System and method for acquiring and determining axis of valve in tire vulcanization mold

Abstract

The invention relates to a system (100) implementing a method for identifying vent holes (150) in a tyre vulcanisation mould (10), said tyre vulcanisation mould (10) comprising one or more sections (10A) and an inner surface (10a), said vent holes being distributed on the inner surface (10a) such that corresponding valves (200) can be inserted therein. The invention also relates to a method implemented by the disclosed system (100).

Description

Technical Field

[0001] The present invention relates to a system and method for inserting a valve into a segment of a tire curing mold. More specifically, the present invention relates to a system and method for identifying vent holes in a tire curing mold, the vent holes of which are distributed to allow corresponding valves to be inserted therein. Background Art

[0002] In the tire art, segmented curing molds are well known. Figure 1 This type of mold is represented by mold 10, which mainly includes two shells (not shown), each shell molding one of the sidewalls of the tire P, and a plurality of segments 12, each of which molds the tread P of the tire P along the inner surface 12a of the segment. 10 The section 12 can be in the open position of the mold 10 (such as Figure 1 The mold of this type may further include at least one clamping ring (not shown) to allow radial movement of the segments. Applicant's patent US10,239,270 discloses an example of this type of mold.

[0003] Manufacturing a tire using this type of mold requires applying pressure to the green tire in order to press it against the inner surface of the mold while applying heat to the mold (for example by electrical induction and / or by magnetic induction, or by a heat transfer fluid, such as pressurized steam). For this reason, this type of mold must be vented so that the green tire expands against the inner surface of the mold segments.

[0004] Therefore, it is also known that this type of mold includes a plurality of ventilation holes (or "vents") to enable such ventilation during the vulcanization cycle. For example, a typical segment mold may include 4,000 to 12,000 substantially cylindrical vents distributed along each segment of the mold. Each vent includes Figure 2 20 (see, for example, patent EP774333B1). The valve 20 comprises a movable insert 22 which rises and falls in a substantially cylindrical housing 24. The movable insert 22 comprises a valve stem 26 which has a valve body 28 which faces an inner cavity 28 (see Figure 2 ) and a flat surface 26b facing the tire surface. The conical portion 26a cooperates with the seat surface 24a of the housing 24 so that during the vulcanization cycle, the valve is closed by the surface of the untreated tire approaching, and after vulcanization, when the tire is removed, the valve is reopened. A gasket (not shown) can be provided between the conical portion 26a and the seat surface 24a in a manner understood by those skilled in the art.

[0005] The valve is in the form of a small tubular and rigid mechanical part (for example of diameter of about 2.5 mm and length of 5 to 12 mm). Furthermore, the installation of the valve in the mould involves forced fitting into a vent hole whose perforated diameter guarantees the regulation and resistance of the valve throughout the life of the mould.

[0006] The installation requires:

[0007] - positioning the vent for insertion of the valve;

[0008] - Clamp the valve in the correct orientation;

[0009] - positioning the valve in the exhaust hole;

[0010] - generate the required force to bring it into regulation; and

[0011] - Apply pressure until the valve is flush.

[0012] The valves are placed individually (either by a human operator or by a mechanical operator such as a robot) in the sections of the mold. This operation is usually performed using a tweezer-type tool that grips the valve and inserts it accurately into the corresponding vent hole of the mold. The valve is then hammered into the vent hole by means of a hammer and chuck. This type of insertion requires a lot of effort and a lot of time. Each insertion represents several seconds of work, which results in a repetitive, tedious task that is not very interesting for a human operator. This leads to fatigue and the risk of forgetting the valve, which calls the correct operation of the mold into question.

[0013] In order to overcome this problem, there are devices for inserting valves into the mold in the prior art. For example, German publication DE102010060901 discloses a tool comprising a tubular guide system in which a valve is arranged. The tubular system is placed in alignment with the exhaust port, where the piston pushes the valve by a force on the axis of the valve so as to press the valve in a guided and regulated manner. The piston rises by a spring and the new valve engages in the tubular system. Therefore, aligning the exhaust port involves automation, but the valves must be precisely positioned in order to find their reference marks.

[0014] Korean patent KR100845093B discloses a system for assembling valves, which includes a machine for manufacturing valves, divided into two parts: a body with a spring mounted and the valve itself. The machine can be reoriented so as to act as a base for the system for mounting valves, so as to insert the valve into the exhaust hole. However, it lacks the ability to adapt to any mold shape, nor the ability to move to position the valve in the section.

[0015] In reality, the perforations that create the vent holes are not always created as shown in the drawings, and there are variations due to deviations in the manufacturing method (e.g., adding vent holes or manually modifying the mold). Since precise knowledge of the location of the vent holes and / or their axes cannot be absolutely guaranteed, it is desirable to develop a system that knows how to operate without this information, just like a human operator who detects and analyzes the situation himself.

