System and method for automatic cleaning of tire tread bottoms

Abstract

The invention relates to a method for cleaning the tread pattern of an identified damaged tire during retreading. The method is implemented by at least one processor comprising a processing module that applies representative data from captured images of the identified tire tread pattern to at least one neural network. The invention also relates to a retreading method comprising the disclosed method. The invention further relates to an automatic cleaning system for cleaning the tread pattern of an identified tire during retreading. Figure for the abstract: Fig. 11

Description

Title of the invention: System and method for automatic cleaning of tire tread bottoms. Technical field

[0001] The invention relates to a system and method for cleaning the tread pattern of a tire during retreading. More particularly, the invention relates to the use of a cleaning device carrying a treatment tool capable of cleaning the tread patterns of tires identified for retreading. Context

[0002] In the field of tire manufacturing and maintenance, retreading is a process known to restore a worn tire to working order by renewing the tread compound and the ply(ies). During the retreading process, the old tread material is removed and replaced with new material. Retreading is a method that is known and regulated in several industries (including, but not limited to, the transportation, aviation, mining, and agricultural industries). The damage could fall into one or more categories (including, but not limited to, tread damage, tear damage, cut damage, and puncture damage).

[0003] Regardless of the type of tires being retreaded, the steps in a retreading process are essentially the same from the moment a tire enters a retreading facility until it exits. During the retreading process, a repair process is carried out in which the tires intended for retreading are examined so that any existing damage can be identified and located. During the repair process, a visual inspection of the damage is performed (either manually or automatically (for example, one or more cameras) or a combination of both). If the focus is on repairing "tread" damage, during the repair process, a cleaning step is performed on the damage using compressed air and a soft brush. Then, a step is performed to remove the remaining rubber using a stiff brush to expose the first top layer of rubber.The repair process continues with a visual inspection for corrosion on the wire cables (either manually, automatically (e.g., using one or more cameras), or a combination of both). If this visual inspection confirms the presence of corrosion, a removal step is carried out using a grinding wheel. These corrosion detection / removal steps are repeated on each successive cable layer. as long as corrosion is detected. After the repair process is completed, the retreading process can proceed in a typical sequence of steps towards the tire leaving the retreading facility (the steps including, but not limited to, a tread application step, a curing step, an inspection step, and a shipping step). Until now, one or more steps in the damage repair process have been carried out manually.

[0004] There are ongoing efforts to automate at least part of the repair process. For example, European patent EP2456612B1 discloses an automatic spraying system for dissolving adhesive during tire retreading. The adhesive spraying occurs after the repair process and just before the tread is applied. The adhesive is sprayed into holes and / or craters automatically detected by acquiring a three-dimensional (or "3D") profile obtained using a profilometer. The sprayed adhesive is then applied to the holes and / or craters by an automatic applicator.

[0005] European patent EP3551443B1 discloses an automatic damage repair system for tire retreading. The disclosed system includes a remote, fixed laser profilometer that scans the entire tread by rotating the carcass. The system also includes a color camera mounted on a robot that takes pictures of damage once the covering rubber has been removed. These color images are used to perform corrosion detection by analyzing the colors in the image (for example, by looking for the color red on cords). The processing algorithm uses the three-dimensional (3D) information from the profilometer to detect damage. The processing used to detect damage includes constructing an average profile and comparing it to this profile. This processing would be sufficient if the carcass could be considered a rigid body.However, a frame is more accurately described as a deformable body that can exhibit numerous deformations, some more localized than others (for example, a more or less pronounced runout). These deformations can be considered as damage deviating from an average profile. Therefore, the disclosed system cannot adapt to all possible damage.

[0006] European patent EP2414153B1 discloses a machine that performs an automatic process for identifying and repairing defects on used tires. The disclosed machine includes a vision system that uses an infrared projector to obtain 3D information on the damage being analyzed. During the process, an automatic scan of the tire's working surface identifies the morphology of the damage and compares the identified morphologies with a reference of known damage. A selection of the work and a corresponding tool is carried out to perform the selected work on the identified damage.

[0007] There is nothing in the prior art that proposes establishing a link between the identification and localization of damage and the cleaning of damage using a processing tool in a completely automated manner. Thus, the disclosed invention relates to the automatic repair of "sculpting" type damage detected by machine learning. An algorithmic process relies on thousands of training images so that the algorithm learns to recognize the points in the image (three-dimensional and color) belonging to a tread pattern on a tire identified for retreading. This approach is scalable and will become increasingly robust because it is possible to integrate images of undetected tread patterns into the training in order to extend the scope of the algorithm's use. Furthermore, the cleaning of the tread patterns does not depend on any predefined procedure because the processing path is defined on the fly according to the shape of the tread pattern. Summary of the invention

[0008] The invention relates to a method for cleaning a tread pattern of an identified damaged tire during retreading, the method implemented by at least one processor comprising a processing module which applies, to at least one neural network, the representative data of the captured images of the tread pattern of the identified tire, characterized in that the method comprises the following steps: - a step of providing an automatic cleaning system of which the processor is a part to automatically recognize an order for processing the damage in the tread pattern of the identified tire; - a calibration step of a cleaning device which is part of the automatic cleaning system, the cleaning device comprising a processing device supported by an elongated pivoting arm which extends to a free end where a processing tool is arranged along a common longitudinal axis; - a step of positioning the identified tire in a processing area where the cleaning device processes it; - a positioning step of at least one camera which is part of the automatic cleaning system at a reference distance from an identified tire tread surface; - a tread pattern detection step including sector-by-sector processing for the identified tire incorporating one or more scans followed by tread pattern processing of the identified tire to recover a color and three-dimensional (3D) image of each scanned sector; - a step of creating the three-dimensional (3D) trajectories that the cleaning device should follow to process the sculptures; - a step of simplifying vertical and horizontal paths during which a 3D point associated with each point of each path is retrieved; - a step of adding approach and withdrawal points to complete the 3D trajectory obtained after the path simplification step, during which, at the beginning and end of each vertical and horizontal path, an approach point and a withdrawal point are added, allowing the cleaning device to start and finish processing the path that corresponds to a tread pattern on the identified tire surface; and - a final step of calculating the orientations of the processing tool during which a 3D orientation of the processing tool is calculated at each point of each vertical and horizontal path; so that the cleaning device is set in motion so that the processing device can place the processing tool to perform the cleaning of the identified tire tread.

[0009] In certain embodiments of the cleaning process of the invention, during the tread detection step, the sector-by-sector treatment for the identified tire comprises: - for each identified tire sector, a scan step of the sector followed by a processing of the scanned sector; - a scan of all successive sectors of the identified tire followed by processing of all scanned sectors; or - a scan of the tire being processed on a different station than the one where the scanned sectors are being processed; so that the cleaning device processes one or more scanned sectors.

[0010] In certain embodiments of the cleaning process of the invention, the calibration step of the cleaning device includes a calibration phase to estimate the pose of the camera relative to a base of the cleaning device.

[0011] In certain embodiments of the cleaning process of the invention, the calibration step of the cleaning device includes a step of positioning a calibration target in the processing space, during which the camera implements a process to estimate the camera's position in the reference frame of the brushing device.

