CN116012767A - Visual detection method for cracks of clutch housing of electrically-controlled silicone oil fan - Google Patents
Visual detection method for cracks of clutch housing of electrically-controlled silicone oil fan Download PDFInfo
- Publication number
- CN116012767A CN116012767A CN202310301726.8A CN202310301726A CN116012767A CN 116012767 A CN116012767 A CN 116012767A CN 202310301726 A CN202310301726 A CN 202310301726A CN 116012767 A CN116012767 A CN 116012767A
- Authority
- CN
- China
- Prior art keywords
- line segment
- target
- line segments
- unit
- unit line
- 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
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to a visual detection method for cracks of a clutch shell of an electrically-controlled silicone oil fan, which comprises the following steps: the method comprises the steps of obtaining gray images of a shell of a silicone oil fan clutch, obtaining a plurality of support domains in a line field, obtaining an extension straight line and the number of unit line segments contained in the extension straight line, obtaining similar line segments, obtaining the blurred degree of the similar line segments and the unit line segments with the same numbers, obtaining the self-adaptive rectangular range of the support domains, and obtaining a crack region. According to the invention, the self-adaptive rectangular range of the supporting domain is obtained in a self-adaptive manner, and then the accurate crack region is obtained.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a visual detection method for cracks of a clutch shell of an electrically-controlled silicone oil fan.
Background
In the production process of the electrically controlled silicone oil fan clutch shell, the produced electrically controlled silicone oil fan clutch shell may have cracks, so that the cracks need to be detected, defective products are screened out, and the product quality is ensured.
The common detection method is an LSD straight line detection algorithm, the extraction of a crack area of the clutch shell of the electric control silicone oil fan can be realized through the LSD straight line detection algorithm, and because a plurality of radiating fins are arranged on the clutch shell of the electric control silicone oil fan, when cracks on the shell appear among the radiating fins, the gray features of the cracks are blocked by the radiating fins to cause blurring, and at the moment, when the LSD straight line detection algorithm is used for detection, the detection accuracy of the LSD straight line detection algorithm is seriously affected, namely the rectangular range of the LSD straight line detection algorithm is not accurate enough, and the acquired crack area is inaccurate.
Disclosure of Invention
The invention provides a visual detection method for cracks of a clutch shell of an electrically controlled silicone oil fan, which aims to solve the problem that the obtained crack area is inaccurate due to the fact that the rectangular range of the conventional LSD straight line detection algorithm is inaccurate.
The invention discloses a visual detection method for cracks of a clutch shell of an electrically-controlled silicone oil fan, which adopts the following technical scheme.
A grayscale image of the housing of the silicone oil fan clutch is acquired.
And obtaining a plurality of support domains in the line field by using a unit line segment perpendicular to the gradient direction of each pixel point in the gray level image.
According to the gradient amplitude value of each unit line segment in each supporting domain, the gray value of the neighborhood pixel point in the neighborhood of the pixel point corresponding to the unit line segment and the gray average value of the normal shell, the possibility that each unit line segment is a crack is obtained, the unit line segment corresponding to the maximum possibility is taken as a target unit line segment, and the number of the unit line segments contained in the extended straight line after the target unit line segment in the supporting domain is obtained.
According to the probability difference value between the target unit line segment and the neighborhood unit line segment in the neighborhood of the target unit line segment, acquiring a first target line segment in the neighborhood unit line segment, taking the first target line segment as a new target unit line segment, acquiring a second target line segment in the neighborhood of the first target line segment, stopping until the number of the acquired target line segments is equal to the number of the unit line segments contained in the extension straight line, and taking all the acquired target line segments as similar line segments.
And numbering all the obtained similar line segments and unit line segments contained in the extension straight line respectively, wherein the numbering of the similar line segments is carried out according to the acquired sequence, the unit line segments contained in the extension straight line are sequentially numbered according to the extension directions of all the similar line segments, and the blurred degree of the similar line segments with the same number and the unit line segments is acquired according to the sum of the distances between the similar line segments with the same number and the corresponding pixel points of the unit line segments and the difference of the possibility between the similar line segments with the same number and the unit line segments.
