CN114882387B - Bearing raceway bruise identification and automatic polishing positioning method in grinding process - Google Patents
Bearing raceway bruise identification and automatic polishing positioning method in grinding process Download PDFInfo
- Publication number
- CN114882387B CN114882387B CN202210811778.5A CN202210811778A CN114882387B CN 114882387 B CN114882387 B CN 114882387B CN 202210811778 A CN202210811778 A CN 202210811778A CN 114882387 B CN114882387 B CN 114882387B
- Authority
- CN
- China
- Prior art keywords
- area
- target
- region
- group
- gray
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention relates to the technical field of data processing, in particular to a bearing raceway bruise identification and automatic polishing positioning method in a grinding process. The method is a method for identifying by using electronic equipment, and polishing positioning is completed by using an artificial intelligence system in the production field. Firstly, acquiring a target bearing image by using a camera; and carrying out data processing on the target bearing image to obtain an initial defect area group, and further obtaining a target defect area from the initial defect area group. According to the invention, the target defect area is obtained after data processing is carried out on the target bearing image acquired by the camera, so that the defect areas belonging to the same collision injury can be obtained under the condition that the target bearing image is not uniformly distributed, and polishing positioning is completed.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a bearing raceway bruise identification and automatic polishing positioning method in a grinding process.
Background
The bearing is a part used for determining the relative motion position of the rotating shaft and other parts and has a supporting or guiding function, and the main function of the bearing is to support the rotating shaft or other moving bodies, guide the rotation or movement and bear the load transmitted by the shaft or parts on the shaft. The bearing is easy to cause the collision damage of the inner ring raceway in the grinding process, and the bearing with the collision damage defect is easy to cause problems in the use process of the hydraulic press, so that serious accidents can be caused.
At present, the method for identifying the defects of the bearing is to identify the defects of the bearing through threshold segmentation and boundary identification technologies. However, since the bearing belongs to a metal product and is greatly influenced by illumination, a defect can be divided into different areas, and the defect identification is inaccurate.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a bearing raceway bruise identification and automatic polishing positioning method in a grinding process, and the adopted technical scheme is as follows:
acquiring an initial bearing image, and identifying a bearing in the initial bearing image to obtain a target bearing image;
performing multi-threshold segmentation on the target bearing image based on the gray value to obtain a plurality of first area groups; matching the first region group according to the distance of the central point and the overlapping area of the regions in the first region group to obtain a plurality of initial region groups;
segmenting the target bearing image to obtain a background area and corresponding background pixel points; screening the initial area group to obtain a plurality of target area groups according to the number proportion of background pixel points in the initial area group; connecting the areas in each target area group to obtain a plurality of characteristic areas;
classifying a plurality of regions of the first region group in the characteristic region according to the area of the region to obtain a second region group; acquiring the gray value, the gray average value and the maximum gray value of each region in the second region group, and constructing a gray difference sequence according to the difference of the gray values of adjacent regions; acquiring the conversion times of positive and negative values in the gray difference value sequence, obtaining the defect probability according to the conversion times, the gray mean value and the maximum gray value, and screening an initial defect area group from a second area group;
and obtaining a target defect area according to the gray scale abnormal condition and the direction abnormal condition of the initial defect area group.
Preferably, the multi-threshold segmentation of the target bearing image to obtain a plurality of first region groups includes:
and performing multi-threshold segmentation on the target bearing image to obtain a plurality of different gray levels, and taking the region with the same gray level as a first region group with the same gray level.
Preferably, the matching the first region group according to the distance between the center points and the overlapping area of the minimum circumscribed rectangle to obtain a plurality of initial region groups includes:
based on the distance between the central points of the areas in the first area groups, matching the areas in different first area groups by using a KM matching algorithm, and taking the obtained maximum KM matching value as the central point matching rate of the two first area groups;
acquiring the minimum circumscribed rectangle of each first region group, and acquiring the overlapping area of the minimum circumscribed rectangles among the first region groups as the range matching rate;
and dividing the first area group with the central point matching rate and the range matching rate both greater than a preset matching rate into the same initial area group.
