Textile defect detection robot based on autonomous learning

Abstract

The invention relates to a textile defect detection robot based on autonomous learning, which relates to the technical field of textile production and comprises a preliminary identification module, a double-sided detection module and a summarization module. According to the application, the preliminary identification module, the double-sided detection module and the summarization module are utilized, and the preliminary identification module is utilized to realize preliminary treatment of textiles, namely defect detection is carried out in the textile production and movement process; then the double-sided detection module is used for re-processing the preliminarily determined defect position in the preliminary identification module to further determine the area and the type of the defect, and finally the summarization module is used for further processing the area and the type of the defect in the double-sided detection module, so that various defects can be summarized and processed conveniently; and subsequent autonomous learning of the detection robot is facilitated.

Description

Technical Field

[0001] The invention relates to the technical field of textile production, and in particular to a textile defect detection robot based on autonomous learning. Background Art

[0002] Textile defect detection is mainly divided into dyeing defects and structural defects, especially dyeing defects. Dyeing defects are usually caused by contamination on the equipment or continuous leakage of contamination on the equipment, which causes the fabric to be dirty and affects the appearance quality of the fabric.

[0003] For example, in the patent document with application number 202210795469.3, a robot detection device for textile defect detection is disclosed, and an image processing method is specifically disclosed, which can use a camera to capture images with a larger field of view.

[0004] In the above technical solution, dye contamination usually comes from contamination on the equipment, so the contaminated surface of the cloth is usually the contact surface with the equipment. During the guiding process of the cloth, direct camera shooting cannot determine the contaminated surface of the cloth, resulting in a certain error in the confirmation of the degree of contamination of the contaminated surface of the cloth. Summary of the invention

[0005] The purpose of the present invention is to provide a textile defect detection robot based on autonomous learning.

[0006] The technical problem solved by the present invention is to solve the problem that the contaminated surface of textiles cannot be determined in the prior art, resulting in a certain error in the confirmation of the contaminated surface of the fabric, making it difficult to trace the defects.

[0007] The present invention can be implemented by the following technical solutions: a textile defect detection robot based on autonomous learning, comprising a preliminary recognition module, a double-sided detection module and a summary module;

[0008] The preliminary recognition module performs visual inspection on the textiles in motion to determine whether there are defects on the textiles.

[0009] The double-sided detection module performs double-sided detection on the defective parts of the textiles based on the defective textiles obtained by the preliminary recognition module to determine the area and type of the defects of the textiles;

[0010] The summary module obtains the defect detection area and type in the double-sided detection module and summarizes the defect problems of the textiles.

[0011] A further technical improvement of the present invention is that the preliminary recognition module includes a mobile visual detection unit, a comparison unit and a classification summary unit;

[0012] The mobile visual inspection unit is arranged on the side of the textile, and the mobile visual inspection unit takes regular pictures of the textile in a moving state, and sends the pictures taken to the comparison unit;

[0013] The comparison unit obtains the captured image, compares the captured image with the original textile form, and determines whether there is any abnormality on the textile;

[0014] The classification summary unit summarizes the textile data based on the data of the abnormal area location and preliminarily determines the type of defects on the textile.

[0015] A further technical improvement of the present invention is that: the double-sided detection module includes a fixing unit, a detection unit and an analysis unit;

[0016] The fixing unit is used to define the position of the textile in a moving state, and to mount the defective textile on the side of the detection unit;

[0017] The detection unit performs photographic detection based on the textile in a fixed state, determines the type of defects on the textile, and analyzes and organizes the defect types through the analysis unit.

[0018] A further technical improvement of the present invention is that the detection unit includes a first visual detection mechanism and a second visual detection mechanism, which are respectively located on both sides of the textile, and the defective positions of the textile are photographed and detected to obtain a first color difference result A1 and a second color difference result A2.

