CN113433137B - Textile cloth surface flaw detection device based on YOLO neural network - Google Patents
Textile cloth surface flaw detection device based on YOLO neural network Download PDFInfo
- Publication number
- CN113433137B CN113433137B CN202110985863.9A CN202110985863A CN113433137B CN 113433137 B CN113433137 B CN 113433137B CN 202110985863 A CN202110985863 A CN 202110985863A CN 113433137 B CN113433137 B CN 113433137B
- Authority
- CN
- China
- Prior art keywords
- cloth
- submodule
- flaw
- module
- image
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
- G01N21/898—Irregularities in textured or patterned surfaces, e.g. textiles, wood
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
Abstract
The invention discloses a textile cloth surface flaw detection device based on a YOLO neural network, which comprises a textile cloth surface flaw detection device and a cloth flaw detection system, and is characterized in that: the textile cloth surface flaw detection device comprises a fixed base plate, rollers are arranged at two ends of the fixed base plate, two groups of rollers are connected with inner bearings, connecting plates are fixed at two ends of each rotating shaft and fixed with the fixed base plate, two groups of supporting rods are fixed at two sides of the fixed base plate, an upper fixed plate is arranged above the fixed base plate and fixed with the supporting rods, two groups of fixed sliding rods are fixed at the bottom of the upper fixed plate, a movable plate is connected to one side of each of the two groups of fixed sliding rods in a sliding mode, a camera is fixed on the bottom surface of the movable plate, and a motor is electrically connected to one side of each roller.
Description
Technical Field
The invention relates to the technical field of textile cloth surface flaw detection devices, in particular to a textile cloth surface flaw detection device based on a YOLO neural network.
Background
At present, textile and clothing production enterprises mainly stand before cloth inspection equipment through professional cloth inspectors to find cloth cover defects through naked eyes and then mark the defects, along with continuous optimization of computer vision detection technology, an intelligent cloth inspection system greatly reduces labor intensity of manual detection, and efficiency and precision of monitoring of fabric quality in a production process are improved.
Disclosure of Invention
The invention aims to provide a textile cloth surface flaw detection device based on a YOLO neural network, so as to solve the problems in the background technology.
In order to solve the technical problems, the invention provides the following technical scheme: weaving cloth surface flaw detection device based on YOLO neural network, including weaving cloth surface flaw detection device and cloth flaw detection system, its characterized in that: the textile cloth surface flaw detection device comprises a fixed bottom plate, rollers are arranged at two ends of the fixed bottom plate, the inner bearings of the two groups of rollers are connected with a rotating shaft, two ends of the rotating shaft are both fixed with connecting plates which are fixed with a fixed bottom plate, two groups of supporting rods are both fixed on two sides of the fixed bottom plate, an upper fixing plate is arranged above the fixing bottom plate and fixed with the supporting rods, two groups of fixing slide bars are fixed at the bottom of the upper fixing plate, a moving plate is connected to one side of each group of fixing slide bars in a sliding manner, a camera is fixed on the bottom surface of the moving plate, one side of each roller is electrically connected with a motor, cloth is arranged on the surfaces of the two groups of rollers, the equal sliding connection in fixed baseplate's both sides has the swing arm, the inside of swing arm is provided with a plurality of knitting needles, and is a plurality of knitting needle is connected with the braided wire of different colours.
According to the technical scheme, the cloth flaw detection system comprises a flaw detection module, a data analysis module and a control module, the flaw detection module comprises a memory sub-module, an irradiation sub-module and a projection receiving sub-module, the projection receiving sub-module comprises a time recording unit and a projection moving detection unit, the control module comprises a pause sub-module, a moving control sub-module and a shooting sub-module, the irradiation sub-module is positioned on two sides of the bottom of an upper fixing plate, the projection receiving sub-module is positioned on the upper surface of a fixing bottom plate, the shooting sub-module is positioned on the surface of a camera, the time recording unit is electrically connected with the pause sub-module, the moving control sub-module is electrically connected with a moving plate, the memory sub-module is used for recording a cloth image without flaws in a neural network, and a comparison effect can be achieved on subsequent work, the illumination submodule is used for irradiating light to the surface of the cloth, the projection receiving submodule is used for receiving the projection of the light finding the surface of the cloth, the time recording unit is used for recording the time of the projection of the cloth to the projection receiving submodule, the suspension submodule is used for suspending the rotation of a roller driven by a motor, the movement control submodule is used for receiving the data of the suspension submodule and controlling the movable plate to move on a fixed sliding rod, the shooting submodule is used for shooting the surface of the cloth and generating the existing cloth image of the cloth, and the projection movement detection unit is used for sensing whether the projection of the cloth moves.
