A method of improved retreading of a tire carcass

Abstract

An method of inspecting a tire carcass and either accepting the tire carcass as a candidate for retread or rejecting the tire carcass by identifying anomalies and or defects using an artificial intelligence deep learning image recognition software, whereby the system is improved by use of two dimensional image data augmented with depth data assigned to an image attribute of the a dimensional image for processing by, for example, a convolutional neural network.

Description

A METHOD OF IMPROVED RETREADING OF A TIRE CARCASSFIELD OF THE INVENTION

[0001] The subject matter of the present invention relates to an improved method of retreading of a tire carcass by using an artificial intelligence algorithm to select a suitable tire carcass for retreading.BACKGROUND OF THE INVENTION

[0002] Retreading is a process by which a tire, having a carcass designed to be durable enough to allow the wearing of more than one tread, has the tread rubber replaced. Such practice is common with heavy truck tires. Since the tire being retread is subject to abuse and obstacles found on and off the road, a wide variety of anomalies, damages, and defects require sophisticated means to detect and classify issues accurately. The current retread industry uses manual visual inspection by trained operators. These operators inspect the full tire, including the tread summit, sidewalls, beads and interior of the tire. It would be desirable to automate the inspection to provide consistency of the tire carcass quality for retread and reduce the time operators have to handle the tire for inspection.

[0003] The advent of artificial intelligence systems and deep learning provides a means to automate the inspection process. Two-dimensional image analysis lacks the sophistication to enable accurate determination of suitability of a tire carcass for retreading. Three-dimensional data requires heavy use of computing resources and thus far has not been shown to be a viable solution for inspection of tire carcasses in a commercial setting.

[0004] The details and benefits of the use of deep learning to detect tire defects are described e.g., in Chinese patent publication CN108711148B, CN110660049, CN113390882, CN113901947, CN109738452, CN110310262, CN109785313 and CN110059751. These publications describe the use of deep learning or neural networks as part of the process to identify a suitable tire for retreading.

[0005] Some of these utilize X-Ray technology which then is fed into an artificial intelligence model for the selection of an appropriate tire for retreading, embodiments of which are described in e.g., Chinese patent publication CN108711148B, CN110660049, CN109738452, CN110310262, and CN109785313 disclose the use of x-ray technology forinspection of the tires for defects. These disclosures require training of the Al system utilizing x-ray technology complicating the process by requiring special x-ray equipment and may not detect defects or anomalies that do not lend themselves to visual inspection.

[0006] CN113390882 is a method of using a system where images of the object are fed through a deep learning algorithm. While the system utilizes visual images, the disclosure does not teach of the use of depth to accurately identify defects below the surface of the tire. CN113901947 is another example where two-dimensional images are used, here disclosing a method for training a neural network to identify tire surface flaws using a small number of samples. It does not disclose depth to accurately identify defects below the surface of the tire.

[0007] What is needed is an efficient method to inspect retread tire carcasses for defects that will allow the system to run more efficiently and identify defects both on the surface and within the tire.SUMMARY OF THE INVENTION

[0008] Aspects and advantages of the invention will be set forth in part in the following description, or may be obvious from the description, or may be learned through practice of the invention.

[0009] The present invention provides a method of improved retreading of a tire carcass, which includes scanning a surface of the tire carcass to obtain a three-dimensional dataset representing the width, height and depth of the surface of the tire, then converting the three-dimensional dataset to a two-dimensional image of height and width while the depth of the image retained as an image attribute, then processing the two-dimensional image using a deep learning algorithm to classify the tire carcass into at least one of two categories, where at least one category is an acceptable tire carcass category representing an acceptable tire carcass for retreading. In at least one embodiment the tire carcass is then retreaded the tire carcass if it belongs the acceptable tire carcass category.

[0010] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.BRIEF DESCRIPTION OF THE DRAWINGS

[0011] A full and enabling disclosure of the present invention, including the best mode thereof, directed to one of ordinary skill in the art, is set forth in the specification, which makes reference to the appended figures, in which:

[0012] FIG. 1 provides a flowchart of the general process of the invention.