Acquiring all final similar line segments by using the blurred degree of each similar line segment and unit line segment with the same number, and acquiring a target area according to all final similar line segments and extension lines; and taking the minimum circumscribed rectangle of the target area as an adaptive rectangle range of the corresponding support domain, and obtaining a crack area by utilizing an LSD linear algorithm based on the adaptive rectangle range of the support domain.
Preferably, the likelihood of each unit line segment being a crack is obtained, including.
And acquiring the gradient amplitude value of each unit line segment in the support domain, and the gradient amplitude value difference value of the maximum gradient amplitude value in the support domain.
And performing quadratic function mapping on the gradient amplitude difference value to obtain a first target value.
And acquiring a gray level difference value between a gray level value of a neighborhood pixel point in the neighborhood of the pixel point corresponding to the unit line segment in the support domain and a gray level average value of the normal shell.
And carrying out negative correlation calculation on the average value of the gray difference absolute values of the gray average values of all neighborhood pixel points in the neighborhood of the pixel point corresponding to the unit line segment and the normal shell to obtain a second target value.
And carrying out normalization calculation on the ratio of the first target value to the second target value to obtain the possibility that each unit line segment is a crack.
Preferably, a first target line segment in the neighborhood unit line segment is obtained, including.
And obtaining the minimum possibility difference value in the possibility difference values of the target unit line segment and each neighborhood unit line segment in the neighborhood of the target unit line segment.
And taking the neighborhood unit line segment corresponding to the minimum probability difference value as a first target line segment.
Preferably, the likelihood difference between the similar line segments with the same number and the unit line segments is obtained, including.
And obtaining the absolute value of the probability difference between the similar line segments with the same numbers and the unit line segments.
The absolute value of the probability difference is taken as the probability difference between the similar line segments with the same number and the unit line segments.
Preferably, the fuzzy degree of the similar line segments and the unit line segments with the same number is obtained, including.
And taking the natural constant as a base, and taking the probability difference between the similar line segments with the same number and the unit line segments as a power to obtain an exponential function value.
And taking the product of the exponential function value, the distances between the similar line segments with the same numbers and the pixel points corresponding to the unit line segments as the initial blurred degree of the similar line segments with the same numbers and the unit line segments.
An initial blurriness difference value of each initial blurriness and two initial blurriness adjacent in number is obtained.
And obtaining the blurred degrees of the similar line segments with the same numbers and the unit line segments according to the average value of the initial blurred degrees of the similar line segments with the same numbers and the unit line segments.
Preferably, all final similar line segments are obtained, including.
And acquiring a target number corresponding to the unit line segment when the blurred degree is larger than a preset blurred degree threshold value.
And obtaining the target distance between the pixel points of the similar line segments corresponding to the numbers adjacent to the target numbers and the pixel points of the unit line segments.
And taking the average value of the two target distances as the distance between the target line segment on the extension line corresponding to the target number and the final similar line segment.
And obtaining the final similar line segment corresponding to the target number of the target line segment on the extension straight line according to the distance between the target line segment on the extension straight line corresponding to the target number and the final similar line segment.
Preferably, all final similar line segments are connected to obtain a final similar line.
Preferably, an area defined by connecting the final similarity line with the adjacent end of the extension straight line is taken as the target area.
Preferably, a grayscale image of the housing of the silicone oil fan clutch is acquired, including.
And acquiring a shell image of the silicone oil fan clutch.
The R, G, B component of the shell image in the RGB color space is projection weighted to obtain a gray scale image of the shell.
The visual detection method for the cracks of the clutch shell of the electrically-controlled silicone oil fan has the beneficial effects that.