Preferably, the screening the initial regional groups to obtain a plurality of target regional groups according to the number ratio of the background pixel points in the initial regional groups includes:
and taking the initial area group with the number ratio of the background pixel points smaller than a preset ratio threshold as a target area group.
Preferably, the connecting the regions in each target region group to obtain a plurality of feature regions includes:
taking the central point of the target area group as an origin point, and making rays in different directions to obtain an outermost layer intersection point of the area in the target area group, wherein the outermost layer intersection point is used as a connection point;
and connecting the connection points according to a Bessel region connection principle to obtain a plurality of characteristic regions.
Preferably, the defect probability is calculated by the following formula:
wherein the content of the first and second substances,
is the defect probability;
the number of times of conversion is;
is the maximum gray value;
is the mean value of the gray levels.
Preferably, the obtaining of the target defect area according to the abnormal gray level condition and the abnormal main direction condition of the initial defect area includes:
performing threshold segmentation on the characteristic region by utilizing an Otsu threshold segmentation method to obtain a segmentation threshold, wherein pixel points with gray values smaller than the segmentation threshold are used as target pixel points, and an initial defect region group with target pixel points with a ratio larger than a preset pixel ratio is used as a second defect region group;
acquiring principal component direction values of all regions on the target bearing image, and constructing a principal component direction value sequence; selecting the principal component direction value with the largest frequency in the principal component direction value sequence as a texture direction value;
and acquiring a second principal component direction of the second defect area group, wherein the second defect area group with the difference value between the second principal component direction and the texture direction value larger than the preset principal component direction difference value is the target defect area.
The embodiment of the invention at least has the following beneficial effects:
the embodiment of the invention utilizes a data processing technology, and the method is a method for identifying by using electronic equipment and finishes polishing and positioning by utilizing an artificial intelligence system in the production field. Firstly, acquiring a target bearing image; performing multi-threshold segmentation on the target bearing image based on the gray value to obtain a plurality of first area groups; matching the first region group according to the central point distance and the overlapping area of the regions in the first region group to obtain a plurality of initial region groups; segmenting the target bearing image to obtain background pixel points; screening the initial regional groups according to the number proportion of the background pixels in the initial regional groups to obtain a plurality of target regional groups; the areas in the target area groups are connected to obtain a plurality of characteristic areas, the characteristic areas are provided to meet the area division of human vision, namely, pixel points with larger gray values or smaller gray values exist in the divided characteristic areas at the same time, and the defect areas belonging to the same defect are divided into the same characteristic area under the condition that the color distribution of the target bearing image collected under the influence of illumination is not uniform. Classifying a plurality of regions of the first region group in the characteristic region according to the area of the region to obtain a second region group; acquiring the gray value, the gray average value and the maximum gray value of each area in the second area group, and constructing a gray difference sequence according to the difference value of the gray values of adjacent areas; acquiring the conversion times of positive and negative values in the gray difference value sequence, obtaining the defect probability according to the conversion times, the gray mean value and the maximum gray value, and screening an initial defect area group from the second area group; and obtaining the target defect area according to the gray scale abnormal condition and the direction abnormal condition of the initial defect area group. According to the invention, the target defect area is obtained after data processing is carried out on the target bearing image acquired by the camera, so that the defect areas belonging to the same collision injury can be obtained under the condition that the target bearing image is not uniformly distributed, and polishing positioning is completed.
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 method for identifying bearing raceway bruise and automatically polishing and positioning in a grinding process 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 the method for identifying bearing race impact and automatically polishing and positioning in grinding process according to the present invention, with reference to the accompanying drawings and preferred embodiments, and its specific implementation, structure, features and effects. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily 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 embodiment of the invention provides a specific implementation method of a bearing raceway bruise identification and automatic polishing positioning method in a grinding process, which is suitable for bearing bruise identification and polishing positioning scenes. Initial bearing images are acquired in the scene by means of an RGB camera looking down. The defect identification method aims to solve the problem that due to the fact that the influence of illumination is large, a defect is divided into different areas, and therefore defect identification is inaccurate. The embodiment of the invention discloses a method for identifying by using electronic equipment, which finishes polishing positioning by using an artificial intelligence system in the production field, particularly divides an area with larger pixel attribute characteristic difference and a shorter distance into the same characteristic area to avoid the problem of dividing defects into different areas, further screens the defect area for multiple times, obtains the target defect area by processing data of a target bearing image acquired by a camera, realizes that the defect areas belonging to the same bruise are obtained under the condition of uneven distribution of the target bearing image, and finishes polishing positioning.