[0016] Therefore, the disclosed invention uses the knowledge of the mold sections to insert the valve repeatedly. The force of inserting the valve can reach about 70 kg, which involves the correct management of the robot's trajectory to avoid damaging the mold. To this end, the disclosed invention uses the coordinates of the vent hole and the detection of its center and normal to provide the robot with the correct approach and push trajectory to facilitate the installation of the valve. Summary of the invention

[0017] The present invention relates to a system for implementing a method for identifying vent holes in a tire vulcanization mold, the tire vulcanization mold comprising one or more sections and an inner surface, the vent holes being distributed on the inner surface to allow corresponding valves to be inserted therein, characterized in that the system comprises:

[0018] - a robot comprising a detection system having one or more sensors that detect the presence of one or more vent holes distributed along the inner surface of a segment of the mold;

[0019] - a communications network that manages the input data into the system from the detection systems; and

[0020] - one or more communication servers, each communication server comprising one or more processors operably connected to a memory, the memory being configured to store an application for analyzing data representing an imaged mold, the one or more processors comprising a module for executing the analysis application for processing the image, the one or more processors being capable of executing programmed instructions stored in the memory to perform the following steps:

[0021] - detecting the presence of the arrangement of vent holes in the field of view of a detection system, said detection system being triggered to capture at least one image of the inner surface of a section of the mould; and

[0022] - A step of searching for the presence of detected vents in the images captured by the inspection system, such that if no vents are detected, the inspection system continues to capture images until the mold search is complete.

[0023] In some embodiments of the system of the present invention, the detection system includes at least one RGB-D type three-dimensional (3D) camera fixed to the robot, and the three-dimensional (3D) camera provides a 3D image represented by a set of 3D points with coordinates (X, Y, Z).

[0024] In some embodiments of the system of the present invention, the one or more processors are capable of executing programmed instructions stored in the memory to perform the following steps:

[0025] - a step of annotating the location of the sample of vents, comprising a step of creating a coordinate reference of the vents searched for in the images captured by the detection system of the system;

[0026] - a step of reconstructing the segment, comprising the step of building an annotation database storing the captured images and the coordinates of the pixels of the captured images;

[0027] - a step of analyzing the profile of the venting hole, performed by a module for executing an analysis application of the system, comprising a step of determining a surface plane by finding the shape that comes closest to a circle representing the searched venting hole, further comprising a step of determining a vector perpendicular to the determined surface plane in order to find the axis of insertion of the valve; and

[0028] - a step of determining the diameter of the exhaust holes allowing a valve having a corresponding diameter to pass therethrough, in which step each exhaust hole is identified by the contour analyzed in the step of analyzing the contour of the exhaust hole, and in which step the corresponding center is identified by a point;

[0029] The system receives the coordinates of the identified exhaust orifice to select a valve having an appropriate diameter.

[0030] In some embodiments of the system of the present invention, a module for executing an analysis application stored in a memory of the system uses annotation software to construct a bounding box around vents that appear on a captured image of a mold.

[0031] In some embodiments of the system of the present invention, a processor of the system continuously forms at least one neural network that outputs a classification of the coordinates of the vents so that the captured images show the locations of the vents.

[0032] In some embodiments of the systems of the invention, the at least one neural network is selected from convolutional neural networks.

[0033] In some embodiments of the system of the present invention, one or more steps use a deformable transformer type neural network.

[0034] In some embodiments of the system of the present invention, the robot comprises a peripheral gripping member supported by a pivotable elongated arm, the peripheral gripping member extending from the elongated arm to a free end at which a gripper is arranged along a common longitudinal axis.

[0035] In some embodiments of the system of the present invention, the clamp includes a pivotable clamp, the pivotable clamp including clamping fingers extending from a platform where the clamp is fixed to the free end of a peripheral clamping member, each finger including a member of a predetermined length extending between an actuating end where movement of the finger occurs and an opposite clamping end where the finger clamps the valve.

[0036] In some embodiments of the system of the present invention, the one or more processors are capable of executing programmed instructions stored in the memory to perform steps of moving the robot so that the robot can place the valve so that the valve can be inserted into the identified vent hole in the section of the mold.

[0037] The present invention also relates to a method implemented by the disclosed system for identifying vent holes in a tire vulcanization mold, the tire vulcanization mold comprising one or more segments and an inner surface, the vent holes being distributed on the inner surface to allow the corresponding valve to be inserted therein, characterized in that the method comprises the following steps:

[0038] - a step of positioning the mould in the field of view of a detection system of the system so that the vents defined along the inner surface of at least one section are visible, during which step the detection system is passed over the mould;

[0039] - detecting the presence of the arrangement of vent holes in the field of view of a detection system, said detection system being triggered to capture at least one image of the inner surface of a section of the mould; and

[0040] - A step of searching for the presence of detected vents in the images captured by the inspection system, such that if no vents are detected, the inspection system continues to capture images until the mold search is complete.

[0041] In some embodiments of the method of the present invention, the method further comprises an inspection step performed after the valve is inserted into the vent hole of the mold.

[0042] In some embodiments of the method of the present invention, the method further comprises a final step of positioning the robot aligned with the identified vent hole on its insertion axis, in which step the robot blows into the valve.

