific example, the histogram analysis unit analyzes the histogram of the image received by the shadow recognition unit from the crack image reception unit. In addition, the shadow determination unit determines whether a shadow is present based on the bimodality analyzed by the histogram analysis unit. In addition, the shadow removal unit performs a shadow removal operation only for the image in which the shadow determined by the shadow determination unit exists. In addition, the crack recognition unit recognizes a crack with respect to a crack image removed by the shadow removal unit or a crack image without shadow. As a result, cracks in the structure can be recognized more quickly and accurately.

Furthermore, the histogram analysis unit calculates a histogram of the image and sets a point at which the variance between classes becomes maximum as a threshold value. Subsequently, the histogram analysis unit estimates a quadratic polynomial for the frequency and variables of the two groups divided based on the threshold. Subsequently, the histogram analysis unit calculates the size and direction of the peaks in each group from the estimated polynomial, and measures the bimodality using the calculated peaks, thereby quickly and accurately determining the presence or absence of a shadow.

Furthermore, the shadow removal operation can be performed only for images classified as shadow images, and the shadow removal process can be omitted for non-shadow images. As a result, it is possible to significantly reduce the processing time required for crack recognition of the structure by reducing the time required for the unnecessary image processing process.

As described above, the present invention has been described with reference to the exemplified drawings, but the present invention is not limited by the examples and drawings disclosed in the present specification, and can be varied by those skilled in the art within the scope of the technical idea of the present invention. It is obvious that modifications can be made. In addition, although the operation and effect according to the configuration of the present invention has not been explicitly described while describing the embodiments of the present invention, it is natural that the predictable effect by the configuration should also be recognized.

S100: crack image receiving step
S200: histogram analysis step
S300: shadow judgment step
S400: shadow removal steps
S500: crack recognition step
100: crack image receiver
200: histogram analysis unit
210: threshold calculation unit
220: polynomial estimation unit
230: peak calculator
240: bimodality calculation unit
300: shadow judgment unit
400: shadow remover
500: crack recognition unit

Claims (20)

A crack image receiver for receiving a crack image of the structure;
A histogram analysis unit to calculate a bimodality by analyzing a histogram of the crack image received from the crack image reception unit;
A shadow determination unit for determining the presence or absence of a shadow in a crack image based on the bimodality calculated by the histogram analysis unit and classifying a shadow image with a shadow and a non-shadow image without a shadow;
A shadow removal unit that performs a shadow removal operation only on the crack image divided into the shadow image by the shadow determination unit;
It includes; a crack recognition unit for recognizing a crack in the shadow removal image in which the shadow removal operation is performed by the shadow removal unit, or for recognizing a crack in the non-shadow image separated by the shadow determination unit;
The histogram analysis unit includes: a threshold calculation unit configured to calculate a histogram of the received crack image and set a point at which variance between classes becomes maximum as a threshold value;
A polynomial estimator for dividing an image into two groups based on the calculated threshold and estimating a quadratic polynomial for variables and frequencies in each group;
A peak calculating unit calculating a peak value including a direction and a magnitude of a peak of each group from the estimated quadratic polynomial; And
It includes; a bimodality calculation unit for calculating a bimodality using the calculated peak value;
In the threshold value calculating unit, the inter-class variance is calculated based on the following equation (1) Crack detection device of a structure.
[Equation 1]

(here

Is the variance between classes, m1 and m2 are averages of each class, and P1 and P2 are the probability of statistically appearing pixel values as weights)
delete delete According to claim 1,
In the threshold calculation unit,
The threshold value for maximizing the variance between classes is calculated based on Equation 2 below.
[Equation 2]

(here

Is a threshold, x is a pixel)
According to claim 4,
In the polynomial estimation unit,
The second-order polynomial of each group divided into two groups based on the threshold is a crack recognition apparatus for a structure, characterized in that it is estimated based on the following equation (3).
[Equation 3]

(here

Wow

Is the second-order polynomial coefficient)
The method of claim 5,
In the peak calculation unit,
The peak value is calculated based on the following equation (4) from the estimated second-order polynomial crack cracking apparatus of the structure.
[Equation 4]

