CN114842011A - Bearing wear detection method and system based on image processing - Google Patents
Bearing wear detection method and system based on image processing Download PDFInfo
- Publication number
- CN114842011A CN114842011A CN202210776652.9A CN202210776652A CN114842011A CN 114842011 A CN114842011 A CN 114842011A CN 202210776652 A CN202210776652 A CN 202210776652A CN 114842011 A CN114842011 A CN 114842011A
- Authority
- CN
- China
- Prior art keywords
- image
- wear
- image block
- template
- defect
- 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 30
- 238000012545 processing Methods 0.000 title claims abstract description 24
- 230000007547 defect Effects 0.000 claims abstract description 109
- 238000009826 distribution Methods 0.000 claims description 25
- 238000000034 method Methods 0.000 claims description 18
- 230000007797 corrosion Effects 0.000 claims description 16
- 238000005260 corrosion Methods 0.000 claims description 16
- 238000005299 abrasion Methods 0.000 claims description 10
- 230000008859 change Effects 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 241000287196 Asthenes Species 0.000 description 2
- 230000009471 action Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000037380 skin damage Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
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
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, in particular to a bearing wear detection method and a bearing wear detection system based on image processing, wherein the detection method comprises the following steps: acquiring a surface image to be measured of a bearing wear surface, wherein the bearing wear surface is a side surface worn by a bearing; traversing and matching the surface image to be detected by utilizing the template image block to obtain a similarity sequence, wherein the area of the surface image to be detected, which corresponds to the similarity greater than a preset similarity threshold value in the similarity sequence, is a wear area; the template image block is a pre-generated image block, and the scale of the template image block is smaller than that of the surface image to be detected; the step of acquiring the template image block comprises the steps of randomly acquiring a wear surface image with bearing wear defects according to history, extracting a defect area in the wear surface image, wherein the defect area is the template image block, and the problem of influence of an irrelevant defect area on defect identification precision caused by using an integral template in the traditional template matching is solved.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a bearing wear detection method and system based on image processing.
Background
At present, most of electric appliances comprise motors, such as an engine, a generator and the like, the motors are the most prone to failure due to long-term high-speed operation, and the main shaft failure is mostly caused by bearing failure. Therefore, can carry out corresponding maintenance to the bearing when main shaft motor breaks down, disassemble main shaft motor and overhaul the trouble one by one, and the wearing and tearing that detect the bearing rely on artifical naked eye to detect mostly, detection efficiency is lower and cause the false detection easily, especially to defects such as the skin damage broken string of stator coil. Patent publication No. CN110246122A proposes that a similarity measurement is performed on a captured bearing image through a template image in an image database, and a defect region is identified by setting a defect threshold.
In practice, the inventors found that the above prior art has the following disadvantages:
the whole bearing image is used as a template image for similarity calculation, compared background features of the whole image have large image matching calculation amount, high requirement on defect threshold precision, sensitive response to noise points and poor detection effect on local small defects.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a bearing wear detection method and system based on image processing, wherein the adopted technical scheme is as follows:
in a first aspect, an embodiment of the present invention provides a bearing wear detection method based on image processing, where the detection method includes:
acquiring a surface image to be measured of a bearing wear surface, wherein the bearing wear surface is a side surface worn by a bearing;
traversing and matching the surface image to be detected by utilizing a template image block to obtain a similarity sequence, wherein the area of the surface image to be detected, which corresponds to the similarity greater than a preset similarity threshold value in the similarity sequence, is a wear area;
the template image block is a pre-generated image block, and the scale of the template image block is smaller than that of the surface image to be detected; the step of obtaining the template image block comprises the steps of extracting a defect area in a wear surface image according to the wear surface image with bearing wear defects which is collected randomly according to history, wherein the defect area is the template image block.
Further, the detection method comprises the step of optimizing the template image block: and obtaining a wear index according to the weighted summation of the texture feature and the color feature of the defect region, and screening the defect region according to the wear index to obtain the optimized template image block.