When the crack appears on the clutch shell due to the interference of the fin on the clutch shell, the edge of the crack is blurred, and the crack cannot be accurately acquired, so that a line field and a supporting domain in an image are firstly acquired by utilizing an LSD algorithm, then a target unit line segment with the maximum possibility of the crack in the supporting domain is acquired, then an extension straight line is acquired in the extension direction of the target unit line segment, namely, as the metal fracture trace is in a broken line shape, but the overall appearance is of a characteristic of extending in a certain direction and is not similar to the shape of a straight line with a relatively flat fin, the extension straight line is acquired in the extension direction of the target unit line segment, because in one supporting domain, the unit line segment existing near the crack can blur the unit line segment belonging to the crack, the degree of blurring is judged by combining with the extension characteristics of the crack form, so that a rectangular range is acquired, namely, firstly, a first target line segment in the neighborhood unit line segment is acquired according to the difference of the possibility of the target unit line segment and the neighborhood unit line segment in the target unit line segment, a second target line segment in the neighborhood of the first target line segment is acquired, until the number of the acquired target line segment is equal to the extension straight line segment in the extension straight line segment, and the extension straight line segment in the extension straight line segment is equal to the extension number of the target line segment in the extension straight line segment, and the extension line segment is equal to the extension line segment in the extension number of the target line segment, and the extension line segment is similar to the extension line segment in the extension line segment, and the extension line segment is similar to the extension line in the unit-like, and the size of the unit line is determined by the extension of the unit-like, and the extension line is similar to the extension line and the size, and the extension line is similar to the size, and the size and the extension line is determined by the extension and the size and the extension line and the size and the extension line and the extension line and the size and the extension line and the, comprehensively acquiring the blurred degree to ensure that the adaptive rectangular range acquired based on the blurred degree is constrained near the real crack range, namely, the crack region acquired based on the adaptive rectangular range is more accurate.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of an embodiment of a visual inspection method for cracks of a clutch housing of an electrically controlled silicone oil fan according to the present invention.
Fig. 2 is a schematic structural diagram of a line field in an embodiment of a visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to the present invention.
Fig. 3 is a schematic view of an extended straight line in a support domain in an embodiment of a visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to the present invention.
Fig. 4 is a schematic diagram of a similar line segment in a support domain in an embodiment of a visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to the present invention.
Fig. 5 is a schematic diagram of an adaptive rectangular range in an embodiment of a visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to the present invention.
Description of the embodiments
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
An embodiment of a visual detection method for cracks of a clutch housing of an electrically controlled silicone oil fan is shown in fig. 1, and comprises the following steps.
S1, acquiring a gray image of a shell of the silicone oil fan clutch.
Specifically, a shell image of a silicone oil fan clutch is collected, and because the components are made of different materials and have different process parameters, the shell color difference exists in different batches of products, so that the color characteristic is lost because the self color characteristic of the products is ignored in the traditional weighted graying is avoided, and the R, G, B component of the shell image in an RGB color space is subjected to projection weighting to obtain a gray image of the shell.
The projection weighting formula is as follows:
in the method, in the process of the invention,
representing +.>
Channel components.
It should be noted that the projection weighting belongs to the prior art, and is used herein to describe that the projection weighting is more suitable for the embodiment scene than the fixed weighting.
S2, acquiring a plurality of support domains in the line field.
Because the position of the radiating fins on the silicone oil fan clutch is higher than the surface of the shell, and the cracks are zigzag lines with irregular shapes, the crack area on the surface of the shell can be influenced by the radiating fins in the process of collecting images, so that the rectangular range of the cracks is different from the rectangular range of the fins, the rectangular range is expanded and reduced monotonously, noise edges are introduced to interfere with crack edge judgment, and the characteristic information is lost, so that the crack pixels are required to be screened by linking the gray level difference of the cracks and the fins, and the LSD rectangular range of the cracks is selected in a self-adaptation mode by combining the shape difference of the cracks and the fins.
Specifically, since the LSD straight line detection algorithm depends on the sub-pixel characteristics generated in the vector direction to select, the line segment support domain is divided under a certain tolerance angle by the direction of the gray edge, and the rectangular approximate range is divided for the line segment support domain, so that the plurality of support domains in the line field are obtained by using the unit line segment perpendicular to the gradient direction of each pixel point in the gray image, that is, the embodiment obtains the gradient direction of each pixel point in the gray image according to the step of the LSD straight line detection algorithm, obtains the unit line segment perpendicular to the gradient direction of each pixel point, and obtains the line field and a plurality of support domains in the line field according to all the unit line segments.
The gradient direction of each pixel point in the gray image is obtained through the existing sobel gradient algorithm, a unit line segment perpendicular to the gradient direction of each pixel point is made, then the gray image is converted into a line field, the line field is shown in fig. 2, finally the unit line segment in the line field is screened according to the preset tolerance degree, a plurality of support domains are obtained, the support domains and the line field are steps in the LSD algorithm, and the embodiment is not changed, so that the repeated description is omitted.