[0019] A further technical improvement of the present invention is that: the analysis unit obtains the first color difference result A1 and the second color difference result A2 of the defective part of the textile, compares the first color difference result A1 and the second color difference result A2 with the color difference reference value AM respectively, records the value exceeding the comparison range as an abnormal value, and records the textile surface corresponding to the abnormal value as an abnormal surface; when it is determined that there is an abnormal surface, a separate comparison is performed based on the first color difference result A1 and the second color difference result A2 to determine the size of the first color difference result A1 and the second color difference result A2, and defines the abnormality of the color result difference value exceeding the set range as a color staining abnormality, and defines the abnormality of the color result difference value not exceeding the set range as a color staining abnormality.

[0020] A further technical improvement of the present invention is that the summary module includes a list unit and a diagnosis unit. The list unit lists the defect data of the textiles into a chart based on the abnormal position and abnormal type on the textiles, calculates the frequency of occurrence of each defect on the chart, and performs feedback adjustment on the equipment according to the instability of the frequency calculation equipment.

[0021] A further technical improvement of the present invention is that it also includes an abnormality library, which includes defect data of textiles in the summary module, records the defect data according to the events that occur, and also includes a maintenance method for each defect data when it occurs.

[0022] A further technical improvement of the present invention is that it also includes a learning module, which obtains defect data and corresponding maintenance methods in the abnormal library, detects the status of the equipment corresponding to the maintenance method, and the detection time interval is extended in an arithmetic progression.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. This application utilizes a preliminary recognition module, a double-sided detection module and a summary module, and utilizes the preliminary recognition module to realize preliminary processing of textiles, that is, to detect defects in the textile production and movement process; then utilizes the double-sided detection module to re-process the defect position preliminarily determined in the preliminary recognition module, and further determine the area and type of the defect; finally, utilizes the summary module to realize further processing of the defect area and type in the double-sided detection module, so as to facilitate summarizing and processing of various defects and facilitate subsequent autonomous learning of the detection robot.

[0025] 2. The detection unit of the present application includes a first visual detection mechanism and a second visual detection mechanism, which are respectively located on both sides of the textile, and photograph and detect the defective position of the textile to obtain a first color difference result A1 and a second color difference result A2. By separately detecting the two sides of the textile, the type of abnormal defect of the textile can be determined, whether it is color contamination abnormality or color diffusion abnormality, which is convenient for subsequent confirmation of the abnormality of the textile. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0027] Figure 1 It is a system principle block diagram of the present invention;

[0028] Figure 2 It is a schematic diagram of the position of the visual detection mechanism of the present invention;

[0029] Figure 3 It is a schematic diagram of the conversion mechanism structure of the present invention;

[0030] Figure 4 It is a schematic diagram of the relative positions of the upper base and the fixed support base of the present invention.

[0031] In the figure: 1. first visual inspection mechanism; 2. positioning mechanism; 3. fixed support base; 4. matching groove; 5. second matching support mechanism; 6. second visual inspection mechanism; 7. first matching support mechanism; 8. conversion mechanism; 9. first rack; 10. intermediate gear; 11. second rack; 12. pressure plate frame; 13. limiting pressure rod; 14. limiting pressure block; 15. limiting pressure groove; 16. vacuum plane; 17. sliding push plate; 18. sliding push rod; 19. sliding base; 20. limiting rod; 21. upper base; 22. upper telescopic rod; 23. upper guide roller; 24. upper bracket; 25. position adjustment roller. DETAILED DESCRIPTION

[0032] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0033] See also Figure 1-4 As shown, a textile defect detection robot based on autonomous learning includes a preliminary recognition module, a double-sided detection module and a summary module, wherein the preliminary recognition module performs detection based on the textile in a moving state to determine whether there are defects on the textile and the type of the defects, and starts the double-sided detection module when there are defects on the textile, and uses the double-sided detection module to realize double-sided detection of the textile to determine the defective surface and defect location of the textile, and finally uses the summary module to realize the summary of the textile defect detection, so as to facilitate tracing the causes of the textile defects.

[0034] First, the preliminary detection module performs detection based on the mobile visual detection mechanism, where the mobile visual detection is used to detect textiles in a moving state. By frame-by-frame comparison, it is determined whether there are defects on the textile, which include damage defects and abnormal dyeing defects. It is preliminarily determined whether the double-sided detection module needs to be started based on the type of defect.