Further, the step of obtaining the texture features of the defect area includes: traversing each pixel point in the defect region to obtain a first distribution probability of the pixel points with the same gray scale feature, and obtaining a texture information entropy of the defect region according to the first distribution probability of each pixel point, wherein the texture information entropy is the texture feature; the gray feature is that the pixel point is taken as a central pixel point, and a domain pixel point of the central pixel point is obtained; forming a first binary group by the central pixel point and the neighborhood pixel points; and obtaining the first distribution probability according to the occurrence times of the first binary group.
Further, the step of obtaining the color characteristics of the defect area comprises: converting the defect area into an HSV color space to obtain an HSV image, traversing each pixel point of the defect area in the HSV image to obtain a second distribution probability of pixel points with the same color characteristic, and obtaining a color information entropy of the defect area according to the second distribution probability of each pixel point, wherein the color information entropy is the color characteristic; forming a second binary group by using hue information and brightness information of each pixel point in the defect area of the HSV image; and obtaining the second distribution probability according to the occurrence times of the second binary group.
Further, after the step of screening the defect area according to the wear indicator to obtain the template image block, the method further comprises an optimization step of: and carrying out corrosion operation on the template image block for multiple times, stopping corrosion when the abrasion index changes after each corrosion operation, and selecting a corrosion image before the current corrosion operation as the optimized template image block.
Further, the step of extracting a defect region in the wear surface image comprises: and performing semantic segmentation on the wear surface image to obtain the defect area.
Further, the step of performing traversal matching on the surface image to be detected by using the template image block to obtain a similarity sequence further comprises an optimization matching step: and adaptively adjusting the matching step length according to the change trend of the similarity.
Further, the step of performing traversal matching on the surface image to be detected by using the template image block to obtain a similarity sequence further comprises an optimization matching step: and increasing the matching step length when the change trends of a plurality of continuous similarity are the same.
Further, the method for obtaining the similarity sequence by traversing and matching the surface image to be detected by using the template image block adopts a normalized cross-correlation algorithm to obtain the similarity between the template image block and each matching block in the surface image to be detected.
In a second aspect, an embodiment of the present invention provides an image processing-based bearing wear detection system, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of any one of the above methods when executing the computer program.
The invention has the following beneficial effects:
according to the embodiment of the invention, the surface image to be detected of the bearing wear surface is acquired, the template image block is utilized to perform traversal matching on the surface image to be detected to obtain the similarity sequence, the area of the surface image to be detected corresponding to the similarity which is greater than the preset threshold in the similarity sequence is the wear area, a standard image library is not required to be constructed, the template matching is performed on the acquired surface image to be detected through the template image block, the defect identification precision is improved, and the influence of an irrelevant defect area on the defect identification precision caused by the utilization of an integral template in the traditional template matching is avoided. The template image blocks obtained by using the historically collected defect images can generate defect template image blocks in a self-adaptive manner according to the defect characteristics of the actual defect images, and meanwhile, the generalization capability of the system can be improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a bearing wear detection method based on image processing according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description will be given to a method and a system for detecting bearing wear based on image processing according to the present invention, with reference to the accompanying drawings and preferred embodiments, and the detailed implementation, structure, features and effects thereof are described below. In the following description, the different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The motor comprises a stator, a rotor and a bearing for driving the rotor to rotate, the bearing precision is reduced due to other reasons in the rotating process of the bearing, the rotor can abrade the surface of the stator under the driving action of the bearing, and obvious abrasion traces can appear on the surface of the stator. For the spindle motor to be overhauled, after disassembly, a surface image of the stator is collected and analyzed in a machine vision mode, and whether the surface image is a defect or not is judged according to the characteristics of the image.
The following describes a specific scheme of a bearing wear detection method and system based on image processing in detail with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a bearing wear detection method based on image processing according to an embodiment of the present invention is shown, where the method includes the following steps:
s001, acquiring a surface image to be measured of a bearing wear surface, wherein the bearing wear surface is a side surface worn by a bearing;
a camera is arranged at an opening of the bearing, the position and posture of the camera can be guaranteed to acquire images of an easily worn area at the edge of the bearing, and a stable light source and a fixed camera position and posture are adopted to acquire images of different bearings.
