Inspection system for monitoring bonded end of film product

Abstract

The utility model relates to an inspection system for monitoring a joint end part of a film product. An inspection system for monitoring a distance between joined ends of a film product having predetermined parameters of length, width, and thickness. The marking unit prints lateral marks on the first end portion and the second end portion at a predetermined distance from the first edge portion, the second edge portion, and the central portion. The detection unit captures an image formed by pixels related to the distance between the lateral marks. The processing unit applies a geometric grid on the surface of the film product representing nodes identifying the first end, the second end, the first edge portion, the second edge portion, and the central portion. The nodes are defined as a function of the calculated distance between the nodes on the lateral mark and the nodes on the lateral mark.

Description

Technical Field

[0001] The present utility model relates to an inspection system for monitoring the joint ends of a film product (in a specific example, monitoring the distance between the joint ends of a film product) to identify joint defects related to material properties. Background Art

[0002] In the prior art, there are several known ways to monitor the joint ends of a sheet product. For example, the availability of rubber between the ends is measured before the ends of the joint tread. The availability of rubber defines the gap, overlap, or perfect joint between the ends. Monitoring can also be performed by measuring the thickness height difference between the joint ends of the tread to avoid offset joints. Therefore, in order to achieve a perfect joint of the tread, it is important to have ends with symmetric geometric properties, including matching thickness, width, and profile, to avoid gaps, overlaps, or offsets between the ends of the tread.

[0003] During tire manufacturing, it may be difficult to characterize the geometry of a rubber product (e.g., a tread in an untreated or unvulcanized state). In fact, the plastic properties of a rubber product in an untreated state are not easily predicted by precise geometric properties for material properties. Due to the plastic and irregular deformation of the ends during cutting, there are difficulties in precisely characterizing the joint ends. Since the profile of the tread (including its thickness and width) varies significantly, this feature also varies for different tire sizes. This poses difficulties in automatically characterizing joints by gap, overlap, and offset defects. Therefore, the characterization of the expansion or compression of the tread around the joint area is crucial for measuring the gap or overlap of offset defects. The expansion or compression of rubber in an untreated state is mainly due to mixture properties (including the processing time of the prepared compound), environmental factors (including temperature and humidity), and various machining steps when defining the shape of the tread.

[0004] Sometimes, transverse marks or marking cuts are added to the joint ends of the tread to determine whether the joint is acceptable once the strip is wound onto the tire manufacturing drum (see, for example, US10913229, JP4537624, and KR10-0746109). In addition, before the joint ends, by using a position sensor on the outer peripheral surface of the tire manufacturing drum, the accurate position of the tread strip can be detected (see, for example, US2017 / 0074645).

[0005] According to the present utility model, the consistency of joints is improved by an inspection system that monitors the distance between the lateral markings on the tread and the joining ends by considering material properties including the expansion and compression of rubber in an untreated state (i.e., unvulcanized). Using neural networks and / or artificial intelligence tools, computer vision is employed to detect changes in these material properties caused by irregular deformations, particularly changes in the material properties in a plastic state. Summary of the Utility Model

[0006] The present utility model relates to an inspection system for monitoring the distance between joining ends of a film product having a profile defined by parameters of a predetermined length, a predetermined width, and a predetermined thickness, wherein the joining ends include a first end, a second end, a first edge portion, a second edge portion, and a central portion defining an end surface of the joining ends along a longitudinal axis, and the inspection system includes:

[0007] - At least one marking unit configured to print a first lateral marking (T1) and a second lateral marking (T2) on the first end and the second end at a first predetermined distance (D1) and a second predetermined distance (D2) from the first edge portion, the second edge portion, and the central portion;

[0008] - At least one detection unit configured to capture one or more images to identify the first lateral marking (T1) and the second lateral marking (T2) on the film product, the images being formed by pixels related to the distances of the first lateral marking (T1) and the second lateral marking (T2) from the first edge portion, the second edge portion, and the central portion;

[0009] - At least one processing unit including one or more software or algorithms or transformation models, configured to process the image of the film product through a geometric grid, so as to be able to define one or more nodes (T 1.1 , T 1.2 , T 1.3 , T 1.4 ) on the first lateral marking (T1) of the first end and one or more nodes (E 1.1 , E 1.2 , E 1.3 , E 1.4 ) on the first edge portion, the second edge portion, and the central portion of the first end;

[0010] - At least one processing unit including one or more software or algorithms or transformation models, configured through a geometric grid so as to be able to define one or more nodes (T 2.1 , T 2.2 , T 2.3 , T 2.4) and one or more nodes on the first edge portion, the second edge portion, and the central portion of the second end (E 2.1 , E 2.2 , E 2.3 , E 2.4 ); and

[0011] - A processing unit configured to define a joint based on the distances (C1, C2, C3, C4) between the nodes (T 1.1 , T 1.2 , T 1.3 , T 1.4 ) on the first transverse mark T1 of the first end and the nodes (T 2.1 , T 2.2 , T 2.3 , T 2.4 ) on the second transverse mark T2 of the second end.