The following describes a specific scheme of a bearing raceway bruise identification and automatic polishing positioning method in a grinding process in detail by combining with the accompanying drawings.
Referring to fig. 1, a flow chart of steps of a method for identifying bearing raceway bruise and automatically polishing and positioning in a grinding process according to an embodiment of the present invention is shown, the method includes the following steps:
and S100, acquiring an initial bearing image, and identifying a bearing in the initial bearing image to obtain a target bearing image.
And (3) utilizing an RGB camera to overlook and acquire an initial bearing image, wherein the initial bearing image is an RGB image.
And identifying and segmenting the bearing in the initial bearing image by utilizing a DNN semantic segmentation network. Specifically, the method comprises the following steps: the data set of the DNN semantic segmentation network is an initial bearing image acquired in a overlooking mode in historical data, wherein the style of the bearing is various. The pixel points needing semantic segmentation are divided into two types, namely, the label labeling process corresponding to the training set is a single-channel semantic label, wherein the bearing area is labeled as 1, and the other areas are labeled as 0, so that a 0-1 mask image is obtained. And multiplying the 0-1 mask image and the initial bearing image to obtain a target bearing image. It should be noted that the target bearing image only contains bearings, and the DNN semantic division network has a task of classifying, and the loss function of the network is a cross entropy loss function.
And graying the obtained target bearing image, and updating the target bearing image, namely the updated target bearing image is a grayscale image.
Step S200, performing multi-threshold segmentation on the target bearing image based on the gray value to obtain a plurality of first area groups; and matching the first region group according to the distance of the central point and the overlapping area of the regions in the first region group to obtain a plurality of initial region groups.
Because the bearing is a metal product, the influence of illumination is large, the threshold segmentation result is not ideal, the boundary identified by the boundary identification method contains the boundary of the bearing, and the complete defect boundary is difficult to extract due to the influence of illumination.
The pixel values with similar gray values on the target bearing image can be identified to the same region through multi-threshold segmentation, the same attribute value is given, and the influence on the identification of the defect region caused by the grinding process due to the system error of the electronic identification device and the error of the metal product caused by the influence of illumination is avoided, so that the accurate positioning of the polishing process is influenced.
Firstly, based on the gray value, multi-threshold segmentation is carried out on the target bearing image to obtain a plurality of different gray levels, and a gray level image is obtained. The gray value of each pixel point in the gray level image is the gray level mean value of the gray level of the original pixel point. In the embodiment of the invention, according to the Fisher criterion, the gray level image is subjected to multi-threshold segmentation by using the principle that the inter-class variance is maximum and the intra-class variance is minimum to obtain different gray levels. The purpose of multi-threshold segmentation is to reduce the influence of uneven illumination, so that the pixel values with similar gray levels are at the same gray level, thereby obtaining a gray level map. The regions belonging to the same gray level are referred to as a first region group of the same gray level, that is, a plurality of first region groups are obtained by multi-threshold division.
And acquiring the minimum circumscribed rectangle corresponding to the first area group. Through calculating the area groups, the first area groups which have attribute value differences and are close to each other in spatial distribution can be classified into the same area group, so that the characteristic areas which accord with the defect areas caused by the grinding process are obtained, namely, a plurality of characteristic areas which accord with the visual segmentation effect of human eyes are obtained through calculation, a good segmentation area is provided for the complete defect area identification in the subsequent steps, and accurate positioning is provided for the polishing process.
And matching the region groups by calculating KM matching of the central points of the regions in the different region groups and the overlapping areas of the different region groups to obtain different region groups.
