CN116958182A - Quick concrete crack detection method based on image data - Google Patents
Quick concrete crack detection method based on image data Download PDFInfo
- Publication number
- CN116958182A CN116958182A CN202311211313.7A CN202311211313A CN116958182A CN 116958182 A CN116958182 A CN 116958182A CN 202311211313 A CN202311211313 A CN 202311211313A CN 116958182 A CN116958182 A CN 116958182A
- Authority
- CN
- China
- Prior art keywords
- line segment
- sub
- level
- edge
- crack
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 23
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 87
- 230000035945 sensitivity Effects 0.000 claims abstract description 52
- 238000000034 method Methods 0.000 claims abstract description 29
- 238000001228 spectrum Methods 0.000 claims abstract description 12
- 238000010606 normalization Methods 0.000 claims description 13
- 238000003708 edge detection Methods 0.000 claims description 11
- 238000010586 diagram Methods 0.000 claims description 4
- 230000007547 defect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005266 casting Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/168—Segmentation; Edge detection involving transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20064—Wavelet transform [DWT]
Abstract
The invention relates to the technical field of image processing, in particular to a concrete crack rapid detection method based on image data. The method comprises the following steps: acquiring edge line segments in a gray level image of a concrete wall; dividing the edge line segments based on the position distribution of the pixel points on the edge line segments to obtain all sub line segments; wavelet decomposition is carried out on the sub-line segments to obtain approximation coefficients, detail coefficients and wavelet spectrograms of the sub-line segments at different levels; determining the optimal decomposition level of each sub-line segment according to the approximation coefficient and the detail coefficient of each sub-line segment in the adjacent level; and obtaining the crack edge confidence degree corresponding to each sub-line segment according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the small spectrum images, the difference of the sensitivity degree and the corresponding optimal decomposition level, and further obtaining the crack region. The invention improves the accuracy of the concrete crack detection result.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a concrete crack rapid detection method based on image data.
Background
The integrity of the concrete outer wall body of the building, which is taken as an important bearing component of the building, plays a key role in the stability and safety of the structure of the building. The crack problem of the wall body can be found and evaluated early by periodically carrying out crack detection on the concrete wall, so that the structural safety of the building is ensured.
Because the wall surface of the concrete outer wall of a building is often rough, the edge detection algorithm is often influenced by rough surface noise in the process of extracting the edge of a crack region, and the extraction of the crack region by detecting excessive edge line segments is interfered, so that the accuracy of a concrete crack detection result is lower.
Disclosure of Invention
In order to solve the problem of lower accuracy in the detection of concrete cracks in the existing method, the invention aims to provide a rapid detection method of concrete cracks based on image data, and the adopted technical scheme is as follows:
the invention provides a rapid concrete crack detection method based on image data, which comprises the following steps:
acquiring a gray level image of a concrete wall of a building to be detected, and carrying out edge detection on the gray level image to obtain each edge line segment;
dividing the edge line segments based on the position distribution of the pixel points on the edge line segments to obtain all sub line segments; wavelet decomposition is carried out on the sub-line segments to obtain approximation coefficients, detail coefficients and wavelet spectrograms of each sub-line segment at different levels; obtaining the sensitivity degree of each sub-line segment at each level according to the approximation coefficient and the detail coefficient of each sub-line segment at the adjacent level, and determining the optimal decomposition level of each sub-line segment based on the sensitivity degree;
obtaining the crack edge confidence level corresponding to each sub-line segment according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the small spectrum images, the difference of the sensitivity degree and the corresponding optimal decomposition level; determining a crack edge line segment based on the crack edge confidence level;
and obtaining a crack region based on all the crack edge line segments.
Preferably, the obtaining the sensitivity degree of each sub-line segment corresponding to each level according to the approximation coefficient and the detail coefficient of each sub-line segment at the adjacent level includes:
for the m-th sub-line segment:
obtaining the decomposition degree of the mth sub-line segment in each level according to the modular length of all approximation coefficients of the mth sub-line segment in each level and the modular length of all detail coefficients;
and determining the normalization result of the absolute value of the difference value between the decomposition degree of the mth sub-line segment at each level and the last level as the sensitivity degree of the mth sub-line segment at each level.
Preferably, the obtaining the decomposition degree of the mth sub-line segment at each level according to the modulo length of all approximation coefficients and the modulo length of all detail coefficients of the mth sub-line segment at each level includes:
for the i-th hierarchy:
the accumulated sum of the modular lengths of all approximation coefficients of the mth sub-line segment at the ith level is recorded as a first characteristic value of the mth sub-line segment at the ith level; the accumulated sum of the modular lengths of all detail coefficients of the mth sub-line segment at the ith level is recorded as a second characteristic value of the mth sub-line segment at the ith level;
and determining a normalization result of the ratio of the first characteristic value to the second characteristic value as the decomposition degree of the mth sub-line segment at the ith level.
Preferably, the determining the optimal decomposition level of each sub-line segment based on the sensitivity degree includes:
for any sub-line segment: and taking the corresponding decomposition level with the maximum sensitivity degree as the optimal decomposition level of the sub-line segment.