[0043] Other aspects of the invention will become apparent from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The nature and various advantages of the present invention will become more apparent from the following detailed description read in conjunction with the accompanying drawings, in which like reference numerals refer to like parts throughout, and in which:

[0045] Figure 1 A perspective view showing an embodiment of a segmented vulcanization mold;

[0046] Figure 2 Shows the insertion Figure 1 Embodiment of the valve in the exhaust hole of the mold;

[0047] Figure 3 A schematic diagram of the system of the present invention is shown which allows the valve to be inserted into a tire curing mold;

[0048] Figure 4 The inner surface of a segment of a tire curing mold is shown, with vent holes formed by Figure 3 System identification;

[0049] Figure 5 shows an example of an annotated bounding box constructed around a vent hole appearing on an image of a mold captured in a method performed by the system of the present invention;

[0050] Figure 6 shows an example of points defining vents appearing on an image of a mold captured in a method performed by the system of the present invention;

[0051] Figure 7 shows examples of points extracted and forming one or more ellipses in the method performed by the system of the present invention;

[0052] Figure 8 showing normal vectors of vent holes identified in an image of a mold captured in a method performed by the system of the present invention;

[0053] Fig. 9 Vent holes identified by contours analyzed in the method performed by the system of the present invention are shown. DETAILED DESCRIPTION

[0054] Reference will now be made to the drawings, wherein like reference numerals represent like elements, and wherein: Figure 3 A system 100 (or "system") of the present invention for inserting a valve is shown. The system 100 is implemented for inserting a valve (e.g. Figure 2 ) is inserted into a tire curing mold (e.g. Figure 1The disclosed method includes a method of the present invention in a section of a mold 10 of the type shown in FIG. 1 and having a section 12. The disclosed method includes a machine learning method, the algorithm used to analyze the inner surface of the mold based on data corresponding to an image obtained from the mold to place and insert the valve into the identified vent hole.

[0055] refer to Figure 3 , the mold 10 is positioned on a table or equivalent support 50 so that it can be processed by the system 100. The support 50 can be configured to move in a rotational manner, in a reciprocating vertical manner, and / or in a reciprocating horizontal manner, thereby enabling processing of a variety of molds.

[0056] Further references Figure 3 In one embodiment of the system 100, the system includes a robot 102 having a peripheral clamping member 104 supported by a pivotable long arm 106. The peripheral clamping member 104 extends from the long arm 106 to a free end 104a, at which a clamp 108 is arranged along a common longitudinal axis. The clamp 108 can be fixed to the peripheral clamping member 104 by tightening an adapter to the free end 104a of the peripheral clamping member. It should be understood that fixing the clamp 108 to the peripheral clamping member 104 can be performed by one or more known fixing means (including but not limited to welding, bonding and equivalent means).

[0057] In one embodiment of the clamp 108, the clamp includes a pivotable clamp 108a including clamping fingers 108b (or "fingers") extending from a platform 108c (where an adapter secures the clamp 108a to the free end 104a of the peripheral clamping member 104). Each finger 108b includes a member having a predetermined length extending between an actuation end (where movement of the finger occurs) and an opposite clamping end (where the finger clamps a valve 200 held by the clamp in a method implemented by the system 100). Each finger 108b has an inner clamping surface and an opposite outer surface, the inner clamping surface engaging the valve 200 when the valve 200 is inserted into the identified vent hole. The fingers 108b are arranged to define a predetermined space between the clamping surfaces, thereby allowing the fingers to move along a common axis in a method implemented by the system 100. Thus, the robot 102 facilitates grasping of various valves without interrupting the linear motion of the fingers.

[0058] The reciprocating movement of the one or more fingers 108b can be achieved by one or more known cylinders, which are actuated by a pressurized fluid (e.g., compressed air) from a pipeline (not shown). Thus, the movement of each finger 108b performs the movement of the finger between a waiting position (where the clamping surfaces remain substantially parallel to the space between them) (not shown) and a clamping position (where the clamping surfaces approach each other to engage the valve 200 and place it in an insertion position relative to the inner surface 10a of the mold 10) (see Figure 3 ) corresponding linear motion between. The one or more cylinders are selected from commercially available cylinders.

[0059] In the method implemented by the system 100, the robot 102 can be moved so that the gripper 108 can grip the valve 200 (described below). By means of the finger 108b, when the gripper is in the gripping position (where the gripper 108 grips the valve selected for insertion into the corresponding identified exhaust hole) (see Figure 3 ) and an insertion position (where the gripper 108 positions the gripped valve so that it can be inserted into the identified vent hole) (not shown), the gripper 108 grips to clamp the valve 200. In an embodiment of the gripper 108 that includes gripping fingers 108b, the gripping position of the gripper 108 means that the fingers are in a position for gripping a selected valve. In all embodiments of the robot 102, the robot can be configured to have six degrees of freedom, thereby allowing it to move along six axes. In all embodiments, the robot 102 can be arranged on a support 55 that is configured to move in a rotational manner, in a reciprocating vertical manner, and / or in a reciprocating horizontal manner, thereby enabling processing of a variety of molds.