According to the technical scheme, the flaw detection module and the control module comprise the following operation steps:
s1, before the woven cloth surface flaw detection device is started, a worker stores the image of the cloth without flaws into a memory sub-module;
s2, starting the woven cloth surface flaw detection device by a worker, driving the roller to rotate by the motor, transporting the cloth, irradiating the light onto the cloth by the irradiation submodule in the cloth transportation process, and receiving the projection by the projection receiving submodule;
s3, the time recording unit is used for counting the time of the cloth projection movement generated by the projection movement detection unit and generating a time value so as to count the running length of the cloth, and the time value is transmitted to the pause submodule when the running moving length of the cloth is long;
s4, the pause submodule is used for receiving the time value and transmitting a signal to the motor to stop the motor from rotating, and the cloth also stops rotating;
s5, the pause submodule transmits a motor pause signal to the movement control submodule so as to control the moving plate 8 to move on the fixed sliding rod;
and S6, when the moving plate moves, the camera drives the shooting submodule to shoot the surface of the cloth which stops running, and an existing cloth image is generated.
According to the technical scheme, the
Is the length of the running cloth.
According to the technical scheme, the data analysis module comprises an image recording submodule, a statistics submodule and an alarm submodule, wherein the image recording submodule comprises a flaw identification unit, an image comparison unit and a position analysis unit, the image recording submodule is electrically connected with the shooting submodule, the memory submodule is electrically connected with the image recording submodule, the alarm submodule is electrically connected with the motor, and the position analysis unit is electrically connected with the statistics submodule;
the image recording submodule is used for receiving the image of the flawless cloth of the memory submodule, the statistics submodule is used for counting the proportion of the flawless area of the cloth and generating the proportion numerical value of the flawed cloth area, the alarm submodule is used for receiving the numerical value of the flawed cloth area, the alarm submodule can give an alarm when the flawed cloth area is large, and transmits a signal to the motor, so that the motor stops running, a worker can conveniently observe the quality of the cloth and can timely replace the cloth, the image comparison unit compares the existing cloth image with the flawless cloth image and generates comparison data, the flaw identification unit is used for receiving the comparison data and identifying whether the flaw is contained in the image of the existing cloth, and the position analysis unit is used for analyzing the position of the flaw.
According to the technical scheme, the data analysis module comprises the following operation steps:
a1, transmitting the flawless cloth image to an image recording submodule, and transmitting the existing cloth image shot by the shooting submodule to the image recording submodule;
a2, comparing the existing cloth image with the flawless image by the image comparison unit, so that the flaw identification unit can identify the flaw area of the existing cloth image and generate a numerical ratio of the flaw area;
a3, the statistical unit receives the flaw area numerical value ratio and transmits the flaw area numerical value to the alarm submodule, when the flaw area numerical value is more, the alarm submodule reports, and when the flaw area numerical value is less, the cloth can be continuously transmitted;
a4, the position analysis unit transmits the flaw positions to the statistics submodule, so that the statistics submodule can perform statistics on the flaw positions on the cloth with fewer flaws, and when the transmission of the cloth with small flaw area values is completed, workers can conveniently find the flaw positions directly to perform repairing;
and A5, when the worker needs to repair the cloth, the edge of the cloth is set as the origin, and the defect position is searched according to the defect position.
According to the above technical solution, in the step a3, the numerical ratio of the flaw area is calculated by the following formula:
in the formula (I), the compound is shown in the specification,
the numerical ratio of the area of the flaw is,
the number of occurrences of the defect area, C the defect area,
is the width of the cloth when
>When 5%, the signal is transmitted to the alarm submodule to stop the motor, and when the signal is detected to be abnormal, the motor is stopped
<And when the percentage of the cloth is 5 percent, the cloth is continuously conveyed.