[0013] FIG. 2 provides an example of the acquisition of three-dimensional data acquisition of the tire selected to potentially be retreaded.

[0014] FIG. 3 provides a two-dimensional image of the exterior of the tire’s first sidewall, crown region and second sidewall.

[0015] FIG. 4 provides a two-dimensional image of the exterior of the tire’s first sidewall, crown region and second sidewall with the color of the image representing depth of the surface features of the tire carcass.

[0016] FIG. 5 provides an example of an interior surface image that may be used for enhancing the determination of the suitability of the tire carcass for retreading.

[0017] The use of identical or similar reference numerals in different figures denotes identical or similar features.DEFINITIONS

[0018] The following terms are defined as follows for this disclosure:

[0019] Artificial Intelligence or “Al” refers to computer algorithms using deep learning or machine learning to process, compare and / or categorize data.

[0020] Convolutional Neural Network or “CNN” refers to a type of Artificial Neural Network used primarily for image recognition and processing, due to its ability to recognize patterns in images.

[0021] Artificial Neural Network, “ANN”, “Neural Network”, or “NN” refer to the branch of Machine Learning models that are built using principles of neuronal organization intended to simulate how biological neural networks are perceived to work in animal brains. Such networks are used to solve problems or recognize characteristics in the data by processing examples of inputs and results.

[0022] Machine Learning is a type of Artificial Intelligence in which machines, usually computers, discover their own algorithms to solve problems.

[0023] Deep Learning is a type of machine learning algorithms that use multiple layers to progressively extract higher-level features from the raw input which allows the computer program to learn to associate an input with an output.DETAILED DESCRIPTION OF THE INVENTION

[0024] The invention disclosed herein improves the method of identifying retread tire carcass defects by reducing three-dimensional surface maps to two-dimensional images which can be processed using image analysis comparison software such as Vision Al (Artificial Intelligence). In this invention, three-dimensional data is preprocessed to "flatten" the overall image, then is further processed to convert three-dimensional data to a rich color "heat map" that translates depth data to a color representation on the two- dimensional image. This allows multiple open source 2D Al CNN (Convolutional Neural Network) models to detect and classify anomalies.

[0025] For purposes of describing the invention, reference now will be made in detail to embodiments and / or methods of the invention, one or more examples of which are illustrated in or with the drawings. Each example is provided by way of explanation of the invention, not limitation of the invention. In fact, it will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the scope or spirit of the invention. For instance, features or steps illustrated or described as part of one embodiment, can be used with another embodiment or steps to yield a still further embodiments or methods. Thus, it is intended that the present invention covers such modifications and variations as come within the scope of the appended claims and their equivalents.

[0026] FIG. 1 provides a flowchart of the process of classifying a tire carcass intended for retreading. The process is broken down into three main stages 10, 30 and 50 prior to the decision phase 60 where the tire is either accepted for retreading or rejected.

[0027] The first part of the process 10 a worn tire carcass is obtained for the retread process and the tire carcass is scanned to obtain a three-dimensional dataset of at least one surface of the tire carcass 20. In at least one embodiment a laser profilometer 100, such as being shown used in FIG. 2 is used to obtain three-dimensional data of the exterior of the tire 110. Two-dimensional images of the tire interior 112, such as shown in FIG. 5, may be obtained using a digital camera. In other embodiments, the three-dimensional data may include the tire interior and the two-dimensional images may also include the tire exterior. In yet additional embodiments, only three-dimensional data scans are obtained of either the exterior of the tire, the interior of the tire or both.

[0028] During the image acquisition phase, in the first embodiment the tire is moved to capture images. In alternative embodiments, the tire may be stationary while the imagingsensors may be moved to capture the surfaces of the tire carcass, or a combination of tire carcass movement and image sensor movement can be made to capture the initial data. The data acquisition may be automated or manual and a variety of methods as known by a person of ordinary skill in the art may be employed to manipulate the tire carcass or equipment to capture the data.