Preferably, the obtaining of the sub-line segment nearest to each sub-line segment includes:
for the m-th sub-line segment:
respectively acquiring Euclidean distances between each pixel point on the mth sub-line segment and each pixel point on the kth sub-line segment, and determining the minimum value of the Euclidean distances between the pixel points on the mth sub-line segment and the pixel points on the kth sub-line segment as the target distance between the mth sub-line segment and the kth sub-line segment;
and determining the sub-line segment with the smallest target distance from the mth sub-line segment in all the sub-line segments except the mth sub-line segment as the sub-line segment nearest to the mth sub-line segment.
Preferably, the obtaining the confidence level of the crack edge corresponding to each sub-line segment according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the small spectrum diagram, the difference of the sensitivity degree and the corresponding optimal decomposition level comprises the following steps:
for the m-th sub-line segment:
for any position in the wavelet spectrum corresponding to the mth sub-line segment: recording the absolute value of the difference between the data values of the corresponding positions in the wavelet spectrogram corresponding to the sub-line segment with the nearest position to the m sub-line segment as a characteristic index of the position;
the average value of the characteristic indexes of all positions in the wavelet spectrogram corresponding to the m-th sub-line segment is recorded as a first difference;
recording the absolute value of the difference between the sensitivity degree of the mth sub-line segment corresponding to the corresponding optimal decomposition level and the sensitivity degree of the sub-line segment nearest to the mth sub-line segment corresponding to the corresponding optimal decomposition level as a second difference;
calculating the product of the target distance of the m-th sub-line segment and the sub-line segment closest to the m-th sub-line segment, the first difference and the second difference, and taking the negative correlation normalized value of the product as the local similarity degree of the m-th sub-line segment;
and determining the normalized result of the product of the local similarity degree of the mth sub-line segment and the optimal decomposition level sequence number corresponding to the mth sub-line segment as the crack edge confidence degree corresponding to the mth sub-line segment.
Preferably, the determining a crack edge line segment based on the crack edge confidence level includes:
and respectively judging whether the confidence degree of the crack edge corresponding to each sub-line segment is larger than a confidence degree threshold value, and if so, determining the corresponding sub-line segment as the crack edge line segment.
Preferably, the dividing the edge line segment based on the position distribution of the pixel points on the edge line segment to obtain each sub-line segment includes:
performing main direction decomposition on edge line segments in the gray level image to obtain corresponding main directions;
for the nth edge line segment: taking the central point of the edge line segment as an origin, taking the main direction as the positive direction of the transverse axis of the rectangular coordinate system, and taking the direction vertical to the transverse axis as the positive direction of the vertical axis, constructing the rectangular coordinate system, and obtaining the coordinates of each pixel point on the edge line segment;
respectively judging whether the number of the pixel points on the edge line segment corresponding to the same abscissa is greater than 1, and if so, marking the pixel point corresponding to the corresponding abscissa as a characteristic pixel point; the continuous line segment formed by the adjacent characteristic pixel points is used as a sub-line segment, and the continuous line segment formed by other pixel points except the characteristic pixel points on the edge line segment is also used as a sub-line segment.
Preferably, obtaining the crack region based on all the crack edge line segments includes:
clustering all crack edge line segments based on target distances among the crack edge line segments to obtain clusters;
and carrying out convex hull detection on pixel points on all crack edge line segments in each cluster to obtain a closed curve, and taking a closed region surrounded by the closed curve as a crack region.
Preferably, the performing edge detection on the gray-scale image to obtain each edge line segment includes:
and carrying out edge detection on the gray level image to obtain a corresponding edge image, and obtaining an edge line segment based on edge pixel points in the edge image.
The invention has at least the following beneficial effects:
according to the method, the edge line segments in the gray level image of the concrete wall of the building to be detected are divided, so that the sub-line segments obtained through division can meet the requirement of wavelet decomposition, the wavelet decomposition is carried out on each sub-line segment to obtain approximation coefficients, detail coefficients and wavelet spectrograms of each sub-line segment at different levels, the sensitivity degree of the edge line segments in the local areas of the edge line segments of a crack area and a noise area in the wavelet decomposition process is considered to be different, the wavelet spectrograms obtained after the wavelet decomposition are also different, and therefore the crack edge confidence degree corresponding to each sub-line segment is obtained according to the difference of the relative distance between each sub-line segment and the nearest sub-line segment, the difference of the wavelet spectrograms and the sensitivity degree and the corresponding optimal decomposition level, the sub-line segments are further screened, the crack edge line segments are obtained, the crack area is determined based on the crack edge line segments, the accuracy of the crack area acquisition result is higher, and the accuracy of the concrete crack detection result is improved, and the safety of the concrete building is ensured.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for rapidly detecting a concrete crack based on image data according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of a concrete crack rapid detection method based on image data according to the invention with reference to the accompanying drawings and the preferred embodiment.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a concrete scheme of the concrete crack rapid detection method based on image data.
The embodiment of the method for rapidly detecting the concrete cracks based on the image data comprises the following steps:
the specific scene aimed at by this embodiment is: the integrity of the concrete outer wall plays a key role in the stability and safety of the building structure, however, crack defects may occur in the concrete outer wall during casting or in the later period, so that the crack defects of the concrete outer wall of the building need to be detected regularly. According to the embodiment, firstly, edge line segments in a gray level image of a concrete wall of a building to be detected are obtained, then the edge line segments are divided, so that the obtained sub-line segments can meet the requirement of wavelet decomposition, each sub-line segment is subjected to wavelet decomposition to obtain a corresponding approximation coefficient, a detail coefficient and a wavelet spectrogram, and then the optimal decomposition level of each sub-line segment is determined; and screening crack edge line segments according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the small spectrum images, the difference of the sensitivity degree and the corresponding optimal decomposition level, and determining a crack region according to the crack edge line segments if the crack edge line segments exist.