According to the above technical solution, in the step a4, the defect position identification formula is:
in the formula, l is the flaw length distance, k is the flaw width distance, x is the length distance of the flaw from the edge of the cloth, y is the width distance of the flaw from the edge of the cloth, and j is 1-n.
According to the technical scheme, the repairing module comprises a positioning submodule, a weaving submodule and an after-treatment submodule, the positioning submodule is electrically connected with the counting submodule, the positioning submodule is electrically connected with the swing arm, the weaving submodule is electrically connected with the flaw identification unit, the positioning submodule is used for receiving flaw position data of the counting submodule, the swing arm moves on the fixed base plate to find the position of the flaw, the weaving submodule receives the flaw of the existing cloth identified by the flaw identification unit, the weaving needle is controlled to repair the position of the cloth flaw, and the after-treatment submodule is used for arranging the edge after sewing so as to guarantee the quality of the cloth image.
According to the technical scheme, the repairing module comprises the following operation steps:
b1, the position analysis unit transmits the position of the flaw to the positioning submodule, so that the positioning submodule drives the swing arm to move on the fixed bottom plate to find the position of the flaw;
b2, the positioning sub-module transmits data to the knitting sub-module after finding the flaw position, and the flaw identification unit also transmits the flaw image of the existing cloth to the knitting sub-module, because the flaw identification unit is compared with the non-flaw cloth, it can be known what color is missing in the flaw place of the existing cloth, so as to control the knitting needles with different color textile lines, and select the knitting needle of the color textile line corresponding to the missing color of the cloth to knit the flaw position of the existing cloth;
b3, after weaving, post-processing the woven edge through a post-processing submodule to ensure the quality of the woven cloth image;
b4, transmitting the data to an irradiation submodule after the flaw positions are woven, and enabling the irradiation submodule to shoot the cloth again;
b5, repeating the steps A1-A4, identifying the positions of the cloth defects again, and checking whether the repaired defects are combined.
Compared with the prior art, the invention has the following beneficial effects: according to the invention, by arranging the roller, when the textile cloth production of the cloth neural network is finished, the device is required to detect the flaws on the surface of the cloth, the cloth is placed on the roller 5, the motor is started to rotate, the roller drives the cloth to be transported, the movable plate moves on the fixed slide bar and drives the camera to shoot the surface of the cloth to obtain the cloth image, the upper fixed plates connected with the support rods on the two sides of the fixed bottom plate can ensure the stability of the whole device, and the connecting plate can be connected with the roller to increase the stability of the roller.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a schematic view of the overall front perspective of the present invention;
FIG. 2 is a schematic view of the overall rear perspective of the present invention;
FIG. 3 is a schematic view of a cloth defect detection system of the present invention;
fig. 4 is a partially enlarged view of the region a of the present invention.
In the figure: 1. fixing the bottom plate; 2. an upper fixing plate; 3. a rotating shaft; 4. a connecting plate; 5. a drum; 6. a support bar; 7. fixing the sliding rod; 8. moving the plate; 9. a camera; 10. swinging arms; 11. and (6) knitting needles.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-4, the present invention provides the following technical solutions: weaving cloth surface flaw detection device based on YOLO neural network, including weaving cloth surface flaw detection device and cloth flaw detection system, its characterized in that: the textile cloth surface flaw detection device comprises a fixed base plate 1, rollers 5 are arranged at two ends of the fixed base plate 1, rotating shafts 3 are connected to internal bearings of two groups of rollers 5, connecting plates 4 are fixed at two ends of each rotating shaft 3, the connecting plates 4 are fixed with the fixed base plate 1, two groups of supporting rods 6 are fixed at two sides of the fixed base plate 1, an upper fixed plate 2 is arranged above the fixed base plate 1, the upper fixed plate 2 is fixed with the supporting rods 6, two groups of fixed sliding rods 7 are fixed at the bottom of the upper fixed plate 2, a movable plate 8 is connected to one side of each group of fixed sliding rods 7 in a sliding manner, a camera 9 is fixed on the bottom surface of the movable plate 8, a motor is electrically connected to one side of each roller 5, cloth is arranged on the surfaces of the two groups of rollers 5, swing arms 10 are connected to two sides of the fixed base plate 1 in a sliding manner, a plurality of knitting needles 11 are arranged in the swing arms 10, and the knitting needles 11 are connected with textile lines with different colors, when the production of textile cloth of a cloth neural network is finished, the device is required to detect flaws on the surface of the cloth, the cloth is placed on the roller 5, the motor is started to rotate, the roller drives the cloth to be transported, the movable plate moves on the fixed slide rod and drives the camera to shoot the surface of the cloth to obtain cloth images, the upper fixed plate connected with the support rods on the two sides of the fixed bottom plate can ensure the stability of the whole device, and the connecting plate can be connected with the roller to increase the stability of the roller;