[0012] In certain embodiments of the cleaning process of the invention, during the positioning step of the identified tire, the identified tire is positioned near the cleaning device on a rotating drum arranged in the processing area, so that, during a repair process of a retreading process, the drum rotates the identified tire so that it can be treated sector by sector.

[0013] In certain embodiments of the cleaning process of the invention, the camera positioning step includes a scan verification step performed by the camera comprising the following steps: - a step of identifying a central horizontal area of ​​the scan which is used to verify that the distance between the identified tire and the camera is correct; - a step of calculating an average distance (Dmoy) of the points inside the horizontal central zone; - a step of preserving the 3D points in the form of a surface mask comprising a binary image representing the points belonging to the rolling surface of the identified tire; - a calculation step again the average distance (Dmoy ) by considering only the points of the horizontal central zone which belong to the rolling surface of the identified tire to obtain a following distance (Dmoy); - a step of comparing the distance (Dmoy surface) with the reference distance of the camera; - a calculation step of the two vertical lines that delimit the beginning and end of the surface mask to allow verification that the camera is centered with respect to the rolling surface of the identified tire; - a calculation step for a central vertical line located in the middle of the vertical lines; and - where the calculated central vertical line deviates from the image's central line beyond a predetermined threshold, a step of applying a lateral correction to the camera position to recenter it relative to the identified tire tread.

[0014] In certain embodiments of the cleaning process of the invention, the sculpture detection step includes a step of applying a neural network to the color and 3D image to determine which points represent sculptures of the identified tire to be treated with the processing tool, during which, at each image, the surface mask is retrieved from the output of the neural network where each white point indicates that the point represents a sculpture to be cleaned.

[0015] In certain embodiments of the cleaning process of the invention, the step of creating the three-dimensional (3D) trajectories comprises the following steps: - a data preparation step, during which points belonging to the contours of all white objects present in the surface mask are extracted to obtain a set of polygons with holes; and - a step of extracting horizontal and vertical two-dimensional (2D) paths, during which holes present inside the polygons extracted during the data preparation step are removed, this step comprising the following steps: - a calculation step of a morphological skeleton of each polygon including the extraction of all the central points of each polygon, in which each morphological skeleton consists of nodes representing the ends of the morphological skeleton and / or the crossing points, and paths including points located between two nodes; - a step of classifying each path into either a horizontal path category or a vertical path category, during which paths of the same category are merged with nearby endpoints and, once the merging is complete, the direction of travel for each path is fixed; and - a 3D trajectory development step which is carried out once the vertical and horizontal paths have been extracted, during which the vertical and horizontal paths are ordered; so that the ordered set of vertical and horizontal paths gives the order of processing of all the sculptures of the scanned sector of the identified tire.

[0016] In certain embodiments of the cleaning process of the invention, during the step of adding the approach and withdrawal points: - a first approach point is constructed by offsetting it by a predetermined distance (L) from the first or last point of the path, and, once offset, the first approach point is moved closer to the camera by the distance (L); and - a second approach point is constructed from the first approach point by bringing this second approach point closer to the camera by a second distance (L2) from the first approach point.

[0017] In certain embodiments of the cleaning process of the invention, the step of calculating the orientations of the processing tool includes a step of calculating the 3D normal at each point as well as the direction vector between each pair of successive points, in which the direction vector gives the direction to follow between two points of each path, where the normal gives the orientation of the Z axis of the processing tool, the direction vector gives the orientation of the X axis of the processing tool, and the Y axis is fixed automatically according to the fixing of the X and Z axes.

[0018] In certain embodiments of the cleaning process of the invention, the processor refers to a table of various tire sizes to perform a determination of one or more parameters of the identified tire.

[0019] In some embodiments of the cleaning process of the invention, the treatment tool includes at least one brush.

[0020] The invention also relates to a retreading process comprising the disclosed cleaning process.

[0021] The invention further relates to an automatic cleaning system for cleaning the tread pattern of an identified tire during retreading, the automatic cleaning system comprising: - at least one cleaning device comprising a robot having a cleaning peripheral supported by an elongated pivoting arm and extending from the elongated arm to a free end where a processing tool is arranged along a common longitudinal axis; - a detection system to gather information about the physical environment around the cleaning device; and - a communication network which manages the data entering the automatic cleaning system from each cleaning device, the communication network comprising at least one communication server enabling the execution of programmed instructions stored in a memory of one or more processors of the automatic cleaning system to implement the cleaning process of the invention; so that the cleaning device is configured on one or more parameters of the identified tire calculated by an image processing module incorporated in the processor's memory to set the cleaning device in motion so that the brushing device can place the processing tool to perform the cleaning of the tread of the identified tire.

[0022] In certain embodiments of the system of the invention, the detection system includes at least one RGB-D type camera fixed to at least one of the extended arm and the processing peripheral of the cleaning device.

[0023] Other aspects of the invention will become evident from the following detailed description. Brief description of the drawings

[0024] The nature and various advantages of the invention will become more evident upon reading the following detailed description, together with the accompanying drawings, in which the same reference numbers designate identical parts throughout, and in which: [Fig.1] Fig.1 represents a schematic cross-sectional view of one embodiment of a known tire. [Fig.2] Fig.2 represents a schematic view of an embodiment of a cleaning device which is part of an automatic cleaning system of the invention. [Fig.3] Fig.3 represents an example of a visualization of a part of a tire obtained during a cleaning process of the invention. [Fig.4] Fig.4 represents a schematic view of the positioning of a camera of the automatic cleaning system of the invention relative to a tread surface of an identified tire during retreading. [Fig.5] Fig.5 represents a schematic view of the input of color and 3D images to a neural network in the creation of surface masks carried out during the cleaning process of the invention. [Fig.6] Fig.6 represents a flow diagram of an algorithm that performs a process of generating 3D trajectories for processing sculptures. [Fig.7] Fig.7 represents an example of an extraction of contours associated with the mask of sculptures created during the cleaning process of the invention. [Fig.8] Fig.8 represents an example of a skeleton created during a 3D trajectory generation process for processing sculptures. [Fig.9] [Fig.9] represents an example of a bounding box for classifying horizontal and vertical paths during a 3D trajectory generation process. [Fig. 10] The [Fig. 10] represents an example of merging paths of the same category during a 3D trajectory generation process. [Fig.ll] [Fig.l 1] represents an example of an order of vertical and horizontal paths retained at the end of a 3D trajectory generation process. [Fig. 12] The [Fig. 12] represents an example of path simplification achieved during the cleaning process of the invention. [Fig. 13] [Fig. 14] [Fig. 13] represents an example of the construction of approach and withdrawal points carried out during the cleaning process of the invention, and [Fig. 14] represents an example of the constructed approach and withdrawal points. [Fig.15] [Fig.15] represents an example of a 3D orientation of a processing tool of the automatic cleaning system of [Fig.2] calculated during the cleaning process of the invention. Detailed description

[0025] When considering the characteristics of a tire for retreading, its geometry must be taken into account. A tire is an object with a known geometry, generally comprising several superimposed layers of rubber (or "layers"), as well as a metallic or textile fiber structure constituting a carcass that reinforces the tire's structure. The type of rubber and the type of reinforcement are chosen according to the desired final characteristics. Figure 1 includes a schematic representation of a tire 10, which, conventionally, includes two circumferential beads designed to allow the tire to be attached to a rim. Each bead includes an annular reinforcing bead. The construction of a tire is typically described by a representation of its components in a meridian plane, that is, a plane containing the tire's axis of rotation.The radial, axial, and circumferential directions respectively refer to the directions perpendicular to the tire's axis of rotation, parallel to the tire's axis of rotation, and perpendicular to any meridian plane. The terms "radially," "axially," and "circumferentially" mean, respectively, "along a radial direction," "along the axial direction," and "along a circumferential direction" of the tire. The terms "radially inside" and "radially outside" mean "closer to, or farther from, the tire's axis of rotation, respectively, along a radial direction."

