CN116758075A - Artificial intelligence-based blower motor operation fault detection method - Google Patents
Artificial intelligence-based blower motor operation fault detection method Download PDFInfo
- Publication number
- CN116758075A CN116758075A CN202311040459.XA CN202311040459A CN116758075A CN 116758075 A CN116758075 A CN 116758075A CN 202311040459 A CN202311040459 A CN 202311040459A CN 116758075 A CN116758075 A CN 116758075A
- Authority
- CN
- China
- Prior art keywords
- stain
- pixel point
- edge
- pixel
- point
- 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 34
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 16
- 230000007547 defect Effects 0.000 claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000000034 method Methods 0.000 claims description 47
- 238000011109 contamination Methods 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 19
- 239000011159 matrix material Substances 0.000 claims description 11
- 230000000739 chaotic effect Effects 0.000 claims description 7
- 238000003708 edge detection Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 239000002689 soil Substances 0.000 claims 3
- 238000004458 analytical method Methods 0.000 abstract description 7
- 238000004519 manufacturing process Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 238000003064 k means clustering Methods 0.000 description 3
- 230000004323 axial length Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 238000007670 refining Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000003796 beauty Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 239000000314 lubricant Substances 0.000 description 1
- 239000010687 lubricating oil Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 239000007769 metal material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000006082 mold release agent Substances 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
- 230000003245 working effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06T7/10—Segmentation; Edge detection
- G06T7/143—Segmentation; Edge detection involving probabilistic approaches, e.g. Markov random field [MRF] modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
- G06T7/44—Analysis of texture based on statistical description of texture using image operators, e.g. filters, edge density metrics or local histograms
-
- 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
- G06T2207/20076—Probabilistic image processing
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image data processing, in particular to an artificial intelligence-based blower motor operation fault detection method, which comprises the following steps: acquiring each stain closed edge in the blade surface gray level image, and performing stain image feature analysis to obtain a first pollution degree and a second pollution degree of each pixel point in each stain closed edge; improving the original crack significant value of each pixel point in each stain closed edge by using the first pollution degree and the second pollution degree to obtain a new significant value; judging whether cracks exist or not by using the new crack significance value; if the crack exists, further judging whether the operation fault of the blower motor occurs. According to the invention, the accuracy of the significant value of the crack defect covered or surrounded by the stain is enhanced, so that the crack detection accuracy of the blade is effectively improved, and the operation fault detection accuracy of the blower motor is improved.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to an artificial intelligence-based blower motor operation fault detection method.
Background
As a personal beauty appliance widely used, a blower has a working effect that directly affects the user's experience. The operation condition of the motor, which is a core component of the blower, directly determines the operation performance of the blower, and if the blower motor fails, the blower cannot operate or the operation efficiency of the blower is obviously reduced. The blower is faulty, and the motor is generally faulty due to factors such as overload operation, bearing damage, winding fault and the like, wherein for the crack defect of the blade, when the blade of the motor has serious crack defect, the motor is usually unstable or uneven in rotation, obvious noise and vibration are caused, the blade cannot effectively rotate, excessive fatigue of the motor cannot normally dissipate heat, and overheating problems can be caused, so that the normal operation of the blower is affected. Therefore, when the operation fault detection of the blower motor is carried out, the crack degree of the motor blade on the production line needs to be monitored, and the phenomenon that the unqualified blower flows into the market and influences the reputation and sales of merchants is avoided.
The crack degree of the motor blade in the traditional blower motor operation fault detection is monitored, so that not only is the human resource wasted, but also the subjectivity dependence on quality inspection personnel is strong, and the fault detection result is inaccurate. The existing blade crack detection based on the image is used for carrying out segmentation and extraction on crack areas in the image, and judging whether the current motor blade crack can have fault influence on the operation of a blower motor or not by analyzing the distribution degree of the crack areas in the whole blade. However, in the production process of the motor blade, friction resistance in the production process is reduced through some greasy dirt or lubricant, and substances may adhere to the surface of the blade to form stains which are difficult to wipe, but the existing blade crack detection method is difficult to identify crack defects covered by the stains or located near the stains, so that the blade crack detection accuracy is poor, and the poor blade crack detection further causes the poor operation fault detection accuracy of the blower motor.
Disclosure of Invention
In order to solve the technical problem of poor accuracy of detection of operation faults of a blower motor caused by the poor accuracy of crack detection of blades, the invention aims to provide an artificial intelligence-based blower motor operation fault detection method, which adopts the following specific technical scheme:
one embodiment of the invention provides an artificial intelligence-based blower motor operation fault detection method, which comprises the following steps:
acquiring a blade surface gray level image of a blower motor to be detected, and performing edge detection on the blade surface gray level image to obtain each edge;
determining the stain probability of each pixel point on each edge according to the gray value, curvature and gradient amplitude of each pixel point on each edge; determining each stain closure edge according to the stain probability;
determining a first pollution degree of each pixel point in each stain closing edge according to the stain probability and the position of each pixel point in each stain closing edge;
acquiring gray scale run matrixes corresponding to the stain closing edges, and determining second pollution degrees of each pixel point in the stain closing edges according to the gray scale run matrixes;
Determining a new crack significance value corresponding to each pixel point in each stain closing edge according to the first pollution degree and the second pollution degree of each pixel point in each stain closing edge and the original crack significance value;
judging whether cracks exist in the blades of the blower motor to be detected or not by using the new crack significant value; if the crack exists, judging whether the blower motor to be detected has faults or not according to the duty ratio of the crack in the gray level image of the blade surface.
Further, determining the stain probability of each pixel point on each edge according to the gray value, curvature and gradient amplitude of each pixel point on each edge comprises:
selecting a pixel point on any one edge as a selected pixel point, taking the selected pixel point as a starting point, selecting a preset number of continuous pixel points on the edge to which the selected pixel point belongs according to a preset direction, and forming a pixel point set corresponding to the selected pixel point;
calculating the gradient amplitude mean value of all the pixel points in the pixel point set corresponding to the selected pixel point; and determining the stain probability of the selected pixel according to the gray value of the selected pixel and the next pixel of the selected pixel, the curvature and gradient amplitude of each pixel in the pixel set corresponding to the selected pixel and the gradient amplitude mean value.
Further, the calculation formula of the stain probability of the selected pixel point is as follows:
wherein L is the stain probability of the selected pixel point,for the gray value of the selected pixel, +.>For the gray value of the next pixel point of the selected pixel point, i is the serial number of each pixel point in the pixel point set corresponding to the selected pixel point, and N is the number of the pixel points in the pixel point set corresponding to the selected pixel point>For the curvature of the (i+1) th pixel point in the pixel point set corresponding to the selected pixel point,/L>For the curvature of the ith pixel point in the pixel point set corresponding to the selected pixel point,/and%>Gradient amplitude of the ith pixel point in the pixel point set corresponding to the selected pixel point,/for the selected pixel point>For the gradient amplitude mean value of all the pixel points in the pixel point set corresponding to the selected pixel point, < +.>For the absolute function, norm is the normalization function.