[0060] The robot 102 moves to place the valve 200 so that the valve 200 can be inserted into the vent hole identified in the section 10A of the mold 10. In one embodiment of the system 100, the robot 102 can form part of a roaming robot that can move by an integrated movement device (e.g., one or more integrated motors) or by a non-integrated movement device (e.g., one or more autonomous movable carts or other equivalent movable devices). In another embodiment of the system 100, the robot 102 can be attached to a ceiling, floor, wall, or any support capable of performing the method implemented by the system 100 (see, e.g., Figure 3 It should be understood that such a robot may be a conventional industrial robot or a collaborative robot, or even a delta robot or a cable robot.

[0061] The robot 102 includes a detection system that utilizes one or more sensors (not shown) to detect information about the physical environment surrounding the robot. In the following description, the terms "sensor," "photographic device," "camera," and "optical sensor" may be used interchangeably and may refer to one or more devices configured to detect two-dimensional (2D) and / or three-dimensional (3D) images to achieve 3D depth perception and / or other types of detection of the physical environment surrounding the robot 102. In an embodiment of the system 100, the sensors of the detection system combined with the robot 102 may be fixed to the robot's long arm 106 (e.g., at the end 104a) and / or the gripper 108.

[0062] One or more sensors of the detection system of the robot 102 detect the presence of one or more vents of the mold. As an example, Figure 4 The inner surface 10a ( Figure 4 A photograph of the inner surface 10a taken with an RGB type camera is shown. A plurality of vent holes 150 are distributed along the inner surface 10a of the section 10A, each vent hole accommodating a corresponding valve 200. It should be understood that each vent hole is substantially cylindrical and all vent holes 150 have substantially similar diameters.

[0063] In some embodiments of the robot 102, the sensor is triggered when a segment of the mold enters the field of view of the camera. In the event that the mold part is not visible in an image acquired by an inspection system of the robot 102 (e.g., a camera of the inspection system), an attachment point can be placed at a known location relative to the sensor (e.g., at a known horizontal distance and a known vertical distance relative to the location of the sensor).

[0064] The detection system can determine information about the physical environment surrounding the mold 10, which can be used by the control system of the system 100 (the control system includes, for example, software for planning the movement of the robot 102). The control system can be located on the robot 102, or can communicate with the robot remotely. In some embodiments of the system 100, one or more 2D or 3D sensors (including but not limited to navigation sensors) mounted on the robot 102 can be integrated to form a digital model of the physical environment (including one or more sides, floors, and ceilings, where applicable). By using the acquired data, the control system can move the robot 102 to navigate between positions for clamping the valve when inserting the valve into the mold 10.

[0065] In one embodiment of the system 100, the detection system includes at least one camera (referred to as an "RGB-D type camera") that provides a 3D image represented as a set of 3D points having coordinates (X, Y, Z) and sometimes having red, green, blue values ​​("RGB" or "RGB-D" format). In this embodiment, the RGB-D type camera is fixed to the robot 102 (e.g., at the end 104a and / or on the gripper 108). Two or more RGB-D cameras can be oriented so that a predetermined overlap is obtained between the fields of view of the cameras. As used herein, the term "camera" includes one or more cameras.

[0066] RGB-D cameras typically provide depth information using a depth map, which is an image that contains the distance between the camera and the corresponding point in space for each pixel. 3D point cloud data from RGB-D cameras has a higher measurement rate than traditional measurement methods (such as manual measurement and other electronic device-based measurements). Using a sparser structure, a point cloud can be constructed from the RGB-D image by calculating the real world (such as coordinates (X, Y, Z)) using the inherent data of the digital camera. Therefore, information about the physical environment surrounding the system 100 is obtained from the 3D point cloud data obtained by detection technologies that can accurately and efficiently capture the 3D surface geometry of the mold. These detection technologies can be selected from commercially available devices (for example, selected from ZIVID AS under the brand name cameras sold by Cognex Corporation, artificial vision systems sold by Cognex Corporation, and their equivalents).

[0067] The term "point cloud" (singular or plural) is used herein to refer to a collection of one or more data points in space. One or more cameras (or one or more equivalent devices) can collect three-dimensional (3D) data and detect the surface of an object (e.g., section 10A of mold 10) with the aid of a series of coordinates. Storing information in the form of a collection of spatial coordinates can save space because many objects do not occupy a large part of the environment. Even if the information is not visual, interpreting the data as a point cloud helps to understand the relationship between multiple variables through classification and segmentation.

[0068] It should be appreciated that one or more cameras may include one or more programming modes, including a learning mode, to provide, modify and train at least one neural network.

[0069] The inspection system of the robot 102 detects the presence of the arrangement of vent holes 150 in the field of view of the inspection system (e.g., the field of view of a camera of the system 100), which triggers the inspection system to capture an image of the inner surface 10a of the section 10A of the mold 10 (see Figure 4). In all embodiments of the system 100, the system "searches" the images acquired by the detection system for the presence of vents "seen" by the robot 102. If no vents are detected, the detection system continues to acquire images until the mold 10 is searched. Points are extracted around the perimeter of each detected vent to determine its center in preparation for inserting the corresponding valve.