[0026] The tire 10 also includes a tread 12, added to the outer surface of the tire. The tread 12 is intended to come into contact with a ground via a running surface 12a. The tire 10 further includes a crown reinforcement comprising a working reinforcement 14 and a reinforcing reinforcement 16, the working reinforcement 14 having working layers represented by layers 14a and 14b. The tire 10 also includes two sidewalls (one sidewall 18 being shown in [Fig. 1]) and two ribs 20 reinforced with a bead 22. A radial carcass layer 24 extends from one rib to the other, surrounding the bead in a known manner. The tread 12 has reinforcements consisting, for example, of superimposed layers having known reinforcing wires. In some embodiments, the tire may include a rubber compound 26 that dissipates the static electricity produced during rolling.

[0027] The tread 12 is bounded, in the radial direction, by two circumferential surfaces, the outermost of which is the tread surface 12a and the innermost of which is called the tread base surface. The tread base surface (or "bottom surface") is defined as the surface of the tread surface translated radially inward by a radial distance equal to the tread depth. It is common for this depth to be progressively smaller. on the outermost axially circumferential portions (called "shoulders") of the tread 12.

[0028] Furthermore, the tread of a tire is delimited, along the axial direction, by two lateral surfaces. The tread is further constituted by one or more rubber compounds. The term "rubber compound" designates a rubber composition comprising at least one elastomer and a filler.

[0029] To achieve good grip on wet surfaces, cutouts are arranged in the tread 12. A cutout is defined as either a well, a groove, an incision, or a circumferential groove, and forms a space opening onto the tread surface 12a. The performance of a tread pattern must remain sufficiently consistent despite wear to ensure the tire's longevity. Consequently, it is necessary to maintain a certain thickness of rubber material between the bottom face of the cutouts (grooves or grooves) and the reinforcing elements to guarantee the tire's durability. For example, grooves must be wide enough to allow the evacuation of liquid present on the ground surface regardless of the stage of wear of the tread (12).

[0030] With reference to the figures, in which the same numbers identify identical elements, [Fig. 2] represents an embodiment of a cleaning device 100 that forms part of an automatic cleaning system (or "system") of the invention. The automatic cleaning system of the invention implements a method of the invention ("method") enabling the cleaning, by means of a processing tool, of a tread pattern on an identified tire during retreading. The disclosed method incorporates a machine learning method based on data corresponding to images obtained from the identified tire. The algorithm used analyzes the tread surface of the identified tire to position and insert the processing tool into a cutout in the identified tire. It is understood that the method of the invention can be incorporated into an existing retreading process to improve its repair process.

[0031] In one embodiment of the cleaning device 100 of the system of the invention, the cleaning device comprises a robot having a processing unit 102 supported by a pivoting extended arm 104. The processing unit 102 extends from the extended arm 104 to a free end 102a where a processing tool is arranged along a common longitudinal axis. In the embodiment of the cleaning device 100 shown in [Fig. 2], the processing tool comprises a brush 106 which is selected from commercially available industrial brushes for cleaning tires during the processes of retreading (for example, industrial rotary brushes made of synthetic or metallic filaments). The following description refers to brush 106, but it is understood that other equivalent processing tools could be used within the scope of the invention (including, without limitation, a knife, rasp, carding tool, milling cutter or grinding wheel).

[0032] The attachment of the brush 106 (or an equivalent processing tool) to the processing device 102 can be achieved by screwing an adapter onto the free end 102a of the processing device. The attachment of the brush 106 is such that it can rotate about its longitudinal axis and also spin around a predefined axis of rotation during a repair process. It is understood that the attachment of the brush 106 to the processing device 102 can be achieved by one or more known fastening methods (including, without limitation, welding, bonding, and equivalent methods). Thus, the robot facilitates the cleaning of a variety of cuts without interrupting the rotation and / or spinning of the brush 106.

[0033] The cleaning device 100 is set in motion so that the processing device 102 can clean the tread pattern of a tire identified P (see [Fig. 2]) by the cleaning device during a process carried out by the system of the invention (as described below). It is understood that the tread pattern of the identified tire incorporates the cuts found on its tread surface. The term "identified tire" (in the singular or plural) is used here to refer to a tire being retreaded and present in the physical environment of the system incorporating the cleaning device 100 (the physical environment being, for example, a retreading plant incorporating the cleaning device 100).

[0034] It is understood that the configuration of the cleaning device 100 is given by way of example. For example, the cleaning device 100 may include a fixed robot installed in a retreading installation, fixed, for example, to a support relative to which the robot extends (for example, a base 110 as shown in [Fig. 2]). In this case, it is understood that the robot may be fixed to a ceiling, a wall, a floor, or any support that allows the process of the invention to be carried out during a retreading process. It is understood that the cleaning device 100 may include at least one mobile robot. By "mobile," it is understood that the gripping device may be set in motion either by integrated means of motion (for example, one or more integrated motors) or by non-integrated means of motion (for example, one or more mobile means, including autonomous mobile means).It is understood that the cleaning device 100 can be a conventional industrial robot, a collaborative robot, or even a delta or cable robot.

[0035] The cutouts that can be cleaned by the cleaning device 100 during the process of the invention include grooves and furrows. The grooves are the cutouts that allow water to be evacuated. A groove has a width such that the opposing material walls delimiting it do not come into contact with each other when the tread passes through the contact patch, under recommended inflation and load conditions as defined in particular by the European standards of the European Tyre and Rim Technical Organization (ETRTO) or ETRTO in its Standards Manual 2020 - Commercial Vehicle Tyres. The compression and shear deformations of the raised elements delimiting the groove determine the pressures in contact with the ground and therefore the wear.

[0036] A groove is a substantially circumferential channel, and the lateral faces are substantially circumferential in the sense that their orientation can vary locally by plus or minus 45° around the circumferential direction. All the motifs belonging to the groove are found all around the tread, forming a substantially continuous whole (i.e., exhibiting discontinuities of less than 10% in length compared to the length of the motifs). The circumferential grooves delimit ribs, each rib of which is composed of the motifs of the tread pattern located between an axial edge of the tire and the nearest outermost circumferential groove, or between two adjacent circumferential grooves.

[0037] The depth of the cut is the maximum radial distance between the rolling surface of the tire and the bottom of the cut.