The embodiment provides a method for rapidly detecting a concrete crack based on image data, as shown in fig. 1, the method for rapidly detecting the concrete crack based on the image data of the embodiment comprises the following steps:
step S1, acquiring a gray image of a concrete wall of a building to be detected, and carrying out edge detection on the gray image to obtain each edge line segment.
In the embodiment, the camera is used for collecting the surface image of the concrete wall of the building to be detected, if the height of the concrete wall of the building to be detected is higher, the unmanned aerial vehicle can be used for carrying the camera, so that the camera is perpendicular to the concrete wall of the building to be detected, namely, the imaging of the concrete wall of the building is prevented from rotating and transforming as much as possible, and the collected surface image of the concrete wall of the building to be detected is an RGB image. And carrying out graying treatment on the acquired surface image of the concrete wall of the building to be detected to obtain a corresponding gray image, carrying out simple denoising treatment on the obtained gray image, and marking the denoised image as the gray image of the concrete wall of the building to be detected. The graying treatment and denoising treatment of the image are all the prior art, and are not repeated here.
The method comprises the steps of carrying out edge detection on a gray image of a concrete wall of a building to be detected by using a Canny edge detection algorithm to obtain a corresponding edge image, wherein the edge image comprises edge pixel points with larger gradients such as a crack region edge and a noise region edge, the edge image can be used for carrying out high-frequency and low-frequency information decomposition on edge line segments formed by the edge pixel points through subsequent wavelet transformation, the edge line segments of the crack region are divided according to structural form information expressed by the edge image, and finally the rapid crack detection of the concrete wall of the building to be detected is completed, so that the edge line segments in the gray image of the concrete wall of the building to be detected are obtained based on the edge pixel points in the edge image, and specifically, a size is constructed according to the obtained edge imageWhen the central pixel point is an edge pixel point and other edge pixel points appear in the 8 neighborhood of the central pixel point, judging that the other edge pixel points and the central pixel point belong to the pixel points on the same edge line segment, and traversing all the edge pixel points to obtain a plurality of edge line segments. The Canny edge detection algorithm is prior art and will not be described in detail here.
Thus, the edge line segments in the gray level image of the concrete wall of the building to be detected are obtained.
Step S2, dividing the edge line segments based on the position distribution of the pixel points on the edge line segments to obtain all sub line segments; wavelet decomposition is carried out on the sub-line segments to obtain approximation coefficients, detail coefficients and wavelet spectrograms of each sub-line segment at different levels; and obtaining the sensitivity degree of each sub-line segment at each level according to the approximation coefficient and the detail coefficient of each sub-line segment at the adjacent level, and determining the optimal decomposition level of each sub-line segment based on the sensitivity degree.
The region attribution type of the divided edge line segments utilizes the difference of the structural morphological characteristics of different edge line segments, and is mainly characterized by the sensitivity degree of wavelet decomposition and the similarity degree of local regions; the sensitivity of wavelet decomposition means that the edge line segments of the crack region and the noise region have different lengths, curvatures and other information, the duty ratio of the high frequency component and the low frequency component is different respectively, and the degree of decomposition is different each time after multiple times of wavelet decomposition; the similarity degree of the local area refers to the fact that adjacent edge line segments of the crack area are adjacent in space positions, and distribution on a small spectrum image of the crack area has similarity; and obtaining the confidence level of the crack edge by combining the sensitivity level of wavelet decomposition and the similarity level of the local area, dividing the edge line segments of the crack area, and finally completing the rapid positioning of the crack area of the concrete wall of the building to be detected.
In this embodiment, the edge line segments in the gray level image of the concrete wall of the building to be detected are already extracted, and since the subsequent wavelet decomposition needs to avoid the situation that the constructed coordinate system has independent variables and dependent variables which are one-to-many, the edge line segments need to be divided into a plurality of single sub-line segments, then wavelet decomposition is performed on all the sub-line segments respectively, each sub-line segment has corresponding approximation coefficient and detail coefficient at each level, and the optimal decomposition level of each sub-line segment is determined according to the approximation coefficient and detail coefficient.
Specifically, since the structural forms of the edge line segments are different, and the wavelet decomposition needs to avoid the situation that independent variables and dependent variables exist in the edge line segments in the constructed coordinate system in one-to-many mode, in this embodiment, the main direction decomposition is performed on the edge line segments in the gray scale image of the concrete wall of the building to be detected by using Principal Component Analysis (PCA), the corresponding main directions are obtained, and the obtained main directions are used for splitting the edge line segments. For the nth edge line segment: taking the central point of the edge line segment as an origin, taking the main direction as the positive direction of the transverse axis of the rectangular coordinate system, and taking the direction vertical to the transverse axis as the positive direction of the vertical axis, constructing the rectangular coordinate system, and obtaining the coordinates of each pixel point on the edge line segment; respectively judging whether the number of the pixel points on the edge line segment corresponding to the same abscissa is greater than 1, and if so, marking the pixel point corresponding to the corresponding abscissa as a characteristic pixel point; the continuous line segment formed by the adjacent characteristic pixel points is used as a sub-line segment, the continuous line segment formed by other pixel points except the characteristic pixel points on the edge line segment is also used as a sub-line segment, namely the nth edge line segment is divided into a plurality of sub-line segments; it should be noted that, if the number of pixel points on the edge line segment corresponding to the same abscissa is less than or equal to 0, the edge line segment is used as a sub-line segment, that is, the edge line segment is not divided.