S002, traversing and matching the surface image to be detected by using the template image block to obtain a similarity sequence, wherein the area of the surface image to be detected, which corresponds to the similarity greater than a preset threshold value in the similarity sequence, is a wear area; the template image block is a pre-generated image block, and the scale of the template image block is smaller than that of the surface image to be detected; the step of obtaining the template image block comprises the steps of randomly collecting a wear surface image with bearing wear defects according to history, and extracting a defect area in the wear surface image, wherein the defect area is the template image block.
In order to extract more accurate defect characteristics, a large number of wear surface images with bearing wear defects are collected, defect areas in the wear surface images are extracted, and the purpose of extracting the characteristics more accurately is achieved by extracting characteristic information of the defect areas.
Wherein the step of extracting the defect region in the wear surface image comprises: and performing semantic segmentation on the wear surface image to obtain a defect area. Specifically, before semantic segmentation, preprocessing is performed on an acquired wear surface image, wherein the preprocessing includes performing projection transformation on the wear surface image to obtain a planar image of the wear surface image, and performing graying processing on the planar image to obtain a grayscale image. Because the image shot by the camera has the characteristic of big-end-up and small-end-up, the image of the worn surface is projected and transformed into a front-view plane image so as to eliminate the phenomenon of big-end-up and small-end-up. Taking gray-scale images corresponding to all wear surface images acquired in history as training set images, and marking the wear defects in the gray-scale images: and marking the wear defect pixel point as 1 and other pixel points as 0. And inputting the marked gray level image into a semantic segmentation network for image convolution to extract a feature vector of the gray level image, and obtaining a defect segmentation image of the gray level image through deconvolution. The semantic segmentation network adopts an Encoder-Decoder structure, and the loss function adopts a cross entropy loss function.
And extracting corresponding defect images in the surface image by taking the defect segmentation image as a mask, taking the defect images corresponding to the surface image as template image blocks, and identifying the defect characteristics of the surface image to be detected by utilizing the template image blocks.
In the embodiment of the present invention, a Normalized cross-correlation (NCC) algorithm is used to obtain the similarity between the template image block and each matching block in the surface image to be measured. The NNC algorithm is a commonly used algorithm in template matching, and in other embodiments, other template matching algorithms capable of achieving the same function may also be used. In the embodiment of the invention, the step length of traversing and matching the surface image to be measured by using the over-template image block by using the NCC algorithm is 3.
Preferably, in order to achieve the purpose of accurately matching and identifying the local defects of the bearing, the template image block is obtained by performing multi-local-area fusion according to the defect characteristics of each local defect area. Specifically, the step of optimizing the template image block includes: and obtaining a wear index according to the weighted summation of the texture feature and the color feature of the defect region, and screening the defect region according to the wear index to obtain the optimized template image block.
The method comprises the following steps of obtaining the texture features of the defect area: traversing each pixel point in the defect region to obtain a first distribution probability of the pixel points with the same gray scale feature, and obtaining a texture information entropy of the defect region according to the first distribution probability of each pixel point, wherein the texture information entropy is a texture feature; the gray feature is that a field pixel point of a central pixel point is obtained by taking the pixel point as the central pixel point; forming a first binary group by the central pixel point and the neighborhood pixel point; and obtaining a first distribution probability according to the occurrence times of the first binary group.
Specifically, the method comprises the following steps: and traversing each pixel point in the defect area, taking the traversed pixel point as a central pixel point, and calculating the gray average value of eight neighborhood pixel points of the central pixel point. The first binary group composed of the gray average of the central pixel and the domain pixel is recorded asWherein, in the step (A),is shown asThe gray value of each of the center pixel points,is shown asAnd the gray average value of the eight neighborhood pixels of each central pixel. Counting frequency doublet corresponding to the occurrence frequency of the first doublet in the current defect regionAccording to the total number of the pixel points contained in the defect area, the distribution probability of different binary groups in the wear area is counted,Is shown asThe distribution probability of the frequency doublet corresponding to the first doublet,is shown asThe frequency of the first binary group corresponding to the frequency bin in the defect area,the subscript that represents the pixel point is,expressing the number of pixel points of the defect area, and obtaining the texture information entropy of the defect area according to the distribution probability of the first binary group. The larger the texture information entropy is, the more complex the texture of the defect region is, the more gray-level pixel points exist, and the larger the corresponding wear degree of the region is.