[0038] The cleaning device 100 of the system also includes a detection system (not shown) for gathering information about the physical environment around the cleaning device. The detection system includes one or more sensors (including one or more cameras) configured to perform two-dimensional (2D) and / or three-dimensional (3D) image detection, 3D depth detection, and / or other types of detection of the physical environment around the gripping device (it is understood that the terms "sensor" and "camera" are used interchangeably). In embodiments of the cleaning device 100 of the type shown in [Fig. 2], the sensor(s) of the detection system are attached to at least one of the extended arm 104 and the processing unit 102 of the robot.In embodiments of the system of the invention, these sensors could be part of an overall detection system which uses these sensors together with one or more sensors positioned in the physical environment in which the cleaning device 100 operates.

[0039] In one embodiment of the cleaning device 100, the detection system includes at least one camera 200 that provides 3D images represented as a set of 3D points with X, Y, Z coordinates and red, green, and blue color values ​​(the “RGB” or “RGB-D” format) (referred to as “an RGB-D type camera”). In this embodiment, an RGB-D type camera is attached to at least one of the robot's extended arm 104 and processing unit 102 (the latter being shown by way of example in [Fig. 2]). Two or more RGB-D cameras can be oriented to obtain a predetermined overlap between the camera fields of view. As used herein, the term “camera” includes one or more cameras.

[0040] RGB-D cameras typically provide depth information using depth maps, which are images where each pixel contains the distance between the camera and the corresponding point in space. Compared to traditional measurement methods such as manual measurement and other measurements based on electronic devices, 3D point cloud data from RGB-D cameras has a much higher measurement rate. By using a more sparse structure, a point cloud can be constructed from RGB-D images by calculating the real world (e.g., X, Y, Z coordinates) using the intrinsic data of a scanning camera.Thus, information about the physical environment around the cleaning device 100 is obtained from 3D point cloud data obtained from detection technologies that are capable of accurately and efficiently capturing the 3D surface geometries of tires. These detection technologies could be chosen from commercially available devices (for example, cameras sold under the PHOTONEO™ brand, cameras sold under the ENSENSO® brand from IDS, cameras sold under the ZIVID® brand from Zivid AS, computer vision systems sold by Cognex Corp., and their equivalents).

[0041] The term "point cloud" (in the singular or plural) is used here to refer to one or more collections of data points in space. One or more cameras (or equivalent devices) collect three-dimensional (3D) data and detect the surfaces of objects (for example, tires identified for processing during a retreading process) using a series of coordinates. Storing information as a collection of spatial coordinates can save space, since many objects do not occupy a large portion of the environment. Even though the information is not visual, interpreting the data as a point cloud helps to understand the relationship between several variables through classification and segmentation.

[0042] It is understood that one or more cameras may include one or more programming modes, including learning, for feeding, modifying, and training at least one neural network. Although the embodiments described here have regard to the use of one or more neural networks (for example, convolutional neural networks, or "CNNs") as a machine learning model, other types of machine learning models may be used.These include, but are not limited to, models using linear regression, logistic regression, decision trees, support vector machines, naive Bayes, nearest neighbor (knn), K stands for grouping, random forest, dimensionality reduction algorithms, gradient descent algorithms, neural networks (e.g., autoencoders, CNNs, RNNs, perceptrons, log short-term memory (LSTM), Hopfield, Boltzmann, deep belief, deconvolution, generative confrontation (GAN), etc.) and their complements and equivalents. The CNN(s) can be trained with ground truth data that is generated using sensor data representative of the movement of the cleaning device 100, including the positioning of the processing device 102 and the processing tool (e.g., the brush 106).

[0043] In certain embodiments of the cleaning device 100, the camera 200 is triggered when an identified segment of the tire enters the camera's field of view. In cases where a part of the tire is not visible in the image obtained by the detection system of the cleaning device 100 (for example, the camera 200), a mounting point can be placed at a known position relative to the sensor (for example, at a known horizontal distance and a known vertical distance from the camera position).

[0044] The detection system of the cleaning device 100 detects the presence of a tread pattern within the field of view of the detection system (for example, the field of view of the camera 200), which triggers it to capture an image of a tread surface of the tire identified for retreading. In all embodiments of the system of the invention, the system "searches" in the image obtained by the detection system for the presence of the damage "seen" by the cleaning device 100. If no damage is detected, the detection system continues to acquire images until the search for the tire is exhausted.

[0045] The detection system can determine information about the physical environment around the identified tire, which can be used by a control system of the invention (the control system comprising, for example, software for planning the movements of the cleaning device 100). The system The control system could be located on the robot or it could be in remote communication with the robot. In embodiments of the system of the invention, one or more 2D or 3D sensors mounted on the robot (including, without limitation, navigation sensors) can be integrated to create a digital model of the physical environment (including, where applicable, the sides, floor, and ceiling). Using the data obtained, the control system can cause the robot to move to navigate between the positions for taking images of the tires during repair processes.

[0046] To properly manage the handling of the cleaning device 100, which ensures clear image capture of the identified tire (for example, the handling of the robot and the positioning of the brush 106 as shown in [Fig. 2]), it is necessary to identify the tire being retreaded and detect the positioning of the relevant damage. Thus, the detection data refers to a plurality of recordings representative of the damage positions of at least one identified tire or a portion of an identified tire tracked over time.For example, detection data may include one or more positions from recordings of the positions of a reference point on a part of the tire (e.g., the tread) over time or at defined time intervals; sensor data taken over time; a video stream that has been processed using a computer vision technique; and / or data indicating the operating status of the Cleaning Device 100 over time. In some cases, detection data may include representative data of one or more continuous movements of the Cleaning Device 100 before it stops to take one or more images of the tires being retreaded. The detection system is therefore configured to generate motion data from the Cleaning Device 100.

[0047] The detection system of the cleaning device 100 can receive a CAD file of the identified tire to match the location of a cut from the CAD file with the identified cut in real time to precisely locate and determine its coordinates. The detection system can receive the CAD file using data transmission methods known to those skilled in the art. The detection system may further include at least one camera and at least one sensor (not shown) to determine the location (i.e., the coordinates) of the damage based on the data collected in real time and / or the contour profile of the identified tire.

[0048] In embodiments of the invention, the detection system of the cleaning device 100 may also include a motion capture device selected from infrared sensors, ultrasonic sensors, accelerometers, Gyroscopes, pressure sensors, and / or equivalent devices. In these embodiments, the system (and particularly the cleaning device 100) learns the movements that perform the image acquisition of the tires without operator intervention during the retreading processes.

[0049] To implement the method of the invention by means of a computer, the system of the invention comprises a communication network (or "network") that manages the data entering the system from various sources (for example, from at least one cleaning device 100 and its associated detection system). The communication network incorporates one or more communication servers (or "servers"), each comprising one or more processors operationally connected to a memory. The memory is configured to store an application for analyzing data representative of the cuts of the imaged tires. The processor(s) include an analysis application execution module that performs image processing, and the processor(s) are capable of executing programmed instructions stored in memory to carry out the steps of the method (as described below).

[0050] The input data to the automatic cleaning system of the invention may include general information concerning the identified tire. General information includes stored data concerning the identification of the identified tire (including, without limitation, its place of production, distribution and / or storage, production date, retreading history if applicable, and its position and mounting history). General information may also include the retreading rank (if applicable) of the identified tire. The data corresponding to a retreading rank of an identified tire is usually managed by the entity that manages the use of one or more identified tires (for example, one or more persons and / or one or more companies) and / or the manufacturer of such tires.