[0035] The preliminary detection module includes a mobile visual detection unit, a comparison unit and a classification and summary unit. First, the mobile visual detection unit takes photos and samples the textiles within a set time based on the textiles in a moving state; the obtained photo samples are compared with the original textile form through the comparison unit to determine whether there is an abnormality on the textile. If so, the position of the abnormal area on the textile is obtained; the textile data is summarized using the data of the abnormal area position based on the classification and summary unit, that is, it is determined whether the textile has a damaged defect or an abnormal dyeing defect based on the change in the light transmittance of the textile, so that the cause of the abnormality of the textile can be obtained.

[0036] When the cause of the textile abnormality is an abnormal dyeing defect, in order to determine the abnormal side of the textile, a double-sided detection module is used to further detect the textile, that is, the double-sided detection module performs another detection based on the cause of the textile abnormality and the abnormal position of the textile obtained in the preliminary detection module, that is, the textile in a moving state reaches the double-sided detection module, and the two sides of the textile are detected by starting the double-sided detection module to determine whether the dyeing abnormality is an externally adhered color or a color abnormality caused by dyeing, so that the cause of the abnormality can be further obtained, which is convenient for tracing the abnormality;

[0037] It also includes an abnormal library, which includes the defect data of the textile in the summary module, records the defect data according to the events that occur, and the abnormal library also includes a maintenance method for each defect data when it occurs

[0038] It also includes a learning module, which obtains defect data and corresponding maintenance methods in the abnormal library, detects the status of the equipment corresponding to the maintenance method, and the detection time interval is extended in an arithmetic progression.

[0039] Specifically, the double-sided detection module includes a first visual detection mechanism 1, which uses a second matching support mechanism 5 to act on a fixed support base 3, and uses the positioning mechanism 2 to fix the position of the textile at the current position. At this time, by using the first visual detection mechanism 1 to shoot the textile in a stationary state, the problem of movement distortion caused by shooting the textile during movement can be avoided, the accuracy of shooting can be ensured, and the first color difference result A1 is obtained. Then, the second matching support mechanism 5 is used to release the matching effect with the fixed support base 3, so that the first matching support mechanism 7 and the fixed support base 3 are matched, and the positioning mechanism 2 is used to fix the position of the textile, so that the second visual detection mechanism 6. Realize the function of photographing and detecting the textile, obtain the color difference result A2 of the other side of the textile, process the color difference result A1 and the color difference result A2, judge the side with the larger difference in color difference result, and judge the side as the color abnormal side, so that the color of the side can be judged as the color abnormality of staining, and it is only necessary to judge whether there is dirt on the equipment. If the color difference result A1 and the color difference result A2 are processed, it is judged that the color difference results are not much different, almost the same, and the values ​​of the color difference result A1 and the color difference result A2 are both large, it means that the area is abnormal due to color staining, so it is necessary to judge whether there is color dripping on the equipment at this time, so as to facilitate the processing of the color;

[0040] As a further embodiment of the present application, when processing the color difference result A1 and the color difference result A2, it includes a reference value acquisition unit, that is, obtaining the reference value AM of the color difference, and comparing the color difference result A1 and the color difference result A2 with the color difference reference value AM respectively, and recording the values ​​exceeding the comparison range as abnormal values, and comparing the color difference result A1 and the color difference result A2 separately to determine whether there is a large gap between them. If so, it indicates that it is a color contamination abnormality, and if there is no large gap, it indicates that the area is an abnormality caused by color contamination.

[0041] Based on the above technical conclusions, the summary module obtains the abnormal positions and types on the textiles, lists them in a chart, calculates the frequency of occurrence of various abnormalities on the textiles according to the chart, and judges the instability of the equipment based on the frequency, and performs feedback processing on the equipment.