the cloth flaw detection system comprises a flaw detection module, a data analysis module and a control module, wherein the flaw detection module comprises a memory submodule, an irradiation submodule and a projection receiving submodule, the projection receiving submodule comprises a time recording unit and a projection moving detection unit, the control module comprises a pause submodule, a moving control submodule and a shooting submodule, the irradiation submodule is positioned on two sides of the bottom of an upper fixing plate 2, the projection receiving submodule is positioned on the upper surface of a fixing bottom plate 1, the shooting submodule is positioned on the surface of a camera 9, the time recording unit is electrically connected with the pause submodule, the moving control submodule is electrically connected with a moving plate 8, the memory submodule is used for recording a cloth image without flaws of a neural network, a comparison effect can be achieved on subsequent work, the irradiation submodule is used for irradiating light to the surface of the cloth, the projection receiving submodule is used for receiving the light to find out projection to the surface of the cloth, the time recording unit is used for recording the time of the cloth projection to the projection receiving submodule, the pause submodule is used for pausing the rotation of a roller driven by a motor, the movement control submodule is used for receiving the data of the pause submodule and controlling the moving plate 8 to move on the fixed sliding rod 7, the shooting submodule is used for shooting the surface of the cloth and generating the existing cloth image, and the projection movement detection unit is used for sensing whether the cloth projection moves or not;
the flaw detection module and the control module comprise the following operation steps:
s1, before the woven cloth surface flaw detection device is started, a worker stores the image of the cloth without flaws into a memory sub-module;
s2, starting the cloth surface flaw detection device by a worker, driving the roller 5 to rotate by the motor, transporting the cloth, irradiating the light onto the cloth by the irradiation submodule in the cloth transportation process, and receiving the projection by the projection receiving submodule;
s3, the time recording unit is used for counting the time of the cloth projection movement generated by the projection movement detection unit and generating a time value so as to count the running length of the cloth, and the time value is transmitted to the pause submodule when the running moving length of the cloth is long;
s4, the pause submodule is used for receiving the time value and transmitting a signal to the motor to stop the motor from rotating, and the cloth also stops rotating;
s5, the pause submodule transmits a motor pause signal to the movement control submodule so as to control the moving plate 8 to move on the fixed slide bar 7;
s6, when the moving plate 8 moves, the camera 9 drives the shooting submodule to shoot the surface of the cloth which stops running, and an existing cloth image is generated;
the data analysis module comprises an image recording submodule, a statistics submodule and an alarm submodule, wherein the image recording submodule comprises a flaw identification unit, an image comparison unit and a position analysis unit, the image recording submodule is electrically connected with the shooting submodule, the memory submodule is electrically connected with the image recording submodule, the alarm submodule is electrically connected with the motor, and the position analysis unit is electrically connected with the statistics submodule;
the image recording submodule is used for receiving an image of flawless cloth of the memory submodule, the statistics submodule is used for counting the proportion of the flawed area of the cloth and generating a proportion numerical value of the flawed cloth area, the alarm submodule is used for receiving the numerical value of the flawed cloth area, when the flawed cloth area is larger, the alarm submodule can give an alarm and transmit a signal to the motor to stop the motor, so that a worker can conveniently observe the cloth quality and can timely replace the cloth, the comparison unit compares the image of the existing cloth with the flawless cloth image and generates comparison data, the flaw identification unit is used for receiving the comparison data and identifying whether the image of the existing cloth contains flaws, and the position analysis unit is used for analyzing the positions of the flaws;