[0051] The term "processor" (or, alternatively, the term "programmable logic circuit") refers to one or more devices capable of processing and analyzing data and comprising one or more software programs for processing such data (for example, one or more integrated circuits known to those skilled in the art as being 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). The processor includes one or more software programs for processing the data captured by the detection system (including the camera 200) of the cleaning device 100 (and the corresponding data obtained) as well as one or more software programs for the identification and localization of variances and the identification of their sources in order to correct them.

[0052] In the system of the invention, the memory may comprise both volatile and non-volatile memory devices. The non-volatile memory may include solid-state memories, such as NAND flash memory, keep-alive memory (or KAM) for saving various operating variables while the processor is powered off, magnetic and optical storage media, or any other suitable data storage device that retains data when the cleaning device is deactivated or loses its power supply. The volatile memory may include static and dynamic RAM that stores program instructions and data, including a learning application.

[0053] In embodiments of the system of the invention, the processor can configure the cleaning device 100 (and in particular the processing unit 102) based on one or more parameters of the identified tire calculated by an image processing module incorporated in the processor's memory. The image processing module analyzes the images of the identified tire and, more specifically, the images of the damage seen by the detection system of the cleaning device 100.

[0054] The processor can also refer to a reference (for example, a table of various tire sizes) to perform a final determination of one or more parameters of the identified tire. The reference can include parameters of known tires corresponding to a plurality of commercially available tires. For example, after the image processing module has calculated one or more parameters of the identified tire, the processor can compare the calculated parameters with the known parameters stored in the reference.The processor can retrieve known tire parameters corresponding to commercially available tires that most closely match the parameters calculated to configure the processing device 102 (and thus position the brush 106 or an equivalent processing tool precisely to clean a corresponding cut in the tread of the identified tire). The tire reference can include measurements corresponding to a plurality of commercially available tires. For example, for a tire size 225 / 50R17, the number "225" identifies the tire's cross-sectional area in millimeters, the number "50" indicates the aspect ratio of the sidewall, and the measurement "R17" represents the rim diameter in inches (being approximately 43.18 centimeters).

[0055] With further reference to [Fig. 2], and moreover to Figures 3 to 15, a detailed description is given by way of example of embodiments of a method of the invention (or "method") for cleaning the tread pattern of a tire during retreading. The method of the invention is implemented by the automatic cleaning system of the invention. It is understood that the system can implement the method of the invention in any physical environment without prior knowledge of the configuration of damage to be repaired on the tires identified for retreading.

[0056] As used herein, the term “method” or “process” may include one or more steps performed by at least one computer system comprising one or more processors to execute instructions that perform the steps. Unless otherwise indicated, any sequence of steps is given by way of example and does not limit the described methods to any particular sequence.

[0057] In the following description, embodiments of the method of the invention are described in which the accuracy of the information obtained by the detection system (for example, the camera 200) differs.

[0058] By implementing the method of the invention, the automatic cleaning system of the invention incorporates a combination of vision and machine learning techniques to correctly and rapidly reconstruct the observed scene from three-dimensional (or "3D") scattered point clouds, derived from a view of the tire identified for retreading. The system of the invention thus achieves continuous improvement in the recognition of tread damage and its relative positioning along the tread surface of the identified tire.

[0059] By initiating an embodiment of the method of the invention, the method includes a calibration step for the cleaning device 100. During this step, the robot is positioned to scan an identified tire. When a scan of the identified tire is performed with the camera 200, three-dimensional (3D) points are acquired, which are expressed by their positioning relative to the camera. In order to control the cleaning device 100 (for example, to indicate the position of each area to be brushed on the tread of the identified tire), these 3D points must be expressed at the robot's base (for example, see the base 110 of the cleaning device 100 in [Fig. 2]). The calibration step of the cleaning device 100 therefore includes a calibration phase to estimate the position of the camera 200 relative to its base 110.

[0060] In one embodiment of the process, the calibration step of the cleaning device 100 includes a step of positioning a calibration target in a processing area where the robot will work. In one embodiment, the calibration target includes a checkerboard pattern that is positioned to allow the realization from a plurality of photos, varying the viewing angle each time. During this step, an optimization algorithm (incorporated into camera 200) is used to estimate the camera 200's position within the coordinate system of the processing device 102. Once calibration is complete, the robot can be positioned as seen by camera 200. This calibration must be performed as soon as the cleaning device 106 is installed in the processing area.

[0061] The method further includes a step of positioning a damaged, identified tire in the processing area where the cleaning device 100 will process it. The identified tire is positioned near the cleaning device 100 on a rotating drum (as known in the prior art). During the repair process of the retreading method, the drum rotates the identified tire so that it can be processed sector by sector. It is understood that the identified tire could be positioned on a worktable or equivalent support so that the system of the invention can process it. The support can be configured to move rotaryly, alternately vertically and / or alternately horizontally, thus allowing the processing of a variety of tires.

[0062] The "sector-by-sector" treatment refers to several embodiments of the method of the invention incorporating one or more scans followed by treatment of the tread pattern of the identified tire. In one embodiment of the method of the invention, the sector-by-sector treatment comprises, for each sector of the tire being treated, a scan of the sector followed by treatment of the scanned sector (for example, brushing the tread pattern of the scanned sector using brush 106). In another embodiment of the method of the invention, the sector-by-sector treatment comprises a scan of all successive sectors of the tire being treated followed by treatment (for example, brushing) of all successive scanned sectors.In another embodiment of the method of the invention, the sector-by-sector treatment includes scanning the tire being treated (either a scan of individual sectors or a scan of all successive sectors) at a station different from the one where treatment (for example, brushing) of the scanned sectors is performed. In the latter case, at least one additional camera (not shown) would be used at the remote station to perform all the scans. Then, the camera 200 of the cleaning device 100 would be used to perform a single scan to locate the sector's position within the field of view of camera 200 among all the scans performed at the remote station. In all these embodiments of the method of the invention, it is understood that treatment would be performed on the sectors requiring treatment. Thus, if no damage requiring treatment is present... identified in a scan of a sector, the cleaning device 100 passes the sector concerned until a scanned sector where one or more damages are identified for treatment.

[0063] The method further includes a step of positioning the camera 200 at a reference distance Dref relative to the tread surface of the identified tire (i.e., the damaged surface). During this step, the camera 200 is centered relative to the tread surface of the identified tire (being a right / left positioning). To do this, the robot first positions the camera 200 in front of the identified tire in a position that allows the camera to see all possible tire dimensions (including diameters and widths). A scan is then performed by the camera 200, and it is subsequently analyzed to verify that the camera 200 is correctly positioned.

[0064] Referring to [Fig. 3], the 3D points are expressed with respect to a coordinate system located at the center of the lens of the camera 200. If this coordinate system is represented in 3D by scanning a section of the tire (see "3D visualization" in [Fig. 3]), the X-axis is directed towards the lens of the camera 200, the Y-axis is directed towards the ground, and the Z-axis is directed towards the identified tire. If this coordinate system is represented in 2D (see "2D visualization" in [Fig. 3]), the origin of the coordinate system is at the center of the image: the X-axis is oriented to the right, the Y-axis is oriented downwards, and the Z-axis is oriented towards the tire.

[0065] With reference to [Fig. 4], the correct positioning of the camera 200 during this step is described. The zones 300 represent the points belonging to the tread surface of the identified tire. First, a horizontal central zone of the scan is considered, which is used to verify that the distance between the identified tire and the camera 200 is correct. An average distance Dmoy of the points is calculated within this horizontal central zone, which is represented by the lines 300A. Then, the 3D points whose Z-distance belongs to an interval [Dmoy margin, Dmoy+margin] are retained. These points represent the points belonging to the tread surface of the identified tire (therefore, the relative distances vary according to the parameters of the identified tire).All these points can be represented as a binary image (identified as the "surface mask" or "binary mask") where each retained point represents a point belonging to the tread surface of the identified tire (these points are represented by the 300 zones in [Fig. 4]). Then, to obtain a mean surface distance Davg, the mean surface distance Davg is calculated again, considering only the points in the horizontal central zone that belong to the tread surface of the identified tire. By comparing this distance with the reference distance Dref of the camera 200 (because the camera 200 must always be positioned at distance Dref relative to the tread surface of the identified tire), the mean surface distance Davg can be calculated. Frontal correction is applied to the camera. Next, the two vertical lines 300B are calculated, which delimit the beginning and end of the surface mask and thus allow verification that the camera 200 is correctly centered relative to the tread of the identified tire (see [Fig. 4]). The central vertical line located midway between the vertical lines 300B is then calculated. If this central vertical line deviates too much from the center line of the image, a lateral correction is applied to the position of the camera 200 to recenter it relative to the tread of the identified tire.

[0066] Considering the cleaning device 106, its initial positioning (and, where applicable, the initial orientation of the processing device 102, and thus the brush 106 or an equivalent processing tool) is determined from data obtained through image acquisition of the system and the physical environment in which the system operates (for example, as shown in [Fig. 2]). The processor's analysis application execution module can employ an automatic and adaptive repositioning algorithm to find an ideal starting position for the cleaning device 100, thereby enabling the execution of programmed instructions stored in memory to perform image acquisition.The algorithm allows for continuous improvement across all image captures, ensuring that the system (and particularly the cleaning device 100) improves based on the experience it acquires, especially regarding the choice of paths to follow when carrying out cleaning processes during tire retreading processes.

[0067] The method further includes a tread pattern detection step, which is performed sector by sector for the identified tire. During this step, for each sector, the cleaning device 100 positions the 3D camera to scan the sector and acquire a color, three-dimensional (3D) image of the sector (it should be noted that sector-by-sector processing is involved during this step). At each point in the image, the color of the point is known, as well as its three-dimensional position relative to the camera 200.

[0068] Once the scan is complete, this sculpture detection step includes applying a neural network to the color and 3D image to determine which points represent sculptures to be cleaned. For each image, the neural network outputs a binary mask where each white point indicates that the point represents a sculpture to be cleaned. Before the neural network can be used, a training phase is carried out to set its parameters. For this, several input images are manually annotated to create hundreds of binary masks and indicate which points in the image represent areas of sculpture to be cleaned (see [Fig. 5]). This training phase is carried out once to set the parameters of the neural network, after which the neural network can be used on new images.

[0069] During the neural network application step, if the input image contains no sculpture, the output will be a black mask only. It is understood that false detections may appear in the neural network outputs (represented by way of example by the circled regions 5A shown in [Fig. 5]).

[0070] The method of the invention further comprises a step of creating the three-dimensional (3D) trajectories that the cleaning device 100 should follow to clean the sculptures. Once the binary mask of the sculptures is obtained during the sculpture detection step, the system processor implements a process for generating the 3D trajectories for processing the sculptures (or "3D trajectory generation process") carried out by an algorithm as shown in [Fig. 6] (the phrase "cleaning mask" as shown in [Fig. 6] refers to the binary mask of the sculptures).

[0071] The 3D trajectory generation process includes a data preparation step. During this step, the binary mask is cleaned to remove any noise and to round the edges. Then, during this step, the points belonging to the edges of all the white objects present in the binary mask are extracted (called "edge extraction"). Thus, a set of polygons with holes is obtained (see [Fig. 7], which shows an example of edge extraction associated with the sculpture mask).

[0072] The 3D trajectory generation step includes a step for extracting horizontal and vertical two-dimensional (2D) paths. During this step, any small holes (possibly present within the various polygons extracted during the data preparation step) are removed. The "morphological skeleton" of each polygon is then calculated during this step. This involves extracting all the central points of each polygon. The morphological skeleton (or "skeleton") consists of nodes (which are the ends of the skeleton or the intersection points) and paths (which are the points located between two nodes). Referring to [Fig. 8], an example of a skeleton is shown where lines 8A represent the points of the polygons and lines 8B represent the points of the skeleton. Nodes 8C are indicated in the skeleton, and paths are located between two nodes 8C.

[0073] During the 3D trajectory generation step, each path is then classified as horizontal or vertical. To do this, its bounding box is calculated (see bounding box 9A in [Fig. 9]), and the orientation of a vector composed of its width and height is compared with the y-axis (see [Fig. 9]). If the angle calculated (represented by angle # in [Fig.9]) exceeds a predetermined threshold, the path is then considered to be horizontal, and vertical otherwise.

[0074] For contours composed of several nodes and paths, the vertical and horizontal paths are classified, and paths of the same category (vertical and horizontal) are merged with nearby endpoints. Referring to the example shown in [Fig. 10], skeleton 1 shows a polygon with six (6) nodes and five (5) paths. In skeleton 2, the paths are classified to obtain three (3) horizontal paths and two (2) vertical paths. In skeleton 3, after merging the paths, two (2) horizontal paths and one (1) vertical path are obtained. Once the merging is complete, the direction of travel for each path must be fixed. This point is crucial to ensure that the cleaning device 100 can traverse all the paths of the segment without jamming.In some embodiments of the process, all vertical paths are traversed from bottom to top and all horizontal paths are traversed from right to left. In other embodiments of the process, all vertical paths are traversed from top to bottom, and all horizontal paths are traversed from left to right.

[0075] The 3D trajectory generation step further includes a 3D trajectory elaboration step, which is performed once the vertical and horizontal paths have been extracted. During this elaboration step (which aims to order the different paths to obtain the order in which the sculptures will be cleaned), the vertical paths are ordered first, then the horizontal paths. It is understood that the method of the invention also works by reversing the order (starting with the horizontal paths, then the vertical paths).

[0076] To order the vertical paths during this processing step, in one embodiment of the process, the process begins with a vertical path whose starting point is located furthest away in the image. Referring to the example shown in [Fig. 1 1], the process begins with the vertical path 1 IA, whose starting point is located furthest to the left in the image (it is understood that the process also works by starting with the vertical path 1 IB, whose starting point is located furthest to the right in the image). The next path will be the one whose starting point is closest to the endpoint of the previous path (for example, the central path 1 IC shown in [Fig. 1 1]). The order for the vertical paths is retained (as shown in [Fig. 1 1] by the vertical dashed lines 11D).

[0077] To order the horizontal paths during this processing step, in one embodiment of the process, one starts with a horizontal path whose starting point is closest to the last ordered vertical path. The same reasoning used for the vertical paths is then applied to arrive at a The order for the horizontal paths (represented by lines 11E in [Fig. 11]) determines the processing order for all treads in the scanned area of ​​the identified tire (this order being represented by the dashed line 11F). This method allows for a processing order based on the parameters of the tire being processed and the identified damage. It is understood that the system could employ other equivalent methods to order the vertical and horizontal paths, thus providing the system with processing flexibility to ensure effective cleaning of all treads.

[0078] The method of the invention further includes a path simplification step during which the 3D point associated with each point of each vertical and horizontal path is retrieved (it is recalled that the camera 200 data represents a color and 3D image). During this step, a simplification algorithm is applied to each path to reduce the number of points. By limiting the number of points, the traversal speed of the cleaning device 100 is increased during a cleaning process. This simplification consists of approximating the path with a set of segments and ensuring that each segment does not deviate too much from the initial path. For this purpose, a threshold in millimeters is used which simplifies each path. An example of simplification is shown in [Fig.12] in which diagram A (left) represents an imprecise simplification with a few points 12A, and diagram B (right) represents a precise simplification with more points 12A. In both cases, we seek to simplify in 3D the path 12B and the set of segments (represented by the dashed lines 12C).

[0079] The method of the invention further comprises a step of adding approach and withdrawal points to complete the 3D path obtained after the simplification step. During this step, two approach and withdrawal points are added at the beginning and end of each vertical and horizontal path, enabling the cleaning device 100 to begin and end the cleaning of the path (which corresponds to a groove on the surface of the tire being treated). These approach and withdrawal points are necessary to ensure that the brush 106 (or equivalent treatment tool) does not damage the tire between the cleaning of two successive paths. Thus, these points must not be too close to the tire to ensure that the brush (or equivalent treatment tool) does not touch the tire between two successive paths.At the same time, these points should not be too far apart to avoid excessively long travel times between two successive paths.

[0080] With reference to [Fig. 13], the principle of constructing the different approach points is illustrated. Approach point 1 is constructed by offsetting it by a The predetermined distance L from the first point of the path or the last point of the path (in this case, the distance L from the last point 13A is shown). The offset is made in the same direction as the first or last segment of the path (in this case, the offset is made in the same direction as the last segment of the path). Once offset, the approach point 1 is moved closer to the camera 200 by the distance L. The approach point 2 is constructed from the approach point 1 by moving this point 2 closer to the camera 200 by a predetermined distance L2 from point 1. It is understood that the parameters L and L2 are adjustable. The approach and withdrawal points 13A, 13B are always located between the tread of the identified tire and the camera 200 (see [Fig. 14]).

[0081] The method of the invention includes a final step of calculating the orientations of the brush 106. During this step, the 3D orientation of the brush 106 is calculated at each point of each vertical and horizontal path. To do this, the 3D normal is calculated at each point, as well as the direction vector between each pair of successive points (this vector gives the direction to follow between two points of each path). The normal gives the orientation of the Z-axis of the brush 106, and the direction vector gives the orientation of the X-axis of the brush (the Y-axis is automatically fixed following the fixing of the X and Z axes). Referring to the example in [Fig. 15], the points 15A represent some points of a path 15B, and the line 15C represents the complete 3D path. At each point 15A, the calculated orientation for the brush 106 is represented by the X-axis, the Y-axis, and the Z-axis (in red).It is understood that the calculation described applies in cases where brush 106 is replaced by an equivalent processing tool.

[0082] The system of the invention can easily repeat the steps of the process in an order to thoroughly clean the grooves of the tire identified during a retreading process.

[0083] A person skilled in the art in this field will recognize that many image processing techniques can be used to select and determine the parameters of target tires. Several commercially available image processing systems can be used.

[0084] In a tire retreading facility, a vision system can be used to detect the presence of a tire in the field of vision of the camera 200, which triggers the camera to capture the image of the identified tire for processing.

[0085] In embodiments of the process of the invention, one or more steps of the process may further include a step of scanning the physical environment containing the cleaning device 100. In embodiments of the process, this step further includes a step of measuring the physical environment to achieve precise positioning of the brush 106 (or equivalent processing tool). During this step, one or more sensors can be used to capture data corresponding to the processing tools and tires in order to determine the shapes and / or positions of individual tires. This information is relevant for enabling accurate modeling of the process of the invention in order to optimize the time of an associated retreading process.

[0086] A method of the invention can be implemented via PLC control and may include pre-programmed management information. For example, a process setting can be associated with the parameters of the tire being retreaded, and / or the properties of tires processed in a retreading plant incorporating the cleaning device 100 (e.g., aircraft tires, truck tires, etc.). The system of the invention (and / or a plant incorporating this system) can easily repeat one or more steps of the method of the invention in a predetermined order to ensure substantially complete cleaning of the tread pattern (and particularly the grooves) of the tire being treated.

[0087] The system of the invention (and / or a retreading plant incorporating this system) may include pre-programmed management information. For example, a process setting may be associated with the parameters of the typical physical environments in which the system operates. In embodiments of the invention, the system (and / or a retreading plant incorporating this system) may receive voice commands or other audio data representing, for example, a start or stop of the cleaning process and / or the loading / unloading of a drum dedicated to tire retreading. A request may be made that includes a request for the current status of an ongoing retreading process cycle. A generated response may be represented audibly, visually, tactilely (for example, using a haptic interface), and / or virtually and / or augmented.This response, along with the corresponding data, can be recorded in a neural network.

[0088] For all embodiments of the system, a monitoring system could be implemented. At least part of the monitoring system can be provided in a portable device such as a mobile network device (e.g., a mobile phone, a laptop computer, one or more portable network-connected devices (including augmented reality and / or virtual reality devices), network-connected wearable clothing, and / or any combination thereof). It is conceivable that detection and comparison steps could be performed iteratively.

[0089] In one embodiment, the method of the invention may include a step of training the system to recognize representative values ​​of the The process can be used to determine the parameters of retreaded tires (e.g., inner and outer diameter values) and to compare them with target values ​​(e.g., to place a tire order that includes retreaded tires). Each training step can include a classification generated by self-learning methods. This classification can include, but is not limited to, the parameters of the selected tires, the durations of the process cycles, and the expected values ​​at the end of a process in progress (e.g., the number of tires retreaded during a cycle to reach a desired level).

[0090] The terms "at least one" and "one or more" are used interchangeably. The ranges presented as being "between a and b" encompass the values ​​"a" and "b".

[0091] Although particular embodiments of the disclosed apparatus have been illustrated and described, it will be understood that various changes, additions, and modifications can be made without departing from the spirit or scope of this disclosure. Therefore, no limitations should be imposed on the scope of the invention described except those set forth in the appended claims.