[0029] During the second part of the process the three-dimensional dataset is converted to a two-dimensional image having relative depth of the surface shown by means of color 30. This creates an image that is in some ways analogous to a “heat map”, though the color represents depth of the features of the surface and not temperature. In the first embodiment, three, two-dimensional images are formed, one image of the exterior crown including the tread, and one of each sidewall, extending from the bead area to the shoulder area of the tire. In other embodiments a single two-dimensional image may be formed. In yet other embodiments, two or more images may be formed from the three-dimensional data for analysis.

[0030] In the first embodiment, during the second part of the process, the image of the sidewall of the tire is “straightened” by converting the ring-like circular sidewall to a stretched out linear representation as shown FIG. 3 showing a two-dimensional image of the first exterior sidewall 120, and the second exterior sidewall 140. This may be accomplished by the data acquisition system or camera system during the acquisition of the data or image or may be accomplished by mapping the radial coordinates of the data to cartesian coordinates and distorting (projecting) the data or image so that the circular sidewall is shown as a single rectangular linear image. In other embodiments the sidewall may remain imaged in a circular form for analysis.

[0031] In the first embodiment, the width and length of the image comes from the X- axis and Y-axis data of the three-dimensional data while the third dimension, the Z-axis, of the three-dimensional data, the “depth” of the surface, is converted to a color spectrum, with each color representing a depth of the surface or relative distance along the Z-axis. As such, the Z-axis dimension is converted to a high-resolution color scale, displaying Z values as a color scale as shown in FIG 4 showing the first sidewall 220, the exterior crown surface 230, and the second sidewall 240. This enables well developed Convolutional Neural Network deep learning models to compare the images and identify characteristics that are typical of defects that would require rejection of the tire carcasses unsuitable for retreading. In other embodiments, the depth data may be converted shades of grey. In yetother embodiments, the depth data may be converted to another image attribute resulting in a two-dimensional image also representing the depth of the surface.

[0032] The resulting image, or color map if as described above by the first embodiment, can then be fed to a CNN deep learning module. Open-source modules may be utilized enabling comparison of several models to determine which models can identify the undesirable anomalies best. A plurality of CNN models can be arranged in processing order or “stacked” to maximize accuracy for all target anomalies.

[0033] Previous attempts to use CNN models directly with three-dimensional data have been attempted, however, none have worked well enough to accurately detect (“presence”) nor classify (“codify”) anomalies in the retread tire carcass. Previous attempts at using two-dimensional imaging systems did not provide images that the CNN models could successfully process to learn to detect or classify anomalies. The authors believe that acceptable “good tires” which are suitable for retreading versus unacceptable “bad tires” which are not suitable for the retread process require the ability to distinguish absolute or relative depth or, in other words, attributes of the tire carcass’s three-dimensional shape.

[0034] An embodiment of the invention presented here captures three-dimensional laser profilometer data, converts it to a rich two-dimensional image. This rich two- dimensional image is in-turn useful utilize current open-source two-dimensional image Al models available for CNN learning.

[0035] There are two stages of the image recognition system development. During a first stage, the image recognition system is taught to recognize the anomalies by feeding the system images of tire carcasses showing known defects and images of tire carcasses showing no defects. As the image recognition system is fed more images, the image recognition system learns to distinguish between images of tires with defects and images of tires without defects. As the image recognition system is taught, it can learn to categorize the anomalies and defects of the tires in addition to categorizing them as suitable or unsuitable for the retread process. Once an acceptable accuracy of the model’s ability to categorize the pictures is achieved, the system is then operated in a second phase where images are categorized and ranked and the resulting tires are either determined to be suitable for retread process, or are determined to be unsuitable and either subject to further processing or rejected outright.

[0036] Direct three-dimensional Al is far less mature than two-dimensional image recognition and classification Al. Most inspection tasks are reasonably successful using two-dimensional Al, when good= no anomalies present. Retreaded tires are unique in thatsmall anomalies below certain dimensions are acceptable, while larger ones are not. This required consideration of three-dimensional mapping. However three-dimensional Al models are not up to the task as yet.

[0037] Selected combinations of aspects of the disclosed technology correspond to a plurality of different embodiments of the present invention. It should be noted that each of the exemplary embodiments presented and discussed herein should not insinuate limitations of the present subject matter. Features or steps illustrated or described as part of one embodiment may be used in combination with aspects of another embodiment to yield yet further embodiments. Additionally, certain features may be interchanged with similar devices or features not expressly mentioned which perform the same or similar function.

[0038] As used herein, the term “method” or “process” refers to one or more steps that may be performed in other ordering than shown without departing from the scope of the presently disclosed invention. As used herein, the term "method" or "process" may include one or more steps performed at least by one electronic or computer-based apparatus. Any sequence of steps is exemplary and is not intended to limit methods described herein to any particular sequence, nor is it intended to preclude adding steps, omitting steps, repeating steps, or performing steps simultaneously. As used herein, the term "method" or "process" may include one or more steps performed at least by one electronic or computer-based apparatus having a processor for executing instructions that carry out the steps.

[0039] The terms "a," "an," and the singular forms of words shall be taken to include the plural form of the same words, such that the terms mean that one or more of something is provided. The terms "at least one" and "one or more" are used interchangeably.

[0040] Every document cited herein, including any cross-referenced or related patent or application is hereby incorporated herein by reference in its entirety unless expressly excluded or otherwise limited. The citation of any document is not an admission that it is prior art with respect to any invention disclosed or claimed herein or that it alone, or in any combination with any other reference or references, teaches, suggests or discloses any such invention. Further, to the extent that any meaning or definition of a term in this document conflicts with any meaning or definition of the same term in a document incorporated by reference, the meaning or definition assigned to that term in this document shall govern.

Claims

WHAT IS CLAIMED IS:

1. A method of improved retreading of a tire carcass, comprising: scanning a surface of the tire carcass to obtain a three-dimensional dataset representing the width, height and depth of the surface of the tire; converting the three-dimensional dataset to a two-dimensional image of height and width while the depth of the image retained as an image attribute; processing the two-dimensional image using a deep learning algorithm to classify the tire carcass into at least one of at least two categories, at least one acceptable tire carcass category representing an acceptable tire carcass for retreading; and retreading the tire carcass if it belongs the acceptable tire carcass category.

2. The method of claim 1 wherein the surface of the tire carcass includes a part surface selected from the group consisting: a first sidewall surface, a second sidewall surface, a crown view surface, a crown innerliner surface, a first sidewall innerliner surface, a second sidewall innerliner surface, a first bead surface, a second bead surface, and a combination of two or more of said part surfaces.

3. The method of claim 1 or 2 wherein said scanning step includes capturing depth data using laser profilometer to capture the three-dimensional dataset.

4. The method of claim 2 wherein the three-dimensional dataset is a three-dimensional surface map.

5. The method of any one of the above claims wherein converting the three- dimensional dataset to a two-dimensional image is completed by flattening the three- dimensional dataset and converting the depth data to color values on the two-dimensional image.

6. The method of any one of the above claims wherein processing the two- dimensional image includes utilizing one or more Convolutional Neural Network deep learning models.

7. The method of claim 6 wherein said one or more Convolutional Neural Networks are trained to identify anomalies on the tire carcass.

8. The method of any one of the above claims wherein the at least one of two categories includes a suspect tire category indicating the tire needs to be further studied and either rejected or further processed.

9. The method of any one of the above claims wherein the method of improved retreading further comprises the step of retreading the acceptable tire carcass.

10. The method of claim 9 wherein the tire is a truck tire carcass.

11. A retread tire produced by the method of any one of the above claims.

Images