[0070] The detection system of system 100 may include a distance measuring device, which is used in the working space of the mold 10 to infer its size. In this embodiment, the distance measuring device includes a scanner (not shown) for scanning the entire inner surface 10a of the mold 10 in real time in the physical environment around the mold. Such a scanner can accurately generate the mold. The scanner can be provided with an observation system (not shown) configured to accurately locate the exhaust hole in the real-time scene based on the 3D profile generated by the scanner.

[0071] The observation system can receive the CAD file from the mold 10 to match the location of the vent from the CAD file with the vent identified in real time to accurately locate and determine its coordinates. The observation system can receive the CAD file using a data transmission method known to those skilled in the art. The observation system can further include at least one camera and at least one sensor (not shown) for determining the location (i.e., coordinates) of the vent based on the data collected in real time and / or based on the contour profile generated by the scanner.

[0072] To implement the method of the present invention by computer, the system 100 includes a communication network (or "network") that manages input data entering the system from various sources (e.g., from at least one robot 102 and associated detection systems). The communication network includes one or more communication servers (or "servers"), each of which includes one or more processors operably connected to a memory. The memory is configured to store an application for analyzing data representing an imaged mold (and sections of a mold). The one or more processors include a module for executing an analysis application for processing an image, wherein the one or more processors are capable of executing programming instructions stored in the memory to perform the steps of the method (described below).

[0073] The term "processor" (or alternatively, the term "programmable logic circuit") refers to one or more devices capable of processing and analyzing data and including one or more software packages for processing data (e.g., one or more integrated circuits included in a computer, 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 known to those skilled in the art). The processor includes one or more software packages for processing data captured by the detection system of system 100 (and corresponding acquisition data), and one or more software packages for identifying and locating discrepancies and identifying the source of the discrepancies to correct the discrepancies.

[0074] In system 100, memory may include both volatile and non-volatile memory devices. Non-volatile memory may include solid-state memory such as NAND flash memory, "keep alive" memory (or "KAM") for saving various operating variables when the processor is turned off, magnetic and optical storage media, or any other suitable data storage device that retains data when system 100 is deactivated or loses its power supply. Volatile memory may include static and dynamic RAM that stores program instructions and data (including learning applications).

[0075] Further references Figures 1 to 4 as well as Figures 5 to 9 , provides a detailed description of an embodiment of the method (or "method") of the present invention implemented by system 100 by way of example. It should be clearly understood that system 100 can implement the method of the present invention in any physical environment without prior knowledge of the configuration of the mold.

[0076] As used herein, the term "method" or "process" may include one or more steps performed by at least one computer system including one or more processors for executing instructions capable of performing the steps. Unless otherwise specified, any sequence of steps is provided by way of example and does not limit the described method to any particular sequence.

[0077] In the following description, embodiments of the method of the present invention are described, in which the accuracy of the information acquired by the detection system (eg a camera) varies.

[0078] In performing the method of the present invention, the system 100 incorporates a combination of vision and machine learning techniques to accurately and quickly reconstruct an observed scene based on a three-dimensional (or "3D") distributed point cloud derived from a view of the section 10A of the mold 10. As a result, the system 100 achieves continuous improvement in identifying vents and their relative distribution along the inner surface 10a of the mold 10.

[0079] At the beginning of an embodiment of the method of the present invention, the method includes the step of positioning the mold 10 in the field of view of the detection system of the robot 102 (e.g., positioning the mold on the support 50, such as Figure 3 The mold 10 is positioned so that the vent holes 150 defined along the inner surface 10a of at least one section 10A are visible in the detection field of the sensor (see Figure 4 ). In this step, the robot 102 (in particular the integrated detection system) passes over the mold 10.

[0080] The method further includes a step of annotating the locations of samples of vents dedicated to learning. In this step, a reference of the coordinates of the vents searched in the image captured by the detection system (e.g., an RGB type camera) of the robot 102 is created. The reference of the coordinates of the vents created in this step includes an expected image corresponding to the vents 150 distributed along the inner surface 10a of the mold 10. This step can be performed before the other steps of the method of the present invention to provide the neural network with the actual coordinates of the vents and their relative positions relative to each other. In this embodiment of the method, at least a portion of the reference of the vents can be created by one or more persons skilled in the art.

[0081] In an embodiment of the method, in this step, a neural network may be trained to identify the actual coordinates of the vents and create a bounding box (or "framed region") around the identified vents. During this training, the coordinates of the bounding box of the identified vents are associated with the coordinates of the searched vents to calculate any displacement between them. The framed region and displacement calculation results are transmitted to a neural network (e.g., one or more CNNs) to jointly learn representations of the vents for various viewing angles of the images captured by the detection system of the robot 102.

[0082] In this embodiment of the method of the present invention, the method further comprises a step of capturing an image of the mold 10 (more specifically, capturing an image of the inner surface 10a of the section 10A of the mold). This step, performed by the robot 102 (particularly by an associated detection system), comprises a step of scanning the detection system of the robot 102 over the section 10A of the mold 10. Each image captured in this step consists of a matrix of pixels, wherein each pixel has a different color and brightness indicating the location of the vent 150 of the mold 10. The acquired images showing one or more locations of the vent 150 drive at least one neural network, which is used to identify all expected locations of the vents in the imaged mold 10. Therefore, these image changes are used as input to the neural network, which outputs a classification of the coordinates of the vent.

[0083] In this step, the purpose of the algorithm of the execution module is to automatically identify and indicate the outer contour of the vent hole and the interface of the periphery of the vent hole (for example, its diameter and the angle of its cylindrical axis relative to the curvature of the inner surface 10a of the mold 10). Therefore, in this step, the execution module uses the annotation software to construct a bounding box around the vent hole 150 appearing on the image of the mold 10 (see Figure 7 ).

[0084] The processor of the system 100 continuously forms a neural network based on new input data of mold images acquired by the inspection system of the robot 102. In order to automatically detect the boundary between the vent hole and the metal material of the surrounding mold, the robot 102 takes images (which may include video) and extracts from the same type (e.g., Figure 4 A set of image data is collected from a plurality of molds of the type represented by the mold 10 in FIG. All image data may be annotated based on operator input data prior to recording to create ground truth data. For example, in some embodiments, all image data are annotated to help the neural network detect and identify the boundaries of the vent hole 150 and / or the metal material of the surrounding mold. The known variations are manually identified based on the knowledge of the mold professional.

[0085] In this step, the processor of the server can use the ground truth data to train and / or develop one or more neural networks to automatically detect the space of the object (e.g., the vent 150 of the mold 10). Therefore, the ground truth data as described herein generally refers to information provided by direct observation of professionals on site, rather than information provided by inference. They can obtain data from several sources (including several professionals located in remote locations) to develop the neural network. The feedback loop of the annotated image can be updated with additional ground truth data over time to improve the accuracy of the system 100.

[0086] In this embodiment of the method of the invention, the method further comprises a step of reconstructing the segment 10A using three-dimensional digitalization with high resolution. In this step, the offset of the 3D camera of the detection system is used in the working space of the mold 10 to instantly reconstruct the volume of the segment 10A and determine its dimensions and define its working area.

[0087] This step consists in constructing an annotation database storing the RGB image, the coordinates (X, Y, Z) of the pixels of the acquired image and the coordinates of the bounding box. In this step, the algorithm of the execution module is aimed at extracting the points defining each exhaust hole 150 (see Figure 8 Each reconstructed vent can be found by using the extracted points and their coordinates (X, Y, Z) and analyzing the surface of the vent defined by these points (again see Figure 8 )'s center C 150 , and thus find the normal vector of the surface plane. The orientation of the reconstructed vent hole 150 relative to the inner surface 10a of the mold 10 indicates the corresponding orientation of the valve so that the valve can be correctly inserted into the mold. Therefore, the position of the mold 10 and its limits, as well as its geometry (including the tread pattern elements), the inner surface 10a, the normal at any point of the mold, the vent hole 150 for inserting the valve can be deduced.

[0088] In this embodiment of the method of the present invention, the method further includes a step of analyzing the profile of each vent hole 150. In this embodiment of the method of the present invention, the 3D camera of the detection system of the robot 102 provides better accuracy in calculating the normal on the homogenized plane on the inner surface 10a of the mold 10 by measuring the width of the spectrum. In this step, each extracted point (including the center C 150 ) have corresponding coordinates.

[0089] The step includes the step of determining the surface plane by finding the shape closest to the circle representing the searched vent hole. Since the vent hole is located along the inner surface 10a of the mold 10 (having a substantially curved portion), it should be understood that the extracted points may form one or more ellipses (see Fig. 9 ellipses provided by example in the example).

[0090] The steps also include the step of determining a vector perpendicular to the surface plane determined in the previous step. Figure 8 The normal vector determined passes through the center C of the exhaust hole 150 150 , to find the insertion axis X of the valve 200 200 (e.g., oriented at an angle α relative to the interior surface 10a of the mold 10). This step will allow the robot 102 to select and orient a valve having an appropriate diameter (e.g., Figure 2to ensure that it is correctly inserted into the corresponding exhaust hole.

[0091] In this embodiment of the method of the present invention, the method further comprises the step of determining the diameter of the exhaust hole that allows a valve having a corresponding diameter to pass through. Fig. 9 Each vent hole 150 is identified by the contour 150A analyzed in the previous step. The corresponding center C 150 Identification by point. In this step, the goal of the algorithm of the execution module is to identify a diameter that corresponds as closely as possible to the diameter of a known vent (for example known in the reference of the vent created in the annotation step of this embodiment of the method).

[0092] In this embodiment of the method of the present invention, the method further includes a final step of positioning the valve 200 in the identified vent having a corresponding diameter. In this step, the robot 102 receives the coordinates of the identified vent so that the robot can select a valve with an appropriate diameter (e.g., Figure 2 ). In this step, the robot 102 can select the valve by means of a tool changer and a valve feed system (both of which are commercially known). In this step, the robot 102 can be aligned with the identified exhaust hole 150 on its insertion axis and can blow air into the valve. Once the valve has been pre-positioned by virtue of its shape (conical or stepped), the robot 102 can press the valve by means of a valve supply head or by pushing the valve using a dedicated area of ​​an effector (not shown) placed at the end 104a.

[0093] The use of neural networks provides reliability and, most importantly, speed when determining the exhaust vents, which avoids long computations. Although the embodiments described herein relate to the use of neural networks (and more specifically, convolutional neural networks or "CNNs") as machine learning models, other types of machine learning models may be used. These machine learning models include, but are not limited to, models utilizing linear regression, logistic regression, decision trees, support vector machines, naive Bayes, K-nearest neighbors (kNN) (where K represents grouping), random forests, dimensionality reduction algorithms, gradient algorithms, neural networks (e.g., autoencoders, RNNs, perceptrons, logarithmic short-term memory (LSTM), Hopfield, Boltzmann, deep belief networks, deconvolution, generative adversarial networks (GANs), etc.) and their supplements and equivalents. One or more CNNs may be formed using ground truth data generated using sensor data representing the motion of the robot 102 (including the positioning of the gripper 108).

[0094] One or more steps of this embodiment of the method of the present invention use a neural network of the deformable transformer type (or "deformable DETR" or "DETR"). By combining a neural network of the CNN type and an encoder-decoder of the "Transformer" type, DETR is used for end-to-end detection of objects. DETR initially detects objects by only being interested in a small set of key sampling points around a reference (e.g., points forming the contour 150A around the exhaust hole 150 of the mold 10) (see Fig. 9 ) to reduce computation. DETR then uses a deformable attention module to aggregate multi-scale features to facilitate the detection of small objects. Therefore, DETR can model the dependencies between long-range objects in the observed scene in order to achieve the ability to detect, localize, and automatically and accurately classify exhaust holes for inserting corresponding valves.

[0095] In all embodiments of the method of the present invention, the method may further include an optional inspection step after the valve 200 is inserted into the vent hole 150 of the mold 10. In this step, an operator may perform a separate manual inspection of all items presented by the robot 102. In this step, a fully automated inspection may be performed, including presence detection and / or probes, to confirm the presence and correct operation of the valve.

[0096] It should be understood that all embodiments of the methods of the present invention may be performed in the same plant (eg, by a single facility comprising system 100).

[0097] By using the system 100 of the present invention to perform the disclosed method, any mold presented to the system 100 is analyzed in the same manner. No CAD file needs to be known in advance and no layout needs to be made in the mold so that it can be perfectly positioned. The system 100 is inherently designed to accommodate changes, which provides the possibility of working with third-party molds and / or hand-finished molds, for example.

[0098] The system 100 of the present invention may include pre-programmed information related to anticipated events.For example, adjustments to the method of the present invention may be associated with parameters of a typical physical environment (eg, a tire production facility) in which the system 100 operates.

[0099] In some embodiments of the invention, system 100 (or another system including system 100) can receive audio commands (including voice commands) or other representative audio data (e.g., commands to start or stop one or more steps of the method of the invention). The request can include a request for the current state of the ongoing method (e.g., the number of valves inserted compared to the number of vents 150 in mold 10 that are configured to receive corresponding vents). The generated response can be represented auditorily, visually, tactilely (e.g., by using a tactile interface), and / or in a virtual and / or augmented manner. The response and corresponding data can be recorded in the neural network.

[0100] It should be understood that the system 100 may include multiple computing devices that perform various aspects of learning. In these embodiments, the processor may configure the system 100 based on one or more parameters of the exhaust vent and its known location. In these embodiments, it should be understood that one or more reinforcement learning devices may be used.

[0101] For all embodiments of system 100, a monitoring system may be implemented. At least a portion of the monitoring or "warning" system may be provided in a portable device, such as a mobile network device (such as a mobile phone, a laptop computer, one or more portable devices connected to a network (including "augmented reality" and / or "virtual reality" devices, wearable clothing connected to a network, and / or any combination thereof and / or any equivalent thereof)). It is contemplated that the detection and comparison steps can be performed iteratively.

[0102] The terms "at least one" and "one or more" can be used interchangeably. Ranges presented as "between a and b" include the values ​​"a" and "b."

[0103] Although specific embodiments of the disclosed device have been shown and described, it will be appreciated that various changes, additions and modifications may be made without departing from the spirit or scope of the invention. Accordingly, no limitations should be placed on the scope of the described invention, except as disclosed in the accompanying claims.

Claims

1. A system (100) for implementing a method for identifying vent holes (150) in a tire vulcanization mold (10), the tire vulcanization mold (10) comprising one or more sections (10A) and an inner surface (10a), the vent holes being distributed on the inner surface (10a) so that a corresponding valve (200) can be inserted Among them, it is characterized in that The system comprises: - a robot (102) comprising a detection system having one or more sensors that detect the presence of one or more vent holes (150) distributed along the inner surface (10a) of the section (10A) of the mold (10); - a communication network that manages the input data from the detection systems into the system (100); and - one or more communication servers, each communication server comprising one or more processors operably connected to a memory, the memory configured to store an application for analyzing data representing an imaged mold, the one or more processors comprising a module for executing the analysis application for processing the image, the one or more processors being capable of executing programmed instructions stored in the memory to perform the following steps: - detecting the presence of the arrangement of vent holes (150) in the field of view of a detection system, said detection system being triggered to capture at least one image of the inner surface (10a) of the section (10A) of the mould (10); and - A step of searching for the presence of detected vents (150) in the images captured by the detection system, such that if no vents are detected, the detection system continues to capture images until the mold (10) has been searched.

2. The system (100) according to claim 1, wherein: The detection system comprises at least one RGB-D type three-dimensional (3D) camera fixed to the robot (102), the three-dimensional (3D) camera providing a 3D image represented by a set of 3D points with coordinates (X, Y, Z).

3. The system (100) according to claim 2, wherein: The one or more processors are capable of executing programmed instructions stored in the memory to perform the following steps: - a step of annotating the location of the sample of vents (150), said step comprising the step of creating a coordinate reference of the vents searched for in the image captured by the detection system of the system (100); - a step of reconstructing the segment (10A), comprising the step of building an annotation database storing the captured images and the coordinates (X, Y, Z) of the pixels of the captured images; - a step of analyzing the profile (150A) of the vent (150) performed by a module for executing an analysis application of the system (100), said step comprising the step of determining a surface plane by finding the shape of a circle that is closest to representing the searched vent (150), said step further comprising determining a vector perpendicular to the determined surface plane to find the insertion axis (X) of the valve (200); 200 ) steps; and - a step of determining the diameter of the exhaust hole that allows the valve (200) having the corresponding diameter to pass therethrough, in which step each exhaust hole (150) is identified by the profile (150A) analyzed in the step of analyzing the profile of the exhaust hole, and in which step the corresponding center (C 150 )Identified by points; The system (100) receives the coordinates of the identified exhaust hole to select a valve (200) having an appropriate diameter.

4. The system (100) according to claim 3, wherein: A module for executing an analysis application stored in the system's memory uses annotation software to construct a bounding box around the vents (150) that appear on the captured image of the mold (10).

5. The system (100) according to claim 4, wherein: The system's processor continuously forms at least one neural network that outputs a classification of the coordinates of the exhaust vent (150) such that the captured image displays the location of the exhaust vent (150).

6. The system (100) according to claim 5, wherein: The at least one neural network is selected from convolutional neural networks (CNNs).

7. The system (100) according to claim 6, wherein: One or more steps use a Deformable Transformer (DETR) type neural network.

8. The system (100) according to any one of claims 1 to 7, wherein: The robot (102) comprises a peripheral gripping member (104) supported by a pivotable elongated arm (106), the peripheral gripping member (104) extending from the elongated arm (106) to a free end (104a) at which a gripper (108) is arranged along a common longitudinal axis.

9. The system (100) according to claim 8, wherein: The clamp (108) includes a pivotable clamp (108a), the pivotable clamp (108a) including clamping fingers (108b) extending from a platform (108c), at which the clamp (108a) is fixed to the free end (104a) of the peripheral clamping member (10), each finger (108b) including a member of a predetermined length extending between an actuating end (108b') where movement of the finger occurs and an opposite clamping end (108b) where the finger clamps the valve (200).

10. The system (100) according to any one of claims 1 to 9, wherein: The one or more processors are capable of executing programmed instructions stored in the memory to perform steps of moving the robot (102) so that the robot (102) can place the valve (200) so that the valve (200) can be inserted into the identified vent hole in the section (10A) of the mold (10).

11. A method for identifying vent holes (150) in a tire vulcanization mold (10) implemented by a system (100) according to any one of claims 1 to 10, wherein the tire vulcanization mold (10) comprises one or more sections (10A) and an inner surface (10a), wherein the vent holes are distributed on the inner surface (10a) so that the corresponding valve (200) can be inserted Among them, it is characterized in that The method comprises the following steps: - a step of positioning the mould (10) in the field of view of a detection system of the system (100) so that the vents (150) defined along the inner surface (10a) of at least one section (10A) are visible, during which the detection system is passed over the mould (10); - a step of detecting the presence of the arrangement of vent holes (150) in the field of view of a detection system, said detection system being triggered to capture at least one image of the inner surface (10a) of the section (10A) of the mould (10); as well as - A step of searching for the presence of detected vents (150) in the images captured by the detection system, such that if no vents are detected, the detection system continues to capture images until the mold (10) has been searched.

12. The method according to claim 11, further comprising an inspection step performed after the valve (200) is inserted into the exhaust hole (150) of the mold (10).

13. The method of claim 11 or claim 12, further comprising positioning the robot (102) at the insertion axis (X 200 ) is aligned with the identified exhaust hole (150), in which step the robot blows air into the valve (200).

Images