S3, obtaining the extension straight line and the number of unit line segments contained in the extension straight line.
Because the supporting domain contains more unit line segments, the unit line segments may belong to cracks and also belong to fins, and the positions of the fins are higher than the shell, so that the gray scale characteristics of the cracks in the image are affected by blurring, the probability that the unit line segments in each supporting domain belong to cracks needs to be obtained by combining the gray scale characteristic differences of the cracks and the fins.
For the range of the support domain in the gray image, the crack position is represented as the position of the broken pixel gray surface of the shell, the top edge position of the fin is higher, the gray edge is sharper, and when the crack is blocked by the fin, the gray edge generated has a certain blur, so that a specific calculation formula of the possibility that the unit line segment is the crack is as follows:
in the method, in the process of the invention,
representing support Domain->
The%>
The individual line segments are the likelihood of cracks.
At the position of
Representing support Domain->
The maximum gradient amplitude value of the gradient amplitude values corresponding to all the pixel points in the array.
e represents a natural constant.
sigmoid represents a normalization function in order to give a value of 0 to 1 for the likelihood of a crack for a unit line segment.
It should be noted that the number of the substrates,
the first target value is represented, the purpose of the target value is to exclude the positions of the fins with overlarge and undersize amplitude and the normal shell through quadratic function mapping, the gradient visibility of crack parts with a certain amplitude but shielded by the fins is reduced, and the gradient visibility is distinguished by using a larger function value, wherein the positions of the fins are provided with a certain amplitude>
Representing a second target value, using an exponential function of the natural constant e, for +.>
Negative correlation mapping is performed, < > on->
The larger the gradation difference between the position of the unit line segment and the normal case is, the smaller the second target value is, that is, the greater the possibility that the unit line segment is a crack is.
Since the greater the likelihood of a crack, the more likely it is that the crack will be, so the unit line segment corresponding to the greatest likelihood will be the target unit line segment, and the crack will be broken by the metal, but the overall appearance will be in the form of extending in a certain direction, and not like the straighter form of fin, therefore, the extension straight line of the unit line segment is obtained along the direction of the unit line segment corresponding to the greatest likelihood, representing the extension direction of the whole crack, namely the extension straight line after the target unit line segment extends in the support domain can be obtained, how many pixel points are on the extension straight line, namely the number of the unit line segments included in the extension straight line.
S4, obtaining a similar line segment.
Specifically, according to the difference value of the likelihood of the target unit line segment and the neighborhood unit line segment in the neighborhood of the target unit line segment, a first target line segment in the neighborhood unit line segment is obtained, the first target line segment is used as a new target unit line segment, a second target line segment in the neighborhood of the first target line segment is obtained, stopping until the number of the obtained target line segments is equal to the number of the unit line segments contained in the extension straight line, and taking all the obtained target line segments as similar line segments.
The obtaining the first target line segment in the neighborhood unit line segment in this embodiment includes: acquiring the minimum possibility difference value of the possibility difference values of each neighborhood unit line segment in the neighborhood of the target unit line segment and the target unit line segment; and taking the neighborhood unit line segment corresponding to the minimum probability difference value as a first target line segment.
All the obtained target line segments are obtained, including: taking each obtained target line segment as a new target unit line segment, for example, according to the possibility difference value between the target unit line segment and the neighborhood unit line segment in the neighborhood of the target unit line segment, obtaining a first target line segment in the neighborhood unit line segment, then taking the first target line segment as a new target unit line segment, obtaining a second target line segment in the neighborhood unit line segment in the neighborhood of the first target line segment, then taking the second target line segment as a new target unit line segment, obtaining a third target line segment in the neighborhood unit line segment in the neighborhood of the second target line segment, stopping until the number of obtained target line segments is equal to the number of unit line segments contained in the extension straight line, obtaining target line segments with the same number as the unit line segments corresponding to the extension straight line at this time, taking all obtained target line segments as the similar line segments, sequentially connecting the obtained target line segments in the order of obtaining the target line segments (namely the similar line segments), as shown in fig. 4, and then connecting the target line segments (namely the similar line segments) in the order of obtaining the target line segments to obtain the similar line segments, namely the initial circumscribed rectangle of the area formed by the similar line segment and the extension straight line as shown in fig. 5.
