CN114897894A - Method for detecting defects of cheese chrysanthemum core - Google Patents
Method for detecting defects of cheese chrysanthemum core Download PDFInfo
- Publication number
- CN114897894A CN114897894A CN202210815245.4A CN202210815245A CN114897894A CN 114897894 A CN114897894 A CN 114897894A CN 202210815245 A CN202210815245 A CN 202210815245A CN 114897894 A CN114897894 A CN 114897894A
- Authority
- CN
- China
- Prior art keywords
- pixel point
- abnormal pixel
- abnormal
- calculating
- 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.)
- Granted
Links
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
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- 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
- G06T2207/30124—Fabrics; Textile; Paper
-
- 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 a method for detecting defects of a cheese chrysanthemum core, and belongs to the field of cheese defect detection. The method comprises the following steps: acquiring an RGB image of the upper surface of the cone yarn, and performing graying processing on the RGB image of the upper surface of the cone yarn to obtain a gray image of the upper surface of the cone yarn; obtaining abnormal pixel points in the grey image of the upper surface of the cone yarn according to the grey gradient of each pixel point in the grey image of the upper surface of the cone yarn; calculating the direction similarity index of each abnormal pixel point according to the gradient direction of each abnormal pixel point; calculating the periodicity index of each abnormal pixel point according to the gray value of each abnormal pixel point; and calculating the probability of the bobbin yarn being the chrysanthemum core defect according to the direction similarity index and the periodicity index of each abnormal pixel point. The invention realizes the automatic detection of the chrysanthemum core defect of the cone yarn.
Description
Technical Field
The invention relates to the field of cheese defect detection, in particular to a cheese chrysanthemum core defect detection method.
Background
The winding process is the last process of spinning in a textile mill, the winding task is to roll the cop into a bobbin, the winding capacity is increased on the basis of changing the winding form, and meanwhile, the high-speed unwinding of the subsequent process is also adapted, and the production efficiency is improved. The cone yarn is a product of a winding processing procedure, and the quality of the cone yarn directly influences the production of the subsequent procedures and finally influences the quality of a textile finished product. Due to various factors in production, the cone yarn may generate various defects in the forming process. The common defects of the cone yarn which have great influence on the production comprise multilayer table cone yarn, gauze yarn, multi-source yarn and the like. In the traditional production line, aiming at various defects of the cone yarn, a manual detection method is mainly adopted for screening. The manual detection mainly has the following defects: firstly, the detection accuracy is low, namely the accuracy depends on the trained level and experience of workers, and false detection and missed detection conditions exist; secondly, the detection speed is low; thirdly, the artificial fatigue period is longer.
For the defects, the defect area is inconsistent with the normal area in color, texture, space and other image characteristics, a template is established through normal textiles, and the product to be detected is compared with the template to realize defect detection. However, the template matching method has poor generalization capability, and is difficult to establish an effective template particularly for chrysanthemum core defects.
The existing method for detecting the chrysanthemum core defect of the cone yarn by depending on manual work has low detection efficiency, and the problem that how to improve the detection efficiency of the chrysanthemum core defect of the cone yarn needs to be solved by cone yarn manufacturers is solved.
Disclosure of Invention
In order to solve the problem of low detection efficiency of the existing method for manually detecting the chrysanthemum core defect of the cone yarn, the invention provides a technical scheme of a method for detecting the chrysanthemum core defect of the cone yarn, which comprises the following steps:
acquiring an RGB image of the upper surface of the cone yarn, and performing graying processing on the RGB image of the upper surface of the cone yarn to obtain a gray image of the upper surface of the cone yarn;
obtaining abnormal pixel points in the grey image of the upper surface of the cone yarn according to the grey gradient of each pixel point in the grey image of the upper surface of the cone yarn;
calculating the direction similarity index of each abnormal pixel point according to the gradient direction of each abnormal pixel point; calculating the periodicity index of each abnormal pixel point according to the gray value of each abnormal pixel point;
and calculating the probability of the bobbin yarn being the chrysanthemum core defect according to the direction similarity index and the periodicity index of each abnormal pixel point.