[0042] In the above technical solution, the double-sided detection device involved in the double-sided detection module includes a fixed support base 3, wherein the fixed support base 3 is fixed on the frame, and the structures on both sides of the fixed support base 3 are symmetrically arranged, wherein the first visual detection mechanism 1 and the second visual detection mechanism 6 are respectively arranged on both sides of the fixed support base 3, which are respectively used to detect the two sides of the textile, and a positioning mechanism is also arranged on the frame, which is used to fix the detection part of the textile on the fixed support base 3, so as to achieve accurate detection of the detection part of the textile and avoid the problem of shooting distortion caused by vibration, so a positioning mechanism is provided on the fixed support base 3. A matching groove 4 for matching with the positioning mechanism 2 is provided with a conversion mechanism 8 on the frame, wherein the conversion mechanism 8 is respectively provided with a first matching support mechanism 7 and a second matching support mechanism 5. When in use, the conversion mechanism 8 is used to control the first matching support mechanism 7 and the second matching support mechanism 5 to respectively match with the fixed support base 3 to support the abnormal surface of the textile, so as to ensure the accuracy of the shooting, and on the basis of this support, the positioning mechanism 2 is controlled to further fix the textile, so as to facilitate the visual inspection of the textile by the first visual inspection mechanism 1 or the second visual inspection mechanism 6.

[0043] The positioning mechanism 2 includes a limiting pressure rod 13, wherein the limiting pressure rod 13 is fixed on the frame, and a pressure plate frame 12 is fixed to the output end of the limiting pressure rod 13, and a limiting pressure block 14 for cooperating with the limiting pressure groove 15 is fixed on the pressure plate frame 12, that is, the limiting pressure groove 15 is provided on the side of the matching groove 4. The limiting pressure rod 13 includes but is not limited to a hydraulic cylinder and a telescopic rod, and the limiting pressure rod 13 is activated to control the pressure plate frame 12 to move longitudinally, so that the limiting pressure block 14 acts on the limiting pressure groove 15, and the fixed cooperation between the pressure plate frame 12 and the fixed support base 3 is realized, thereby realizing the limited support of the textile to prevent the textile from moving or wrinkling.

[0044] The conversion mechanism 8 includes an intermediate gear 10, which is driven by a driving motor, wherein the driving motor is fixed on the frame, and a first rack 9 and a second rack 11 are symmetrically meshed on both sides of the intermediate gear 10, wherein a first matching support mechanism 7 is installed on the first rack 9, and a second matching support mechanism 5 is arranged on the second rack 11, wherein the first matching support mechanism 7 and the second matching support mechanism 5 are respectively connected to the frame in a sliding manner, so when in use, the driving motor is used to control the rotation of the intermediate gear 10, so that the first matching support mechanism 7 and the second matching support mechanism 5 are controlled to act on the fixed support base 3 respectively, so as to realize separate detection of the two sides of the textile.

[0045] The first matching support mechanism 7 and the second matching support mechanism 5 have the same structure. Taking the first matching support mechanism 7 as an example, the first matching support mechanism 7 includes a limit rod 20, which is fixedly connected to the first rack 9, wherein the limit rod 20 is slidably connected to the fixed support base 3, and a sliding base 19 is fixedly installed at the end of the limit rod 20, wherein the sliding base 19 is slidably connected to the frame, and a sliding push rod 18 is fixed on the sliding base 19, and a sliding push plate 17 is fixed at the output end of the sliding push rod 18, and a vacuum plane 16 is fixed at the end of the sliding push plate 17. The vacuum plane 16 is used to contact the textile fabric and adsorb the textile fabric to avoid the formation of wrinkles. Therefore, when in use, the first rack 9 is used to drive the sliding base 19 to move, so that the sliding base 19 moves to the fixed support base 3, and then the sliding push rod 18 is started to push the sliding push plate 17 to move longitudinally, so that the vacuum platform 16 moves to the fixed support base 3 until the vacuum plane 16 presses the textile downward until the textile contacts the vacuum plane 16, and the vacuum plane 16 is used to adsorb the textile.

[0046] When the textile passes over the fixed support base 3, the second cooperating support mechanism 5 moves upward to the inside of the fixed support base 3 until the vacuum plane 16 of the second cooperating support mechanism 5 acts on the textile, and the first visual inspection mechanism 1 is used to inspect the textile on the second cooperating support mechanism 5. When the other side of the textile is inspected, the first cooperating support mechanism 7 moves downward, which will cause the textile on the fixed support base 3 to pass through the fixed support base 3 under the action of the first cooperating support mechanism 7 until the textile passes through the fixed support base 3, which is convenient for inspection by the second visual inspection mechanism 6. However, in this process, the inspection result of the second visual inspection mechanism 6 will cause certain errors due to the wrinkles generated at the edge of the textile. Therefore, the present application is achieved by setting an upper base 21 for use in conjunction with the first cooperating support mechanism 7.