So far, the embodiment divides all edge line segments to obtain a plurality of sub-line segments, and then the embodiment adopts a Haar wavelet function to respectively carry out wavelet decomposition on each sub-line segment to obtain approximation coefficients, detail coefficients and a wavelet spectrogram of each sub-line segment at different levels; the number of layers of wavelet decomposition is set to 5 in this embodiment, and in a specific application, the implementer can set according to the specific situation. It should be noted that: both the approximation coefficients and the detail coefficients are in the form of vectors.
The approximation coefficients refer to the low frequency portion of the signal, which contains most of the energy of the signal. In wavelet decomposition, approximation coefficients are obtained by a low-pass filter, which filters out the high-frequency part of the signal, and only retains the low-frequency part, and approximation coefficients can be used to represent the overall trend of the signal, which can be used to analyze the stationarity and periodicity of the signal. The detail coefficient refers to a high-frequency part of the signal, which contains detail information of the signal, and in wavelet decomposition, the detail coefficient is obtained through a high-pass filter, which can filter out a low-frequency part of the signal, only the high-frequency part is reserved, and the detail coefficient can be used for representing local characteristics of the signal and can be used for analyzing the change rate and noise level of the signal. Because the edge line segments of the crack region and the noise region have different information such as length, curvature and the like, the ratio of the sub-line segments of the crack region to the sub-line segments of the noise region in high and low frequency components is different, and the ratio of the modulus length of the approximation coefficient to the modulus length of the detail coefficient in different levels is different. According to the embodiment, the decomposition degree of each sub-line segment at each level is obtained according to the modulo length of all approximation coefficients and the modulo length of all detail coefficients of each sub-line segment at each level, and the sensitivity degree of each sub-line segment at each level is further determined.
For the m-th sub-line segment:
for the i-th hierarchy: the accumulated sum of the modular lengths of all approximation coefficients of the mth sub-line segment at the ith level is recorded as a first characteristic value of the mth sub-line segment at the ith level; the accumulated sum of the modular lengths of all detail coefficients of the mth sub-line segment at the ith level is recorded as a second characteristic value of the mth sub-line segment at the ith level; and determining a normalization result of the ratio of the first characteristic value to the second characteristic value as the decomposition degree of the mth sub-line segment at the ith level. In this embodiment, the normalization processing is performed by using a maximum and minimum normalization method, which is the prior art and will not be described in detail here. By adopting the method, the decomposition degree of the mth sub-line segment at the levels except the first level can be obtained.
And determining the normalization result of the absolute value of the difference value between the decomposition degree of the mth sub-line segment at each level and the last level as the sensitivity degree of the mth sub-line segment at each level. Specifically, for the i-th hierarchy: and determining the normalization result of the absolute value of the difference value between the ith hierarchy and the decomposition degree of the mth sub-line segment at the ith-1 hierarchy as the sensitivity degree of the mth sub-line segment corresponding to the ith hierarchy. The normalization processing is still performed by using a maximum and minimum normalization method, which is the prior art and will not be described in detail here. It should be noted that, in this embodiment, the optimal decomposition levels selected in the following are all other levels except the first level, so that the sensitivity level of the sub-line segment in the first level does not need to be calculated, and the sensitivity level of the sub-line segment in each level mentioned in the following is the sensitivity level of the other levels except the first level.
By adopting the method, the corresponding sensitivity degree of each sub-line segment at each level is obtained, and the greater the sensitivity degree is, the better the decomposition effect of the corresponding sub-line segment at the level is indicated, so that for any sub-line segment: and taking the corresponding decomposition level with the maximum sensitivity degree as the optimal decomposition level of the sub-line segment. By adopting the method, the optimal decomposition level of each sub-line segment can be obtained.
To this end, an optimal decomposition level for each sub-line segment is obtained.
Step S3, obtaining the crack edge confidence level corresponding to each sub-line segment according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the small spectrum diagrams, the difference of the sensitivity degree and the corresponding optimal decomposition level; and determining a crack edge line segment based on the crack edge confidence level.
Next, the present embodiment will evaluate the similarity between each sub-line segment and its local sub-line segment according to the relative distance between each sub-line segment and its nearest sub-line segment, the difference of small spectrum diagrams, the difference of sensitivity and the corresponding optimal decomposition level, and the more similar each sub-line segment is to its local sub-line segment, the more likely the corresponding sub-line segment is to be a crack edge line segment.
For the m-th sub-line segment:
the Euclidean distance between each pixel point on the mth sub-line segment and each pixel point on the kth sub-line segment is respectively obtained, and it is to be noted that: for any pixel point on the mth sub-line segment, a Euclidean distance is reserved between the pixel point and each pixel point on the kth sub-line segment; and determining the minimum value of the Euclidean distances between the pixel points on the m-th sub-line segment and the pixel points on the k-th sub-line segment as the target distance between the m-th sub-line segment and the k-th sub-line segment. By adopting the method, the target distance between the mth sub-line segment and each other sub-line segment can be obtained. And determining the sub-line segment with the smallest target distance from the mth sub-line segment in all the sub-line segments except the mth sub-line segment as the sub-line segment nearest to the mth sub-line segment.