Has the advantages that: according to the method, the abnormal pixel points in the gray level image of the upper surface of the cone yarn are judged according to the gray level gradient of the pixel points in the gray level image of the upper surface of the cone yarn, and the abnormal pixel points are considered that if the abnormal pixel points are chrysanthemum core defect pixel points, the gradient direction of the abnormal pixel points is generally vertical to the direction pointing to the center of a circular ring, and the pixel points at the chrysanthemum core defect parts have certain regularity.
Further, the calculating the direction similarity index of each abnormal pixel point according to the gradient direction of each abnormal pixel point includes:
for any abnormal pixel point:
acquiring the direction of a connecting line which takes the abnormal pixel point as a starting point and the central pixel point of the gray image on the upper surface of the cone yarn as an end point;
and calculating the direction similarity index of the abnormal pixel point according to the gradient direction of the abnormal pixel point and the direction of the connecting line.
Further, the direction similarity index of each abnormal pixel point is calculated by using the following calculation formula:
wherein the content of the first and second substances,
is the direction similarity index of the ith abnormal pixel point,
is the gradient direction of the ith abnormal pixel point,
and the direction of the connecting line corresponding to the ith abnormal pixel point.
Further, the calculating the periodicity index of each anomalous pixel point according to the gray value of each anomalous pixel point includes:
calculating the gray level of a normal pixel point corresponding to the gray image on the upper surface of the cone yarn;
for any abnormal pixel point: constructing a pixel gray value sequence with a set length by taking the abnormal pixel as a central point, constructing a reconstruction sequence corresponding to the pixel gray value sequence according to whether each gray value in the pixel gray value sequence is greater than the gray level of a normal pixel, and counting the jumping times of the reconstruction sequence; and calculating the periodicity index of the abnormal pixel point according to the jumping times.
Further, the step of constructing a reconstruction sequence corresponding to the pixel gray value sequence according to whether each gray value in the pixel gray value sequence is greater than the gray level of the normal pixel, and counting the hopping times of the reconstruction sequence includes:
sequentially judging whether each gray value in the gray value sequence of the pixel point is greater than the gray level of the normal pixel point, and if so, recording a corresponding element in the reconstruction sequence as 1; if not, marking the corresponding element in the reconstruction sequence as 0;
and counting the total times of the elements in the reconstruction sequence from 1 to 0 and from 0 to 1, and recording the total times as the transition times of the reconstruction sequence.
Further, the calculating the periodicity index of the abnormal pixel point according to the hopping times includes:
wherein the content of the first and second substances,
is a periodic index of the abnormal pixel point, tanh is a hyperbolic tangent function,
in order to be a hyper-parameter,
and the jump times of the reconstruction sequence corresponding to the abnormal pixel point are calculated.
Further, the probability of the bobbin yarn being a chrysanthemum core defect is calculated by using the following formula, comprising the following steps:
wherein N represents the number of abnormal pixel points,
the periodicity index of the nth abnormal pixel point is represented,
the direction similarity index of the nth abnormal pixel point is expressed,
the probability of the bobbin yarn being a chrysanthemum core defect.
Further, the obtaining of the abnormal pixel points in the grey image of the upper surface of the cone yarn according to the grey gradient of each pixel point in the grey image of the upper surface of the cone yarn comprises:
calculating each pixel point in gray level image of cone yarn by using sobel operator
And
a gradient of direction;
according to each pixel point
And
calculating the gradient amplitude of each pixel point by the directional gradient;
and judging whether the gradient amplitude of each pixel point is greater than or equal to a set gradient threshold value, if so, judging the corresponding pixel point as an abnormal pixel point.
Drawings
FIG. 1 is a schematic representation of a cheese chrysanthemum core defect;
FIG. 2 is a schematic illustration of a qualified package;
FIG. 3 is a flow chart of a method for detecting defects in a cheese chrysanthemum core according to the present invention;
fig. 4 is a schematic image acquisition of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention.