the data analysis module comprises the following operation steps:
a1, transmitting the flawless cloth image to an image recording submodule, and transmitting the existing cloth image shot by the shooting submodule to the image recording submodule;
a2, comparing the existing cloth image with the flawless image by the image comparison unit, so that the flaw identification unit can identify the flaw area of the existing cloth image and generate a numerical ratio of the flaw area;
a3, the statistical unit receives the flaw area numerical value ratio and transmits the flaw area numerical value to the alarm submodule, when the flaw area numerical value is more, the alarm submodule reports, and when the flaw area numerical value is less, the cloth can be continuously transmitted;
a4, the position analysis unit transmits the flaw positions to the statistics submodule, so that the statistics submodule can perform statistics on the flaw positions on the cloth with fewer flaws, and when the transmission of the cloth with small flaw area values is completed, workers can conveniently find the flaw positions directly to perform repairing;
a5, when a worker needs to repair cloth, the edge of the cloth is set as an origin, and a flaw position is found according to the flaw position;
in step a3, the numerical ratio of the flaw area is calculated as:
in the formula (I), the compound is shown in the specification,
the numerical ratio of the area of the flaw is,
the number of occurrences of the defect area, C the defect area,
is the width of the cloth when
>When 5%, the signal is transmitted to the alarm submodule to stop the motor, and when the signal is detected to be abnormal, the motor is stopped
<When the percentage of the cloth is 5 percent, the cloth is continuously conveyed;
in step a4, the flaw position identification formula is:
wherein l is a flaw length distance, k is a flaw width distance, x is a length distance from the flaw to the edge of the cloth, y is a width distance from the flaw to the edge of the cloth, and j is 1-n;
the repairing module comprises a positioning sub-module, a weaving sub-module and a post-processing sub-module, the positioning sub-module is electrically connected with the counting sub-module, the positioning sub-module is electrically connected with the swing arm 10, the weaving sub-module is electrically connected with the flaw identification unit, the positioning sub-module is used for receiving flaw position data of the counting sub-module, the swing arm 10 moves on the fixed bottom plate 1 to find the position of the flaw, the weaving sub-module receives the flaw of the existing cloth identified by the flaw identification unit and controls the weaving needle 11 to repair the flaw position of the cloth, and the post-processing sub-module is used for arranging the sewn edge well to ensure the quality of the cloth image;
the repairing module comprises the following operation steps:
b1, the position analysis unit transmits the position of the flaw to the positioning submodule, so that the positioning submodule drives the swing arm 10 to move on the fixed bottom plate 1 to find the position of the flaw;
b2, the positioning sub-module transmits data to the knitting sub-module after finding the flaw position, and the flaw identification unit also transmits the flaw image of the existing cloth to the knitting sub-module, because the flaw identification unit is compared with the non-flaw cloth, it can be known what color is missing in the flaw place of the existing cloth, so as to control the knitting needles 11 with different color textile lines, and the knitting needles 11 with color textile lines corresponding to the missing color of the cloth are selected to knit the flaw position of the existing cloth;
b3, after weaving, post-processing the woven edge through a post-processing submodule to ensure the quality of the woven cloth image;
b4, transmitting the data to an irradiation submodule after the flaw positions are woven, and enabling the irradiation submodule to shoot the cloth again;
b5, repeating the steps A1-A4, identifying the positions of the cloth defects again, and checking whether the repaired defects are combined.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. Weaving cloth surface flaw detection device based on YOLO neural network, including weaving cloth surface flaw detection device and cloth flaw detection system, its characterized in that: the textile cloth surface flaw detection device comprises a fixed base plate (1), rollers (5) are arranged at two ends of the fixed base plate (1), two groups of rollers (5) are connected with a rotating shaft (3) through internal bearings, connecting plates (4) are fixed at two ends of the rotating shaft (3), the connecting plates (4) are fixed with the fixed base plate (1), two groups of supporting rods (6) are fixed at two sides of the fixed base plate (1), an upper fixing plate (2) is arranged above the fixed base plate (1), the upper fixing plate (2) is fixed with the supporting rods (6), two groups of fixed sliding rods (7) are fixed at the bottom of the upper fixing plate (2), a movable plate (8) is connected to one side of each group of the fixed sliding rods (7) in a sliding mode, a camera (9) is fixed on the bottom surface of the movable plate (8), and a motor is electrically connected to one side of each roller (5), cloth is arranged on the surfaces of the two groups of rollers (5), swing arms (10) are connected to the two sides of the fixed base plate (1) in a sliding mode, a plurality of knitting needles (11) are arranged inside the swing arms (10), and the knitting needles (11) are connected with textile threads with different colors;