[0012] In a specific embodiment of the inspection system, if the distances (C1, C2, C3, C4) between the nodes (T 1.1 , T 1.2 , T 1.3 , T 1.4 ) on the first transverse mark T1 of the first end and the nodes (T 2.1 , T 2.2 , T 2.3 , T 2.4 ) on the second transverse mark T2 of the second end calculated by the processing unit are equal to the sum of a first predetermined distance and a second predetermined distance (D1 + D2), then the defined joint is a perfect joint (P) without defects.

[0013] In a specific embodiment of the inspection system, if the distances (C1, C2, C3, C4) between the nodes (T 1.1 , T 1.2 , T 1.3 , T 1.4 ) on the first transverse mark T1 of the first end and the nodes (T 2.1 , T 2.2 , T 2.3 , T 2.4 ) on the second transverse mark T2 of the second end calculated by the processing unit are greater than the sum of a first predetermined distance and a second predetermined distance (D1 + D2), then the defined joint is an imperfect joint due to a gap defect (G).

[0014] In a specific embodiment of the inspection system, if the distances (C1, C2, C3, C4) between the nodes (T 1.1 , T 1.2 , T 1.3 , T 1.4)The distances (C1, C2, C3, C4) between the nodes (T 2.1 , T 2.2 , T 2.3 , T 2.4 ) on the second transverse mark T2 of the second end part and the sum of the first predetermined distance and the second predetermined distance (D1 + D2), the defined joint is an imperfect joint due to an overlap defect (O).

[0015] In a specific implementation of the inspection system, if the nodes (T 1.1 , T 1.2 , T 1.3 , T 1.4 ) on the first transverse mark T1 of the first end part and the nodes (E 1.1 , E 1.2 , E 1.3 , E 1.4 ) on the edge part and the nodes (T 2.1 , T 2.2 , T 2.3 , T 2.4 ) on the second transverse mark T2 of the second end part and the nodes (E 2.1 , E 2.2 , E 2.3 , E 2.4 ) on the edge part are not aligned, the defined joint is an imperfect joint due to an offset defect (J).

[0016] In a specific implementation of the inspection system, the defined joint is a perfect joint (P) or an imperfect joint due to a combination of at least one of a gap defect (G), an overlap defect (O), and an offset defect (J).

[0017] In a specific implementation of the inspection system, a processing unit including one or more software or algorithms or transformation models is trained based on a perfect joint (P) or an identified defect including at least one of a gap defect (G), an overlap defect (O), and an offset defect (J) to predict the quality of subsequent joints or expected defects.

[0018] In a specific implementation of the inspection system, the processing unit includes an image processing module, and the image processing module deploys one or more transformation models defined based on the CenterNet neural network architecture for object detection as points.

[0019] In a specific implementation of the inspection system, the processing unit includes an image processing module, and the image processing module deploys one or more transformation models including an encoder - decoder configured to reduce or increase the resolution of the captured image.

[0020] In a specific embodiment of the inspection system, the processing unit includes an image processing module, and the image processing module is configured to apply geometric correction to compensate for the non-planarity of the measurement area.

[0021] In a specific embodiment of the inspection system, a distance sensor is located at the detection unit to measure the distance between the detection unit and the surface of the film product, so as to convert the pixel area into a distance suitable for the diameter change of the surface of the film product. Brief Description of the Drawings

[0022] By reading the following detailed description in conjunction with the accompanying drawings, the nature and various advantages of the present utility model will become more apparent. In the drawings, the same reference numerals always denote the same components, wherein:

[0023] Figure 1 A schematic diagram showing an embodiment of the inspection system according to the present utility model.

[0024] Figure 2 A schematic diagram showing an embodiment of a joint end in a perfect joint state according to the present utility model.

[0025] Figure 3 Showing according to the present utility model Figure 2 A perspective view of an embodiment of a joint end in a perfect joint state.

[0026] Figure 4 A schematic diagram showing an embodiment of a joint end in an imperfect joint state due to a gap defect according to the present utility model.

[0027] Figure 5 Showing according to the present utility model Figure 4 A perspective view of an embodiment of a joint end in an imperfect joint state due to a gap defect.

[0028] Figure 6 A schematic diagram showing an embodiment of a joint end in an imperfect joint state due to an overlap defect according to the present utility model.

[0029] Figure 7 Showing according to the present utility model Figure 6 A perspective view of an embodiment of a joint end in an imperfect joint state due to an overlap defect.

[0030] Figure 8 A schematic diagram showing an embodiment of a joint end in an imperfect joint state due to a combination of gap and overlap defects and in a perfect joint state according to the present utility model.

[0031] Figure 9Schematic diagram of an embodiment of a joining end in an imperfect joint state due to an offset joint according to the present invention.