Based on the distance between the central points of the first area groups, the areas in different first area groups are matched by using a KM matching algorithm to obtain KM matching values, and the obtained maximum KM matching value is used as the central point matching rate of the two first area groups. It should be noted that KM matching is a bipartite graph matching method, that is, each edge of a bipartite graph has a weight, and KM matching is a complete matching scheme, so that the sum of the weights of all matching edges is maximized and is recorded as the optimal perfect matching. The KM matching result is one-to-one matching, and the matching numbers of the left and right sides are required to be consistent. Therefore, when two different first region groups are matched, the region groups with a large number of regions need to be combined to obtain a plurality of region combinations, for example, three regions b1, b2 and b3 are included in the region groups, two or two of the three regions are combined into (b 1, b 2), (b 1, b 3) and (b 2 and b 3), and then the matching results of the two first region groups are calculated through KM matching.
And calculating one-to-one weight of every two regions belonging to different first region groups in the two first region groups, wherein the weight refers to the Euclidean distance of the central point of the corresponding region. When two first region groups are matched, the first region group with the larger number of regions in the first two region groups and the first region group with the smaller number of regions in the first two region groups are matched in a combined manner, for example, if three regions b1, b2 and b3 exist in one first region group, and two regions a1 and a2 exist in the other first region group, the KM matching values of (a 1, a 2), (b 1, b 2), (a 1, a 2) and (b 1, b 3), (a 1, a 2) and (b 2, b 3) are respectively calculated, and the KM matching values are also the sum of matched edge weights. The maximum matching value among all KM matching results is taken as the KM matching value of the two area groups, and the calculated KM matching value may also be referred to as the center point matching rate. And normalizing the central point matching rate, and updating the central point matching rate to be the normalized central point matching rate.
Further, acquiring a spatial range where the first area group is located, specifically: and acquiring the minimum circumscribed rectangle corresponding to the first area group as the space range of the area group. And acquiring the spatial range of different first region groups, and taking the overlapping area of the minimum circumscribed rectangles of the different first region groups as the range matching rate. And normalizing the range matching rate, and updating the range matching rate into the normalized range matching rate.
And matching the first region group according to the distance of the central point of the region in the first region group and the overlapping area to obtain a plurality of initial region groups. Specifically, the method comprises the following steps: and dividing the first area group with the range matching rate and the center point matching rate both greater than the preset matching rate into the same area group. In the embodiment of the present invention, the predetermined matching rate is 0.8, and in other embodiments, the implementer may adjust the value according to the actual situation.
Step S300, segmenting the target bearing image to obtain a background area and corresponding background pixel points; screening the initial area groups according to the number proportion of background pixel points in the initial area groups to obtain a plurality of target area groups; and connecting the areas in each target area group to obtain a plurality of characteristic areas.
After a plurality of initial area groups are obtained, characteristic areas are further obtained, the characteristic areas refer to different areas which are obtained by identification from the initial area groups and are divided through boundary lines, the different areas obtained by identification accord with the characteristics of the defect areas caused by the grinding process, the defect areas can be completely and accurately identified through the characteristic areas, and then the polishing process is carried out on the bearing.
In the target bearing image, the area with a smaller gray value, i.e., a darker color, is considered as a target area, i.e., a foreground area. Due to the influence of illumination, areas with darker colors and areas with lighter colors in the target bearing image often exist at intervals, so that defect areas with the same defect are often difficult to identify as the same defect, the characteristic area is provided to meet the area division of human vision, namely pixel points with larger gray values or smaller gray values exist in the divided characteristic area at the same time, and the defect areas belonging to the same defect are divided into the same characteristic area under the condition that the color distribution of the target bearing image collected under the influence of illumination is not uniform.
And (3) carrying out an Otsu threshold segmentation method on the target bearing image to obtain a background threshold, taking pixel points larger than the background threshold as background pixel points, and obtaining a corresponding background area. And obtaining the number ratio of background pixel points in each initial area group, and screening the initial area groups to obtain a plurality of target area groups. Specifically, the method comprises the following steps: and taking the initial area group with the number ratio of the background pixel points smaller than a preset ratio threshold as a target area group. In the embodiment of the present invention, the preset duty threshold is 0.4, and in other embodiments, the implementer may adjust the value according to the actual situation.
Furthermore, the regions in each target region group are connected to obtain a plurality of feature regions. Specifically, the method comprises the following steps:
taking the central point of the target area group as an origin point, and making rays in different directions to obtain an outermost layer intersection point of the area in the target area group, wherein the outermost layer intersection point is used as a connection point.
And acquiring the central point of each target area group, taking the central point as an origin point, and making rays in different directions to obtain an area intersection point with the target area group, wherein the area intersection point is an outermost layer intersection point of the boundary of the area in the target area group, which is obtained by continuously extending the corresponding rays along the direction. In the embodiment of the present invention, the different directions are 360 directions centered on the origin, i.e., one-degree-one directions, so that the outermost layer intersections obtained are 360 in total, which are also referred to as connection points.
And step two, connecting the connection points according to a Bessel region connection principle to obtain a plurality of characteristic regions.
For the obtained multiple connecting points, connecting lines, also called boundary lines, are obtained according to the Bessel region connection principle, namely boundary lines of different target region groups are obtained, and the boundary lines divide the target bearing image into multiple different characteristic regions.
Step S400, classifying a plurality of regions of a first region group in the characteristic region according to the area size of the region to obtain a second region group; acquiring the gray value, the gray average value and the maximum gray value of each region in the second region group, and constructing a gray difference sequence according to the difference of the gray values of adjacent regions; and acquiring the conversion times of positive and negative values in the gray difference value sequence, obtaining the defect probability according to the conversion times, the gray mean value and the maximum gray value, and screening an initial defect area group from the second area group.
The bearing roller path is impacted to form an impact area group, the impact area with the largest stress is taken as the central area of the target area group, the gray value is increased outwards, the color of the central area is deepest, and the corresponding gray value is minimum.
Firstly, a plurality of regions of a first region group in a characteristic region are preliminarily classified according to the area size of the region, and a plurality of second region groups are obtained. Specifically, the method comprises the following steps: based on the areas of different regions in the same first region group, multiple area levels are obtained by multi-threshold segmentation, so that regions with similar areas are divided into the same second region group. According to the Fisher criterion, the area of the region is subjected to multi-threshold segmentation by using the principle that the inter-class variance is maximum and the intra-class variance is minimum.
That is, the subsequent step is to perform calculation based on the region groups divided into the same second region group with similar areas.
And acquiring the gray value, the average gray value and the maximum gray value of each second area group. And constructing a gray difference value sequence according to the difference value of the gray values of the adjacent areas in the second area group. And acquiring turning points of positive and negative numbers in the gray difference sequence to obtain conversion times, wherein the gray difference sequence and the corresponding conversion times reflect the fluctuation degree of the gray values of the adjacent areas in the second area group.
The closer the number of transitions is to 1, the greater the probability that the corresponding second region group is a defective region, and the closer the position of the turning point is to the center position, the greater the probability that the second region group is a defective region.
And obtaining the defect probability according to the conversion times, the gray average value and the maximum gray value.
Probability of the defect
The calculation formula of (2) is as follows:
wherein the content of the first and second substances,
the number of times of conversion;
is the maximum gray value;
is the mean value of the gray levels.
And taking the second area group with the defect probability greater than the preset defect probability threshold as an initial defect area group, wherein the value of the preset defect probability threshold is 0.7 in the embodiment of the invention, and the implementer can adjust the value according to the actual situation in other embodiments.
Step S500, obtaining the target defect area according to the gray scale abnormal condition and the direction abnormal condition of the initial defect area group.
And a plurality of initial defect area groups may exist in one characteristic area at the same time, the direction and the gray level condition of the area groups are further considered for the initial defect area groups, and the initial defect area groups are screened again to obtain target defect areas.
The coordinates of each pixel point of each area in the initial defect area group are obtained, the principal component direction of the pixel points in the area is obtained by using Principal Component Analysis (PCA), a plurality of principal component directions are obtained, the principal component direction with the largest characteristic value is obtained, the largest principal component direction is taken as the principal direction of the area, and the direction with the largest projection variance of the data, namely the main distribution direction of the data, is represented. Each region within each initial defect region group corresponds to a main direction.
And then obtaining a principal direction line through the central point of the area and the principal component direction, wherein the intersection point of the principal direction line and the boundary of the area is called an end point, and convex hull calculation is carried out on the end point data of all the areas belonging to the same initial defect area group to obtain an external graph range surrounding the areas, wherein the external graph range is a dividing line of the initial defect area group.