the cloth flaw detection system comprises a flaw detection module, a data analysis module, a control module and a repair module, wherein the flaw detection module comprises a memory sub-module, an irradiation sub-module and a projection receiving sub-module, the projection receiving sub-module comprises a time recording unit and a projection moving detection unit, the control module comprises a pause sub-module, a moving control sub-module and a shooting sub-module, the irradiation sub-module is positioned at two sides of the bottom of an upper fixing plate (2), the projection receiving sub-module is positioned on the upper surface of a fixed bottom plate (1), the shooting sub-module is positioned on the surface of a camera (9), the time recording unit is electrically connected with the pause sub-module, the moving control sub-module is electrically connected with a moving plate (8), the memory sub-module is used for recording a flawless cloth image of a neural network, and can achieve a comparison effect on subsequent work, the irradiation submodule is used for irradiating light to the surface of the cloth, the projection receiving submodule is used for receiving the light and finding the projection of the surface of the cloth, the time recording unit is used for recording the time of projecting the cloth to the projection receiving submodule, the suspension submodule is used for suspending the rotation of a roller driven by a motor, the movement control submodule is used for receiving the data of the suspension submodule and controlling a moving plate (8) to move on a fixed sliding rod (7), the shooting submodule is used for shooting the surface of the cloth and generating the existing cloth image, and the projection movement detection unit is used for sensing whether the projection of the cloth moves.
2. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 1, wherein: the flaw detection module and the control module comprise the following operation steps:
s1, before the woven cloth surface flaw detection device is started, a worker stores the image of the cloth without flaws into a memory sub-module;
s2, starting the woven cloth surface flaw detection device by a worker, driving the roller (5) to rotate by the motor, transporting the cloth, irradiating the light onto the cloth by the irradiation submodule in the cloth transportation process, and receiving the projection by the projection receiving submodule;
s3, the time recording unit is used for counting the time of the cloth projection movement generated by the projection movement detection unit and generating a time value so as to count the running length of the cloth, and the time value is transmitted to the pause submodule when the running moving length of the cloth is long;
s4, the pause submodule is used for receiving the time value and transmitting a signal to the motor to stop the motor from rotating, and the cloth also stops rotating;
s5, the pause submodule transmits a motor pause signal to the movement control submodule so as to control the moving plate (8) to move on the fixed slide bar (7);
s6, when the moving plate (8) moves, the camera (9) drives the shooting submodule to shoot the surface of the cloth which stops running, and an existing cloth image is generated.
3. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 2, wherein: the data analysis module comprises an image recording submodule, a statistics submodule and an alarm submodule, wherein the image recording submodule comprises a flaw identification unit, an image comparison unit and a position analysis unit, the image recording submodule is electrically connected with the shooting submodule, the memory submodule is electrically connected with the image recording submodule, the alarm submodule is electrically connected with the motor, and the position analysis unit is electrically connected with the statistics submodule;
the image recording submodule is used for receiving an image of flawless cloth of the memory submodule, the statistics submodule is used for counting the proportion of the flawless area of the cloth and generating the proportion numerical value of the flawed cloth area, the alarm submodule is used for receiving the proportion numerical value of the flawed cloth area, the alarm submodule can give an alarm when the flawed cloth area is large, and transmits a signal to the motor, so that the motor stops running, a worker can observe the quality of the cloth conveniently, the cloth can be replaced in time, the image comparison unit compares the existing cloth image with the flawless cloth image and generates comparison data, the flaw identification unit is used for receiving the comparison data and identifying whether the flaw is contained in the image of the existing cloth, and the position analysis unit is used for analyzing the position of the flaw.
4. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 3, wherein: the data analysis module comprises the following operation steps:
a1, transmitting the flawless cloth image to an image recording submodule, and transmitting the existing cloth image shot by the shooting submodule to the image recording submodule;
a2, comparing the existing cloth image with the flawless image by the image comparison unit, so that the flaw identification unit can identify the flaw area of the existing cloth image and generate a numerical ratio of the flaw area;
a3, the statistical unit receives the flaw area numerical value ratio and transmits the flaw area numerical value ratio to the alarm submodule, when the flaw area is large, the alarm submodule gives an alarm, and when the flaw area is small, cloth is continuously transmitted;
a4, the position analysis unit transmits the flaw positions to the statistics submodule, so that the statistics submodule can perform statistics on the flaw positions on the cloth with fewer flaws, and when the transmission of the cloth with small flaw area values is completed, workers can conveniently find the flaw positions directly to perform repairing;
and A5, when the worker needs to repair the cloth, the edge of the cloth is set as the origin, and the defect position is searched according to the defect position.
5. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 4, wherein: in step a3, the numerical ratio of the defect area is calculated as:
in the formula (I), the compound is shown in the specification,
the numerical ratio of the area of the flaw is,
the number of occurrences of the defect area, C the defect area,
the width of the cloth is the width of the cloth,
for running the length of the cloth when
>When 5%, the signal is transmitted to the alarm submodule to stop the motor, and when the signal is detected to be abnormal, the motor is stopped
<And when the percentage of the cloth is 5 percent, the cloth is continuously conveyed.
6. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 5, wherein: in step a4, the flaw position identification formula is:
wherein l is the defect length distance, k is the defect width distance,
the length distance from the jth defect to the edge of the cloth,
the width distance from the jth flaw to the edge of the cloth is defined as j 1-n.
7. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 6, wherein: the repair module is including the location submodule piece, weave submodule piece and aftertreatment submodule piece, the location submodule piece is connected for the electricity with the statistics submodule piece, the location submodule piece is connected for the electricity with swing arm (10), it is connected for the electricity with flaw recognition unit to weave the submodule piece, the location submodule piece is used for receiving the flaw position data of statistics submodule piece, makes swing arm (10) move on PMKD (1), finds flaw place position, weave the flaw that the submodule piece received the current cloth of flaw recognition unit discernment, control knitting needle (11) mends in cloth flaw position, the aftertreatment submodule piece is used for managing the edge after mending to guarantee the quality of cloth image.
8. The YOLO neural network-based textile cloth surface flaw detection device as claimed in claim 7, wherein: the repairing module comprises the following operation steps:
b1, the position analysis unit transmits the position of the flaw to the positioning submodule, so that the positioning submodule drives the swing arm (10) to move on the fixed bottom plate (1) to find the position of the flaw;
b2, the positioning sub-module transmits data to the knitting sub-module after finding the flaw position, and the flaw identification unit also transmits the flaw image of the existing cloth to the knitting sub-module, because the flaw identification unit is compared with the non-flaw cloth, the flaw identification unit can know what color is missing in the flaw place of the existing cloth, so as to control the knitting needles (11) with different color textile lines, and the knitting needles (11) of the color textile lines corresponding to the missing color of the cloth are selected to knit the flaw position of the existing cloth;
b3, after weaving, post-processing the woven edge through a post-processing submodule to ensure the quality of the woven cloth image;
b4, transmitting the data to an irradiation submodule after the flaw positions are woven, and enabling the irradiation submodule to shoot the cloth again;