For any position in the wavelet spectrum corresponding to the mth sub-line segment: and recording the absolute value of the difference between the data values of the corresponding positions in the wavelet spectrogram corresponding to the sub-line segment with the nearest position to the m sub-line segment as the characteristic index of the position. By adopting the method, the characteristic index of each position in the wavelet spectrogram corresponding to the mth sub-line segment can be obtained, and each position in the wavelet spectrogram corresponding to the mth sub-line segment corresponds to one characteristic index.
The average value of the characteristic indexes of all positions in the wavelet spectrogram corresponding to the m-th sub-line segment is recorded as a first difference; and recording the absolute value of the difference between the sensitivity degree of the mth sub-line segment corresponding to the corresponding optimal decomposition level and the sensitivity degree of the sub-line segment nearest to the mth sub-line segment corresponding to the corresponding optimal decomposition level as a second difference. And calculating the product of the target distance between the m-th sub-line segment and the sub-line segment closest to the m-th sub-line segment, the first difference and the second difference, and taking the negative correlation normalized value of the product as the local similarity degree of the m-th sub-line segment. The specific calculation formula of the local similarity degree of the m-th sub-line segment is as follows:
；
wherein, the liquid crystal display device comprises a liquid crystal display device,for the local similarity of the mth sub-line segment, < >>For the mth sub-line segment and the target distance of the sub-line segment nearest thereto, +.>For the mean value of the characteristic indexes of all positions in the wavelet spectrogram corresponding to the mth sub-line segment, ++>For the absolute value of the difference between the sensitivity level of the mth sub-line segment corresponding to the corresponding optimal decomposition level and the sensitivity level of the sub-line segment nearest to the mth sub-line segment corresponding to the corresponding optimal decomposition level>Is an exponential function based on natural constants.
The first difference is represented, and reflects the difference between the wavelet spectrogram corresponding to the m-th sub-line segment and the wavelet spectrogram corresponding to the sub-line segment nearest to the m-th sub-line segment, and the larger the first difference is, the larger the difference between the wavelet spectrogram corresponding to the m-th sub-line segment and the wavelet spectrogram corresponding to the sub-line segment nearest to the m-th sub-line segment is. />And representing a second difference, and reflecting the difference condition between the sensitivity degree of the m-th sub-line segment corresponding to the corresponding optimal decomposition level and the sensitivity degree of the sub-line segment closest to the m-th sub-line segment corresponding to the corresponding optimal decomposition level, wherein the larger the second difference is, the larger the sensitivity degree difference between the m-th sub-line segment corresponding to the corresponding optimal decomposition level and the sensitivity degree of the sub-line segment closest to the m-th sub-line segment corresponding to the corresponding optimal decomposition level is. When the closer the target distance between the m-th sub-line segment and the sub-line segment closest thereto is, the smaller the mean value of the characteristic indexes of all positions in the wavelet spectrogram corresponding to the m-th sub-line segment is, and the smaller the absolute value of the difference value between the sensitivity degree of the m-th sub-line segment corresponding to the optimal decomposition level and the sensitivity degree of the m-th sub-line segment closest thereto is, the more similar the m-th sub-line segment and the closest sub-line segment thereof are, namely the greater the local similarity degree of the m-th sub-line segment is.
And determining the normalized result of the product of the local similarity degree of the mth sub-line segment and the optimal decomposition level sequence number corresponding to the mth sub-line segment as the crack edge confidence degree corresponding to the mth sub-line segment. In this embodiment, the normalization processing is performed by using a maximum and minimum normalization method, which is the prior art and will not be described in detail here.
By adopting the method, the crack edge confidence degree corresponding to each sub-line segment can be obtained, and the greater the crack edge confidence degree is, the more likely the corresponding sub-line segment is to be the edge of the crack region is, so that the embodiment respectively judges whether the crack edge confidence degree corresponding to each sub-line segment is greater than a confidence degree threshold value, and if so, the corresponding sub-line segment is determined to be the crack edge line segment. In this embodiment, the confidence level threshold is 0.7, and in a specific application, the practitioner may set according to a specific situation.
So far, by adopting the method provided by the embodiment, the crack edge line segments are screened out from all the sub-line segments, and the crack area is determined based on the crack edge line segments. It should be noted that, if the confidence degrees of the crack edges corresponding to all the sub-line segments are smaller than or equal to the confidence degree threshold, then no crack edge line segment exists in the sub-line segments, and at this time, it is indicated that no crack exists on the concrete wall of the building to be detected, so that it is not necessary to obtain a crack area on the concrete wall of the building to be detected.
And S4, obtaining a crack region based on all the crack edge line segments.
In this embodiment, in step S3, a crack edge line segment in the gray scale image of the concrete wall of the building to be detected is selected, and then a crack region is determined based on all the crack edge line segments.
Specifically, based on the target distances among the crack edge line segments, clustering is performed on all the crack edge line segments by using a mean shift clustering algorithm, namely, the target distances among the crack edge line segments serve as indexes for measuring the similarity of the crack edge line segments, a plurality of clusters are obtained, and the mean shift clustering algorithm is the prior art and is not repeated here.
And (3) taking all crack edge line segments in each cluster as a whole to perform convex hull detection, namely performing convex hull detection on pixel points on all crack edge line segments in each cluster to obtain a closed curve, and taking a closed region enclosed by the closed curve as a crack region, wherein each cluster corresponds to one crack region. Convex hull detection is prior art and will not be described in detail here.