Claims

Demands

1. A method for cleaning the tread pattern of an identified damaged tire during retreading, the method implemented by at least one processor comprising a processing module that applies, to at least one neural network, representative data from the captured images of the tread pattern of the identified tire, characterized in that the method comprises the following steps: - a step of providing an automatic cleaning system of which the processor is a part for automatically recognizing a damage treatment order in the tread pattern of the identified tire; - a step of calibrating a cleaning device (100) which is part of the automatic cleaning system, the cleaning device (100) comprising a processing device (102) supported by an elongated pivoting arm (104) which extends to a free end (102a) where a processing tool is disposed along a common longitudinal axis;- a step of positioning the identified tire in a processing space where the cleaning device (100) processes it; - a step of positioning at least one camera (200) which is part of the automatic cleaning system at a reference distance from a tread surface of the identified tire; - a tread detection step comprising sector-by-sector processing for the identified tire incorporating one or more scans followed by tread processing of the identified tire to recover a color and three-dimensional (3D) image of each scanned sector, in which, during the tread detection step, the sector-by-sector processing for the identified tire includes: - for each sector of the identified tire, a sector scan step followed by processing of the scanned sector;- a scan of all successive sectors of the identified tire followed by processing of all successive scanned sectors; or - a scan of the tire being processed on a station different from the one where processing of the scanned sectors is carried out; so that a processing of one or more scanned sectors is carried out by the cleaning device (100); - a step of creating the three-dimensional (3D) trajectories that the cleaning device (100) should follow to process the treads; - a step of simplifying the vertical and horizontal paths during which a 3D point associated with each point of each path is retrieved; - a step of adding the approach and withdrawal points to complete the 3D trajectory obtained after the path simplification step, during which, at the beginning and end of each vertical and horizontal path, an approach point and a withdrawal point are added allowing the cleaning device (100) to start and finish the processing of the path that corresponds to a tread on the identified tire tread;and - a final step of calculating the orientations of the processing tool during which a 3D orientation of the processing tool is calculated at each point of each vertical and horizontal path; so that the cleaning device (100) is set in motion so that the processing device (102) can place the processing tool to carry out the cleaning of the tread of the identified tire.;

2. The cleaning method of claim 1, wherein the calibration step of the cleaning device (100) includes a calibration phase to estimate the pose of the camera (200) relative to a base (110) of the cleaning device (100).

3. The cleaning method of claim 2, wherein the calibration step of the cleaning device (100) includes a positioning step of a calibration target in the processing space, during which the camera (200) implements a process to estimate the pose of the camera (200) in the reference frame of the brushing device (102).

4. The cleaning method of claim 3, wherein, during the positioning step of the identified tire, the identified tire is positioned in the vicinity of the cleaning device (100) on a rotating drum disposed in the processing space, such that, in the course of a repair process of a process of During retreading, the drum rotates the identified tire so that it can be treated sector by sector.

5. The cleaning method of any one of claims 1 to 4, wherein the camera (200) positioning step includes a scan verification step performed by the camera (200) comprising the following steps: - a step of identifying a horizontal central zone of the scan which is used to verify that the distance between the identified tire and the camera (200) is correct; - a step of calculating an average distance (Dmoy) of the points within the horizontal central zone; - a step of storing the 3D points in the form of a surface mask comprising a binary image representing the points belonging to the tread surface of the identified tire; - a step of recalculating the average distance (Dmoy) by considering only the points of the horizontal central zone that belong to the tread surface of the identified tire to obtain a following distance (Dmoy);- a step comparing the distance (Davg surface) with the reference distance of the camera (200); - a step calculating the two vertical lines (300B) that delimit the beginning and end of the surface mask to allow verification that the camera (200) is centered with respect to the tread surface of the identified tire; - a step calculating a central vertical line located in the middle of the vertical lines (300B); and - where the calculated central vertical line deviates from the center line of the image beyond a predetermined threshold, a step applying a lateral correction to the position of the camera (200) to recenter it with respect to the tread surface of the identified tire.

6. The cleaning method of claim 5, wherein the tread detection step includes a step of applying a neural network to the color and 3D image to determine which points represent treads of the identified tire to be processed with the processing tool, during which, for each image, the surface mask is retrieved as output from the neural network where Each white dot indicates that the dot represents a sculpture to be cleaned.

7. The cleaning method of claim 6, wherein the three-dimensional (3D) trajectory creation step comprises the following steps: - a data preparation step, during which points belonging to the contours of all white objects present in the surface mask are extracted to obtain a set of polygons with holes;and - a step of extracting two-dimensional (2D) horizontal and vertical paths, during which the holes present inside the polygons extracted during the data preparation step are removed, this step comprising the following steps: - a step of calculating a morphological skeleton of each polygon including the extraction of all the central points of each polygon, in which each morphological skeleton consists of nodes representing the ends of the morphological skeleton and / or the crossing points, and of paths comprising points located between two nodes; - a step of classifying each path into the category horizontal path or category vertical path, during which paths of the same category are merged with nearby ends and, once the merging is carried out, the direction of traversal of each path is fixed;and - a 3D trajectory development step which is carried out once the vertical and horizontal paths have been extracted, during which the vertical and horizontal paths are ordered; so that the ordered set of vertical and horizontal paths gives the processing order for all the treads in the scanned sector of the identified tire.

8. The cleaning method of any one of claims 1 to 7, wherein, during the step of adding and withdrawing the approach and withdrawal points: - a first approach point is constructed by offsetting it by a predetermined distance (L) from the first or last point of the path, and, once offset, the first approach point is brought closer to the camera (200) by the distance (L); and - a second approach point is constructed from the first approach point by bringing this second approach point closer to the camera (200) by a distance (L2) from the first approach point.

9. The cleaning method of any one of claims 1 to 8, wherein the step of calculating the orientations of the processing tool includes a step of calculating the 3D normal at each point as well as the direction vector between each pair of successive points, wherein the direction vector gives the direction to follow between two points of each path, where the normal gives the orientation of the Z axis of the processing tool, the direction vector gives the orientation of the X axis of the processing tool, and the Y axis is fixed automatically according to the fixing of the X and Z axes.

10. The cleaning method of any one of claims 1 to 9, wherein the processor refers to a table of various tire sizes to perform a determination of one or more parameters of the identified tire.

11. The cleaning method of any one of claims 1 to 10, wherein the processing tool comprises at least one brush (106).

12. A retreading process comprising the cleaning process of any one of claims 1 to 11.

13. An automatic cleaning system for cleaning the tread pattern of an identified tire during retreading, the automatic cleaning system comprising: - at least one cleaning device (100) comprising a robot having a cleaning peripheral (102) supported by a pivoting extended arm (104) extending from the extended arm (104) to a free end (102a) where a processing tool is disposed along a common longitudinal axis; - a detection system for gathering information on the physical environment around the cleaning device (100), wherein the detection system comprises at least one RGB-D camera (200) fixed to at least one of the extended arm (104) and the processing peripheral (102) of the cleaning device (100); and - a communication network that manages the incoming data to the automatic cleaning system from each device. cleaning (100), the communication network comprising at least one communication server enabling the execution of programmed instructions stored in a memory of one or more processors of the automatic cleaning system to implement the cleaning process of any one of claims 1 to 12; so that the cleaning device (100) is configured on one or more parameters of the identified tire calculated by an image processing module incorporated in the processor memory to set the cleaning device (100) in motion so that the brushing device (102) can place the processing tool to perform the cleaning of the tread of the identified tire.

Images