b5, repeating the steps A1-A4, identifying the positions of the cloth defects again, and checking whether the repaired defects are qualified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110985863.9A CN113433137B (en) | 2021-08-26 | 2021-08-26 | Textile cloth surface flaw detection device based on YOLO neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110985863.9A CN113433137B (en) | 2021-08-26 | 2021-08-26 | Textile cloth surface flaw detection device based on YOLO neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113433137A CN113433137A (en) | 2021-09-24 |
| CN113433137B true CN113433137B (en) | 2022-01-18 |
Family
ID=77798011
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110985863.9A Active CN113433137B (en) | 2021-08-26 | 2021-08-26 | Textile cloth surface flaw detection device based on YOLO neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113433137B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113578764B (en) * | 2021-09-28 | 2021-12-10 | 南通宝硕纺织品有限公司 | Non-woven fabric production detection device based on Internet of things |
| CN114047201A (en) * | 2021-10-25 | 2022-02-15 | 上海仪电智能科技有限公司 | AI intelligent technology-based multiple fabric flaw detection device and operation method |
| CN116575206A (en) * | 2022-01-13 | 2023-08-11 | 江正荣 | Automatic feeding system for textile bleaching |
| CN115110230B (en) * | 2022-02-25 | 2023-11-21 | 耿艳 | Textile cloth bleaching device based on big data technology of Internet of things |
| CN115508282B (en) * | 2022-10-18 | 2023-05-23 | 广东纱纤亿信息科技股份有限公司 | Online intelligent cloth inspection detection system |
| CN115611053B (en) * | 2022-10-18 | 2023-12-26 | 海安冠益纺织科技有限公司 | Device is tailor in weaving based on big data technology of thing networking |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11263493A (en) * | 1998-03-19 | 1999-09-28 | Daishowa Paper Mfg Co Ltd | Paper machine control device |
| US6259526B1 (en) * | 1999-03-01 | 2001-07-10 | Sara Lee Corporation | Operator-free fabric web inspection system |
| FR3005042B1 (en) * | 2013-04-26 | 2016-01-01 | Snecma | WEAVING OR WINDING MACHINE FOR FIBROUS TEXTURE FOR ANOMALY CONTROL BY IMAGE ANALYSIS |
| CN106959299A (en) * | 2017-03-16 | 2017-07-18 | 淄博卓意玻纤材料有限公司 | EVS cloth cover detecting systems |
| CN112215824A (en) * | 2020-10-16 | 2021-01-12 | 南通大学 | YOLO-v 3-based cloth cover defect detection and auxiliary device and method |
| CN112726174A (en) * | 2020-12-22 | 2021-04-30 | 杭州聚森科技有限公司 | Cloth surface defect automatic detection device and cloth surface defect detection method |
-
2021
- 2021-08-26 CN CN202110985863.9A patent/CN113433137B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113433137A (en) | 2021-09-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113433137B (en) | Textile cloth surface flaw detection device based on YOLO neural network | |
| CN205484100U (en) | Flaw off -line monitoring system of cloth | |
| CN113166987B (en) | Machine and method for controlling textile quality | |
| CN104764750B (en) | Elevator balanced compensated chain automatic detection device for quality and method based on machine vision | |
| CN202330297U (en) | Device for automatically detecting texture defects | |
| CN202770781U (en) | Fabric defect online detection device based on machine vision | |
| KR102445162B1 (en) | Fabric inspection system | |
| CN106841222A (en) | A kind of image detecting system for automatic cloth inspection | |
| CN105044121A (en) | Online flaw detection device and online flaw detection method of circular knitting machine | |
| CN109385876A (en) | A kind of the intelligent quality management system and its method of woven fabric | |
| CN111862069A (en) | Circular weft knitting horizontal bar defect online detection method based on gray gradient method | |
| CN110318245A (en) | A kind of machine vision fabric inspecting system being integrated on loom | |
| CN113592852A (en) | Carbon fiber carbon wire quality online detection system and method | |
| CN111707675A (en) | Cloth surface flaw on-line monitoring device and monitoring method thereof | |
| CN205484106U (en) | Flaw off -line monitoring device of cloth | |
| CN208733293U (en) | Intelligent cloth cover monitors system | |
| Divyadevi et al. | Survey of automated fabric inspection in textile industries | |
| CN107144570A (en) | A kind of tufting machine row's yarn error-detecting method based on machine vision | |
| CN204807470U (en) | Two -sided perching device | |
| CN115508282B (en) | Online intelligent cloth inspection detection system | |
| CN208520778U (en) | A kind of cloth quality verifying attachment based on machine vision | |
| CN109827963A (en) | A kind of automanual cloth examination device | |
| CN112391731B (en) | Online detection method for yarn breakage during weaving of warp knitting machine | |
| CN115142187A (en) | Knitting circular knitting machine manufacturing quality optimization control system | |
| CN114622311A (en) | Yarn breakage detection method and device and spinning machine |
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 |