(Where z _k is the peak value of the two groups)
The method of claim 6,
In the bimodality calculation unit,
The bimodality is a crack recognition device for a structure, characterized in that calculated by substituting the peak value in Equation 5 below.
[Equation 5]

The method of claim 7,
The shadow determining unit,
When the bimodality is greater than the threshold, a crack recognition apparatus for a structure, characterized in that the image is classified as the shadow image.
According to claim 1,
The shadow removal unit,
A crack recognition apparatus for a structure, characterized in that the shadow removal operation is performed based on Equation 6 below.
[Equation 6]

(here

Is the shadow image before the shadow removal operation,

Is the median filtering result of the shadow image,

Is a shadow removal image after the shadow removal operation)
According to claim 1,
The crack recognition unit,
A crack is recognized in the shadow removal image or the non-shadow image,
The crack recognition is to extract a crack candidate through binarization based on contrast, and then recognize the crack as a crack when the set condition is satisfied by using the size and circularity of the extracted object. .
A method for recognizing cracks in a structure using a crack recognition device of any one of claims 1, 4 to 10,
(a) receiving, at the crack image receiving unit, a crack image of the structure;
(b) calculating a bimodality by analyzing the histogram of the crack image received from the crack image receiver in the histogram analysis unit;
(c) in the shadow determination unit, determining the presence or absence of a shadow in a crack image based on the bimodality calculated from the histogram analysis unit, and classifying the shadow image into a shadow image and a non-shadow image without a shadow. ;
(d) in the shadow removing unit, performing the shadow removing operation on the crack image divided into the shadow image by the shadow determining unit; And
(e) in the crack recognition unit, recognizing a crack in the shadow removal image in which the shadow removal operation has been performed, or recognizing a crack in the non-shadow image separated in the shadow determination unit;
The step (b) may include: (b-1) calculating, by the threshold calculating unit, a histogram of the received crack image to set a point at which the variance between classes becomes maximum as a threshold;
(b-2) estimating a quadratic polynomial for variables and frequencies in each group by dividing the image into two groups based on the calculated threshold value in the polynomial estimator;
(b-3) calculating a peak value including a direction and a magnitude of a peak of each group from the estimated quadratic polynomial in the peak calculator; And
(b-4) calculating, by the bimodality calculating unit, the bimodality using the calculated peak value;
In step (b-1), in the threshold calculation unit, the inter-class variance is calculated based on Equation 1 below.
[Equation 1]

(here

Is the variance between classes, m1 and m2 are averages of each class, and P1 and P2 are the probability of statistically appearing pixel values as weights)
delete delete The method of claim 11,
In step (b-1),
In the threshold calculation unit, the threshold value for the variance between classes to be the maximum is calculated based on the following equation (2) crack cracking method of the structure.
[Equation 2]

(here

Is a threshold, x is a pixel)
The method of claim 14,
In step (b-2),
In the polynomial estimator, the second-order polynomial of each group divided into two groups based on the threshold is estimated based on Equation 3 below.
[Equation 3]

(here

Wow

Is the second-order polynomial coefficient)
The method of claim 15,
In step (b-3),
In the peak calculating unit, the peak value is calculated based on the following equation (4) from the estimated second-order polynomial crack cracking method of the structure.
[Equation 4]

(Where z _k is the peak value of the two groups)
The method of claim 16,
In step (b-4),
In the bimodality calculator, the bimodality is calculated by substituting the peak value into Equation 5 below.
[Equation 5]

The method of claim 17,
In step (c),
In the shadow determination unit, when the bimodality is greater than the threshold, the corresponding image is classified as the shadow image.
How to recognize cracks in structures.
The method of claim 11,
In step (d),
In the shadow removal unit, the method for crack recognition of a structure characterized in that the shadow removal operation is performed based on the following equation (6).
[Equation 6]

(here

Is the shadow image before the shadow removal operation,

Is the median filtering result of the shadow image,

Is a shadow removal image after the shadow removal operation)
The method of claim 11,
In step (e),
The crack recognition unit recognizes a crack of the shadow removal image or the non-shadow image, and the crack recognition extracts a crack candidate through binarization based on contrast, and then uses the size and circularity of the extracted object. A method for recognizing cracks in a structure, characterized in that it is recognized as a crack when the set condition is satisfied.

Images