The invention aims to process the acquired RGB images of the cone yarn by using a computer vision technology and detect the chrysanthemum core defect with unobvious characteristics. The qualified cone yarn is shown in figure 1, the cone yarn with chrysanthemum core defects is shown in figure 2, the chrysanthemum core defects are shown as specific textures, are three-dimensional deformation defects, and are radial bulges and depressions formed by yarn layers around the cone yarn tube by taking the cone yarn tube as the center; the main forming reason is that the outer layer yarn generates larger pressure due to mechanical reasons during production, so that the inner layer yarn layer is extruded to generate deformation, and the shape similar to a chrysanthemum core is formed in the areas of two ends of the cone yarn close to the cone core. The chrysanthemum core defect can make the range of yarn layer disorder, causes the yarn embedding of a certain layer other yarn layer, brings the influence to the high-speed backing-off of subsequent handling, influences production efficiency, and the quality of section of thick bamboo yarn quality can also influence the quality of textile products simultaneously.
Specifically, as shown in fig. 3, the method for detecting a defect of a chrysanthemum core of a cheese in the embodiment includes the following steps:
(1) acquiring an RGB image of the upper surface of the cone yarn, and performing graying processing on the RGB image of the upper surface of the cone yarn to obtain a gray image of the upper surface of the cone yarn;
the purpose of this embodiment is to detect the chrysanthemum core defect of section of thick bamboo yarn upper surface, needs to gather section of thick bamboo yarn upper surface image earlier. In order to collect the image of the upper surface of the cheese, as shown in fig. 4, a camera is arranged right above the cheese, the image of the surface of the cheese is collected in a overlooking mode, an LED white light bar-shaped light source is adopted as the light source, and the light source is arranged obliquely above the cheese.
The collected images of the upper surface of the cone yarn not only have the cone yarn, but also have a background. To avoid the effect of the background on the detection of surface defects on the package, the present embodiment first uses the DNN technique to identify the package in the image. The data set used in DNN network training is various cone yarn upper surface image data sets; the pixels needing to be segmented are divided into two types, namely the labeling process of the corresponding labels of the training set is as follows: the semantic label of the single channel, the pixel of the corresponding position belongs to the background class and is marked as 0, and the pixel of the corresponding position belongs to the cone yarn and is marked as 1; the task of the DNN network is to classify, and all the loss functions used are cross entropy loss functions.
The image acquired by the camera is an RGB image, the RGB image of the upper surface of the cheese, which only comprises the cheese and does not comprise the background, is acquired by DNN network classification, and the RGB image of the upper surface of the cheese is converted into a gray image of the upper surface of the cheese by a gray processing method.
(2) Obtaining abnormal pixel points in the grey image of the upper surface of the cone yarn according to the grey gradient of each pixel point in the grey image of the upper surface of the cone yarn;
the chrysanthemum core defects are not as sharply characterized in the image as defects in multilayer table yarns, multi-source yarns, and the like. The process of judging the chrysanthemum core defect in the embodiment is more through the change of gray scale. Because the texture image of the top surface of the cheese is collected by the active illumination of the oblique light source, the convex part in the cheese chrysanthemum core is strongly illuminated, and the gray value is higher; whereas the recessed portions in the chrysanthemum core produce shadows with lower gray values. If the annular image is unfolded into a rectangular image, the chrysanthemum core parts are changed into vertical stripes which are sequentially staggered and periodically arranged according to the color depth (namely the gray value) in the horizontal direction, and the normal cone yarn has uniform texture after being unfolded and does not have a periodic stripe image. Because the chrysanthemum core defect has light and shadow change and has bulges and depressions in space, the gray gradient exists on the pixel points of the defect area, and the pixel points of the chrysanthemum core defect part have certain regularity, so the more the number of the pixel points with larger gray gradient amplitude in the image is, the more the gray change of the pixel points has periodicity, and the higher the probability of the chrysanthemum core defect is when the pixel points accord with the specific rule.
Because the upper surface of the qualified cone yarn is relatively flat, the pixel points with larger gradient difference account for less, and because the cone yarn is circular, the gradient direction of the pixel points with larger gradient is generally directed to the center of the circular ring; the cone yarn with the chrysanthemum core defect is formed by radial bulges and depressions formed by the yarn layers surrounding the cone yarn tube by taking the cone yarn tube as the center, so that the gradient direction of pixel points with larger gradient is usually vertical to the direction pointing to the center of the circular ring.