By adopting the method provided by the embodiment, the crack defect of the concrete wall of the building to be detected can be rapidly detected, the information such as the area and the position of the crack region in the gray level image of the concrete wall of the building to be detected can be obtained, and the worker can repair the crack in time according to the detection result.
According to the method, the edge line segments in the gray level image of the concrete wall of the building to be detected are firstly divided, so that the sub-line segments obtained through division can meet the requirement of wavelet decomposition, each sub-line segment is subjected to wavelet decomposition to obtain approximation coefficients, detail coefficients and wavelet spectrograms of each sub-line segment at different levels, the fact that the sensitivity degree of the edge line segments in the local area of the edge line segments of a crack area and a noise area in the wavelet decomposition process is different is considered, the wavelet spectrograms obtained after the wavelet decomposition are also different is considered, and therefore the accuracy and the reliability of the concrete crack detection result are improved according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the wavelet spectrograms, the sensitivity degree and the corresponding optimal decomposition level, the crack edge confidence degree corresponding to each sub-line segment are obtained, the sub-line segments are further screened, the crack edge line segments are obtained, the crack area is determined based on the crack edge line segments, the accuracy of the crack area acquisition result is higher, and the concrete crack detection result is improved.
It should be noted that: the foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. The rapid concrete crack detection method based on the image data is characterized by comprising the following steps of:
acquiring a gray level image of a concrete wall of a building to be detected, and carrying out edge detection on the gray level image to obtain each edge line segment;
dividing the edge line segments based on the position distribution of the pixel points on the edge line segments to obtain all sub line segments; wavelet decomposition is carried out on the sub-line segments to obtain approximation coefficients, detail coefficients and wavelet spectrograms of each sub-line segment at different levels; obtaining the sensitivity degree of each sub-line segment at each level according to the approximation coefficient and the detail coefficient of each sub-line segment at the adjacent level, and determining the optimal decomposition level of each sub-line segment based on the sensitivity degree;
obtaining the crack edge confidence level corresponding to each sub-line segment according to the relative distance between each sub-line segment and the sub-line segment closest to the sub-line segment, the difference of the small spectrum images, the difference of the sensitivity degree and the corresponding optimal decomposition level; determining a crack edge line segment based on the crack edge confidence level;
and obtaining a crack region based on all the crack edge line segments.
2. The method for quickly detecting the concrete cracks based on the image data according to claim 1, wherein the step of obtaining the sensitivity degree of each sub-line segment corresponding to each level according to the approximation coefficient and the detail coefficient of each sub-line segment at the adjacent level comprises the following steps:
for the m-th sub-line segment:
obtaining the decomposition degree of the mth sub-line segment in each level according to the modular length of all approximation coefficients of the mth sub-line segment in each level and the modular length of all detail coefficients;
and determining the normalization result of the absolute value of the difference value between the decomposition degree of the mth sub-line segment at each level and the last level as the sensitivity degree of the mth sub-line segment at each level.
3. The method for rapidly detecting concrete cracks based on image data according to claim 2, wherein obtaining the decomposition degree of the mth sub-line segment at each level according to the modular length of all approximation coefficients and the modular length of all detail coefficients of the mth sub-line segment at each level comprises:
for the i-th hierarchy:
the accumulated sum of the modular lengths of all approximation coefficients of the mth sub-line segment at the ith level is recorded as a first characteristic value of the mth sub-line segment at the ith level; the accumulated sum of the modular lengths of all detail coefficients of the mth sub-line segment at the ith level is recorded as a second characteristic value of the mth sub-line segment at the ith level;
and determining a normalization result of the ratio of the first characteristic value to the second characteristic value as the decomposition degree of the mth sub-line segment at the ith level.
4. The method for quickly detecting concrete cracks based on image data according to claim 1, wherein the determining an optimal decomposition level of each sub-line segment based on the sensitivity degree comprises:
for any sub-line segment: and taking the corresponding decomposition level with the maximum sensitivity degree as the optimal decomposition level of the sub-line segment.
5. The method for rapidly detecting concrete cracks based on image data according to claim 1, wherein the acquisition of the sub-line segment nearest to each sub-line segment comprises:
for the m-th sub-line segment:
respectively acquiring Euclidean distances between each pixel point on the mth sub-line segment and each pixel point on the kth sub-line segment, and determining the minimum value of the Euclidean distances between the pixel points on the mth sub-line segment and the pixel points on the kth sub-line segment as the target distance between the mth sub-line segment and the kth sub-line segment;
and determining the sub-line segment with the smallest target distance from the mth sub-line segment in all the sub-line segments except the mth sub-line segment as the sub-line segment nearest to the mth sub-line segment.