The step of acquiring the color characteristics of the defect area comprises the following steps: converting the defect area into an HSV color space to obtain an HSV image, traversing each pixel point of the defect area in the HSV image to obtain a second distribution probability of the pixel points with the same color characteristic, and obtaining a color information entropy of the defect area according to the second distribution probability of each pixel point, wherein the color information entropy is the color characteristic; forming a second binary group by using hue information and brightness information of each pixel point in a defect area of the HSV image; and obtaining a second distribution probability according to the occurrence times of the second binary group.
Specifically, the gray scale inverse transformation is carried out on the defect area image to obtain an RGB image, and HSV color space transformation is carried out on the RGB image to obtain an HSV image. Each pixel point of the corresponding wear area has three channel values of hue (H), saturation (S) and lightness (V), and the bearing wear characteristics comprise: the surface of the bearing has red rust spots, a part of smooth area generated by abrasion is slightly abraded, a missing surface generated by abrasion is seriously abraded, the abrasion characteristic of the bearing can show that the surface of the bearing abrasion area can present various hue values and has lightness differences of different degrees, and hue information and lightness information of each pixel point in the defect area of the HSV image form a second binary groupIs marked asWherein, in the step (A),is shown asThe hue of each of the pixels is determined,is shown asThe brightness of each pixel. Counting the frequency binary group corresponding to the occurrence frequency of the second binary group in the current defect areaAccording to the total number of the pixel points contained in the defect area, the distribution probability of different second tuples in the wear area is counted,Is shown asThe distribution probability of the frequency doublet corresponding to the second doublet,is shown asThe frequency bin corresponding to the second bin is the frequency bin of the defect area,representing pixel pointsThe subscript of (a) is,the number of pixels representing the defective area. Obtaining the color information entropy of the defect area according to the spatial distribution characteristics and the color richness of the color. The larger the entropy of the color information is, the more complicated the color of the defect area is, the more likely there is combined wear of a plurality of wear degrees, and the larger the wear degree of the corresponding curve area is.
Further, since the texture information entropy may only reflect the gray distribution change of the surface under different wear degrees, the wear at different levels can be evaluated more accurately by means of the color complexity. Thus, combining the two characteristics to obtain a wear indicator, the wear indicator is labeled asThen, there are:
wherein the content of the first and second substances,is the weight of the texture feature and is,is the weight of the color feature.
Screening all defect regions according to the wear index to obtain all defectsHaving greatest wear index in the regionAnd taking the corresponding abrasion area as an optimized template image, and matching the surface image to be detected by using the optimized template image.
Preferably, because the surface of the stator has uniformly distributed grains, the groove part of the grains is not easy to wear. In order to remove the evaluation influence of the correlation region between the adjacent defect regions, the precision of subsequent template matching is improved. After the step of screening the defect area according to the wear index to obtain the template image block, the method also comprises an optimization step: and carrying out corrosion operation on the template image block for multiple times, stopping corrosion when the abrasion index changes after each corrosion operation, and selecting a corrosion image before the current corrosion operation as the optimized template image block. And when the abrasion index is not changed after each corrosion operation, continuously corroding for at most four times, and taking the corroded image as a final optimized template image.
Preferably, in order to improve the matching efficiency of the template and realize the rapid defect detection of the surface image to be detected, the step of performing traversal matching on the surface image to be detected by using the template image block to obtain a similarity sequence further comprises the step of optimizing matching: and the matching step length is adjusted in a self-adaptive mode according to the change trend of the similarity.
Specifically, based on the characteristics of bearing wear, when a certain area is worn, the adjacent areas are worn to different degrees, so that a defect area can be an aggregation area; accordingly, the adjacent area of the normal area is not easily worn. In order to improve the speed of template matching, in the process of template image block matching, when the change trends of a plurality of continuous similarity degrees are the same, the step length of matching is increased. Specifically, according to the similarity between the template image block and the corresponding region of the surface image to be measuredThe size of the step length is self-adaptively adjusted and matched; when traversed in the same direction, the phases are continuousChanging the step length of traversal in the process of continuously increasing the similarity obtained by the image blocks of the adjacent template, and continuing to perform matching; and when the image similarity is smaller than the step size adjusting threshold value, increasing the step size. Specifically, the adjusted step length is recorded asThe regulated amount is recorded asThen adjusted step lengthWherein, in the embodiments of the present inventionThe step size adjustment threshold is 0.1.
In order to ensure the matching precision of the template image blocks, rolling back the matching step length of the template image blocks according to the change trend of the similarity; when the similarity corresponding to the continuous adjacent template image blocks meets the process of continuously increasing again, the template image blocks roll back by one step length, the matching of the template image blocks is carried out again, and the initial step length is recovered。
When the template image block is matched, the area corresponding to the template image block in the plane image is marked as a defect area when the similarity corresponding to the template image block is greater than the similarity threshold value, so that the defect area is quickly detected.
In the embodiment of the present invention, the value of the similarity threshold is 0.5, and in other embodiments, the similarity threshold may be set as needed.
And defect image analysis and feature extraction are not required to be carried out on all bearing images, and rapid defect detection can be realized by utilizing a minimized optimization template.
To sum up, in the embodiment of the present invention, the surface image to be detected of the bearing wear surface is acquired, the template image block is used to perform traversal matching on the surface image to be detected to obtain the similarity sequence, the region of the surface image to be detected corresponding to the similarity greater than the preset threshold in the similarity sequence is the wear region, a standard image library is not required to be constructed, the template matching is performed on the acquired surface image to be detected through the template image block, the defect identification precision is improved, and the influence of an irrelevant defect region on the defect identification precision caused by using an integral template in the conventional template matching is avoided. The template image blocks obtained by using the historically collected defect images can generate defect template image blocks in a self-adaptive manner according to the defect characteristics of the actual defect images, and meanwhile, the generalization capability of the system can be improved.
Based on the same inventive concept as the method embodiment, the embodiment of the present invention further provides an image processing-based bearing wear detection system, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the image processing-based bearing wear detection method according to any one of the above embodiments when executing the computer program. A bearing wear detection method based on image processing has been described in detail in the above embodiments, and is not described again.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (8)
1. A bearing wear detection method based on image processing is characterized by comprising the following steps:
acquiring a surface image to be measured of a bearing wear surface, wherein the bearing wear surface is a side surface worn by a bearing;
traversing and matching the surface image to be detected by utilizing a template image block to obtain a similarity sequence, wherein the area of the surface image to be detected, which corresponds to the similarity greater than a preset similarity threshold value in the similarity sequence, is a wear area;
the template image block is a pre-generated image block, and the scale of the template image block is smaller than that of the surface image to be detected; the acquiring step of the template image block comprises the steps of extracting a defect area in a wear surface image according to the wear surface image with bearing wear defects which is randomly acquired according to history, wherein the defect area is the template image block;
the detection method comprises the following steps of optimizing the template image block:
obtaining a wear index according to the weighted summation of the texture feature and the color feature of the defect region, and screening the defect region according to the wear index to obtain the optimized template image block;
the step of acquiring the color characteristics of the defect area comprises the following steps:
converting the defect area into an HSV color space to obtain an HSV image, traversing each pixel point of the defect area in the HSV image to obtain a second distribution probability of pixel points with the same color characteristic, and obtaining a color information entropy of the defect area according to the second distribution probability of each pixel point, wherein the color information entropy is the color characteristic;
forming a second binary group by using hue information and brightness information of each pixel point in the defect area of the HSV image; and obtaining the second distribution probability according to the occurrence times of the second binary group.
2. The image processing-based bearing wear detection method according to claim 1, wherein the step of obtaining the texture features of the defect region comprises:
traversing each pixel point in the defect area to obtain a first distribution probability of the pixel points with the same gray scale feature, and obtaining a texture information entropy of the defect area according to the first distribution probability of each pixel point, wherein the texture information entropy is the texture feature;
the gray feature is that the pixel point is taken as a central pixel point, and a domain pixel point of the central pixel point is obtained; forming a first binary group by the central pixel point and the neighborhood pixel points; and obtaining the first distribution probability according to the occurrence times of the first binary group.
3. The method according to claim 1, wherein after the step of screening the defect area according to the wear indicator to obtain the template image block, the method further comprises an optimization step of:
and carrying out corrosion operation on the template image block for multiple times, stopping corrosion when the abrasion index changes after each corrosion operation, and selecting a corrosion image before the current corrosion operation as the optimized template image block.
4. The image processing-based bearing wear detection method according to claim 1, wherein the step of extracting the defect region in the wear surface image comprises: and performing semantic segmentation on the wear surface image to obtain the defect area.
5. The image processing-based bearing wear detection method according to claim 1, wherein the step of performing traversal matching on the surface image to be detected by using the template image block to obtain the similarity sequence further comprises an optimization matching step: and adaptively adjusting the matching step length according to the change trend of the similarity.
6. The image processing-based bearing wear detection method according to claim 1, wherein the step of performing traversal matching on the surface image to be detected by using the template image block to obtain the similarity sequence further comprises an optimization matching step: and increasing the matching step length when the change trends of a plurality of continuous similarity are the same.
7. The method for detecting bearing wear based on image processing as claimed in claim 1, wherein the method for obtaining the similarity sequence by traversing and matching the surface image to be detected with the template image block employs a normalized cross-correlation algorithm to obtain the similarity between the template image block and each matching block in the surface image to be detected.
8. An image processing based bearing wear detection system comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor when executing the computer program implementing the steps of the image processing based bearing wear detection method according to any one of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210776652.9A CN114842011B (en) | 2022-07-04 | 2022-07-04 | Bearing wear detection method and system based on image processing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210776652.9A CN114842011B (en) | 2022-07-04 | 2022-07-04 | Bearing wear detection method and system based on image processing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114842011A true CN114842011A (en) | 2022-08-02 |
CN114842011B CN114842011B (en) | 2022-09-09 |
Family
ID=82574809
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210776652.9A Active CN114842011B (en) | 2022-07-04 | 2022-07-04 | Bearing wear detection method and system based on image processing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114842011B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115063620A (en) * | 2022-08-19 | 2022-09-16 | 启东市海信机械有限公司 | Bit layering-based Roots blower bearing wear detection method |
CN115082466A (en) * | 2022-08-22 | 2022-09-20 | 江苏庆慈机械制造有限公司 | PCB surface welding spot defect detection method and system |
CN116152249A (en) * | 2023-04-20 | 2023-05-23 | 济宁立德印务有限公司 | Intelligent digital printing quality detection method |
CN117664984A (en) * | 2023-12-01 | 2024-03-08 | 上海宝柏新材料股份有限公司 | Defect detection method, device, system and storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140193065A1 (en) * | 2013-01-09 | 2014-07-10 | Kla-Tencor Corporation | Detecting Defects on a Wafer Using Template Image Matching |
CN106952257A (en) * | 2017-03-21 | 2017-07-14 | 南京大学 | A kind of curved surface label open defect detection method based on template matches and Similarity Measure |
CN111079556A (en) * | 2019-11-25 | 2020-04-28 | 航天时代飞鸿技术有限公司 | Multi-temporal unmanned aerial vehicle video image change area detection and classification method |
CN113538429A (en) * | 2021-09-16 | 2021-10-22 | 海门市创睿机械有限公司 | Mechanical part surface defect detection method based on image processing |
CN113538433A (en) * | 2021-09-17 | 2021-10-22 | 海门市创睿机械有限公司 | Mechanical casting defect detection method and system based on artificial intelligence |
-
2022
- 2022-07-04 CN CN202210776652.9A patent/CN114842011B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140193065A1 (en) * | 2013-01-09 | 2014-07-10 | Kla-Tencor Corporation | Detecting Defects on a Wafer Using Template Image Matching |
CN106952257A (en) * | 2017-03-21 | 2017-07-14 | 南京大学 | A kind of curved surface label open defect detection method based on template matches and Similarity Measure |
CN111079556A (en) * | 2019-11-25 | 2020-04-28 | 航天时代飞鸿技术有限公司 | Multi-temporal unmanned aerial vehicle video image change area detection and classification method |
CN113538429A (en) * | 2021-09-16 | 2021-10-22 | 海门市创睿机械有限公司 | Mechanical part surface defect detection method based on image processing |
CN113538433A (en) * | 2021-09-17 | 2021-10-22 | 海门市创睿机械有限公司 | Mechanical casting defect detection method and system based on artificial intelligence |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115063620A (en) * | 2022-08-19 | 2022-09-16 | 启东市海信机械有限公司 | Bit layering-based Roots blower bearing wear detection method |
CN115063620B (en) * | 2022-08-19 | 2023-11-28 | 启东市海信机械有限公司 | Bit layering based Roots blower bearing wear detection method |
CN115082466A (en) * | 2022-08-22 | 2022-09-20 | 江苏庆慈机械制造有限公司 | PCB surface welding spot defect detection method and system |
CN115082466B (en) * | 2022-08-22 | 2023-09-01 | 倍利得电子科技(深圳)有限公司 | PCB surface welding spot defect detection method and system |
CN116152249A (en) * | 2023-04-20 | 2023-05-23 | 济宁立德印务有限公司 | Intelligent digital printing quality detection method |
CN117664984A (en) * | 2023-12-01 | 2024-03-08 | 上海宝柏新材料股份有限公司 | Defect detection method, device, system and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN114842011B (en) | 2022-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114842011B (en) | Bearing wear detection method and system based on image processing | |
US10963670B2 (en) | Object change detection and measurement using digital fingerprints | |
CN114937055B (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
CN109596634B (en) | Cable defect detection method and device, storage medium and processor | |
CN110263192B (en) | Abrasive particle morphology database creation method for generating countermeasure network based on conditions | |
CN109598287B (en) | Appearance flaw detection method for resisting network sample generation based on deep convolution generation | |
CN116205919B (en) | Hardware part production quality detection method and system based on artificial intelligence | |
CN113592845A (en) | Defect detection method and device for battery coating and storage medium | |
Chen et al. | Is overfeat useful for image-based surface defect classification tasks? | |
CN110070531B (en) | Model training method for detecting fundus picture, and fundus picture detection method and device | |
CN111968098A (en) | Strip steel surface defect detection method, device and equipment | |
CN109740572A (en) | A kind of human face in-vivo detection method based on partial color textural characteristics | |
CN109087330A (en) | It is a kind of based on by slightly to the moving target detecting method of smart image segmentation | |
CN115375690B (en) | Classification and identification method for greasy tongue coating | |
CN113743421B (en) | Method for segmenting and quantitatively analyzing anthocyanin developing area of rice leaf | |
CN116385435B (en) | Pharmaceutical capsule counting method based on image segmentation | |
CN111707672A (en) | Method for detecting surface defects of wind power rotary supporting piece | |
He et al. | CBAM‐YOLOv5: A promising network model for wear particle recognition | |
CN111242898A (en) | Train pantograph abrasion detection method and system based on deep neural network | |
CN115855961A (en) | Distribution box fault detection method used in operation | |
KR20120040004A (en) | System for color clustering based on tensor voting and method therefor | |
García et al. | Pollen grains contour analysis on verification approach | |
Fu et al. | On the quality and diversity of synthetic face data and its relation to the generator training data | |
Papenberg et al. | Visualization of relevant areas of milling tools for the classification of tool wear by machine learning methods | |
de Oliveira et al. | Defect inspection in stator windings of induction motors based on convolutional neural networks |
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 |