And screening out the target defect area according to the gray scale abnormal condition and the direction abnormal condition of the external graph range corresponding to the initial defect area group. It should be noted that the target defect region is a defect inside the feature region, and a relatively complete defect region caused by poor grinding, which is less affected by illumination, can be obtained by calculating different feature regions, so as to perform polishing control.
And performing threshold segmentation on the characteristic region by utilizing an Otsu threshold segmentation method to obtain a segmentation threshold, taking the pixel points with the gray values of the pixel points smaller than the segmentation threshold as target pixel points, and taking the region group with the target pixel point ratio larger than the preset pixel ratio in the initial defect region group as a second defect region group. In the embodiment of the present invention, the value of the predetermined pixel ratio is 0.8, and in other embodiments, an implementer may adjust the value according to an actual situation. That is, the purpose of screening the initial defect area group according to the gray level abnormality of the initial defect area group is completed.
And calculating principal component direction values of all regions on the target bearing image, constructing a principal component direction value sequence, and selecting the principal component direction value with the maximum occurrence frequency as the texture direction value of the target bearing image.
And obtaining coordinates of each pixel point in the second defect area group, and obtaining a second principal component direction of the second defect area group by using a principal component direction analysis method, wherein if the number of the second defect area group is K, the number of the second principal component directions can be K.
And acquiring a difference value between the second principal component direction of each second defect area group and the texture direction value, wherein the second defect area group with the difference value larger than the preset principal component direction difference value is a target defect area group, and the external graphic range corresponding to the target defect area group is a target defect area. In the embodiment of the present invention, the principal component direction difference is preset to be 20, and in other embodiments, the value may be adjusted by an implementer according to an actual situation. That is, the purpose of screening the initial defective area group according to the direction abnormality of the initial defective area group is completed.
And after the target defect areas in different characteristic areas are obtained through calculation, the polishing position of the polishing process is determined, and the identified target defect areas are subjected to an automatic polishing process to eliminate defects.
In summary, the embodiment of the present invention utilizes a data processing technology, which is a method for identifying by using an electronic device, and utilizes an artificial intelligence system in the production field to complete polishing positioning. Firstly, acquiring a target bearing image; performing multi-threshold segmentation on the target bearing image based on the gray value to obtain a plurality of first area groups; matching the first region group according to the central point distance and the overlapping area of the regions in the first region group to obtain a plurality of initial region groups; segmenting the target bearing image to obtain background pixel points; screening the initial area groups according to the number proportion of background pixel points in the initial area groups to obtain a plurality of target area groups; connecting the areas in each target area group to obtain a plurality of characteristic areas; classifying a plurality of regions of the first region group in the characteristic region according to the area size of the region to obtain a second region group; acquiring the gray value, the gray average value and the maximum gray value of each region in the second region group, and constructing a gray difference sequence according to the difference of the gray values of adjacent regions; acquiring the conversion times of positive and negative values in the gray difference value sequence, obtaining the defect probability according to the conversion times, the gray mean value and the maximum gray value, and screening an initial defect area group from the second area group; and obtaining the target defect area according to the gray scale abnormal condition and the direction abnormal condition of the initial defect area group. According to the invention, the target defect area is obtained after data processing is carried out on the target bearing image acquired by the camera, so that the defect areas belonging to the same collision injury can be obtained under the condition that the target bearing image is not uniformly distributed, and polishing positioning is completed.