S5, obtaining the blurred degrees of the similar line segments and the unit line segments with the same numbers.
Specifically, the obtained similar line segments and the unit line segments included in the extension straight line are numbered respectively, the number of the similar line segments is that the number of the unit line segments included in the extension straight line is that the unit line segments included in the extension straight line is sequentially numbered according to the extension direction of the similar line segments, and the fuzzy degree of the similar line segments and the unit line segments which are the same in number is obtained according to the sum of distances between pixel points corresponding to the similar line segments and the unit line segments which are the same in number and the difference of possibility between the similar line segments and the unit line segments which are the same in number, wherein the number of the similar line segments is that the sequence of the similar line segments is obtained, and the number of the unit line segments on the extension straight line is marked along the number traversal direction of the similar line segments.
Specifically, obtaining the blurred degrees of the similar line segments and the unit line segments with the same numbers includes: taking a natural constant as a base, and taking the difference of the possibility between the similar line segments with the same number and the unit line segments as a power to obtain an exponential function value; taking the product of the exponential function value and the distance between the similar line segments with the same numbers and the pixel points corresponding to the unit line segments as the initial blurred degree of the similar line segments with the same numbers, wherein the calculation formula of the initial blurred degree is as follows:
in the method, in the process of the invention,
the number of the extension line YS is +.>
Unit line segment of (2) and number +.>
The initial blurriness of the similar line segments.
e represents a natural constant.
The difference in the possibility
The larger the number is, the stronger the characteristic interference exists between the similar line segments with the same number and the unit line segments, namely the greater the degree of blurring is, the Euclidean distance is +.>
The larger the description is, the more discrete the similar line segment and the unit line segment are, i.e., the greater the degree of blurring is.
Specifically, obtaining the blurred degrees of the similar line segments and the unit line segments with the same numbers includes: acquiring an initial blurred degree difference value between each initial blurred degree and two initial blurred degrees adjacent in number; according to the difference value of the two initial blurred degrees, the average value of the initial blurred degrees of the similar line segments and the unit line segments with the same numbers is obtained, and the blurred degree of each similar line segment and the unit line segment with the same numbers is obtained, wherein the calculation formula of the blurred degree of the similar line segments and the unit line segments with the same numbers is as follows:
in the method, in the process of the invention,
representing the number +.>
The degree of blurring of the corresponding two unit line segments.
It should be noted that, for the position with large degree of blurring, the constraint is needed by the rest positions with small degree of blurring, so as to achieve the purpose of accurate rectangular range, namely, by calculation
Number->
And number->
Corresponding initial blurred degree difference +.>
Number->
And number->
Corresponding initial blurriness difference values, the greater the two initial blurriness difference values, the explanation of blurriness +.>
The larger the rectangle range should be constrained when performing the LSD straight line detection algorithm, and secondly, the XS in the formula is the similarity line in FIG. 4, with all the similarity line segments lying on the similarity line.
S6, acquiring a self-adaptive rectangular range of the support domain, and obtaining a crack region.
Because the greater the degree of blurring, the more the rectangular range of the target region is constrained, all final similar line segments are obtained according to the degree of blurring of similar line segments and unit line segments with the same number, and the target region is obtained according to all final similar line segments and extension straight lines; and taking the minimum circumscribed rectangle of the target area as an adaptive rectangle range of the corresponding support domain, and obtaining a crack area by utilizing an LSD linear algorithm based on the adaptive rectangle range of the support domain.
Specifically, obtaining a final similar line segment includes: taking a preset blurred degree threshold value to be 0.7 according to the detection precision requirement, and then acquiring a target number corresponding to a unit line segment when the blurred degree is greater than the preset blurred degree threshold value; acquiring a target distance between a pixel point of a similar line segment corresponding to a number adjacent to a target number and a pixel point of a unit line segment; taking the average value of the two target distances as the distance between the target line segment on the extension line corresponding to the target number and the final similar line segment; and obtaining the final similar line segment corresponding to the target number of the target line segment on the extension straight line according to the distance between the target line segment on the extension straight line corresponding to the target number and the final similar line segment.
As shown in fig. 5, the acquisition of the adaptive rectangular range includes: and connecting all final similar line segments to obtain final similar lines, taking a region formed by connecting the final similar lines with adjacent ends of the extended straight lines as a target region, taking the minimum circumscribed rectangle of the target region, taking the minimum circumscribed rectangle as an adaptive rectangular range of a corresponding support region, obtaining the straight line based on the adaptive rectangular range of the support region and by utilizing an LSD straight line algorithm, and judging whether the straight line is a crack straight line or not.
The invention relates to a visual inspection method for cracks of a clutch shell of an electrically controlled silicone oil fan, which is characterized in that when the cracks appear on the clutch shell due to the interference of fins on the clutch shell, the edges of the cracks are blurred, and the cracks cannot be accurately acquired, so that a line field and a supporting domain in an image are firstly acquired by utilizing an LSD algorithm, then a target unit line segment with the maximum possibility of the cracks in the supporting domain is acquired, then an extension straight line is acquired in the extension direction of the target unit line segment, namely, the extension straight line is acquired along the extension direction of the target unit line segment because a metal fracture trace is in a fold line shape but generally presents the characteristic of extending in a certain direction and is not similar to a shape of which the fins are relatively straight, the unit line segment existing near the cracks generates blur to the unit line segment belonging to the cracks in one supporting domain, and the degree of the blurring is judged by combining with the extension characteristics of the crack shape, so that a rectangular range is acquired, firstly, according to the probability difference value between the target unit line segment and the neighborhood unit line segment in the neighborhood of the target unit line segment, obtaining a first target line segment in the neighborhood unit line segment, taking the first target line segment as a new target unit line segment, obtaining a second target line segment in the neighborhood of the first target line segment, stopping until the number of the obtained target line segments is equal to the number of the unit line segments contained in the extension straight line, taking all the obtained target line segments as similar line segments, then obtaining the blurred degree of each similar line segment with the same number and the unit line segment according to the distance between the similar line segments with the same number and the pixel points corresponding to the unit line segments and the probability difference between the similar line segments with the same number and the unit line segments, determining the final similar line segment, determining the self-adaptive rectangular range based on all the final similar line segments and the extension straight line, the blurred degree is comprehensively obtained through the possibility and the extension characteristics of the crack morphology, so that the adaptive rectangular range obtained based on the blurred degree is constrained near the real crack range, namely, the crack region obtained based on the adaptive rectangular range is more accurate.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (9)
1. The visual detection method for the cracks of the clutch shell of the electrically-controlled silicone oil fan is characterized by comprising the following steps of:
acquiring a gray image of a shell of the silicone oil fan clutch;
obtaining a plurality of support domains in a line field by using a unit line segment perpendicular to the gradient direction of each pixel point in the gray level image;
according to the gradient amplitude value of each unit line segment in each supporting domain, the gray value of the neighborhood pixel point in the neighborhood of the pixel point corresponding to the unit line segment and the gray average value of the normal shell, the possibility that each unit line segment is a crack is obtained, the unit line segment corresponding to the maximum possibility is taken as a target unit line segment, and the number of the unit line segments contained in the extended straight line after the target unit line segment in the supporting domain is obtained;
according to the probability difference value between the target unit line segment and the neighborhood unit line segment in the neighborhood of the target unit line segment, acquiring a first target line segment in the neighborhood unit line segment, taking the first target line segment as a new target unit line segment, acquiring a second target line segment in the neighborhood of the first target line segment, stopping until the number of the acquired target line segments is equal to the number of the unit line segments contained in the extension straight line, and taking all the acquired target line segments as similar line segments;
numbering all the obtained similar line segments and unit line segments contained in the extension straight line respectively, wherein the numbering of the similar line segments is carried out according to the acquired sequence, the unit line segments contained in the extension straight line are sequentially numbered according to the extension direction of all the similar line segments, and the blurred degree of the similar line segments with the same number and the unit line segments is acquired according to the sum of the distances between the similar line segments with the same number and the corresponding pixel points of the unit line segments and the difference of the possibility between the similar line segments with the same number and the unit line segments;
acquiring all final similar line segments by using the blurred degree of each similar line segment and unit line segment with the same number, and acquiring a target area according to all final similar line segments and extension lines; and taking the minimum circumscribed rectangle of the target area as an adaptive rectangle range of the corresponding support domain, and obtaining a crack area by utilizing an LSD linear algorithm based on the adaptive rectangle range of the support domain.
2. The visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein obtaining the probability that each unit line segment is a crack comprises:
acquiring a gradient amplitude value of each unit line segment in the support domain, and a gradient amplitude value difference value of the maximum gradient amplitude value in the support domain;
performing quadratic function mapping on the gradient amplitude difference value to obtain a first target value;
acquiring a gray level difference value of a gray level average value of a neighborhood pixel point in the neighborhood of the pixel point corresponding to a unit line segment in the support domain and a normal shell;
carrying out negative correlation calculation on the average value of the gray difference absolute values of the gray average values of all neighborhood pixel points in the neighborhood of the pixel point corresponding to the unit line segment and the normal shell to obtain a second target value;
and carrying out normalization calculation on the ratio of the first target value to the second target value to obtain the possibility that each unit line segment is a crack.
3. The visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein obtaining a first target line segment in a neighborhood unit line segment comprises:
acquiring the minimum possibility difference value of the possibility difference values of each neighborhood unit line segment in the neighborhood of the target unit line segment and the target unit line segment;
and taking the neighborhood unit line segment corresponding to the minimum probability difference value as a first target line segment.
4. The visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein obtaining a difference in likelihood between like line segments of the same number and unit line segments comprises:
acquiring the absolute value of the probability difference between the similar line segments with the same number and the unit line segments;
the absolute value of the probability difference is taken as the probability difference between the similar line segments with the same number and the unit line segments.
5. The visual inspection method for cracks of a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein obtaining the degree of blurring of similar line segments and unit line segments with the same number comprises:
taking a natural constant as a base, and taking the difference of the possibility between the similar line segments with the same number and the unit line segments as a power to obtain an exponential function value;
taking the product of the exponential function value, the distances between the similar line segments with the same numbers and the pixel points corresponding to the unit line segments as the initial blurred degree of the similar line segments with the same numbers and the unit line segments;
acquiring an initial blurred degree difference value between each initial blurred degree and two initial blurred degrees adjacent in number;
and obtaining the blurred degrees of the similar line segments with the same numbers and the unit line segments according to the average value of the initial blurred degrees of the similar line segments with the same numbers and the unit line segments.
6. The visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein obtaining all final similar line segments comprises:
acquiring a target number corresponding to a unit line segment when the blurred degree is greater than a preset blurred degree threshold value;
acquiring a target distance between a pixel point of a similar line segment corresponding to a number adjacent to a target number and a pixel point of a unit line segment;
taking the average value of the two target distances as the distance between the target line segment on the extension line corresponding to the target number and the final similar line segment;
and obtaining the final similar line segment corresponding to the target number of the target line segment on the extension straight line according to the distance between the target line segment on the extension straight line corresponding to the target number and the final similar line segment.
7. The visual inspection method for cracks in a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein all final similar line segments are connected to obtain a final similar line.
8. The visual inspection method for cracks of a clutch housing of an electrically controlled silicone oil fan according to claim 1, wherein an area defined by connecting a final similarity line with an adjacent end of an extension straight line is used as a target area.