6. The method for rapidly detecting concrete cracks based on image data according to claim 5, wherein obtaining the confidence level of the crack edges corresponding to each sub-line segment according to the relative distance between each sub-line segment and the sub-line segment nearest to each sub-line segment, the difference of the small spectrum diagrams, the difference of the sensitivity degree and the corresponding optimal decomposition level, comprises:
for the m-th sub-line segment:
for any position in the wavelet spectrum corresponding to the mth sub-line segment: recording the absolute value of the difference between the data values of the corresponding positions in the wavelet spectrogram corresponding to the sub-line segment with the nearest position to the m sub-line segment as a characteristic index of the position;
the average value of the characteristic indexes of all positions in the wavelet spectrogram corresponding to the m-th sub-line segment is recorded as a first difference;
recording the absolute value of the difference between the sensitivity degree of the mth sub-line segment corresponding to the corresponding optimal decomposition level and the sensitivity degree of the sub-line segment nearest to the mth sub-line segment corresponding to the corresponding optimal decomposition level as a second difference;
calculating the product of the target distance of the m-th sub-line segment and the sub-line segment closest to the m-th sub-line segment, the first difference and the second difference, and taking the negative correlation normalized value of the product as the local similarity degree of the m-th sub-line segment;
and determining the normalized result of the product of the local similarity degree of the mth sub-line segment and the optimal decomposition level sequence number corresponding to the mth sub-line segment as the crack edge confidence degree corresponding to the mth sub-line segment.
7. The method for quickly detecting a concrete crack based on image data according to claim 1, wherein the determining a crack edge line segment based on the crack edge confidence level comprises:
and respectively judging whether the confidence degree of the crack edge corresponding to each sub-line segment is larger than a confidence degree threshold value, and if so, determining the corresponding sub-line segment as the crack edge line segment.
8. The method for quickly detecting the concrete cracks based on the image data according to claim 1, wherein the dividing the edge line segments based on the position distribution of the pixel points on the edge line segments to obtain the sub line segments comprises the following steps:
performing main direction decomposition on edge line segments in the gray level image to obtain corresponding main directions;
for the nth edge line segment: taking the central point of the edge line segment as an origin, taking the main direction as the positive direction of the transverse axis of the rectangular coordinate system, and taking the direction vertical to the transverse axis as the positive direction of the vertical axis, constructing the rectangular coordinate system, and obtaining the coordinates of each pixel point on the edge line segment;
respectively judging whether the number of the pixel points on the edge line segment corresponding to the same abscissa is greater than 1, and if so, marking the pixel point corresponding to the corresponding abscissa as a characteristic pixel point; the continuous line segment formed by the adjacent characteristic pixel points is used as a sub-line segment, and the continuous line segment formed by other pixel points except the characteristic pixel points on the edge line segment is also used as a sub-line segment.
9. The method for rapidly detecting concrete cracks based on image data according to claim 5, wherein obtaining crack areas based on all the crack edge segments comprises:
clustering all crack edge line segments based on target distances among the crack edge line segments to obtain clusters;
and carrying out convex hull detection on pixel points on all crack edge line segments in each cluster to obtain a closed curve, and taking a closed region surrounded by the closed curve as a crack region.
10. The method for rapidly detecting concrete cracks based on image data according to claim 1, wherein the step of performing edge detection on the gray scale image to obtain each edge line segment comprises the steps of:
and carrying out edge detection on the gray level image to obtain a corresponding edge image, and obtaining an edge line segment based on edge pixel points in the edge image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311211313.7A CN116958182B (en) | 2023-09-20 | 2023-09-20 | Quick concrete crack detection method based on image data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311211313.7A CN116958182B (en) | 2023-09-20 | 2023-09-20 | Quick concrete crack detection method based on image data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116958182A true CN116958182A (en) | 2023-10-27 |
| CN116958182B CN116958182B (en) | 2023-12-08 |
Family
ID=88446495
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311211313.7A Active CN116958182B (en) | 2023-09-20 | 2023-09-20 | Quick concrete crack detection method based on image data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116958182B (en) |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7295695B1 (en) * | 2002-03-19 | 2007-11-13 | Kla-Tencor Technologies Corporation | Defect detection via multiscale wavelets-based algorithms |
| CN102331425A (en) * | 2011-06-28 | 2012-01-25 | 合肥工业大学 | Textile defect detection method based on defect enhancement |
| KR20160054151A (en) * | 2014-11-05 | 2016-05-16 | 한밭대학교 산학협력단 | System and Method for Automatically Detecting a Mura Defect using Mask Filtering in Wavelet Transform |
| CN109523479A (en) * | 2018-11-10 | 2019-03-26 | 东莞理工学院 | A kind of bridge pier surface gaps visible detection method |
| CN110598166A (en) * | 2019-09-18 | 2019-12-20 | 河海大学 | Wavelet denoising method for adaptively determining wavelet hierarchical level |