[0047] At this time, the first visual inspection mechanism 1 acts on the fixed support base 3, the second visual inspection mechanism 6 cooperates with the upper base 21, and the upper base 21 is fixed on the upper bracket 24, wherein the upper bracket 24 is fixedly installed at the output end of the upper telescopic rod 22, wherein the output end of the upper telescopic rod 22 is also rotatably provided with an upper guide roller 23, and a position standard roller 25 is also rotatably provided on the frame, and the longitudinal height of the position standard roller 25 is adjustable.

[0048] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A textile defect detection robot based on autonomous learning, characterized in that: It includes a preliminary recognition module, a two-sided detection module and a summary module; The preliminary recognition module performs visual inspection on the textile in a moving state to determine whether there are defects on the textile; The double-sided detection module performs double-sided detection on defective parts of the textile based on the defective textile obtained by the preliminary recognition module to determine the area and type of the defect of the textile; The summary module obtains the defect detection area and type in the double-sided detection module and summarizes the defect problems of the textile.

2. The textile defect detection robot based on autonomous learning according to claim 1, characterized in that: The preliminary recognition module includes a mobile visual detection unit, a comparison unit and a classification summary unit; The mobile visual detection unit is arranged on the side of the textile, and the mobile visual detection unit takes regular pictures of the textile in a moving state, and sends the taken pictures to the comparison unit; The comparison unit obtains the photographed picture, compares the photographed picture with the original textile form, and determines whether there is any abnormality on the textile; The classification and summarization unit summarizes the textile data based on the data of the abnormal area position and preliminarily determines the type of defects on the textile.

3. The textile defect detection robot based on autonomous learning according to claim 1, characterized in that: The double-sided detection module includes a fixing unit, a detection unit and an analysis unit; The fixing unit is used to limit the position of the textile in a moving state, and to mount the defective textile on the side of the detection unit; The detection unit performs photographing detection based on the textile in a fixed state, determines the defect type existing on the textile, and analyzes and organizes the defect type through the analysis unit.

4. The textile defect detection robot based on autonomous learning according to claim 3, characterized in that: The detection unit includes a first visual detection mechanism and a second visual detection mechanism, which are respectively located on two sides of the textile, and photograph and detect defective positions of the textile to obtain a first color difference result A1 and a second color difference result A2.

5. The textile defect detection robot based on autonomous learning according to claim 4, characterized in that: The analysis unit obtains a first color difference result A1 and a second color difference result A2 of a defective part of the textile, compares the first color difference result A1 and the second color difference result A2 with the color difference reference value AM respectively, records a value exceeding the comparison range as an abnormal value, and records the textile surface corresponding to the abnormal value as an abnormal surface; when determining that an abnormal surface exists, a separate comparison is performed based on the first color difference result A1 and the second color difference result A2 to determine the size of the first color difference result A1 and the second color difference result A2, defines the abnormality of the color result difference value exceeding the set range as a color staining abnormality, and defines the abnormality of the color result difference value not exceeding the set range as a color soaking abnormality.

6. The textile defect detection robot based on autonomous learning according to claim 5, characterized in that: The summary module includes a list unit and a diagnosis unit. The list unit lists the defect data of the textile into a chart based on the abnormal position and abnormal type on the textile, calculates the frequency of occurrence of each defect on the chart, and performs feedback adjustment on the equipment according to the instability of the frequency calculation equipment.

7. The textile defect detection robot based on autonomous learning according to claim 6, characterized in that: It also includes an abnormal library, which includes defect data of textiles in the summary module, records the defect data according to the events that occur, and also includes a maintenance method for each defect data when it occurs.

8. The textile defect detection robot based on autonomous learning according to claim 7, characterized in that: It also includes a learning module, which obtains defect data and corresponding maintenance methods in the abnormal library, detects the status of the equipment corresponding to the maintenance method, and the detection time interval is extended in an arithmetic progression.

Images