Further, determining respective stain closure edges according to the stain probabilities, comprising:
for any one edge, if the stain probability of any one pixel point on the edge is larger than the stain threshold value, judging that the pixel point is a stain closed edge pixel point, otherwise, judging that the pixel point is not a stain closed edge pixel point; traversing all the stain closing edge pixel points in the blade surface gray level image by utilizing an edge tracking algorithm to obtain each stain closing edge to be selected;
Judging whether the closed edges of any stains to be selected are overlapped or not through intersection judgment, and if so, marking the overlapped edges; when judging the internal and external relation between any pixel point on the closed edge of the stain to be selected and other closed edges of the stain to be selected by using a ray method, checking whether an intersection point is generated between the pixel point on the ray path and the marked coincident edge, and if the intersection point is generated, introducing rays in the direction instead of the intersection point until the intersection point does not exist between the pixel point and the coincident edge; counting the number of intersection points when the pixel points on the closed edge of the stain to be selected intersect with other closed edges of the stain to be selected, and if the number of intersection points is even, judging that the closed edge of the stain to be selected is the closed edge of the stain; and if the number of the intersection points is odd, discarding the closed edge of the stain to be selected.
Further, determining a first contamination level for each pixel point within the respective stain closure edge based on the stain probability and the location of each pixel point within the respective stain closure edge, comprising:
optionally selecting a pixel point on any stain closed edge as a target point, constructing a coordinate system by taking the target point as an origin, and making all straight lines of the origin; according to the coordinate position of each pixel point on the stain closing edge, calculating the distance between two intersection points where the straight line and the stain closing edge intersect, and obtaining a maximum distance value and a minimum distance value corresponding to the target point;
Determining four axis-edge intersection points of two coordinate axes of the coordinate system and the stain closed edge, and calculating the distance between the target point and each axis-edge intersection point;
and determining the first pollution degree of the target point on the stain closed edge according to the maximum distance value and the minimum distance value corresponding to the target point, the distance between the target point and each axis edge intersection point and the stain probability of each axis edge intersection point.
Further, the calculation formula of the first pollution level is as follows:
in the method, in the process of the invention,for the first pollution level of the target point, j is the serial number of the intersection point of the axis and the side>For the distance between the target point and the intersection of the j-th axis edge,/>For the minimum distance value corresponding to the target point, +.>For the maximum distance value corresponding to the target point, +.>Is the dirty probability of the j-th axis edge intersection point.
Further, determining a second contamination level for each pixel point within each stain closure edge according to the gray scale run matrix, comprising:
clustering the gray scale run matrix corresponding to each stain closing edge to obtain a gray scale range corresponding to a target cluster of each stain closing edge; the target cluster is the cluster with the largest run length;
selecting pixel points belonging to a gray level range corresponding to a target cluster in the stain closing edge as centers for any stain closing edge, and constructing a local window area with a preset size; and determining the local confusion degree of the central pixel point in the local window area according to the gray value of the central pixel point in the local window area, the gray average value of all the pixel points and the local texture characteristic value of the central pixel point, and determining the local confusion degree as the second pollution degree of the central pixel point.
Further, determining the local confusion degree of the central pixel point in the local window area according to the gray value of the central pixel point in the local window area, the gray average value of all the pixel points and the local texture characteristic value of the central pixel point comprises the following steps:
calculating the ratio of the gray value of the central pixel point to the gray average value of all the pixel points, carrying out data processing on the ratio by utilizing a logarithmic function, and determining the absolute value of the ratio after the data processing as a first chaotic factor of the central pixel point; determining the local texture characteristic value of the central pixel point as a second chaotic factor; the product of the first clutter factor and the second clutter factor is determined as the local clutter level of the center pixel.
Further, the calculation formula of the new crack significance value is as follows:
in the method, in the process of the invention,the inner coordinate of the closed edge for stains is +.>New crack significance values corresponding to pixel points of (1), x is an abscissa, y is an ordinate, and Norm is a normalization function; />The inner coordinate of the closed edge for stains is +.>An exponential function of the pixel points of +.>The inner coordinate of the edge closed by the stain is +.>Whether the gray value of the pixel point of (2) is within the gray level range corresponding to the target cluster +.>Judging internally, if yes- >Is a value of 1, if not, +.>Is a value of 0; e is a natural constant, < >>The inner coordinate of the closed edge for stains is +.>First contamination level of the pixel of +.>The inner coordinate of the closed edge for stains is +.>Second contamination level of the pixel of (2),/and (c)>The inner coordinates of the closed edge for stains areThe pixel points of (3) correspond to the original crack significance values.
Further, determining whether a crack exists in a blade of the blower motor to be detected using the new crack significance value includes:
substituting the new crack significant value into a significant value detection algorithm, and detecting the significant value of the blade surface gray level image; judging whether pixel points with the significant value larger than the defect threshold exist in the gray level image of the surface of the blade, if so, judging that cracks exist in the blade of the blower motor to be detected, and determining the pixel points with the significant value larger than the defect threshold as crack pixel points.
The invention has the following beneficial effects:
the invention provides a blower motor operation fault detection method based on artificial intelligence, which comprises the steps of firstly, obtaining a blade surface gray level image for facilitating subsequent image feature extraction and analysis, and carrying out edge detection on the blade surface gray level image to obtain each edge; in order to reduce the significant value of each pixel point in the stain closing edge, image analysis is carried out on each stain closing edge, the characteristic information of the stain closing edge is extracted, the stain closing edge is required to be screened out from a plurality of edges, and the stain probability with higher accuracy can be obtained by analyzing the gray value, curvature and gradient amplitude of each pixel point on each edge, so that each stain closing edge is determined; the first pollution degree of each pixel point in each stain closing edge can be obtained by analyzing the relative position of each pixel point in each stain closing edge from the edge, and the second pollution degree of each pixel point in each stain closing edge can be obtained by analyzing the local texture characteristics of each pixel point in each stain closing edge; the significance value calculation is carried out based on the first pollution degree and the second pollution degree, so that the significance of crack defects surrounded or covered by stains is improved, the significance of a stain area is reduced, the original significance value calculation can be used for a conventional area, so that the crack defects which cannot be detected due to the fact that the voltage blade is wrapped by the stains have higher precision, and the robustness and the precision of the existing significance algorithm are further enhanced; the invention improves the accuracy of detecting the cracks of the motor blade of the blower, and further improves the accuracy of detecting the operation faults of the motor of the blower through accurate crack detection of the blade, and is mainly applied to the field of detecting the operation faults of the motor.