| CN110956183A (en) * | 2019-11-04 | 2020-04-03 | 东南大学 | Asphalt pavement crack morphology extraction method |
| CN111311508A (en) * | 2020-01-21 | 2020-06-19 | 东南大学 | Noise reduction method for pavement crack image with noise |
| CN112862760A (en) * | 2021-01-19 | 2021-05-28 | 浙江大学 | Bearing outer ring surface defect area detection method |
| CN113808149A (en) * | 2021-07-26 | 2021-12-17 | 南京理工大学智能计算成像研究院有限公司 | Automobile scratch detection method based on self-adaptive threshold |
| CN114820438A (en) * | 2022-03-15 | 2022-07-29 | 扬州新扬通风设备有限公司 | Method and system for detecting surface material defects of silencer |
| US20220301136A1 (en) * | 2019-08-23 | 2022-09-22 | Hitachi High-Tech Corporation | Defect Inspection Method and Defect Inspection Device |
| CN115908411A (en) * | 2023-01-06 | 2023-04-04 | 山东水利建设集团有限公司 | Concrete curing quality analysis method based on visual detection |
| CN116385434A (en) * | 2023-06-02 | 2023-07-04 | 同济检测(济宁)有限公司 | Intelligent detection method for precast beam cracks |
| CN116703910A (en) * | 2023-08-07 | 2023-09-05 | 威海丰荟建筑工业科技有限公司 | Intelligent detection method for quality of concrete prefabricated bottom plate |
-
2023
- 2023-09-20 CN CN202311211313.7A patent/CN116958182B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7295695B1 (en) * | 2002-03-19 | 2007-11-13 | Kla-Tencor Technologies Corporation | Defect detection via multiscale wavelets-based algorithms |
| CN102331425A (en) * | 2011-06-28 | 2012-01-25 | 合肥工业大学 | Textile defect detection method based on defect enhancement |
| KR20160054151A (en) * | 2014-11-05 | 2016-05-16 | 한밭대학교 산학협력단 | System and Method for Automatically Detecting a Mura Defect using Mask Filtering in Wavelet Transform |
| CN109523479A (en) * | 2018-11-10 | 2019-03-26 | 东莞理工学院 | A kind of bridge pier surface gaps visible detection method |
| US20220301136A1 (en) * | 2019-08-23 | 2022-09-22 | Hitachi High-Tech Corporation | Defect Inspection Method and Defect Inspection Device |
| CN110598166A (en) * | 2019-09-18 | 2019-12-20 | 河海大学 | Wavelet denoising method for adaptively determining wavelet hierarchical level |
| CN110956183A (en) * | 2019-11-04 | 2020-04-03 | 东南大学 | Asphalt pavement crack morphology extraction method |
| CN111311508A (en) * | 2020-01-21 | 2020-06-19 | 东南大学 | Noise reduction method for pavement crack image with noise |
| CN112862760A (en) * | 2021-01-19 | 2021-05-28 | 浙江大学 | Bearing outer ring surface defect area detection method |
| CN113808149A (en) * | 2021-07-26 | 2021-12-17 | 南京理工大学智能计算成像研究院有限公司 | Automobile scratch detection method based on self-adaptive threshold |
| CN114820438A (en) * | 2022-03-15 | 2022-07-29 | 扬州新扬通风设备有限公司 | Method and system for detecting surface material defects of silencer |
| CN115908411A (en) * | 2023-01-06 | 2023-04-04 | 山东水利建设集团有限公司 | Concrete curing quality analysis method based on visual detection |
| CN116385434A (en) * | 2023-06-02 | 2023-07-04 | 同济检测(济宁)有限公司 | Intelligent detection method for precast beam cracks |
| CN116703910A (en) * | 2023-08-07 | 2023-09-05 | 威海丰荟建筑工业科技有限公司 | Intelligent detection method for quality of concrete prefabricated bottom plate |
Non-Patent Citations (3)
| Title |
|---|
| RAVI NIGAM ET AL.: "Crack detection in a beam using wavelet transform and photographic measurements", 《STRUCTURES》, vol. 25, pages 436 - 447 * |
| 姚建军;: "基于多数据融合和小波分解的油气管道缺陷检测方法研究", 石油化工自动化, no. 02, pages 28 - 52 * |
| 苏泽斌;高敏;李鹏飞;景军锋;张缓缓;: "融合小波分解的LBPV数码印花缺陷检测方法", 传感器与微系统, no. 09, pages 126 - 128 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116958182B (en) | 2023-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114862862B (en) | Pump body cold shut defect identification method and system based on image processing | |
| CN115170576B (en) | Aluminum pipe surface defect detection method based on machine vision | |
| EP1233374B1 (en) | Apparatus and method for extracting objects based on feature matching between segmented regions in images | |
| CN109559324B (en) | Target contour detection method in linear array image | |
| CN114140462A (en) | Bearing wear degree evaluation method based on image processing | |
| CN115861291B (en) | Chip circuit board production defect detection method based on machine vision | |
| CN111738342B (en) | Pantograph foreign matter detection method, storage medium and computer equipment | |
| CN115311275B (en) | Method and system for detecting surface defects of electronic component | |
| CN113191979B (en) | Non-local mean denoising method for partitioned SAR (synthetic aperture radar) image | |
| CN114782329A (en) | Bearing defect damage degree evaluation method and system based on image processing | |
| CN111681174A (en) | Joint filtering method for inhibiting salt and pepper/Gaussian noise mixing target image | |
| CN116883408B (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
| CN115511907B (en) | Scratch detection method for LED screen | |
| CN116152115B (en) | Garbage image denoising processing method based on computer vision | |
| CN116012357A (en) | Automobile circular trailer hole air line defect detection method based on machine vision | |
| CN115639248A (en) | System and method for detecting quality of building outer wall | |
| CN117115117B (en) | Pathological image recognition method based on small sample, electronic equipment and storage medium | |
| CN115797374B (en) | Airport runway extraction method based on image processing | |
| CN116958182B (en) | Quick concrete crack detection method based on image data | |
| CN117291843A (en) | Efficient management method for image database | |
| CN115564705B (en) | Burr detection precision enhancement method and system for precision standard component | |
| CN115512310A (en) | Vehicle type recognition method and system based on face features under video monitoring | |
| CN111473944B (en) | PIV data correction method and device for observing complex wall surface in flow field | |
| CN114140742A (en) | Track foreign matter intrusion detection method based on light field depth image | |
| CN109886133B (en) | Ship detection method and system based on optical remote sensing image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |