CN113643293B - High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence - Google Patents

High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence Download PDF

Info

Publication number
CN113643293B
CN113643293B CN202111198460.6A CN202111198460A CN113643293B CN 113643293 B CN113643293 B CN 113643293B CN 202111198460 A CN202111198460 A CN 202111198460A CN 113643293 B CN113643293 B CN 113643293B
Authority
CN
China
Prior art keywords
sub
brush roller
region
abnormal
regions
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
Application number
CN202111198460.6A
Other languages
Chinese (zh)
Other versions
CN113643293A (en
Inventor
沈拥军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Xiangshun Fabric Co ltd
Original Assignee
Jiangsu Xiangshun Fabric Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangsu Xiangshun Fabric Co ltd filed Critical Jiangsu Xiangshun Fabric Co ltd
Priority to CN202111198460.6A priority Critical patent/CN113643293B/en
Publication of CN113643293A publication Critical patent/CN113643293A/en
Application granted granted Critical
Publication of CN113643293B publication Critical patent/CN113643293B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06T7/001—Industrial image inspection using an image reference approach
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
    • G01D21/00—Measuring or testing not otherwise provided for
    • G01D21/02—Measuring two or more variables by means not covered by a single other subclass
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/40—Analysis of texture
    • G06T7/41—Analysis of texture based on statistical description of texture
    • G06T7/44—Analysis of texture based on statistical description of texture using image operators, e.g. filters, edge density metrics or local histograms
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C3/00—Registering or indicating the condition or the working of machines or other apparatus, other than vehicles
    • G07C3/005—Registering or indicating the condition or the working of machines or other apparatus, other than vehicles during manufacturing process
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Manufacturing & Machinery (AREA)
  • Probability & Statistics with Applications (AREA)
  • Quality & Reliability (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Finish Polishing, Edge Sharpening, And Grinding By Specific Grinding Devices (AREA)

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a high-speed positioning method for an abnormal area of a textile brush roller based on artificial intelligence. The method comprises the steps that a cylindrical surface image of a brush roller is divided into a plurality of sub-areas, and after the brush roller rotates 90 degrees each time, texture change strength and a load sequence of bristles corresponding to each sub-area are obtained; and calculating the similarity distance between every two sub-regions according to the texture change intensity and the load sequence to obtain a suspected abnormal sub-region, and when the suspected abnormal sub-region is the same sub-region continuously for multiple times, determining that the sub-region is an abnormal sub-region. The method comprises the following steps of dividing the brush roller into sub-regions to obtain anisotropic response and load sequences of texture changes of bristles corresponding to the sub-regions so as to reduce calculated amount and improve response time, confirming abnormal sub-regions through the anisotropic response and the load sequences, and giving out warning in time to avoid cloth damage.

Description

High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a high-speed positioning method for an abnormal area of a textile brush roller based on artificial intelligence.
Background
An industrial brush roller is a common brush type and is mainly used in the industries of textile, printing and dyeing and the like, and the types of the industrial brush roller comprise a polishing brush, an abrasive brush, a cleaning and cleaning type brush roller, a safety type brush roller and the like.
The brush of the brush roller is made of a cluster of nylon wires, horsehair, pig bristles, wool and other materials through a hair planting process, and the hair planting machine plants the bristles into the brush roller in a bundle, so that the brush with uniformly distributed texture is obtained. In the hair planting process of the industrial brush roller, the nylon material is mainly fixed by injection molding, animal hair is fixed by the processes of springs, pressing sheets and the like, and the hardness of the root of the bristles is far greater than that of the bristles in any material.
In the textile field, the brush roller is mainly used in the sweeping and sanding process. In the working process of the brush roller, the bristles can experience complex vibration, the long-time work of the bristles can lead to the aging of the fixed structure, so that the bristles fall off outwards or are inclined, the cleaning or polishing effect of the bristles is abnormal, the bristles are slightly cleaned and polished insufficiently, and the bristles fall off seriously. When the steel wires at the tail parts of the bristles and the welded hard plastics are in direct contact with the cloth, the fabric can be scratched, and the cloth is easily damaged.
The brush roller is a high-speed rotating mechanism, when the brush bristles are inclined or fall off, the defect detection system with high real-time performance and high accuracy is not available at present. The main problem lies in that the equipment cost based on image detection brush roller is higher at present, and the required refresh rate of gathering the brush roller image is higher.
Therefore, the traditional visual detection method cannot detect the fault position immediately when the bristles are inclined and fall off on the premise of reducing cost and improving calculation efficiency.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a high-speed positioning method for an abnormal area of a textile brush roller based on artificial intelligence, which adopts the following technical scheme:
the embodiment of the invention provides a high-speed positioning method for an abnormal area of a textile brush roller based on artificial intelligence, which comprises the following specific steps of:
acquiring a cylindrical image of the brush roller and a corresponding motor load at each rotation angle; segmenting the cylindrical image into a plurality of sub-regions, obtaining texture change intensity of bristles corresponding to each sub-region after the brush roller rotates 90 degrees each time, and obtaining a load sequence of each sub-region through the motor load;
calculating similarity distances between every two sub-regions according to the texture change strength and the load sequence, and dividing the sub-regions into normal samples and abnormal samples according to the similarity distances; obtaining a suspected abnormal subarea according to the similarity distance in the abnormal sample;
and when the suspected abnormal subareas are the same subarea continuously for multiple times, confirming that the subarea is an abnormal subarea.
Preferably, the method for dividing the lenticular image into a plurality of sub-regions includes:
averagely dividing the cylindrical image into sub-cylindrical images corresponding to a plurality of sub-cylinders based on the cylinder shape of the brush roller;
and dividing each sub-cylinder image into a plurality of sub-areas according to the angle average of the sub-cylinders.
Preferably, the method for acquiring the texture variation strength of the corresponding bristles of each sub-area comprises the following steps:
carrying out difference according to the new texture and the old texture in the sub-region to obtain a brightness difference image, wherein the brightness difference image represents the brightness change value of the sub-region before and after the rotation angle change;
and after the brightness difference image is subjected to Sobel operator processing in the X direction and the Y direction respectively, obtaining the texture change strength of the brush hair in the transverse direction and the longitudinal direction respectively according to pixel values in the processed brightness difference image, wherein the X direction represents the axial direction of the brush roller, and the Y direction represents the rotation direction of the brush roller.
Preferably, the method for acquiring the motor load includes:
obtaining the reference power of the motor according to the rotational inertia of the brush roller;
and acquiring the output power of the motor at each rotation angle, and combining the reference power and the output power to obtain the motor load.
Preferably, the method for obtaining a suspected abnormal sub-region according to the similarity distance in the abnormal sample includes:
setting a similarity distance threshold, reserving the sub-region corresponding to the abnormal sample, wherein the similarity distance is greater than the similarity distance threshold, and taking the reserved sub-region as the suspected abnormal sub-region.
Preferably, the calculation formula for calculating the similarity distance between each two sub-regions from the texture variation strength and the load sequence is as follows:
Figure 796810DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 307426DEST_PATH_IMAGE002
is as follows
Figure 637563DEST_PATH_IMAGE003
In the sub-cylindrical image
Figure 95089DEST_PATH_IMAGE004
Each of said sub-regions;
Figure 50407DEST_PATH_IMAGE005
is as follows
Figure 420339DEST_PATH_IMAGE006
In the sub-cylindrical image
Figure 156214DEST_PATH_IMAGE007
Each of said sub-regions;
Figure 214169DEST_PATH_IMAGE008
is the similarity distance;
Figure 430518DEST_PATH_IMAGE009
is the cosine similarity;
Figure 892723DEST_PATH_IMAGE010
is the sum of the L2 distances after dynamic time warping of the two load sequences;
Figure 506107DEST_PATH_IMAGE011
is the loading sequence;
Figure 711961DEST_PATH_IMAGE012
varying the intensity for the texture.
Preferably, the method for obtaining the motor load by combining the reference power and the output power includes:
and taking the ratio of the reference power to the output power as the motor load.
The embodiment of the invention at least has the following beneficial effects: the method comprises the following steps of dividing a brush roller into subareas to obtain anisotropic response and load sequences of texture changes of bristles corresponding to the subareas so as to reduce the calculated amount and improve the response time; and then, classifying the subareas according to the similarity of anisotropic response and the aperiodic motor load characteristic to obtain a suspected abnormal subarea, so that the abnormal subarea is determined, and a warning is given out in time to avoid cloth damage.
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 flow chart of steps of a method for high-speed positioning of abnormal areas of a textile brush roller based on artificial intelligence according to an embodiment of the invention;
FIG. 2 is a schematic diagram of a linear scanning camera for image acquisition of a brush roller according to an embodiment of the present invention;
fig. 3 is a schematic diagram of dividing sub-regions with respect to a cylindrical image according to an embodiment of the present invention.
Detailed Description
In order 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 specific implementation, structure, features and effects of the artificial intelligence based high-speed positioning method for the abnormal area of the textile brush roller according to the present invention with reference to the accompanying drawings and preferred embodiments. 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 specific scheme of the high-speed positioning method for the abnormal area of the textile brush roller based on artificial intelligence is specifically described below by combining the attached drawings.
Referring to fig. 1, a flow chart of the steps of a method for high-speed positioning of an abnormal area of a textile brush roller based on artificial intelligence according to an embodiment of the present invention is shown, wherein the method comprises the following steps:
s001, acquiring a cylindrical image of the brush roller and a corresponding motor load at each rotation angle; the cylindrical image is divided into a plurality of sub-regions, after the brush roller rotates 90 degrees each time, the texture change intensity of the bristles corresponding to each sub-region is obtained, and the load sequence of each sub-region is obtained through the load of the motor.
Specifically, an absolute value encoder is installed on the shaft side of the brush roller for obtaining the rotation angle of the current brush roller in real time, and the rotation angle range is [0,360] or [ -180,180], and the range is determined by the implementer in the actual implementation.
It should be noted that, since the brush roller is driven by the motor, the absolute value encoder may be mounted on the electrode.
Based on the angle of rotation
Figure 779886DEST_PATH_IMAGE013
And obtaining a cylindrical image of the brush roller by using a linear scanning camera. Calibrating rotation angle
Figure 68785DEST_PATH_IMAGE013
Is offset amount of
Figure 779252DEST_PATH_IMAGE014
And the angle corresponding to one side of the brush roller attached to the cloth is aligned with the corresponding angle during image acquisition so as to ensure the data loss.
Preferably, the offset in the embodiment of the present invention
Figure 601846DEST_PATH_IMAGE015
The implementer can set the setting according to the self-requirement.
Referring to fig. 2, based on the position of the linear scanning camera 1, when the brush roller rotates, the linear scanning camera 1 collects the cylindrical image area of the brush roller 2 in real time to update the line data in the cylindrical image, and records the brightness of the new line and the brightness difference between the new line and the old line, and the specific process is as follows:
(1) in the embodiment of the invention, the line division rate of the linear scanning camera is 2048, and the resolution of the cylindrical image collected by the brush roller is 2048 lines. When the absolute value encoder enters the next angle, the linear scanning camera newly acquires a row of pixels to be superimposed on the cylindrical image so as to update the cylindrical image.
(2) Recording the current pixel value and the pixel value before updating according to the cylindrical image areas before and after the change of the rotation angle
Figure 838792DEST_PATH_IMAGE016
。
(3) And in the process of rotating the brush roller for one circle, obtaining a new cylindrical image and an old cylindrical image in real time according to the change of the rotating angle, namely the corresponding cylindrical images before and after the pixel value is updated.
It should be noted that, when the cylindrical image is unfolded, since the two ends cannot be connected together, several rows of adjacent rows need to be additionally assigned, so that the features at the seams are continuous.
Preferably, in the embodiment of the present invention, the seams of the cylindrical images are supplemented with 32 rows of data at the other end.
Further, when the brush roller rotates, the linear scanning camera collects and updates the cylindrical image area in real time, and meanwhile collects the corresponding motor load under the current rotation angle
Figure 908379DEST_PATH_IMAGE011
Then, the method for calculating the motor load is as follows:
(1) because the brush roller has certain rotational inertia, when the motor drives the brush roller to be in a working state, a reference work is obtained according to the rotational inertia of the brush rollerRate of change
Figure 184771DEST_PATH_IMAGE017
。
(2) When the bristles of the brush roller fall off or are displaced, additional resistance is generated when the bristles impact the cloth. Because the motor controller is controlled by PID at a constant speed, the power of the motor driving the brush roller can be changed, the power of the collected motor is slightly higher than the reference power, and the closed-loop control of the PID can cause the power to generate certain oscillation. In order to analyze the power change characteristics conveniently, the motor load is obtained by combining the output power of the current motor controller and the reference power, and the ratio of the output power obtained in real time to the reference power is used as the corresponding motor load under the rotation angle, namely the motor load
Figure 122640DEST_PATH_IMAGE018
Wherein, in the step (A),
Figure 89459DEST_PATH_IMAGE019
is the output power.
Because the resistance of the motor for driving the brush roller can be instantly increased after the bristles impact the cloth, but due to the rotational inertia and the hysteresis characteristic of PID closed-loop control, the corresponding rotation angle of the obtained motor load P when the first increase occurs and the corresponding rotation angle of the bristles when the cloth impact event occurs generate certain delay, so that the motor load cannot be analyzed only
Figure 408576DEST_PATH_IMAGE011
The size of the image data is used to determine whether an abnormal brushing region or an abnormal angle interval occurs, so the embodiment of the invention divides the cylindrical image into a plurality of sub-regions to obtain the anisotropic response of the texture change of each sub-region and the corresponding load change so as to reduce the calculated amount and improve the response time, and the specific process is as follows:
(1) the cylindrical image of the brush roller is a cluster of bristles, the texture of the brush roller has a certain fixed repetition rule, and when the bristles fall off or shift, a suspected abnormal bristle area can be obtained from the cylindrical image. And because the distance between the bristles is large, the bristles are long, and irregular shaking can occur due to the elasticity of the bristles during the rotation of the brush roller, so that the collected cylindrical images cannot directly analyze the positions of abnormal falling or deviation of the bristles through textures, the cylindrical images are averagely divided into sub-cylindrical images corresponding to a plurality of sub-cylinders based on the cylinder shape of the brush roller, and each sub-cylindrical image is averagely divided into a plurality of sub-regions according to the angle of the sub-cylinder.
Specifically, referring to fig. 3, in the embodiment of the present invention, the cylindrical image is divided into 8 sub-cylindrical images, which may also be understood as 8 sub-cylinders, and each sub-cylinder is further divided into 4 sub-regions according to an angle.
It should be noted that the implementer of the number of the sub-cylindrical images can select the number according to the rotating speed of the brush roller in practical application; the number of sub-areas is selected according to the actual performance of the motor controller, and for a general brushless motor, the number of sub-areas is not suitable to be too large, otherwise the motor load can be reduced
Figure 952690DEST_PATH_IMAGE011
Too large hysteresis of (a) results in the motor load characteristic represented by the sub-zone not being able to correspond, typically 4 sub-zones are suitable.
(2) For 4 sub-regions of each sub-cylinder, after each 90 degrees rotation of the brush roller, then the anisotropic response and corresponding load variation of the texture variation for each sub-region is:
in particular, in the following
Figure 569616DEST_PATH_IMAGE003
In the subsidiary cylinder
Figure 484218DEST_PATH_IMAGE004
Image corresponding to sub-region
Figure 98870DEST_PATH_IMAGE020
For example, for an image
Figure 599122DEST_PATH_IMAGE020
The new and old textures in the image are differentiated to obtain brightnessThe brightness difference image represents the brightness change value of the sub-region before and after the rotation angle change, Sobel operator processing in the X direction and Sobel operator processing in the Y direction are respectively carried out on the brightness difference image, the size of a convolution kernel is 3X 3, and the texture change strength of the transverse direction and the longitudinal direction of the brush hair is obtained
Figure 488580DEST_PATH_IMAGE021
For characterizing the texture of bristles, wherein the texture varies in intensity
Figure 508620DEST_PATH_IMAGE022
Is the sum of the pixel values of all the pixel points in the convolved luminance difference image,
Figure 481124DEST_PATH_IMAGE023
the magnitude of the response representing the change in texture in the X direction,
Figure 953825DEST_PATH_IMAGE024
representing the response magnitude of the Y-direction texture change. Meanwhile, a load sequence of the corresponding position of the sub-area is obtained according to the motor load obtaining method, namely the load sequence is obtained
Figure 646975DEST_PATH_IMAGE025
Wherein, the length of the loading sequence is related to the rotation angle of the brush roller and can be set by an implementer according to the requirements of the implementer.
The X direction indicates the axial direction of the brush roller, and the Y direction indicates the rotational direction of the brush roller.
(3) By using the method in step (2), the texture variation strength and load sequence of all the sub-regions can be obtained.
Step S002, calculating the similarity distance between every two sub-regions according to the texture change intensity and the load sequence, and dividing the sub-regions into normal samples and abnormal samples based on the similarity distance; and obtaining a suspected abnormal subarea according to the similarity distance in the abnormal sample.
Specifically, the sub-regions are compared with each other through the texture change strength and the load sequence to determine a suspected abnormal sub-region, that is, a suspected abnormal bristle region, and the method includes:
(1) calculating the similarity distance between every two subregions according to the texture change intensity and the load sequence, namely:
Figure 567526DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 396417DEST_PATH_IMAGE002
is as follows
Figure 277786DEST_PATH_IMAGE003
In the sub-cylindrical image
Figure 368101DEST_PATH_IMAGE004
A sub-region;
Figure 221788DEST_PATH_IMAGE005
is as follows
Figure 958931DEST_PATH_IMAGE006
In the sub-cylindrical image
Figure 327595DEST_PATH_IMAGE007
A sub-region;
Figure 752760DEST_PATH_IMAGE008
is the similarity distance;
Figure 523270DEST_PATH_IMAGE009
is the cosine similarity;
Figure 431315DEST_PATH_IMAGE010
is the sum of the L2 distances after dynamic time warping of the two load sequences;
Figure 677488DEST_PATH_IMAGE011
is a loading sequence;
Figure 597689DEST_PATH_IMAGE012
the intensity is varied for the texture.
It should be noted that the DTW for the load sequence is to: the motor load of the brush roller in the rotating process has certain high-frequency periodicity due to manufacturing reasons in normal work, so that if the load L2 distances of two different sub-areas are directly calculated, samples are not aligned, and the problem that the sum of the L2 distances is too large occurs, the sum of the L2 distances after dynamic time warping processing can represent the aperiodic load of the brush roller area, namely:
Figure 222706DEST_PATH_IMAGE026
larger means that the load is different between this sub-zone and the other sub-zones, whereas the load pattern is similar.
It is noted that two sub-regions are calculated
Figure 816498DEST_PATH_IMAGE027
The meaning of (A) is: since the brush is oscillated during normal rotation, the luminance difference image of each sub-region is subjected to Sobel operator processing, so that a directional response in which a luminance change occurs between the axial direction (X direction) and the rotational direction (Y direction) is obtained. Between normal subregions
Figure 35121DEST_PATH_IMAGE028
Tends to 1 because the bristle distribution patterns of the two sub-regions are similar, so the anisotropic response between the vectors is more uniform; whereas it means that there is a sub-area in which an exception has occurred,
Figure 5351DEST_PATH_IMAGE028
tending to 0.
(2) Based on the similarity distance, classifying the sub-regions by utilizing a binary K-Means algorithm to obtain two clusters, wherein one cluster is a normal sample, and the other cluster is an abnormal sample. Setting similarity distance threshold
Figure 484874DEST_PATH_IMAGE029
For similarity distance in abnormal sample
Figure 62617DEST_PATH_IMAGE030
And (3) removing the sub-region, reserving the sub-region larger than the similarity distance threshold, considering that the reserved sub-region is a suspected abnormal sub-region, and counting the sub-region.
It should be noted that, in terms of objective rules, for a brush roller, when an abnormal region occurs, the specific conditions are as follows: only one sub-region of the whole brush roller is abnormal, so that once the similarity distance between every two sub-regions is smaller than the similarity distance threshold value, the two sub-regions are considered to be ignored.
And step S003, when the suspected abnormal subareas are the same subarea continuously for multiple times, confirming that the subarea is the abnormal subarea.
Specifically, a counting threshold value N is set, a suspected abnormal sub-region which can be confirmed once can be obtained when the brush roller rotates for one circle, and when the continuous counting value of one sub-region is smaller than the counting threshold value N, the sub-region is considered not to be an abnormal sub-region; on the contrary, when the continuous counting value of a certain sub-region is larger than or equal to the counting threshold value N, the sub-region is considered to be an abnormal sub-region, the abnormal condition that the bristles fall off or shift occurs at the position corresponding to the brush roller in the sub-region is described, and an implementer can close the brush roller motor and cloth feeding according to the confirmation result, so that the equipment gives a warning.
Preferably, in the embodiment of the present invention, the count threshold N = 10.
In summary, the embodiment of the invention provides a high-speed positioning method for an abnormal area of a textile brush roller based on artificial intelligence, which is used for acquiring a cylindrical image of the brush roller and a corresponding motor load at each rotation angle; the cylindrical surface image is divided into a plurality of sub-regions, after the brush roller rotates 90 degrees each time, the texture change intensity of the bristles corresponding to each sub-region is obtained, and the load sequence of each sub-region is obtained through the load of the motor; calculating the similarity distance between every two sub-regions according to the texture change intensity and the load sequence, and dividing the sub-regions into normal samples and abnormal samples based on the similarity distance; obtaining a suspected abnormal subarea according to the similarity distance in the abnormal sample; and when the suspected abnormal subarea is the same subarea continuously for multiple times, confirming that the subarea is the abnormal subarea. The method comprises the following steps of dividing a brush roller into subareas to obtain anisotropic response and load sequences of texture changes of bristles corresponding to the subareas so as to reduce the calculated amount and improve the response time; and then, classifying the subareas according to the similarity of anisotropic response and the aperiodic motor load characteristic to obtain a suspected abnormal subarea, so that the abnormal subarea is determined, and a warning is given out in time to avoid cloth damage.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (1)

1. A high-speed positioning method for abnormal areas of a textile brush roller based on artificial intelligence is characterized by comprising the following steps:
acquiring a cylindrical surface image of a brush roller and a corresponding motor load at each rotation angle, wherein the motor load is obtained by reference power of a motor according to the rotational inertia of the brush roller, the output power of the motor at each rotation angle is acquired, and the ratio of the reference power to the output power is the motor load; averagely dividing the cylindrical image into sub-cylindrical images corresponding to a plurality of sub-cylinders based on the cylinder shape of the brush roller; equally dividing each sub-cylindrical surface image into a plurality of sub-regions according to the angle of the sub-cylinder, obtaining texture change intensity of bristles corresponding to each sub-region after the brush roller rotates for 90 degrees each time, and obtaining a load sequence of each sub-region through the motor load;
calculating similarity distances between every two sub-regions according to the texture change strength and the load sequence, and dividing the sub-regions into normal samples and abnormal samples according to the similarity distances; setting a similarity distance threshold, reserving the sub-region corresponding to the similarity distance greater than the similarity distance threshold in the abnormal sample, and taking the reserved sub-region as a suspected abnormal sub-region;
when the suspected abnormal subareas are the same subarea continuously for multiple times, confirming that the subarea is an abnormal subarea;
the method for acquiring the texture variation intensity of the bristles corresponding to each sub-area comprises the following steps: carrying out difference according to the new texture and the old texture in the sub-region to obtain a brightness difference image, wherein the brightness difference image represents the brightness change value of the sub-region before and after the rotation angle change; after Sobel operator processing is carried out on the brightness difference image in the X direction and the Y direction respectively, the texture change strength of the brush hair in the transverse direction and the longitudinal direction is obtained respectively through pixel values in the processed brightness difference image, wherein the X direction represents the axial direction of the brush roller, and the Y direction represents the rotating direction of the brush roller;
wherein the calculation formula for calculating the similarity distance between each two sub-regions by the texture change strength and the load sequence is as follows:
Figure 100918DEST_PATH_IMAGE002
in the formula (I), the compound is shown in the specification,
Figure DEST_PATH_IMAGE003
is as follows
Figure 221321DEST_PATH_IMAGE004
In the sub-cylindrical image
Figure DEST_PATH_IMAGE005
Each of said sub-regions;
Figure 328954DEST_PATH_IMAGE006
is as follows
Figure DEST_PATH_IMAGE007
In the sub-cylindrical image
Figure 534283DEST_PATH_IMAGE008
Each of said sub-regions;
Figure DEST_PATH_IMAGE009
is the similarity distance;
Figure 676551DEST_PATH_IMAGE010
is the cosine similarity;
Figure DEST_PATH_IMAGE011
is the sum of the L2 distances after dynamic time warping of the two load sequences;
Figure 312063DEST_PATH_IMAGE012
is the loading sequence;
Figure DEST_PATH_IMAGE013
varying the intensity for the texture.
CN202111198460.6A 2021-10-14 2021-10-14 High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence Active CN113643293B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111198460.6A CN113643293B (en) 2021-10-14 2021-10-14 High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111198460.6A CN113643293B (en) 2021-10-14 2021-10-14 High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence

Publications (2)

Publication Number Publication Date
CN113643293A CN113643293A (en) 2021-11-12
CN113643293B true CN113643293B (en) 2022-02-11

Family

ID=78426860

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111198460.6A Active CN113643293B (en) 2021-10-14 2021-10-14 High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence

Country Status (1)

Country Link
CN (1) CN113643293B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114166858B (en) * 2022-02-11 2022-04-12 海门市芳华纺织有限公司 Method for detecting leather scratching area of textile brush roller based on artificial intelligence
CN115129760B (en) * 2022-08-30 2022-11-15 南通博纳纺织品有限公司 Artificial intelligence-based method and system for determining feeding rate of brushing machine

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110334722A (en) * 2019-03-29 2019-10-15 上海联影智能医疗科技有限公司 Image classification method, device, computer equipment and storage medium

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110182495A1 (en) * 2010-01-26 2011-07-28 General Electric Company System and method for automatic defect recognition of an inspection image

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110334722A (en) * 2019-03-29 2019-10-15 上海联影智能医疗科技有限公司 Image classification method, device, computer equipment and storage medium

Also Published As

Publication number Publication date
CN113643293A (en) 2021-11-12

Similar Documents

Publication Publication Date Title
CN113643293B (en) High-speed positioning method for abnormal area of textile brush roller based on artificial intelligence
CN113638104B (en) Intelligent yarn cleaning control method and system for bobbin winder
CN113724241B (en) Broken filament detection method and device for carbon fiber warp-knitted fabric and storage medium
JP4152853B2 (en) Surface treatment apparatus and method for manufacturing liquid crystal display device
CN113570598B (en) Textile brush roller wear analysis and service life prediction method based on artificial intelligence
CN112634198B (en) Machine vision-based dynamic recognition and detection method for Taylor cone by near-field electrospinning direct writing
TW570843B (en) Cleaning tool
CN103243532A (en) Fine cleaning roller
CN210314988U (en) Automatic change cleaning device for spinning machine
US4308649A (en) Apparatus to pattern brush pile fabric
CN205523053U (en) Circular screen printer device with automatically cleaning water washing device
CN115294165B (en) Intelligent operation method for textile singeing process based on machine vision
Peng et al. A fast detection scheme for original fabric based on Blob, canny and rotating integral algorithm
CN210682799U (en) Yarn pulling device for knitted fabric
TWI806845B (en) Carding machine with card flat
EP1178145A3 (en) Machine for softening the surface of a fabric
CN113718477A (en) Washing control method of washing machine
CN213389309U (en) Hair falling device
CN117188135A (en) Fabric weft straightening method based on machine vision
CN214724232U (en) Intelligent cleaning robot system guided by 3D vision
CN111667534B (en) Quick subpixel dent positioning method and device applied to high-speed drawing-in machine
CN114166858B (en) Method for detecting leather scratching area of textile brush roller based on artificial intelligence
CN115323562B (en) Multifunctional environment-friendly warp knitted fabric spinning equipment and method
CN219824250U (en) Carding device for fur processing
CN117173162B (en) Textile flaw detection method and system

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