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 flow chart of a method for detecting operation faults of a blower motor based on artificial intelligence according to the present invention;
fig. 2 is a schematic diagram corresponding to a coordinate system constructed by taking a target point as an origin in an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
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.
Application scenario for which this embodiment is aimed: and detecting blade cracks on the production line of the blower motor, and judging whether the crack degree of the blade of the blower motor to be detected can cause the failure of the blower motor to be detected. When the existing crack image segmentation algorithm is used for blade crack analysis, crack defects covered by stains or positioned near the stains are easily ignored, so that a crack degree detection result is inaccurate, and further, the operation fault detection accuracy of the blower motor is poor.
In order to improve the accuracy of detecting the operation fault of the blower motor, the embodiment provides an artificial intelligence-based blower motor operation fault detection method, as shown in fig. 1, comprising the following steps:
s1, acquiring a blade surface gray level image of a blower motor to be detected, and performing edge detection on the blade surface gray level image to obtain each edge.
First, acquiring a blade surface gray image of a blower motor to be detected.
In this embodiment, an industrial camera of a Charge-Coupled Device (CCD) is selected to photograph a blade of a blower motor to be detected, and when photographing, the industrial camera should be perpendicular or nearly perpendicular to the front of the blade, and the photographing distance is moderate, so as to ensure that a clear and complete product image can be obtained. The product image refers to a blade surface image of the blower motor to be detected.
In order to facilitate the subsequent image data analysis, the blade surface image is subjected to graying treatment, so that a blade surface gray image can be obtained, and the graying treatment is realized by the following methods including but not limited to: the implementation process of the graying process is a prior art, and will not be described in detail here. In order to improve the definition of the gray level image on the surface of the blade, the gray level image on the surface of the blade is subjected to image enhancement processing, namely, the image enhancement processing is performed by utilizing median filtering denoising and self-adaptive histogram to perform balanced enhancement, so that the influence of most of noise in the image is eliminated, and the implementation process of image enhancement is the prior art and is not described in detail herein. It should be noted that, the images used in the subsequent image analysis are all images subjected to the image enhancement processing.
And secondly, carrying out edge detection on the gray level image of the blade surface to obtain each edge.
In this embodiment, the clean and nondestructive blade is smoother, if stains exist on the blade, the whole image is more obvious, and the image features of the stained area in the grayscale image on the surface of the blade can be analyzed, so that the accuracy of crack defect detection is improved. And carrying out edge detection on the gray level image of the blade surface by using a Canny edge operator to obtain each initial edge, wherein the implementation process of the Canny edge operator is the prior art and is not described in detail herein. And refining each initial edge by using Zhang-suen, and refining the initial edges with uneven thickness into an edge skeleton with single pixel points, so as to obtain each edge in the gray level image of the blade surface, wherein the implementation process of Zhang-suen is the prior art and is not described in detail herein.
To this end, the present embodiment obtains the respective edges in the blade surface gradation image.
S2, determining the stain probability of each pixel point on each edge according to the gray value, curvature and gradient amplitude of each pixel point on each edge; each stain closure edge is determined based on the stain probabilities.
When producing the blower motor blade, the lubricating oil used may adhere to the blade surface, which may cause stains that are difficult to clean after adsorbing other impurities, and residues such as mold release agent and glue used during molding of the blade may be transferred to the blade surface, thereby forming stains that are difficult to clean. The accumulation of stains can mask the production crack defects of the motor blade, and the traditional image processing can hardly detect the masked crack defects, so that the stain areas on the surface of the blade are subjected to heavy detection analysis when the motor blade is detected.
First, determining the stain probability of each pixel point on each edge according to the gray value, curvature and gradient amplitude of each pixel point on each edge.
In order to screen out the edges of the stain region from the edges, the probability that each pixel point on each edge is the pixel point on the edge of the stain region is determined based on the image characteristics of the stain edges. The specific implementation steps can include:
The first substep, selecting a pixel point on any edge as a selected pixel point, taking the selected pixel point as a starting point, selecting a preset number of continuous pixel points on the edge to which the selected pixel point belongs according to a preset direction, and forming a pixel point set corresponding to the selected pixel point.
In this embodiment, the preset direction is a randomly selected direction, and the preset directions of the same edge are the same; the number of the selected preset number of pixel points can be 10, and the number of the selected continuous pixel points can be set by an implementer according to specific practical conditions; the pixel point set corresponding to the selected pixel point comprises the selected pixel point per se, and the number of the pixel points in the pixel point set of the selected pixel point is 10.
A second sub-step of calculating the gradient amplitude mean value of all the pixels in the pixel set corresponding to the selected pixel; and determining the stain probability of the selected pixel according to the gray value of the selected pixel and the next pixel of the selected pixel, the curvature and gradient amplitude value of each pixel in the pixel set corresponding to the selected pixel and the gradient amplitude value mean value.
As an example, the calculation formula of the stain probability of the selected pixel point may be:
Wherein L is the stain probability of the selected pixel point,for the gray value of the selected pixel, +.>For the gray value of the next pixel point of the selected pixel point, i is the serial number of each pixel point in the pixel point set corresponding to the selected pixel point, and N is the number of the pixel points in the pixel point set corresponding to the selected pixel point>For the curvature of the (i+1) th pixel point in the pixel point set corresponding to the selected pixel point,/L>For the curvature of the ith pixel point in the pixel point set corresponding to the selected pixel point,/and%>Gradient amplitude of the ith pixel point in the pixel point set corresponding to the selected pixel point,/for the selected pixel point>For the gradient amplitude mean value of all the pixel points in the pixel point set corresponding to the selected pixel point, < +.>For the absolute function, norm is the normalization function.
In the calculation formula of the probability of stains,can be used for a watchThe gray level difference between the selected pixel point and the adjacent edge pixel point is characterized in that the gray level difference of the adjacent pixel points on the edge of partial stains is larger, so that the larger the gray level difference is, the larger the probability that the selected pixel point is the pixel point on the edge of the stains is; />The method can be used for representing the summation of N-1 adjacent curvature differences corresponding to N continuous edge pixel points, and the larger the summation of N-1 adjacent curvature differences is due to the irregular shape of part of the stain edge, the larger the probability that the selected pixel point is the pixel point on the stain edge is; / >The method can be used for representing the difference summation between the gradient amplitude value of each pixel point in the pixel point set corresponding to the selected pixel point and the gradient amplitude value average value of all the pixel points, and the larger the difference summation of the gradient amplitude values is, the larger the probability that the selected pixel point is the pixel point on the edge of the dirty point is indicated because the interior of the blade is smooth and the dirty area is distributed roughly. Referring to the calculation process of the stain probability of the selected pixel points, the stain probability of each pixel point on each edge can be obtained.
And secondly, determining the closed edge of each stain according to the stain probability.
The first substep, for any one edge, judging that the pixel point is a stain closed edge pixel point if the stain probability of any one pixel point on the edge is larger than a stain threshold value, otherwise, judging that the pixel point is not the stain closed edge pixel point; and traversing all the stain closing edge pixel points in the blade surface gray level image by utilizing an edge tracking algorithm to obtain each stain closing edge to be selected.
In this embodiment, the stain threshold may take a tested value of 0.7, determine the stain edge for each pixel point on each edge, extract and mark the pixels on all the stain edges, and obtain all the stain closed edge pixels. The edge of the stain area on the blade is a closed area, so an edge tracking algorithm can be used to check whether the tracking path is closed during tracking, and the closed edge is determined to be the stain closed edge to be selected. The implementation process of the edge tracking algorithm is the prior art, and is not in the scope of the present invention, and will not be described in detail here.
A second sub-step, judging whether the closed edges of the stains to be selected are overlapped or not through intersection judgment, and if so, marking the overlapped edges; when judging the internal and external relation between any pixel point on the closed edge of the stain to be selected and other closed edges of the stain to be selected by using a ray method, checking whether an intersection point is generated between the pixel point on the ray path and the marked coincident edge, and if the intersection point is generated, introducing rays in the direction instead of the intersection point until the intersection point does not exist between the pixel point and the coincident edge; counting the number of intersection points when the pixel points on the closed edge of the stain to be selected intersect with other closed edges of the stain to be selected, and if the number of intersection points is even, judging that the closed edge of the stain to be selected is the closed edge of the stain; and if the number of the intersection points is odd, discarding the closed edge of the stain to be selected.
In this embodiment, there may be a sub-annular edge within the stain-closing edge to be selected, which indicates that the stain-closing edge to be selected contains different levels of contamination, but only the outermost edge is selected as the stain-closing edge in this embodiment. In order to eliminate the special condition that the overlapping edges exist between two adjacent stain closing edges to be selected, firstly judging whether the stain closing edges to be selected are overlapped by utilizing an intersection judgment existing algorithm so as to obtain more accurate intersection point number, judging whether the stain closing edges to be selected are sub-annular edges, and when the intersection point number is even, indicating that the stain closing edges to be selected are positioned on the outermost layer edges, so that the stain closing edges to be selected can be directly judged to be stain closing edges; when the number of the intersection points is odd, the closed edges of the stains to be selected are located on the sub-annular edges, namely, the closed edges of the stains to be selected are the sub-annular edges, and the closed edges of the stains to be selected can be directly discarded. The detailed implementation process of intersection judgment and ray method is the prior art, and is not in the scope of the present invention, and will not be described in detail here.
To this end, this embodiment obtains individual stain closing edges in the blade surface grayscale image.
S3, determining the first pollution degree of each pixel point in each stain closing edge according to the stain probability and the position of each pixel point in each stain closing edge.
It should be noted that, for the stain on the motor blade, the closer to the center of the stain area, the higher the degree of contamination, the closer to the edge of the stain area, the lower the degree of contamination, and all the pixel points in the stain closing edge may form the stain area. By analyzing the position of each pixel point in the stain-closing edge, the first contamination level of each pixel point in the respective stain-closing edge is analyzed, and the specific implementation steps may include:
the first step, selecting a pixel point on any stain closed edge as a target point, constructing a coordinate system by taking the target point as an origin, and making all straight lines of the origin; and calculating the distance between two intersection points of the straight line and the stain closing edge according to the coordinate position of each pixel point on the stain closing edge, and obtaining a maximum distance value and a minimum distance value corresponding to the target point.
In this embodiment, a schematic diagram corresponding to a coordinate system constructed by taking a target point as an origin is shown in fig. 2, x represents an abscissa, y represents an ordinate, a position between two intersection points may be a euclidean distance between two points, and a calculation process of the euclidean distance between two points is a prior art, which is not described in detail herein. The number of all straight lines passing through the origin is not particularly limited in this embodiment, and the main purpose of this step is to obtain the maximum distance value and the minimum distance value corresponding to the target point on the stain-closing edge.
And secondly, determining the intersection points of the two coordinate axes of the coordinate system and the four axis edges of the stain closed edge, and calculating the distance between the target point and each axis edge intersection point.
In this embodiment, the distance between the target point and each intersection of the axis edges is determined to measure the distance between the target point and the edge, and the closer the distance between the target point and each intersection of the axis edges is, the closer the target point is to the stain closing edge, which means that the lower the pollution level of the target point is. The two coordinate axes are respectively an abscissa axis and an ordinate axis, the intersection points of four axis edges are the intersection points of a positive transverse axis, a negative transverse axis, a positive longitudinal axis and a negative longitudinal axis which take a target point as an origin and the closed edge of a stain to which the target point belongs, the distance between the two points is the Euclidean distance, the calculation process of the Euclidean distance is the prior art, and the calculation process is not described in detail in the protection scope of the invention.
And thirdly, determining the first pollution degree of the target point on the stain closed edge according to the maximum distance value and the minimum distance value corresponding to the target point, the distance between the target point and each axis edge intersection point and the stain probability of each axis edge intersection point.
As an example, the calculation formula of the first pollution level of the target point may be:
In the method, in the process of the invention,for the first pollution level of the target point, j is the serial number of the intersection point of the axis and the side>For the distance between the target point and the intersection of the j-th axis edge,/>For the minimum distance value corresponding to the target point, +.>For the maximum distance value corresponding to the target point, +.>Is the dirty probability of the j-th axis edge intersection point.
In the calculation formula of the first pollution level,denominator>Can also be +.>The denominator has the main effect of representing the proportion of the distance from the pixel point to the stain closing edge to the sum of the longest axial length and the shortest axial length of the whole stain area, and reflecting the relative position of the pixel point from the stain closing edge; for the stain area, the higher the internal pollution degree is, the larger the stain area is, and the maximum distance value and the minimum distance value corresponding to the target point are larger; distance between target point and jth axis intersection +.>Is +.>Is positive correlation, is->The greater the first degree of contamination->The larger; the stain probability of the intersection point of the shaft edge is used as the weight, so that the mapping of the edge to the pollution degree in the edge can be better reflected, and the pollution condition in the closed edge of the stain can be more accurately reflected; stain probability of axis edge intersection->With a first degree of pollution Is positive correlation, is->The bigger the->The larger; referring to the determination of the first contamination level of the target point, the first contamination level of each pixel point within the respective stain-closing edge may be obtained.