It should be noted that: the sequence of the above embodiments of the present invention is only for description, and does not represent the advantages or disadvantages of the embodiments. And that specific embodiments 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 (5)
1. A bearing raceway bruise identification and automatic polishing positioning method in a grinding process is characterized by comprising the following steps:
acquiring an initial bearing image, and identifying a bearing in the initial bearing image to obtain a target bearing image;
performing multi-threshold segmentation on the target bearing image based on the gray value to obtain a plurality of first area groups; matching the first region group according to the distance of the central point and the overlapping area of the regions in the first region group to obtain a plurality of initial region groups;
segmenting the target bearing image to obtain a background area and corresponding background pixel points; screening the initial area group to obtain a plurality of target area groups according to the number proportion of background pixel points in the initial area group; connecting the areas in each target area group to obtain a plurality of characteristic areas;
classifying a plurality of regions of the first region group in the characteristic region according to the area of the region to obtain a second region group; acquiring the gray value, the gray average value and the maximum gray value of each region in the second region group, and constructing a gray difference sequence according to the difference of the gray values of adjacent regions; acquiring the conversion times of positive values and negative values in the gray difference value sequence, obtaining the defect probability according to the conversion times, the gray mean value and the maximum gray value, and screening an initial defect area group from a second area group;
obtaining a target defect area according to the gray scale abnormal condition and the direction abnormal condition of the initial defect area group;
the method for acquiring the plurality of initial area groups comprises the following steps: based on the distance between the central points of the areas in the first area groups, matching the areas in different first area groups by using a KM matching algorithm, and taking the obtained maximum KM matching value as the central point matching rate of the two first area groups; acquiring the minimum circumscribed rectangle of each first region group, and acquiring the overlapping area of the minimum circumscribed rectangles among the first region groups as the range matching rate; dividing the first area group with the center point matching rate and the range matching rate both greater than a preset matching rate into the same initial area group;
obtaining a target defect area according to the gray scale abnormal condition and the direction abnormal condition of the initial defect area group, wherein the target defect area is obtained by the following steps: performing threshold segmentation on the characteristic region by utilizing an Otsu threshold segmentation method to obtain a segmentation threshold, wherein pixel points with gray values smaller than the segmentation threshold are used as target pixel points, and an initial defect region group with target pixel points with a ratio larger than a preset pixel ratio is used as a second defect region group; acquiring principal component direction values of all regions on the target bearing image, and constructing a principal component direction value sequence; selecting the principal component direction value with the largest frequency in the principal component direction value sequence as a texture direction value; and acquiring a second principal component direction of the second defect area group, wherein the second defect area group with the difference value between the second principal component direction and the texture direction value larger than the preset principal component direction difference value is the target defect area.
2. The method for identifying the collision damage of the bearing raceway and automatically polishing and positioning the bearing raceway in the grinding process as claimed in claim 1, wherein the performing multi-threshold segmentation on the target bearing image to obtain a plurality of first area groups comprises:
and performing multi-threshold segmentation on the target bearing image to obtain a plurality of different gray levels, and taking the region with the same gray level as a first region group with the same gray level.
3. The method for identifying the bearing raceway damage and automatically polishing and positioning in the grinding process according to claim 1, wherein the step of screening the initial area group to obtain a plurality of target area groups according to the number ratio of background pixel points in the initial area group comprises the steps of:
and taking the initial area group with the number ratio of the background pixel points smaller than a preset ratio threshold as a target area group.
4. The method for identifying bearing raceway bruise and automatically polishing and positioning as claimed in claim 1, wherein said connecting regions in each of said target region groups to obtain a plurality of characteristic regions comprises:
taking the central point of the target area group as an origin point, and taking rays in different directions to obtain an outermost layer intersection point of the area in the target area group, wherein the outermost layer intersection point is taken as a connection point;
and connecting the connection points according to a Bessel region connection principle to obtain a plurality of characteristic regions.