In the embodiment, each pixel point in the gray level image of the cheese is calculated by using the sobel operator
And
gradient of direction, respectively
And
. The gradient amplitude of the pixel point is
With a corresponding gradient direction of
。
The pixel points are marked and classified according to the gradient amplitude of the pixel points, and the gray values of the pixel points are similar because the upper surface of the qualified cone yarn is smooth, so that a small gradient exists. When the gradient amplitude of a certain pixel point is larger, the pixel point may be noise, and may also be a pixel point belonging to the chrysanthemum core defect, so that the pixel points are screened,setting a gradient magnitude threshold
When is coming into contact with
Time (the setting of the embodiment)
And the adjustment can be carried out according to the actual situation in the actual application), marking the pixel points, and judging the pixel points as abnormal pixel points; when in use
And then, judging the pixel points as normal pixel points.
(3) Calculating the direction similarity index of each abnormal pixel point according to the gradient direction of each abnormal pixel point; calculating the periodicity index of each abnormal pixel point according to the gray value of each abnormal pixel point;
abnormal pixel points in the gray image of the upper surface of the cone yarn are obtained in the steps, and the abnormal pixel points are not necessarily all the pixel points of the chrysanthemum core defect and possibly are noise; however, the gradient direction of the pixel points of the chrysanthemum core defect is generally perpendicular to the direction pointing to the center of the circular ring, and the pixel points of the chrysanthemum core defect have certain regularity, so the direction similarity index and the periodicity index of each abnormal pixel point are calculated in the embodiment, and the specific process is as follows.
Direction similarity index of different abnormal pixel points
Calculating the gradient direction of the abnormal pixel point
The direction of the connection line between the pixel point and the central pixel point
The direction similarity of the yarn layer around the bobbin tube is that the yarn layer around the bobbin tube forms radial bulges and depressions by taking the bobbin tube as the center, so that the defective pixelThe gradient direction of the points is generally perpendicular to the direction pointing towards the center of the circle; if the package is a good product, the gradient of the texture present on its upper surface is generally parallel to the direction pointing towards the center of the ring. The present embodiment calculates the direction similarity index of the abnormal pixel point by using the following formula
：
In the formula (I), the compound is shown in the specification,
the direction similarity index of the ith abnormal pixel point is expressed,
the gradient direction of the ith abnormal pixel point is shown,
the direction of the ith abnormal pixel point pointing to the center of the circular ring (namely the direction of the connecting line of the pixel point and the center point of the circular ring) is represented; when the gradient direction of the abnormal pixel point is more parallel to the direction of the pixel point pointing to the center of the circular ring,
the larger the value of (A), the closer to 1; when the gradient direction of the abnormal pixel point is more vertical to the direction of the pixel point pointing to the center of the circular ring,
the smaller the value of (b), the closer to 0.
The larger the value of (b), the more likely the ith abnormal pixel point is to be a pixel point with chrysanthemum core defect, and the more unlikely the ith abnormal pixel point is to be noise.
Periodic index of each abnormal pixel point
Because the top surface of the cone yarn is annular, the annular is unfolded into a rectangular image of a Cartesian coordinate system according to a polar coordinate unfolding mode, and the annular image is converted into the rectangular image, so that the burden of a special-shaped area on periodic calculation can be avoided; on the other hand, the normal texture can be unfolded into a horizontal line, and the subsequent processing is simplified. The method for converting the annular image into the rectangular image is the prior art, and is not described herein again.
In order to calculate the periodicity index of the abnormal pixel point, the gray level of the normal pixel point is calculated, then the continuous pixel point gray value sequence with the length L and the abnormal pixel point as the center point is obtained, the gray value sequence is encoded, and the periodicity index of the gray value sequence is calculated.
The method for calculating the gray level of the normal pixel point in the embodiment is as follows:
as described above, the gray image of the upper surface of the cone yarn in this embodiment includes only the cone yarn and does not include the background; establishing a gray level histogram for the gray level image of the upper surface of the cone yarn, and calculating the frequency of each gray level, namely:
in the formula
Indicating the frequency of occurrence of the jth gray level,
and B represents the frequency of the pixel points corresponding to the jth gray level, and B represents the total number of the image pixel points.