It should be noted that the larger longest and shortest axis lengths may indicate a flattened shape of the contaminated area, indicating that the stain distribution may be evenly distributed or trending. The higher the internal contamination level, the larger the size and morphology of the contaminated area, and the greater the computational impact on the internal pixel points. Thus, there is a need to quantify the edge morphology and size of the stain closure edge, i.e., determine the first degree of contamination.
Thus, the present embodiment obtains a first degree of contamination for each pixel point within the respective stain-closing edge.
S4, acquiring gray scale run matrixes corresponding to the stain closed edges, and determining the second pollution degree of each pixel point in the stain closed edges according to the gray scale run matrixes.
When a metal material is subjected to stress or environmental factors, cracks may be generated on the surface thereof. Because crack formation is a breaking behavior caused by the grain structure and grain boundaries of the metal, it affects crack propagation and propagation, resulting in the formation of specific morphology and paths of the cracks, such that the cracks exhibit significant structure and ordering. Therefore, the pixel point texture details of the spot area are disordered, the pixel point texture details of the crack area are strong in regularity, and obvious pixel distribution characteristics exist.
The texture details of the stained area are disordered, so that the gray level run length of the stained area is shorter, and the pixel level is disordered; in contrast, the texture details of the crack region are regular, so the run length of the crack region is generally long and the pixel level is not very different. Based on the gray scale run matrix corresponding to each stain closing edge, analyzing the texture detail confusion degree and the pixel level confusion condition inside each stain closing edge to determine the second pollution degree, wherein the specific implementation steps can include:
first, acquiring gray scale run matrixes corresponding to the closed edges of all stains.
In this embodiment, 256 gray values are divided equally into 64 gray levels, each pixel point in each stain-closing edge is traversed based on the gray values of all pixel points in each stain-closing edge, and the same gray level is counted continuouslyCan obtain a pixel number of a size ofThe elements in the two-dimensional matrix can represent the lengths travelled by the pixels of different grey levels, i.e. the pixels with a plurality of grey levels occurring consecutively within the stain-closing edge. The number of the halving gray levels may be set by the practitioner according to the specific practical situation, and is not particularly limited here.
And a second step of determining a second pollution degree of each pixel point in each stain closing edge according to the gray scale run matrix.
And a first sub-step, clustering the gray scale run matrix corresponding to each stain closed edge to obtain the gray scale range corresponding to the target cluster of each stain closed edge.
In this embodiment, the K-means clustering is used to cluster the gray scale run matrix corresponding to each stain closed edge, and the euclidean distance of the run length is used to measure the absolute difference of the run lengths in the clustering process, and the run lengths of the gray scale run matrix are subsequently divided into long run, medium run and short run, so the number of clusters in the K-means clustering is set to be 3. Selecting a gray level range corresponding to the target cluster from the three clusters, wherein the gray level values of all pixel points of the crack part can be considered to be in the corresponding gray level range in the gray level range corresponding to the target cluster; each stain closure edge has its corresponding gray value range. The target cluster is the cluster with the largest run length. The implementation process of K-means clustering is the prior art and is not within the scope of the present invention, and will not be described in detail herein.
A second sub-step of selecting pixel points belonging to a gray level range corresponding to the target cluster in the stain closing edge as the center for any stain closing edge to construct a local window area with a preset size; and determining the local confusion degree of the central pixel point in the local window area according to the gray value of the central pixel point in the local window area, the gray average value of all the pixel points and the local texture characteristic value of the central pixel point, and determining the local confusion degree as the second pollution degree of the central pixel point.
Firstly, selecting pixel points belonging to a gray level range corresponding to a target cluster in a stain closed edge as a center for a gray level run matrix of any stain closed edge, and constructing a local window area with a preset size.
In this embodiment, in order to analyze the local confusion degree of the pixels whose gray values in the closed edges of different stains belong to the gray level range corresponding to the target cluster, the size corresponding to the selected pixels is reconstructed to beWhich can be used to calculate the local clutter level of the pixel. Where M should be smaller, the checked value may be 1.
And secondly, determining the local confusion degree of the central pixel point in the local window area according to the gray value of the central pixel point in the local window area, the gray average value of all the pixel points and the local texture characteristic value of the central pixel point, and determining the local confusion degree as the second pollution degree of the central pixel point.
Calculating the ratio of the gray value of the central pixel point to the gray average value of all the pixel points, performing data processing by using the logarithmic function comparison value, and determining the absolute value of the ratio after the data processing as a first chaotic factor of the central pixel point; determining the local texture characteristic value of the central pixel point as a second chaotic factor; the product of the first clutter factor and the second clutter factor is determined as the local clutter level of the center pixel.
As an example, the calculation formula of the local confusion degree may be:
in the method, in the process of the invention,is the local confusion degree of the central pixel point, ln is a logarithmic function, ++>Is the gray value of the center pixel point,is the gray level average value of all pixel points in the local window area,/for the pixel points>Local texture feature value for center pixel, < +.>As a function of absolute value.
In the calculation formula of the degree of local confusion,the difference between the gray value of the central pixel point and the gray average value of all the pixel points in the local window can be represented, and the gray value is->The larger the pixel distribution in the local window is, the more chaotic the pixel distribution is; local texture feature value +.>Can be used for capturing local texture characteristics, and if the texture of the central pixel point is changed greatly, local texture characteristic valuesThe greater the local degree of confusion +. >The larger; degree of local confusion->The influence of the gray value of the central pixel point, the gray average value in the local window and the texture characteristic in the local window to which the central pixel point belongs reflects the local regularity of the central pixel point, and the greater the local confusion degree is, the worse the local regularity of the central pixel point is, and the greater the pollution degree is; to avoidSpecial cases where there is a negative number, use absolute function pair +.>Performing numerical processing; the local confusion degree of each pixel point in each stain closing edge can be obtained by referring to the calculation process of the local confusion degree of the central pixel point.
It should be noted that, if no crack defect exists in the stain-closing edge, the local window pixel distribution of the central pixel point in the corresponding stain-closing edge is more disordered, and the local texture characteristic value is higherLarger gray scale difference->Will also be greater, the degree of local confusion +.>The larger; if there is a crack defect in the stain closure edge, the local clutter of the pixel point on the crack defect is less.
Thus far, the present embodiment achieves a second degree of contamination for each pixel point within the respective stain-closing edge.
S5, determining new crack significance values corresponding to each pixel point in each stain closing edge according to the first pollution degree and the second pollution degree of each pixel point in each stain closing edge and the original crack significance values.
In this embodiment, the original crack salient value of each pixel point in each stain closed edge can be obtained by performing image processing on the crack defect in the blade surface gray level image using a history-based Contrast salient value detection algorithm. The implementation process of the saliency detection algorithm is the prior art and is not within the scope of the present invention, and will not be described in detail here.
Based on the first pollution level and the second pollution level, improving the original crack significance value of each pixel point in each stain closing edge, so as to obtain a new crack significance value corresponding to each pixel point in each stain closing edge, wherein a calculation formula of the new crack significance value can be as follows:
in the method, in the process of the invention,the inner coordinate of the closed edge for stains is +.>New crack significance values corresponding to pixel points of (1), x is an abscissa, y is an ordinate, and Norm is a normalization function; />The inner coordinate of the closed edge for stains is +.>An exponential function of the pixel points of +.>The inner coordinate of the edge closed by the stain is +.>Whether the gray value of the pixel point of (2) is within the gray level range corresponding to the target cluster +.>Judging internally, if yes->Is a value of 1, if not, +.>Is a value of 0; e is a natural constant, < > >The inner coordinate of the closed edge for stains is +.>First contamination level of the pixel of +.>The inner coordinate of the closed edge for stains is +.>Second contamination level of the pixel of (2),/and (c)>The inner coordinates of the closed edge for stains areThe pixel points of (3) correspond to the original crack significance values.
In the calculation formula of the new crack significance value, an exponential functionThe gray level range can be determined by the target cluster corresponding to the stain closed edge, if a crack part exists in the stain closed edge, the gray level values of all the pixel points corresponding to the crack part are indicated to be in the gray level range corresponding to the target cluster, and the pixel points outside the gray level range can be temporarily removed without analysis; first degree of contamination->Can be used for characterizing the relative position of the pixel point from the stain closing edge and the stain probability of the intersection point of the axis edge, analyzing the stain closing edge, and enabling the first degree of stain to be +.>The larger the pollution degree of the spot area where the pixel point is located is, the larger the pollution degree is, and in order to detect a crack area, the pixel point with the larger pollution degree is given with smaller weight; second degree of contamination->Can characterize the structure and the order presented by the pixel points, analyze the stain closing edge, and second pollution degree +. >The larger, the more complex and random the texture around the pixel, the lack of a distinct ordering pattern, which may indicate that the pixel is less likely to be a crack region, ">And will decrease as well.
Thus, the embodiment obtains a new crack significance value corresponding to each pixel point in each stain-closing edge.
S6, judging whether cracks exist in the blades of the blower motor to be detected or not by using the new crack significant value; if the crack exists, judging whether the blower motor to be detected has faults or not according to the duty ratio of the crack in the gray level image of the blade surface.
And firstly, judging whether cracks exist in the blades of the blower motor to be detected or not by using the new crack significant value.
Substituting the new crack significant value into a significant value detection algorithm, and detecting the significant value of the blade surface gray level image; judging whether pixel points with the significant value larger than the defect threshold exist in the gray level image of the surface of the blade, if so, judging that cracks exist in the blade of the blower motor to be detected, and determining the pixel points with the significant value larger than the defect threshold as crack pixel points.
In this embodiment, the empirical value of the defect threshold may be 1.5, and the pixel point with the significant value greater than the defect threshold is determined to be a significant pixel point, that is, a crack pixel point, so that all crack pixel points in the gray scale image of the blade surface may be obtained, and if there is a crack in the blade of the blower motor to be detected, it may be directly determined that there is a crack. The magnitude of the defect threshold may be set by the practitioner according to a specific practical situation, and is not particularly limited herein.
And secondly, judging whether the motor of the blower to be detected has faults or not according to the duty ratio of the cracks in the gray level image of the surface of the blade.
According to the embodiment, the method has stronger pertinence to the area around or covered by the stain, the crack defect in the area can be accurately detected, the number of pixels with the significant value larger than the defect threshold is counted, the proportion of the number of pixels with the significant value larger than the defect threshold in the number of all pixels in the whole blade surface gray level image is calculated, if the proportion is larger than the proportion threshold, the empirical value of the proportion threshold can be 0.1%, the blade surface crack curve is considered to generate faults for the operation of the blower motor, and the detection is completed.
The present invention has been completed.
The invention provides an artificial intelligence-based blower motor operation fault detection method, which is characterized in that a stain closed edge is obtained by detecting an edge based on a curvature analysis result of the edge, local confusion is built based on texture analysis in the stain closed edge, and an improved result of a significant value is obtained based on the local confusion and edge expression, so that the blower motor operation fault is detected, and the blower motor operation fault detection accuracy is effectively improved.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention and are intended to be included within the scope of the invention.
Claims (10)
1. The blower motor operation fault detection method based on artificial intelligence is characterized by comprising the following steps:
acquiring a blade surface gray level image of a blower motor to be detected, and performing edge detection on the blade surface gray level image to obtain each edge;
determining the stain probability of each pixel point on each edge according to the gray value, curvature and gradient amplitude of each pixel point on each edge; determining each stain closure edge according to the stain probability;
determining a first pollution degree of each pixel point in each stain closing edge according to the stain probability and the position of each pixel point in each stain closing edge;
Acquiring gray scale run matrixes corresponding to the stain closing edges, and determining second pollution degrees of each pixel point in the stain closing edges according to the gray scale run matrixes;
determining a new crack significance value corresponding to each pixel point in each stain closing edge according to the first pollution degree and the second pollution degree of each pixel point in each stain closing edge and the original crack significance value;
judging whether cracks exist in the blades of the blower motor to be detected or not by using the new crack significant value; if the crack exists, judging whether the blower motor to be detected has faults or not according to the duty ratio of the crack in the gray level image of the blade surface.
2. The method of claim 1, wherein determining the probability of a soil for each pixel on each edge based on the gray level, curvature, and gradient magnitude of each pixel on each edge comprises:
selecting a pixel point on any one edge as a selected pixel point, taking the selected pixel point as a starting point, selecting a preset number of continuous pixel points on the edge to which the selected pixel point belongs according to a preset direction, and forming a pixel point set corresponding to the selected pixel point;
Calculating the gradient amplitude mean value of all the pixel points in the pixel point set corresponding to the selected pixel point; and determining the stain probability of the selected pixel according to the gray value of the selected pixel and the next pixel of the selected pixel, the curvature and gradient amplitude of each pixel in the pixel set corresponding to the selected pixel and the gradient amplitude mean value.