5. The method for identifying bearing raceway bruise and automatically polishing and positioning according to claim 1, wherein the obtaining of the defect probability from the conversion times, the gray-scale mean value and the maximum gray-scale value comprises:
the calculation formula of the defect probability is as follows:
wherein the content of the first and second substances,
is the defect probability;
the number of times of conversion is;
is the maximum gray value;
is the mean value of the gray levels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210811778.5A CN114882387B (en) | 2022-07-12 | 2022-07-12 | Bearing raceway bruise identification and automatic polishing positioning method in grinding process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210811778.5A CN114882387B (en) | 2022-07-12 | 2022-07-12 | Bearing raceway bruise identification and automatic polishing positioning method in grinding process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114882387A CN114882387A (en) | 2022-08-09 |
| CN114882387B true CN114882387B (en) | 2022-09-16 |
Family
ID=82683470
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210811778.5A Active CN114882387B (en) | 2022-07-12 | 2022-07-12 | Bearing raceway bruise identification and automatic polishing positioning method in grinding process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114882387B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115294410B (en) * | 2022-10-08 | 2023-10-17 | 加乐新材料(南通)有限公司 | Plastic product molding control method based on pattern recognition |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104990925A (en) * | 2015-06-23 | 2015-10-21 | 泉州装备制造研究所 | Defect detecting method based on gradient multiple threshold value optimization |
| WO2018000731A1 (en) * | 2016-06-28 | 2018-01-04 | 华南理工大学 | Method for automatically detecting curved surface defect and device thereof |
| CN112784847A (en) * | 2021-01-28 | 2021-05-11 | 中国空气动力研究与发展中心超高速空气动力研究所 | Segmentation and identification method for ultra-high-speed impact damage detection image |
| CN114202541A (en) * | 2022-02-17 | 2022-03-18 | 湖北中海电线电缆有限公司 | Cable defect detection method based on artificial intelligence |
-
2022
- 2022-07-12 CN CN202210811778.5A patent/CN114882387B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104990925A (en) * | 2015-06-23 | 2015-10-21 | 泉州装备制造研究所 | Defect detecting method based on gradient multiple threshold value optimization |
| WO2018000731A1 (en) * | 2016-06-28 | 2018-01-04 | 华南理工大学 | Method for automatically detecting curved surface defect and device thereof |
| CN112784847A (en) * | 2021-01-28 | 2021-05-11 | 中国空气动力研究与发展中心超高速空气动力研究所 | Segmentation and identification method for ultra-high-speed impact damage detection image |
| CN114202541A (en) * | 2022-02-17 | 2022-03-18 | 湖北中海电线电缆有限公司 | Cable defect detection method based on artificial intelligence |
Non-Patent Citations (1)
| Title |
|---|
| 基于灰度直方图拟合曲线的数字图像多阈值分割技术研究;梁忠伟等;《现代制造工程》;20070918(第09期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114882387A (en) | 2022-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115082683B (en) | Injection molding defect detection method based on image processing | |
| CN114140462B (en) | Bearing wear degree assessment method based on image processing | |
| CN107437245B (en) | High-speed railway contact net fault diagnosis method based on deep convolutional neural network | |
| CN115311277B (en) | Pit defect identification method for stainless steel product | |
| CN103543394A (en) | Discharge ultraviolet imaging quantization parameter extraction method of high-voltage electric equipment | |
| CN109523529B (en) | Power transmission line defect identification method based on SURF algorithm | |
| CN115345885A (en) | Method for detecting appearance quality of metal fitness equipment | |
| CN109145964B (en) | Method and system for realizing image color clustering | |
| CN114240888A (en) | Furniture assembly paint spraying defect repairing method and system based on image processing | |
| CN102306307B (en) | Positioning method of fixed point noise in color microscopic image sequence | |
| CN113256624A (en) | Continuous casting round billet defect detection method and device, electronic equipment and readable storage medium | |
| CN114820625B (en) | Automobile top block defect detection method | |
| CN114882387B (en) | Bearing raceway bruise identification and automatic polishing positioning method in grinding process | |
| CN113888536B (en) | Printed matter double image detection method and system based on computer vision | |
| CN116883408B (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
| CN115100174B (en) | Ship sheet metal part paint surface defect detection method | |
| CN116012291A (en) | Industrial part image defect detection method and system, electronic equipment and storage medium | |
| CN111259893A (en) | Intelligent tool management method based on deep learning | |
| CN117094975A (en) | Method and device for detecting surface defects of steel and electronic equipment | |
| CN114627116A (en) | Fabric defect identification method and system based on artificial intelligence | |
| JP2004030188A (en) | Method, apparatus and program for dividing image into areas | |
| CN107967690B (en) | Self-adaptive ferrographic abrasive particle image binarization processing method | |
| CN116977931A (en) | High-altitude parabolic identification method based on deep learning | |
| CN114833648B (en) | Surface damage identification and grinding control method in hydraulic pump bearing grinding process | |
| CN115830319A (en) | Strabismus iris segmentation method based on attention mechanism and verification method |
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 |