[0032] Figure 10 Schematic diagram of an embodiment of a tire showing geometric correction for compensating for the non-planarity of a measurement area according to the present invention. Detailed description

[0033] The present invention relates to an inspection system for monitoring the distance between joining ends of a tread. The inspection system is adapted to monitor various profiles of products of different thicknesses and widths while taking into account material properties including rubber expansion and compression (especially due to the plastic properties or irregular deformation of rubber in the untreated or unvulcanized state). The plastic properties of rubber in the untreated state are not easily predicted by only considering the geometry of the joining ends or only considering the profile of the tread. The expansion or compression of rubber in the untreated state is mainly due to mixture properties (including the processing time for preparing the compound), environmental factors (including temperature and humidity), and various machining steps when defining the profile of the tread. Due to the significant variations in the profile of the tread (including its thickness and width) during processing, this also varies depending on different tire sizes. In several examples, the film product can be the tread of a tire, a rubber track, or a film product of a similar nature. For the purposes of the present invention, the film product can be referred to as the tread, however, it should be understood that the present invention is not limited to the tread of a tire. Due to the inspection method of the present invention, efficient and qualitative joining of the ends of the tread can be achieved.

[0034] The disclosed inspection system monitors two joining ends of a tread band of a tread forming a joint. Generally, the tread band is a length portion of the tread applied to a tire carcass, and the tread band has a predetermined length, a predetermined width, and a predetermined thickness. The tread band can include a tread having tread patterns on the ground-engaging side of the tread. The tread patterns can include any known tread features, such as transverse lugs and / or longitudinal ribs separated by grooves and / or channels. The joining ends of a single tread can be joined to form a continuous tread band. By further example, separate treads (i.e., tread segments, sections, or lengths) can be joined to form a single tread including multiple joined tread segments. Although the present invention relates to a tread in the untreated state, the present invention is also applicable to a tread band that is at least partially cured.

[0035] Each joining end of the tread band includes an end surface (or "joining end surface") that forms the end of the tread. The joining end surface includes a cross-sectional surface of the tread that extends laterally across the width of the tread and has a height that extends through the thickness of the tread. The joining end surface can extend across the width of the tread in a direction orthogonal to the length direction of the tread or at any other angle that is biased towards the length direction. Additionally, the joining end surface can extend laterally in a linear or non-linear path. Similarly, the height of the joining end surface can extend through the tread thickness in any direction, including for example a direction orthogonal to the length direction of the tread, and along any linear or non-linear path. Since the tread can include tread patterns that extend into the thickness of the tread, the tread end surface can include voids that are disposed within the tread thickness and are arranged inwardly from the outer cross-sectional profile.

[0036] The length positions for forming the joining ends are selected along each tread portion to provide joining ends whose profiles generally match the profile of another joining end with which it will be joined. For example, joining ends that are generally matched (i.e., their cross-sections, profiles, and / or perimeters generally match) can be selected and formed or otherwise provided. This provides a tread joint that is consistent with the adjacent portions or features of the joined tread and the overall tread pattern of the joined tread. In other words, by forming the joining ends to generally match, the joined or assembled tread can include a tread pattern that is substantially uninterrupted at the joint, where the tread features of the opposing joining ends are generally aligned with respect to the height (i.e., thickness) and width of the tread.

[0037] The arrangement of the joining ends provides a joined or assembled tread in which the joined tread portions generally extend in the same length direction, which can extend linearly as a tread strip or extend annularly in the form of a tread band. Alternatively, determining the positions for forming the joining ends along any tread length can select not only to generally match the joining ends of the tread to be joined but also to select positions that enable the treads to be properly joined.

[0038] Reference Figure 1, shows an inspection system 100 according to the present utility model. The inspection system 100 is defined to monitor the distance between the joining ends 102 of the tread 104. The joining ends 102 of the tread 104 include a first end 106 and a second end 108 with predetermined profile parameters, the profile parameters including but not limited to length, width, and thickness. In addition, each of the first end 106 and the second end 108 of the tread 104 includes an end surface defining a first first edge portion 110, a second first edge portion 116, a first second edge portion 112, a second second edge portion 118, and a first central portion 114, a second central portion 120. The inspection system 100 includes at least one marking unit configured to print a first lateral mark T1 and a second lateral mark T2 on the first end 106 and the second end 108 at a first predetermined distance D1 and a second predetermined distance D2 from the first first edge portion 110, the second first edge portion 116, the first second edge portion 112, the second second edge portion 118, and the first central portion 114, the second central portion 120 after the manufacture and cutting of the tread 104 (see Figure 2 ). In addition, the tread 104 may include one or more tread segments.

[0039] The first lateral mark T1 and the second lateral mark T2 may be printed along the total width of the tread 104, as Figure 1 shown. The first predetermined distance D1 and the second predetermined distance D2 between the first lateral mark T1 and the second lateral mark T2 and the joining ends 102 are fixed by a cutting blade and the marking unit. The marking unit may include one or more marking rollers configured to apply the first lateral mark T1 and the second lateral mark T2 at the first predetermined distance D1 and the second predetermined distance D2 between the first lateral mark T1 and the second lateral mark T2 before the tread 104 is cut into segments by the cutting blade. Although two lateral marks (the first lateral mark T1 and the second lateral mark T2) are marked on the tread 104, those skilled in the art will recognize that multiple lateral or linear marks may be marked to illustrate the distance between the marks and the joining ends 102.

[0040] The inspection system 100 further includes a detection unit 122 that employs one or more sensors (not shown) that capture information about the physical environment around the tread. In the following description, the terms "sensor", "imaging device", "camera", and "optical sensor" may be used interchangeably and may refer to one or more devices configured to detect two-dimensional (2-D) and / or three-dimensional (3-D) images to enable 3-D depth perception of the physical environment and / or other types of detection. In an example, the sensor of the detection unit 122 may be any commercially available RGB-D camera with a frame resolution of 1920×1080 and a frame rate of 30 fps or frames per second. In another example, one or more sensors may be a laser profiler. The sensor of the detection unit 122 may be mounted away from the tread 104, fixed to a stationary bracket, or fixed to a movable bracket that covers the field of view of the inspection system 100. Additionally, the sensor of the detection unit 122 may be aligned with the center of the tire manufacturing drum and / or the center of the engaging end 102.

[0041] One or more sensors of the detection unit 122 of the inspection system 100 detect the presence of the arrangement of the tread 104 in the field of view of the sensor, which triggers the sensor to capture an image of the tread 104 and its profile parameters. In certain embodiments of the inspection system 100, the sensor is triggered when the tread 104 enters the field of view against the background of the captured image. The captured image is formed by pixels related to the distances of the first lateral marker T1 and the second lateral marker T2 from the first first edge portion 110, the second first edge portion 116, the first second edge portion 112, the second second edge portion 118, and the first central portion 114, the second central portion 120. The sensor of the detection unit 122 sends the captured Figure 1 profile parameters of the tread 104 shown.

[0042] In an embodiment of the present utility model, the inspection system 100 includes a distance sensor adapted to be positioned at the lens of the camera of the detection unit 122 to measure the distance between the camera and the surface of the tread 104. Then, the processing unit can adjust the conversion rate of pixels to distance (e.g., in millimeters). This principle of the distance sensor, the focal length of the camera of the detection unit 122, and the height of the pixels enables the inspection system 100 to adapt to changes in the diameter of the surface of the tread 104 caused by process dispersion or dimensional variations.

[0043] The sensors of the detection unit 122 are configured to capture one or more images of the tread 104 including the first lateral mark T1, the second lateral mark T2, and the distances D1, D2 from the joining end 102. The one or more captured images are transmitted and stored as captured images in the memory of the processing unit. The processing unit, which executes the instructions of the image processing module of the processing unit, analyzes the images to determine one or more parameters of the imaged tread 104. The term "processor" (alternatively, the term "programmable logic circuit") refers to one or more devices (e.g., one or more integrated circuits, one or more controllers, one or more microcontrollers, one or more microcomputers, one or more programmable logic controllers (or PLCs), one or more application-specific integrated circuits, one or more neural networks, and / or one or more other known equivalent programmable circuits included in a computer as known to those skilled in the art) capable of processing and analyzing data and having one or more software packages for its processing. The processing unit includes software for processing data (and the corresponding data obtained) captured by subsystems associated with the inspection system 100, and software for identifying and locating differences and identifying their sources for correction.

[0044] The processing unit, which includes one or more software, algorithms, and / or models, is configured to process the image of the tread 104 by starting geometric meshing of the captured image of the tread 104 using the prior art. In an embodiment of the present utility model (see Figure 2 and Figure 3 ), geometric meshing can define a first node T for the first lateral mark T1 on the first end 106 1.1 , a second node T 1.2 , a third node T 1.3 , a fourth node T 1.4 . In addition, a fifth node E is defined on the edge portion of the first end 106 1.1 , a sixth node E 1.2 , a seventh node E 1.3 , an eighth node E 1.4 . Similarly, the geometric mesh can define an eleventh node T for the second lateral mark T2 on the second end 108 2.1 , a twelfth node T 2.2 , a thirteenth node T 2.3 , a fourteenth node T 2.4 . In addition, a fifteenth node E is defined on the edge portion of the second end 108 2.1 , a sixteenth node E 2.2 , a seventeenth node E 2.3 , an eighteenth node E 2.4Although, in an example of the present utility model, the processing unit defines 16 nodes, those of ordinary skill in the art can understand that based on the contour of the tread 104 or processing techniques using various software or algorithms or models, N nodes can be defined.

[0045] In the embodiments of the present utility model as shown in Figure 2 and Figure 3 if the distances C1, C2, C3, C4 between the first node T 1.1 , the second node T 1.2 , the third node T 1.3 , the fourth node T 1.4 on the first lateral mark T1 of the first end portion 106 and the eleventh node T 2.1 , the twelfth node T 2.2 , the thirteenth node T 2.3 , the fourteenth node T 2.4 on the second lateral mark T2 of the second end portion 108 are equal to the sum (D1 + D2) of the first predetermined distance and the second predetermined distance defined after the manufacturing and cutting of the tread 104 into segments, then the joint between the first end portion 106 and the second end portion 108 is called a perfect joint P. In addition, the first node T 1.1 , the second node T 1.2 , the third node T 1.3 , the fourth node T 1.4 on the first lateral mark T1 of the first end portion 106 and the fifth node E 1.1 , the sixth node E 1.2 , the seventh node E 1.3 , the eighth node E 1.4 on the edge portion and the eleventh node T 2.1 , the twelfth node T 2.2 , the thirteenth node T 2.3 , the fourteenth node T 2.4 on the second lateral mark T2 of the second end portion 108 and the fifteenth node E 2.1 , the sixteenth node E 2.2 , the seventeenth node E 2.3 , the eighteenth node E 2.4Alignment. In such a perfect joint P, there is a sufficient amount of available rubber between the first end and the second end, and the rubber does not expand or compress, so that no gap or overlap defect or offset defect is generated. In the example, if the first predetermined distance D1 and the second predetermined distance D2 between the first transverse mark T1 and the second transverse mark T2 and the first end 106 and the second end 108 are 75 mm respectively, the calculated distances C1, C2, C3, C4 between the nodes on the first transverse mark T1 and the second transverse mark T2 are 150 mm, and there is no gap or overlap between the joint ends 102.

[0046] In the embodiments of the present invention as shown in Figure 4 and Figure 5 if the first node T 1.1 on the first transverse mark T1 of the first end 106, the second node T 1.2 the third node T 1.3 the fourth node T 1.4 calculated by the processing unit and the eleventh node T 2.1 the twelfth node T 2.2 the thirteenth node T 2.3 the fourteenth node T 2.4 on the second transverse mark T2 of the second end 108, the distance C1, C2, C3, C4 between them is greater than the sum (D1 + D2) of the first predetermined distance and the second predetermined distance defined after the manufacture and segmentation of the tread 104, then the joint between the first end 106 and the second end 108 is called an imperfect joint due to the gap defect G. In addition, the first node T 1.1 the second node T 1.2 the third node T 1.3 the fourth node T 1.4 and the fifth node E on the edge part 1.1 the sixth node E 1.2 the seventh node E 1.3 the eighth node E 1.4 on the first transverse mark T1 of the first end 106 and the eleventh node T 2.1 the twelfth node T 2.2 the thirteenth node T 2.3 the fourteenth node T 2.4 and the fifteenth node E on the edge part 2.1 the sixteenth node E 2.2 the seventeenth node E 2.3 the eighteenth node E 2.4Alignment. In such an imperfect joint, there is a lack of available rubber between the first end 106 and the second end 108, resulting in a gap defect G. The gap defect G is mainly caused by the plastic properties of the rubber, which causes the rubber to expand, resulting in a gap between the first end 106 and the second end 108. In the example, if the first transverse mark T1 and the second transverse mark T2 are at a first predetermined distance D1 and a second predetermined distance D2 of 75 mm from the first end 106 and the second end 108 respectively, the calculated distances C1, C2, C3, C4 between the nodes on the first transverse mark T1 and the second transverse mark T2 are 154 mm, leaving a gap between the joint ends 102.

[0047] In an embodiment of the present invention as shown in Figure 6 and Figure 7 if the first node T 1.1 on the first transverse mark T1 of the first end 106, the second node T 1.2 the third node T 1.3 the fourth node T 1.4 calculated by the processing unit and the eleventh node T 2.1 the twelfth node T 2.2 the thirteenth node T 2.3 the fourteenth node T 2.4 on the second transverse mark T2 of the second end 108, and the distances C1, C2, C3, C4 between them are less than the sum (D1 + D2) of the first predetermined distance and the second predetermined distance defined after the manufacture and segmentation of the tread 104, then the joint between the first end 106 and the second end 108 is called an imperfect joint due to an overlap defect O. In addition, the first node T 1.1 the second node T 1.2 the third node T 1.3 the fourth node T 1.4 and the fifth node E 1.1 the sixth node E 1.2 the seventh node E 1.3 the eighth node E 1.4 on the first transverse mark T1 of the first end 106 and the eleventh node T 2.1 the twelfth node T 2.2 the thirteenth node T 2.3 the fourteenth node T 2.4 and the fifteenth node E 2.1 the sixteenth node E 2.2 the seventeenth node E 2.3 the eighteenth node E 2.4Alignment. In such an imperfect joint, there is an excessive amount of available rubber between the first end 106 and the second end 108, resulting in an overlap defect O. The overlap defect O is mainly due to the plastic properties of the rubber causing compression of the rubber, resulting in an overlap between the first end 106 and the second end 108. In the example, if the first transverse mark T1 and the second transverse mark T2 are at a first predetermined distance D1 and a second predetermined distance D2 from the first end 106 and the second end 108 respectively of 75 mm, then due to the overlapping joint between the joined ends 102, the calculated distances C1, C2, C3, C4 between the nodes on the first transverse mark T1 and the second transverse mark T2 are 148 mm.

[0048] In an embodiment of the present invention as shown in Figure 8 , the joint between the first end 106 and the second end 108 is also referred to as a perfect joint P or an imperfect joint caused by a combination of at least one of a gap defect G and an overlap defect O. Due to the non-uniform expansion or compression of the rubber at several nodes on the first transverse mark T1 and the second transverse mark T2, if the first node T 1.1 on the first transverse mark T1 of the first end 106, the second node T 1.2 , the third node T 1.3 , the fourth node T 1.4 calculated by the processing unit and the eleventh node T 2.1 on the second transverse mark T2 of the second end 108, the twelfth node T 2.2 , the thirteenth node T 2.3 , the fourteenth node T 2.4 and the distance C1, C2, C3, C4 between them is greater than or less than the sum of the first predetermined distance and the second predetermined distance (D1 + D2), then an imperfect joint is determined.

[0049] In this specific example of this embodiment showing a combination of defects as shown in Figure 8 , if the calculated distance C1 between the nodes T 1.1 and T 2.1 is greater than the sum of the first predetermined distance and the second predetermined distance (D1 + D2), then a gap defect G may be caused. In addition, if the calculated distances C2, C3 between the second node T 1.2 , the third node T 1.3 and the twelfth node T 2.2 , the thirteenth node T 2.3 are equal to the sum of the first predetermined distance and the second predetermined distance (D1 + D2), then a defect-free perfect joint P is produced. In a similar example, if the calculated distance between the fourth node T 1.4 and the fourteenth node T 2.4If the distance C4 between them is less than the sum of the first predetermined distance and the second predetermined distance (D1 + D2), it may result in an overlap defect O. In an imperfect joint caused by a combination of defects, several fifth nodes E on the edge portion of the first end 106 1.1 , sixth nodes E 1.2 , seventh nodes E 1.3 , eighth nodes E 1.4 and fifteenth nodes E on the edge portion of the second end 108 2.1 , sixteenth nodes E 2.2 , seventeenth nodes E 2.3 , eighteenth nodes E 2.4 have excessive rubber, lack of rubber, or a sufficient amount of available rubber.

[0050] In an embodiment of the present invention as shown in Figure 9 , the joint between the first end 106 and the second end 108 is also referred to as an imperfect joint due to the misalignment of the nodes on the joint end 102. If the first node T on the first lateral mark T1 of the first end 106 1.1 , second node T 1.2 , third node T 1.3 , fourth node T 1.4 and fifth nodes E on the edge portion 1.1 , sixth nodes E 1.2 , seventh nodes E 1.3 , eighth nodes E 1.4 and eleventh nodes T on the second lateral mark T2 of the second end 108 2.1 , twelfth nodes T 2.2 , thirteenth nodes T 2.3 , fourteenth nodes T 2.4 and fifteenth nodes E on the edge portion 2.1 , sixteenth nodes E 2.2 , seventeenth nodes E 2.3 , eighteenth nodes E 2.4 are not aligned, the imperfect joint between the first end 106 and the second end 108 will result in an offset joint J. Such an imperfect joint may be due to the change in the thickness of the first end 106 and the second end 108 caused by the plastic properties of the rubber resulting in compression or expansion.

[0051] In an embodiment of the present invention, in the case of an inconsistent or imperfect joint caused by at least one of a gap defect G, an overlap defect O, or an offset defect J, the joining operation is interrupted and an alarm can be issued to the technician. Depending on the defect, the technician can repair the imperfect joint if possible, or can replace the defective tread 104 with a new tread 104. A processing unit including one or more software or algorithms or transformation models can be trained based on perfect joints or identified defects (including gap defect G, overlap defect O, offset defect J) to predict the quality of subsequent joints or expected defects. Such training techniques can establish optimizations or adjustments to the algorithms to change the parameters of the joining machine in order to anticipate or correct such performance.

[0052] The image processing module can deploy one or more machine learning models using the parameters of the tread 104 captured from the sensor to identify the joining end 102 of the tread 104. Although embodiments regarding the use of neural networks (specifically convolutional neural networks (CNNs)) as machine learning models are described herein, other types of machine learning models can also be used. These machine learning models include but are not limited to models employing linear regression, logistic regression, decision trees, support vector machines, naive Bayes, K-nearest neighbors (kNN) (where K represents the grouping), random forests, dimensionality reduction algorithms, gradient algorithms, neural networks (such as autoencoders, CNNs, RNNs, perceptrons, long short-term memory (LSTM), Hopfield, Boltzmann, deep belief networks, deconvolution, generative adversarial network (GAN), etc.) and their complements and equivalents.

[0053] In an embodiment of the present invention, the image processing module of the processing unit can include a transformation model designed to identify a first lateral mark T1 and a second lateral mark T2 on the tread 104. In one example, the Hough transformation model can be applied to detect lines or marks in the captured image of the tread 104.

[0054] In a preferred embodiment of the present utility model, the image processing module may deploy one or more machine learning models based on a neural network architecture, which is defined as object as points detection and was proposed as CenterNet in 2014. The model "CenterNet: Object as Points" is one of the milestones of anchor-free object detection algorithms. Anchor-free object detection is a technique for directly predicting the bounding boxes of an image relative to fixed reference objects in the image. An object is modeled as the center point of its bounding box. The size of the bounding box and other object attributes are inferred or regressed based on the key point features of the center (see https: / / arxiv.org / pdf / 1904.07850.pdf).

[0055] In one aspect of the present utility model, the CenterNet model includes an encoder-decoder for processing the input image of the tread 104. The encoder is configured to iteratively reduce the resolution of the captured image and may increase the depth of the image. In a specific example, if the resolution of the captured image is 1024×1024×3 (RGB image => 3), the encoder may iteratively reduce the spatial dimension to 16×16 using convolution operations, but increase the depth to 1024. Convolution operations are defined as applying filters to retain the features of interest in the captured image. This lower-resolution "image" with a specific resolution is called a "feature map". The decoder is configured to enlarge the feature map using transposed convolution. To reduce the computational cost, the feature map may be enlarged at half the resolution (512×512). In addition, a geometric grid may be applied at half the size resolution of the captured input image.

[0056] In an embodiment according to the present utility model, since there is only one object in each image, a centre detection head and a relative detection head may not be required. The centre detection head is defined as identifying key points at the center of each of multiple objects in the image. The relative detection head is defined as N key points associated with the central key point. In addition, the CenterNet model may include two heads, namely for 16 key nodes (T 1.1 、T 1.2 、T 1.3 、T 1.4 、E 1.1 、E 1.2 、E 1.3 、E 1.4 、T 2.1 、T 2.2 、T 2.3 、T 2.4 、E 2.1 、E 2.2 、E2.3 , E 2.4 ), and an offset prediction head for 32 points, which is twice the number of key points, capable of predicting offsets related to the reduction in resolution due to downsampling in the neural network and may be necessary due to memory limitations.

[0057] The processing unit can configure the inspection system 100 according to one or more parameters of the joint end 102 of the tread 104 calculated by the image processing module. The processing unit can also refer to reference data (e.g., size tables of various treads) to finally determine one or more target tread parameters. The reference data can include known tread parameters corresponding to a plurality of known commercially available treads. For example, after the image processing module calculates one or more tread parameters, the processing unit can compare the calculated tread parameters with the known tread parameters recorded in the reference data. The image processing module can also be trained by data augmentation techniques, which artificially increase the training set by creating modified copies of the dataset using existing data. This includes making minor changes to the dataset or using deep learning to generate new data points. In the present utility model, the image of the tread 104 can be transformed by geometric transformation and color space transformation. The geometric transformation includes flipping, cropping, rotating, translating, and kernel filters to sharpen or blur the image, and the color space transformation includes changing the RGB color channels, enhancing any color, and changing the brightness. By performing data augmentation techniques, the algorithm can be trained and made more robust to environmental changes.

[0058] Identifying the joint end 102 of the tread 104 is related to the representation and can be found by post-processing the previously generated segmentation of the tread 104. For example, methods can be used to determine whether a pixel is a candidate pixel for the region of the joint end 102, which includes the first end 106, the second end 108, the first first edge part 110, the second first edge part 116, the first second edge part 112, the second second edge part 118, and the first central part 114, the second central part 120. For example, an active contour model can be applied together with path planning and distance transformation to extract parts of the tread 104. A morphological-based level set model can be used to perform the extraction of the region of the tread 104 by learning the structural pattern of an object similar to the target tread 104 and estimating the joint end 102 of the object as a path. Therefore, the present utility model utilizes artificial intelligence (or "AI")-based methods and tools to supplement the partial information provided by perception.

[0059] In such as Figure 10In the embodiment of the present invention shown, the image processing module is configured to apply geometric correction to compensate for the non-planarity of the measurement area. Due to the radius of the tread 104, the calculated distance measured between the first lateral mark T1 and the second lateral mark T2 on the first end 106 and the second end 108 may be curved. Since the captured image detected on the image captured by the camera is planar, geometric correction needs to be applied.

[0060] As Figure 10 shown, let us consider 3 surfaces A0, A3, and A4 and the relationship A0 = A3 + A4

[0061] A0 = (D + R).C

[0062] A3 = R.Cos(α).C

[0063] A4 = (D + r).C

[0064] C = Cord / 2

[0065] Therefore, using the relationship A0 = A3 + A4

[0066] (D + R).C = R.Cos(α).C + (D + r).C

[0067] r = R.(1 - R.Cos(α)) - (Equation 1)

[0068] Constraint on Cord (C):

[0069] (D + r).Tan(θ) = R.Sin(α) - (Equation 2)

[0070] By substituting r from Equation 1 into Equation 2

[0071] (D + R.(1 - R.Cos(α))).Tan(θ) = R.Sin(α)

[0072] (D + r).Tan(θ) = R.Sin(α) + R.Cos(α).Tan(θ) - (Equation 3)

[0073] By multiplying by Cos(θ) in Equation 3

[0074] (D + r).Sin(θ) = R.Sin(α).Cos(θ) + R.Cos(α).Sin(θ) - (Equation 4)

[0075] According to the Sin(a + b) formula for the right-hand side term in Equation 4

[0076] (D + r).Sin(θ) = R.Sin(α + θ) - (Equation 5)

[0077] Therefore,

[0078] α = Arcsin(((D + R) / R).Sin(θ)) – θ - (Equation 6)

[0079] Y Image = D.Tan(θ) - (Equation 7)

[0080] Y Tire = R.[(((D + R) / R).Sin(θ)) - θ] - (Equation 7)

[0081] Therefore, the final equation

[0082] Y Tire = R.[Arcsin(((D + R) / R).Sin(Arctan(Y Image / D)) – Arctan

[0083] (Y Image / D))] – (Equation 8)

[0084] Although specific embodiments of the disclosed device have been shown and described, it should be understood that various changes, additions, and modifications can be made without departing from the spirit and scope of the present utility model. Therefore, except for the limitations set forth in the appended claims, the scope of the described utility model should not be limited.

Claims

1. An inspection system (100) for monitoring a joint end (102) of a film product (104), the film product (104) having a profile defined by parameters of a predetermined length, a predetermined width and a predetermined thickness, characterized in that: The joint end portion (102) includes a first end portion (106), a second end portion (108), a first first edge portion (110), a second first edge portion (116), a first second edge portion (112), a second second edge portion (118), and a first central portion (114), a second central portion (120) defining an end surface of the joint end portion (102) along a longitudinal axis, and the inspection system (100) includes: - at least one marking unit configured to print a first transverse mark (T1) and a second transverse mark (T2) on the first end portion (106) and the second end portion (108) at a first predetermined distance and a second predetermined distance from the first first edge portion (110), the second first edge portion (116), the first second edge portion (112), the second second edge portion (118) and the first central portion (114), the second central portion (120); - at least one detection unit (122) configured to capture one or more images to identify a first transverse mark (T1) and a second transverse mark (T2) on the film product (104), the image being formed by pixels associated with distances of the first transverse mark (T1) and the second transverse mark (T2) from the first first edge portion (110), the second first edge portion (116), the first second edge portion (112), the second second edge portion (118) and the first central portion (114), the second central portion (120); - at least one processing unit comprising one or more software or algorithms or transformation models configured to process the image of the film product (104) through a geometric grid so as to define one or more nodes on the first transverse mark (T1) of the first end (106) and one or more nodes on the first first edge portion (110), the first second edge portion (112) and the first central portion (114) of the first end (106); - at least one processing unit comprising one or more software or algorithms or transformation models configured by means of a geometric grid so as to be able to define one or more nodes on a second transverse mark (T2) of the second end (108) and one or more nodes on a second first edge portion (116), a second second edge portion (118) and a second central portion (120) of the second end (108); and - a processing unit configured to define a joint based on a calculated distance between a node on a first transverse marking (T1) of the first end (106) and a node on a second transverse marking (T2) of the second end (108).

2. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1, characterized in that: If the distance between the node on the first transverse mark (T1) of the first end (106) and the node on the second transverse mark (T2) of the second end (108) calculated by the processing unit is equal to the sum of the first predetermined distance and the second predetermined distance, the defined joint is a perfect joint (P) without defects.

3. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1, characterized in that: If the distance between a node on a first transverse mark (T1) of a first end (106) and a node on a second transverse mark (T2) of a second end (108) calculated by the processing unit is greater than the sum of a first predetermined distance and a second predetermined distance, the defined joint is an imperfect joint due to a gap defect (G).

4. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1, characterized in that: If the distance between a node on a first transverse mark (T1) of a first end (106) and a node on a second transverse mark (T2) of a second end (108) calculated by the processing unit is less than the sum of a first predetermined distance and a second predetermined distance, the defined joint is an imperfect joint due to an overlap defect (O).

5. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1, characterized in that: If the nodes on the first transverse mark (T1) and the nodes on the edge portion of the first end (106) are not aligned with the nodes on the second transverse mark (T2) and the nodes on the edge portion of the second end (108), the defined joint is an imperfect joint due to an offset defect (J).

6. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1, characterized in that: The defined joint is a perfect joint (P) or an imperfect joint due to a combination of at least one of a gap defect (G), an overlap defect (O), and an offset defect (J).

7. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1, characterized in that: A processing unit including one or more software or algorithms or transformation models is trained based on a perfect joint (P) or an identified defect including at least one of a gap defect (G), an overlap defect (O), and an offset defect (J) to predict the quality or expected defects of subsequent joints.

8. The inspection system (100) for monitoring the joint end (102) of a film product (104) according to claim 1 or 6, characterized in that: The processing unit includes an image processing module that deploys one or more transformation models defined for object as point detection based on a CenterNet neural network architecture.

9. The inspection system (100) for monitoring a joint end (102) of a film product (104) according to claim 1, characterized in that: The processing unit includes an image processing module that deploys one or more transformation models including an encoder-decoder configured to reduce or increase the resolution of a captured image.

10. The inspection system (100) for monitoring a joint end (102) of a film product (104) according to claim 1, characterized in that: The processing unit comprises an image processing module configured to apply a geometric correction to compensate for non-planarity of a measurement area.

11. The inspection system (100) for monitoring a joint end (102) of a film product (104) according to claim 1, characterized in that: The distance sensor is located at the detection unit (122) to measure the distance between the detection unit (122) and the surface of the film product (104), thereby converting the pixel area into a distance suitable for the diameter change of the surface of the film product (104).

Images