3. The method for detecting the operation fault of the blower motor based on artificial intelligence according to claim 2, wherein the calculation formula of the contamination probability of the selected pixel point is:
wherein L is the stain probability of the selected pixel point,for the gray value of the selected pixel, +.>For the gray value of the next pixel point of the selected pixel point, i is the serial number of each pixel point in the pixel point set corresponding to the selected pixel point, and N is the number of the pixel points in the pixel point set corresponding to the selected pixel point>For the curvature of the (i+1) th pixel point in the pixel point set corresponding to the selected pixel point,/L>For the curvature of the ith pixel point in the pixel point set corresponding to the selected pixel point,/and%>Gradient amplitude of the ith pixel point in the pixel point set corresponding to the selected pixel point,/for the selected pixel point>For the gradient amplitude mean value of all the pixel points in the pixel point set corresponding to the selected pixel point, < +. >For the absolute function, norm is the normalization function.
4. The method of claim 1, wherein determining each soil closing edge based on soil probabilities comprises:
for any one edge, if the stain probability of any one pixel point on the edge is larger than the stain threshold value, judging that the pixel point is a stain closed edge pixel point, otherwise, judging that the pixel point is not a stain closed edge pixel point; traversing all the stain closing edge pixel points in the blade surface gray level image by utilizing an edge tracking algorithm to obtain each stain closing edge to be selected;
judging whether the closed edges of any stains to be selected are overlapped or not through intersection judgment, and if so, marking the overlapped edges; when judging the internal and external relation between any pixel point on the closed edge of the stain to be selected and other closed edges of the stain to be selected by using a ray method, checking whether an intersection point is generated between the pixel point on the ray path and the marked coincident edge, and if the intersection point is generated, introducing rays in the direction instead of the intersection point until the intersection point does not exist between the pixel point and the coincident edge; counting the number of intersection points when the pixel points on the closed edge of the stain to be selected intersect with other closed edges of the stain to be selected, and if the number of intersection points is even, judging that the closed edge of the stain to be selected is the closed edge of the stain; and if the number of the intersection points is odd, discarding the closed edge of the stain to be selected.
5. The method of claim 1, wherein determining the first contamination level for each pixel point in each stain closure edge based on the stain probability and location for each pixel point in each stain closure edge comprises:
optionally selecting a pixel point on any stain closed edge as a target point, constructing a coordinate system by taking the target point as an origin, and making all straight lines of the origin; according to the coordinate position of each pixel point on the stain closing edge, calculating the distance between two intersection points where the straight line and the stain closing edge intersect, and obtaining a maximum distance value and a minimum distance value corresponding to the target point;
determining four axis-edge intersection points of two coordinate axes of the coordinate system and the stain closed edge, and calculating the distance between the target point and each axis-edge intersection point;
and determining the first pollution degree of the target point on the stain closed edge according to the maximum distance value and the minimum distance value corresponding to the target point, the distance between the target point and each axis edge intersection point and the stain probability of each axis edge intersection point.
6. The method for detecting an operational failure of a blower motor based on artificial intelligence according to claim 5, wherein the first pollution level is calculated by the formula:
In the method, in the process of the invention,for the first pollution level of the target point, j is the serial number of the intersection point of the axis and the side>For the distance between the target point and the intersection of the j-th axis edge,/>For the minimum distance value corresponding to the target point, +.>For the maximum distance value corresponding to the target point, +.>Is the dirty probability of the j-th axis edge intersection point.
7. The method of claim 1, wherein determining a second contamination level for each pixel point in each stain closure edge based on a gray scale run matrix comprises:
clustering the gray scale run matrix corresponding to each stain closing edge to obtain a gray scale range corresponding to a target cluster of each stain closing edge; the target cluster is the cluster with the largest run length;
selecting pixel points belonging to a gray level range corresponding to a target cluster in the stain closing edge as centers for any stain closing edge, and constructing a local window area with a preset size; and determining the local confusion degree of the central pixel point in the local window area according to the gray value of the central pixel point in the local window area, the gray average value of all the pixel points and the local texture characteristic value of the central pixel point, and determining the local confusion degree as the second pollution degree of the central pixel point.
8. The method for detecting an operation fault of a blower motor based on artificial intelligence according to claim 7, wherein determining the local confusion degree of the central pixel point in the local window area according to the gray value of the central pixel point in the local window area, the gray average value of all the pixel points and the local texture feature value of the central pixel point comprises:
calculating the ratio of the gray value of the central pixel point to the gray average value of all the pixel points, carrying out data processing on the ratio by utilizing a logarithmic function, and determining the absolute value of the ratio after the data processing as a first chaotic factor of the central pixel point; determining the local texture characteristic value of the central pixel point as a second chaotic factor; the product of the first clutter factor and the second clutter factor is determined as the local clutter level of the center pixel.
9. The method for detecting the operation fault of the blower motor based on artificial intelligence according to claim 1, wherein the new crack significance value is calculated by the following formula:
in the method, in the process of the invention,closing the edge into the stainCoordinates of->New crack significance values corresponding to pixel points of (1), x is an abscissa, y is an ordinate, and Norm is a normalization function; / >The inner coordinate of the closed edge for stains is +.>An exponential function of the pixel points of +.>The inner coordinate of the edge closed by the stain is +.>Whether the gray value of the pixel point of (2) is within the gray level range corresponding to the target cluster +.>Judging internally, if yes->Is a value of 1, if not, +.>Is a value of 0; e is a natural constant, < >>The inner coordinate of the closed edge for stains is +.>First contamination level of the pixel of +.>The inner coordinate of the closed edge for stains is +.>Second contamination level of the pixel of (2),/and (c)>The inner coordinate of the closed edge for stains is +.>The pixel points of (3) correspond to the original crack significance values.