When calculating the gray level of the normal pixel point, the maximum frequency value or the mean value of the gray histogram is used to represent that the gray level of the normal pixel point has a larger deviation (when the defect is serious, the maximum frequency value of the gray level may be the defect), but the required ideal gray level of the normal pixel point is located between the maximum frequency value and the mean value of the gray histogram, so that the gaussian distribution is established, and the mean value of the gaussian distribution are located between the maximum frequency value and the mean value of the gray histogramVariance is the maximum grey value with frequency
And mean value of gray value
Mean value of gray values between
At the maximum gray value of frequency
And mean value of gray value
The variance of gray values therebetween is variance
The correlation calculation formula is:
the probability calculation formula of the gray level belonging to the gray level of the normal pixel point is as follows:
selecting
Gray level corresponding to maximum value
The gray level of the pixel point is normal.
After the gray level of the normal pixel point is obtained, the gray level of the normal pixel point is used
As a threshold, comparing the gray value of the gray value sequence of the continuous pixels with the length of L and the abnormal pixel as the center point, and making the gray value larger than the threshold
The position of the pixel point is marked as 1, otherwise, the position is marked as 0.
Therefore, the reconstruction sequence of the continuous pixel gray value sequence with the abnormal pixel point as the center point and the length of L can be obtained in the form of
. The reconstructed sequence is again counted with the number of transitions, where a transition from 0 to 1 or from 1 to 0 may represent a cycle, and the number of transitions of the reconstructed sequence is counted, for example, 000000 includes 0 transition, 00001111 includes 1 transition from 0 to 1, 100011 includes 1 transition from 1 to 0 and 1 transition from 0 to 1, and the number of transitions is recorded as 2. Recording the jumping times, calculating the periodicity index of the continuous pixel gray value sequence with the length of the center point of the abnormal pixel as L, and assigning the periodicity value to the abnormal pixel, namely:
in the formula
The periodicity index of the abnormal pixel point is represented, tanh represents a hyperbolic tangent function which plays a role of normalization,
is a root of Chao ShenAnd (4) counting. Periodic index of abnormal pixel
The more the value of (1) is close to 1, the higher the probability that the abnormal pixel point is the chrysanthemum core defect pixel point is.
In the embodiment, the grey pattern of the upper surface of the cheese is an annular image, and in order to simplify the calculation amount of the periodic indexes of various abnormal pixel points, the grey pattern of the upper surface of the cheese is converted into a rectangular image, and the continuous pixel points which are correspondingly obtained and take the abnormal pixel points as the central points are the pixel points on the same line; as another embodiment, the gray level graph on the upper surface of the cheese is not converted into a rectangular image, and then the correspondingly obtained continuous pixels with the abnormal pixel as the center point are pixels on the same circle.
(4) And calculating the probability of the bobbin yarn being the chrysanthemum core defect according to the direction similarity index and the periodicity index of each abnormal pixel point.
The direction similarity index and the periodicity index of the abnormal pixel point are in positive correlation with the probability that the abnormal pixel point is the chrysanthemum core defect. The formula for calculating the probability p of a defect with a cheese being a chrysanthemum core in this example is as follows:
wherein N represents the number of abnormal pixel points,
the periodicity index of the nth abnormal pixel point is represented,
and expressing the direction similarity index of the nth abnormal pixel point. For the nth abnormal pixel point, if the direction similarity index is
The smaller the periodicity index, the larger the anomalyThe greater the probability that the pixel point is the chrysanthemum core defect pixel point, the corresponding one
The greater the value of (A); if the nth abnormal pixel point is noise, the probability that the direction similarity index is smaller and the periodicity index is larger is lower, and the corresponding abnormal pixel point is noise
The value of (c) is also smaller. Therefore, when the number of the abnormal pixel points is more, and the direction similarity index of most of the abnormal pixel points is smaller and the periodicity index is larger, the probability that the chrysanthemum core exists in the cheese is larger. Setting a chrysanthemum core defect judgment threshold value according to the empirical value when
And judging that the bobbin yarn has the defect of chrysanthemum core.
The quality of the cone yarn and the severity of the chrysanthemum core defect are determined according to the defect probability in the embodiment, which specifically comprises the following steps:
when in use
And if the abnormal pixel points are noise points or normal cone yarn textures, the cone yarns are qualified products and are not processed.
When in use
The defect of the chrysanthemum core of the cone yarn is a serious defect, the cone yarn can influence the subsequent production efficiency and the quality of textiles, and the products are marked as defective products and need to be recycled and rebuilt.
The embodiment sets up the core of the chrysanthemumThe trap determination threshold value is
As another embodiment, the threshold may be adjusted according to actual needs. This embodiment is as follows
The yarn is adjusted and then continuously put into use
The cone yarn is recycled and rebuilt; as a further embodiment, the manner of handling the package in different situations can be adapted, for example directly
The cone yarn is recycled and rebuilt.
The abnormal pixel points in the gray scale image of the upper surface of the bobbin yarn are obtained according to the gray scale gradient of the pixel points in the gray scale image of the upper surface of the bobbin yarn, and the fact that the gradient direction of the abnormal pixel points is generally perpendicular to the direction pointing to the center of the circular ring if the abnormal pixel points are chrysanthemum core defect pixel points and the pixel points of the chrysanthemum core defect positions have certain regularity is considered, so that the probability that the bobbin yarn is the chrysanthemum core defect is calculated according to the direction similarity index and the periodicity index of each abnormal pixel point, automatic detection of the chrysanthemum core defect existing in the bobbin yarn is achieved, and the problem that the existing method for detecting the chrysanthemum core defect existing in the bobbin yarn manually is low in detection efficiency is solved.
It should be noted that while the preferred embodiments of the present invention have been described, additional variations and modifications to these embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
Claims (2)
1. The method for detecting the defects of the flower cores of the crossostephium glaucocalyx is characterized by comprising the following steps of:
acquiring an RGB image of the upper surface of the cone yarn, and performing graying processing on the RGB image of the upper surface of the cone yarn to obtain a gray image of the upper surface of the cone yarn;
obtaining abnormal pixel points in the grey image of the upper surface of the cone yarn according to the grey gradient of each pixel point in the grey image of the upper surface of the cone yarn;
calculating the direction similarity index of each abnormal pixel point according to the gradient direction of each abnormal pixel point; calculating the periodicity index of each abnormal pixel point according to the gray value of each abnormal pixel point;
calculating the probability of the bobbin yarn as the chrysanthemum core defect according to the direction similarity index and the periodicity index of each abnormal pixel point;
the method for calculating the direction similarity index of each abnormal pixel point according to the gradient direction of each abnormal pixel point comprises the following steps:
for any abnormal pixel point:
acquiring the direction of a connecting line which takes the abnormal pixel point as a starting point and the central pixel point of the gray image on the upper surface of the cone yarn as an end point;
calculating the direction similarity index of the abnormal pixel point according to the gradient direction of the abnormal pixel point and the direction of the connecting line;
calculating the direction similarity index of each abnormal pixel point by using the following calculation formula:
wherein, the first and the second end of the pipe are connected with each other,
is the direction similarity index of the ith abnormal pixel point,
is the gradient direction of the ith abnormal pixel point,
is the ith abnormal imageThe direction of the connecting line corresponding to the prime point;
the calculating the periodicity index of each abnormal pixel point according to the gray value of each abnormal pixel point comprises the following steps:
calculating the gray level of a normal pixel point corresponding to the gray image on the upper surface of the cone yarn;
for any abnormal pixel point: constructing a pixel gray value sequence with a set length by taking the abnormal pixel as a central point, constructing a reconstruction sequence corresponding to the pixel gray value sequence according to whether each gray value in the pixel gray value sequence is greater than the gray level of a normal pixel, and counting the jumping times of the reconstruction sequence; calculating the periodicity index of the abnormal pixel point according to the jumping times;
the method for constructing the reconstruction sequence corresponding to the pixel gray value sequence according to whether each gray value in the pixel gray value sequence is greater than the gray level of the normal pixel, and counting the jumping times of the reconstruction sequence comprises the following steps:
sequentially judging whether each gray value in the gray value sequence of the pixel point is greater than the gray level of the normal pixel point, and if so, recording a corresponding element in the reconstruction sequence as 1; if not, marking the corresponding element in the reconstruction sequence as 0;
counting the total times of changing elements in the reconstruction sequence from 1 to 0 and changing the elements in the reconstruction sequence from 0 to 1, and recording the total times as the hopping times of the reconstruction sequence;
the calculating the periodicity index of the abnormal pixel point according to the jumping times comprises the following steps:
wherein the content of the first and second substances,
is a periodic index of the abnormal pixel point, tanh is a hyperbolic tangent function,
in order to be a hyper-parameter,
the jumping times of the reconstruction sequence corresponding to the abnormal pixel point are counted;
calculating the probability that the cone yarn is the chrysanthemum core defect by using the following formula, comprising the following steps:
wherein N represents the number of abnormal pixel points,
the periodicity index of the nth abnormal pixel point is represented,
the direction similarity index of the nth abnormal pixel point is expressed,
the probability of the bobbin yarn being a chrysanthemum core defect.
2. The method for detecting defects of a cheese chrysanthemum flower core according to claim 1, wherein the step of obtaining abnormal pixel points in the gray image of the upper surface of the cheese according to the gray gradient of each pixel point in the gray image of the upper surface of the cheese comprises the following steps:
calculating each pixel point in gray level image of cone yarn by using sobel operator
And
a gradient of direction;
according to each pixel point
And
calculating the gradient amplitude of each pixel point by the directional gradient;
and judging whether the gradient amplitude of each pixel point is greater than or equal to a set gradient threshold value, if so, judging the corresponding pixel point as an abnormal pixel point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210815245.4A CN114897894B (en) | 2022-07-11 | 2022-07-11 | Method for detecting defects of cheese chrysanthemum core |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210815245.4A CN114897894B (en) | 2022-07-11 | 2022-07-11 | Method for detecting defects of cheese chrysanthemum core |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114897894A true CN114897894A (en) | 2022-08-12 |
| CN114897894B CN114897894B (en) | 2022-09-09 |
Family
ID=82729851
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210815245.4A Active CN114897894B (en) | 2022-07-11 | 2022-07-11 | Method for detecting defects of cheese chrysanthemum core |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114897894B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082467A (en) * | 2022-08-22 | 2022-09-20 | 山东亿昌装配式建筑科技有限公司 | Building material welding surface defect detection method based on computer vision |
| CN115272301A (en) * | 2022-09-20 | 2022-11-01 | 江苏新世嘉家纺高新科技股份有限公司 | Automatic cheese defect detection method based on robot |
| CN115311303A (en) * | 2022-10-12 | 2022-11-08 | 南通富兰妮纺织品有限公司 | Textile warp and weft defect detection method |
| CN115375588A (en) * | 2022-10-25 | 2022-11-22 | 山东旗胜电气股份有限公司 | Power grid transformer fault identification method based on infrared imaging |
| CN116823718A (en) * | 2023-04-17 | 2023-09-29 | 南通大学 | Method for classifying cone yarn defect images based on deep learning |
| CN117252874A (en) * | 2023-11-16 | 2023-12-19 | 深圳市京鼎工业技术股份有限公司 | Visual detection method for production quality of plastic mold |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002303584A (en) * | 2001-04-05 | 2002-10-18 | Toray Ind Inc | Surface fouling inspection method for wound yarn package |
| CN1814533A (en) * | 2005-02-01 | 2006-08-09 | 村田机械株式会社 | Method and device for unwinding yarn from package |
| CN101175878A (en) * | 2005-05-16 | 2008-05-07 | 村田机械株式会社 | Wadding detecting method and device in core yarn spinning |
| CN107102009A (en) * | 2017-05-04 | 2017-08-29 | 武汉理工大学 | A kind of method of the cylinder spool quality testing based on machine vision |
| CN109919939A (en) * | 2019-03-27 | 2019-06-21 | 王合山 | A kind of yarn faults detection method and device based on genetic algorithm |
-
2022
- 2022-07-11 CN CN202210815245.4A patent/CN114897894B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002303584A (en) * | 2001-04-05 | 2002-10-18 | Toray Ind Inc | Surface fouling inspection method for wound yarn package |
| CN1814533A (en) * | 2005-02-01 | 2006-08-09 | 村田机械株式会社 | Method and device for unwinding yarn from package |
| CN101175878A (en) * | 2005-05-16 | 2008-05-07 | 村田机械株式会社 | Wadding detecting method and device in core yarn spinning |
| CN107102009A (en) * | 2017-05-04 | 2017-08-29 | 武汉理工大学 | A kind of method of the cylinder spool quality testing based on machine vision |