9. The visual inspection method of cracks in a housing of an electrically controlled silicone oil fan clutch according to claim 1, wherein acquiring a gray scale image of the housing of the silicone oil fan clutch comprises:
acquiring a shell image of a silicone oil fan clutch;
the R, G, B component of the shell image in the RGB color space is projection weighted to obtain a gray scale image of the shell.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310301726.8A CN116012767B (en) | 2023-03-27 | 2023-03-27 | Visual detection method for cracks of clutch housing of electrically-controlled silicone oil fan |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310301726.8A CN116012767B (en) | 2023-03-27 | 2023-03-27 | Visual detection method for cracks of clutch housing of electrically-controlled silicone oil fan |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116012767A true CN116012767A (en) | 2023-04-25 |
| CN116012767B CN116012767B (en) | 2023-10-13 |
Family
ID=86037756
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310301726.8A Active CN116012767B (en) | 2023-03-27 | 2023-03-27 | Visual detection method for cracks of clutch housing of electrically-controlled silicone oil fan |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116012767B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117274288A (en) * | 2023-09-27 | 2023-12-22 | 河海大学 | Shaft part sub-pixel edge detection method based on improved LSD algorithm |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109389644A (en) * | 2017-08-10 | 2019-02-26 | 纵目科技（上海）股份有限公司 | Parking stall line detecting method based on direction gradient enhancing |
| CN114140484A (en) * | 2021-11-24 | 2022-03-04 | 中山大学 | High-robustness sea-sky-line extraction method based on photoelectric sensor |
| CN114757898A (en) * | 2022-03-31 | 2022-07-15 | 广州科语机器人有限公司 | Image line segment detection method based on LSD algorithm |
| CN114998350A (en) * | 2022-08-04 | 2022-09-02 | 山东泗水泰和石材有限公司 | Stone defect detection method based on image processing |
-
2023
- 2023-03-27 CN CN202310301726.8A patent/CN116012767B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109389644A (en) * | 2017-08-10 | 2019-02-26 | 纵目科技（上海）股份有限公司 | Parking stall line detecting method based on direction gradient enhancing |
| CN114140484A (en) * | 2021-11-24 | 2022-03-04 | 中山大学 | High-robustness sea-sky-line extraction method based on photoelectric sensor |
| CN114757898A (en) * | 2022-03-31 | 2022-07-15 | 广州科语机器人有限公司 | Image line segment detection method based on LSD algorithm |
| CN114998350A (en) * | 2022-08-04 | 2022-09-02 | 山东泗水泰和石材有限公司 | Stone defect detection method based on image processing |
Non-Patent Citations (1)
| Title |
|---|
| 徐辛超等: "基于多通道梯度特征的彩色图像直线提取方法", 《测绘工程》, vol. 31, no. 3, pages 40 - 45 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117274288A (en) * | 2023-09-27 | 2023-12-22 | 河海大学 | Shaft part sub-pixel edge detection method based on improved LSD algorithm |
| CN117274288B (en) * | 2023-09-27 | 2024-05-03 | 河海大学 | Shaft part sub-pixel edge detection method based on improved LSD algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116012767B (en) | 2023-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116309570B (en) | Titanium alloy bar quality detection method and system | |
| CN110866924B (en) | Line structured light center line extraction method and storage medium | |
| CN108629775B (en) | Thermal state high-speed wire rod surface image processing method | |
| CN115829883B (en) | Surface image denoising method for special-shaped metal structural member | |
| CN111563889B (en) | Liquid crystal screen Mura defect detection method based on computer vision | |
| CN110008969B (en) | Method and device for detecting image saliency region | |
| CN109272489B (en) | Infrared weak and small target detection method based on background suppression and multi-scale local entropy | |
| CN108805023B (en) | Image detection method, device, computer equipment and storage medium | |
| CN111915704A (en) | Apple hierarchical identification method based on deep learning | |
| CN115170669B (en) | Identification and positioning method and system based on edge feature point set registration and storage medium | |
| CN110111387B (en) | Dial plate characteristic-based pointer meter positioning and reading method | |
| CN116012767B (en) | Visual detection method for cracks of clutch housing of electrically-controlled silicone oil fan | |
| CN116205823A (en) | Ultrasonic image denoising method based on spatial domain filtering | |
| CN110996096B (en) | Tone mapping image quality evaluation method based on structural similarity difference | |
| CN116245880B (en) | Electric vehicle charging pile fire risk detection method based on infrared identification | |
| CN116958125B (en) | Electronic contest host power supply element defect visual detection method based on image processing | |
| CN109064479B (en) | Sea-sky-line detection method based on gray dynamic features of adjacent video frames | |
| CN113192059A (en) | On-orbit ship detection method based on infrared remote sensing image | |
| CN116468640B (en) | Video image enhancement method for Internet teaching | |
| CN114549441A (en) | Sucker defect detection method based on image processing | |
| CN115330646A (en) | Metal component inspection method for fitness equipment paint spraying process | |
| CN114092682A (en) | Small hardware fitting defect detection algorithm based on machine learning | |
| CN115797327A (en) | Defect detection method and device, terminal device and storage medium | |
| CN112489052A (en) | Line structure light central line extraction method under complex environment | |
| CN113159027A (en) | Seven-segment type digital display instrument identification method based on minimum external rectangle variant |
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 |