10. The method for detecting an operational failure of a blower motor based on artificial intelligence according to claim 1, wherein determining whether a crack exists in a blade of the blower motor to be detected using a new crack significance value comprises:
substituting the new crack significant value into a significant value detection algorithm, and detecting the significant value of the blade surface gray level image; judging whether pixel points with the significant value larger than the defect threshold exist in the gray level image of the surface of the blade, if so, judging that cracks exist in the blade of the blower motor to be detected, and determining the pixel points with the significant value larger than the defect threshold as crack pixel points.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311040459.XA CN116758075B (en) | 2023-08-18 | 2023-08-18 | Artificial intelligence-based blower motor operation fault detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311040459.XA CN116758075B (en) | 2023-08-18 | 2023-08-18 | Artificial intelligence-based blower motor operation fault detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116758075A true CN116758075A (en) | 2023-09-15 |
| CN116758075B CN116758075B (en) | 2023-11-24 |
Family
ID=87959479
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311040459.XA Active CN116758075B (en) | 2023-08-18 | 2023-08-18 | Artificial intelligence-based blower motor operation fault detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116758075B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117011386A (en) * | 2023-09-27 | 2023-11-07 | 天津水科机电有限公司 | Pollution discharge effect evaluation method based on backwashing water filter |
| CN117315289A (en) * | 2023-11-28 | 2023-12-29 | 苏州翰微材料科技有限公司 | Aeroengine blade contour edge detection method based on image processing |
| CN117593294A (en) * | 2024-01-18 | 2024-02-23 | 中化学西部新材料有限公司 | Centrifugal fan blade crack visual detection method based on image feature analysis |
| CN117893533A (en) * | 2024-03-14 | 2024-04-16 | 自贡市第一人民医院 | Image feature-based heart-chest ratio intelligent detection method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006162477A (en) * | 2004-12-08 | 2006-06-22 | Ohbayashi Corp | Crack detecting method on structural surfaces |
| JP2018128309A (en) * | 2017-02-07 | 2018-08-16 | 大成建設株式会社 | Crack detection method |
| KR20200031860A (en) * | 2018-09-17 | 2020-03-25 | 윈디텍 주식회사 | System and method for managing safety of blade for wind power generator |
| CN112465776A (en) * | 2020-11-26 | 2021-03-09 | 常州信息职业技术学院 | Intelligent crack detection method based on fuzzy image on surface of wind turbine |
| CN113406091A (en) * | 2021-06-09 | 2021-09-17 | 东方电气集团科学技术研究院有限公司 | Unmanned aerial vehicle system for detecting fan blade and control method |
| CN115294119A (en) * | 2022-10-08 | 2022-11-04 | 江苏三通科技有限公司 | Machine vision-based method for detecting stains in inner groove of head of plum-blossom-shaped thread |
| US11580647B1 (en) * | 2021-12-24 | 2023-02-14 | Guangzhou University | Global and local binary pattern image crack segmentation method based on robot vision |
-
2023
- 2023-08-18 CN CN202311040459.XA patent/CN116758075B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006162477A (en) * | 2004-12-08 | 2006-06-22 | Ohbayashi Corp | Crack detecting method on structural surfaces |
| JP2018128309A (en) * | 2017-02-07 | 2018-08-16 | 大成建設株式会社 | Crack detection method |
| KR20200031860A (en) * | 2018-09-17 | 2020-03-25 | 윈디텍 주식회사 | System and method for managing safety of blade for wind power generator |
| CN112465776A (en) * | 2020-11-26 | 2021-03-09 | 常州信息职业技术学院 | Intelligent crack detection method based on fuzzy image on surface of wind turbine |
| CN113406091A (en) * | 2021-06-09 | 2021-09-17 | 东方电气集团科学技术研究院有限公司 | Unmanned aerial vehicle system for detecting fan blade and control method |
| US11580647B1 (en) * | 2021-12-24 | 2023-02-14 | Guangzhou University | Global and local binary pattern image crack segmentation method based on robot vision |
| CN115294119A (en) * | 2022-10-08 | 2022-11-04 | 江苏三通科技有限公司 | Machine vision-based method for detecting stains in inner groove of head of plum-blossom-shaped thread |
Non-Patent Citations (2)
| Title |
|---|
| ZHONGHE REN.ET AL: "State of the Art in Defect Detection Based on Machine Vision", SPRINGER, pages 661 * |
| 仇梓峰;王爽心;李蒙;: "基于无人机图像的风力发电机叶片缺陷识别", 发电技术, no. 03, pages 277 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117011386A (en) * | 2023-09-27 | 2023-11-07 | 天津水科机电有限公司 | Pollution discharge effect evaluation method based on backwashing water filter |
| CN117011386B (en) * | 2023-09-27 | 2024-01-26 | 天津水科机电有限公司 | Pollution discharge effect evaluation method based on backwashing water filter |
| CN117315289A (en) * | 2023-11-28 | 2023-12-29 | 苏州翰微材料科技有限公司 | Aeroengine blade contour edge detection method based on image processing |
| CN117315289B (en) * | 2023-11-28 | 2024-02-09 | 苏州翰微材料科技有限公司 | Aeroengine blade contour edge detection method based on image processing |
| CN117593294A (en) * | 2024-01-18 | 2024-02-23 | 中化学西部新材料有限公司 | Centrifugal fan blade crack visual detection method based on image feature analysis |
| CN117593294B (en) * | 2024-01-18 | 2024-04-09 | 中化学西部新材料有限公司 | Centrifugal fan blade crack visual detection method based on image feature analysis |
| CN117893533A (en) * | 2024-03-14 | 2024-04-16 | 自贡市第一人民医院 | Image feature-based heart-chest ratio intelligent detection method and system |
| CN117893533B (en) * | 2024-03-14 | 2024-05-28 | 自贡市第一人民医院 | Image feature-based heart-chest ratio intelligent detection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116758075B (en) | 2023-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116758075B (en) | Artificial intelligence-based blower motor operation fault detection method | |
| CN109141232B (en) | Online detection method for disc castings based on machine vision | |
| CN115082467B (en) | Building material welding surface defect detection method based on computer vision | |
| CN111383209B (en) | Unsupervised flaw detection method based on full convolution self-encoder network | |
| CN115082418A (en) | Precise identification method for automobile parts | |
| Neogi et al. | Defect detection of steel surfaces with global adaptive percentile thresholding of gradient image | |
| CN115222733A (en) | Metal component surface defect detection method based on image recognition | |
| CN107490582B (en) | Assembly line workpiece detection system | |
| CN111598801B (en) | Identification method for weak Mura defect | |
| CN111415339B (en) | Image defect detection method for complex texture industrial product | |
| CN113469951B (en) | Hub defect detection method based on cascade region convolutional neural network | |
| CN107991309A (en) | Product quality detection method, device and electronic equipment | |
| CN117011292B (en) | Method for rapidly detecting surface quality of composite board | |
| CN115272336A (en) | Metal part defect accurate detection method based on gradient vector | |
| CN114549441A (en) | Sucker defect detection method based on image processing | |
| CN116883408B (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
| CN115619708A (en) | Method for judging fault based on image recognition of oil quality change of main shaft bearing lubricating oil | |
| CN115272303A (en) | Textile fabric defect degree evaluation method, device and system based on Gaussian blur | |
| CN113516619A (en) | Product surface flaw identification method based on image processing technology | |
| CN111814852A (en) | Image detection method, image detection device, electronic equipment and computer-readable storage medium | |
| CN117274251B (en) | Tablet quality detection method in medicine production process based on image data | |
| CN114881961A (en) | Glass fiber plate crack detection method and device | |
| CN115797361B (en) | Aluminum template surface defect detection method | |
| CN116612112A (en) | Visual inspection method for surface defects of bucket | |
| CN116542963A (en) | Float glass defect detection system and detection method based on machine learning |
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 |