| CN109919939A (en) * | 2019-03-27 | 2019-06-21 | 王合山 | A kind of yarn faults detection method and device based on genetic algorithm |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082467A (en) * | 2022-08-22 | 2022-09-20 | 山东亿昌装配式建筑科技有限公司 | Building material welding surface defect detection method based on computer vision |
| CN115272301A (en) * | 2022-09-20 | 2022-11-01 | 江苏新世嘉家纺高新科技股份有限公司 | Automatic cheese defect detection method based on robot |
| CN115311303A (en) * | 2022-10-12 | 2022-11-08 | 南通富兰妮纺织品有限公司 | Textile warp and weft defect detection method |
| CN115375588A (en) * | 2022-10-25 | 2022-11-22 | 山东旗胜电气股份有限公司 | Power grid transformer fault identification method based on infrared imaging |
| CN115375588B (en) * | 2022-10-25 | 2023-02-07 | 山东旗胜电气股份有限公司 | Power grid transformer fault identification method based on infrared imaging |
| CN116823718A (en) * | 2023-04-17 | 2023-09-29 | 南通大学 | Method for classifying cone yarn defect images based on deep learning |
| CN116823718B (en) * | 2023-04-17 | 2024-02-23 | 南通大学 | Method for classifying cone yarn defect images based on deep learning |
| CN117252874A (en) * | 2023-11-16 | 2023-12-19 | 深圳市京鼎工业技术股份有限公司 | Visual detection method for production quality of plastic mold |
| CN117252874B (en) * | 2023-11-16 | 2024-03-19 | 深圳市京鼎工业技术股份有限公司 | Visual detection method for production quality of plastic mold |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114897894B (en) | 2022-09-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114897894B (en) | Method for detecting defects of cheese chrysanthemum core | |
| CN113643289B (en) | Fabric surface defect detection method and system based on image processing | |
| CN110403232B (en) | Cigarette quality detection method based on secondary algorithm | |
| CN114170228B (en) | Computer image edge detection method | |
| CN113920086B (en) | Yarn state detection method and device in carbon fiber weaving process | |
| CN109540917B (en) | Method for extracting and analyzing yarn appearance characteristic parameters in multi-angle mode | |
| CN114842007A (en) | Textile wear defect detection method based on image processing | |
| CN113638104B (en) | Intelligent yarn cleaning control method and system for bobbin winder | |
| CN114998321B (en) | Textile material surface hairiness degree identification method based on optical means | |
| CN116309671B (en) | Geosynthetic fabric quality detection system | |
| CN114998341B (en) | Gear defect detection method and system based on template matching | |
| CN115330772A (en) | Online welding spot quality detection method | |
| CN116977342B (en) | PCB circuit detection method based on image segmentation | |
| CN115131348A (en) | Method and system for detecting textile surface defects | |
| CN115719344A (en) | Fabric defect analysis and identification method for textile fabric quality inspection | |
| CN116523899A (en) | Textile flaw detection method and system based on machine vision | |
| CN117237354B (en) | Visual detection method for defects of textile clothes | |
| CN114581376B (en) | Automatic sorting method and system for textile silkworm cocoons based on image recognition | |
| CN114923912A (en) | Machine vision-based method for detecting outer surface defects of nylon spinning cakes | |
| TW201211348A (en) | Yarn detecting device and method thereof | |
| CN114693652B (en) | Fabric Defect Detection Method Based on Gaussian Mixture Model | |
| CN115294097A (en) | Textile surface defect detection method based on machine vision | |
| CN104881652A (en) | Row number automatic detection algorithm based on corn ear convexity characteristic | |
| CN110857919A (en) | Tail yarn defect detection method for package filament | |
| CN113393444B (en) | Polyester DTY network point detection method based on image processing technology |
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 |