Method and system for determining a tyre condition, equipment, method for training a neural network to estimate the depth of a tyre groove and corresponding computer-readable memories

Abstract

The present invention relates to a method and a system for determining a tyre condition by estimating and / or indirectly measuring the depth of one or more grooves of the tyre(s) of interest, by submitting an image of the tread of the tyre(s) of interest to a machine learning model. By unequivocally linking the depth of each groove to its position on the tyre tread, it is possible to make a series of comparisons and determine individual and relative tyre conditions, as well as instructions on how to proceed such as continue with the tyre as it is, discard it, retread it, calibrate it, check its alignment, and change its position. In addition, the present invention relates to a method for training a neural network to estimate the depth of a tyre groove, as well as to equipment, and to corresponding computer-readable memories.

Description

METHOD AND SYSTEM FOR DETERMINING A TIRE CONDITION, EQUIPMENT, METHOD FOR TRAINING A NEURAL NETWORK TO ESTIMATE THE DEPTH OF A TIRE GROOVE AND CORRESPONDING COMPUTER-READ MEMORIES Field of application

[0001] The present invention pertains to the field of devices, systems, and methods for testing vehicles and their respective components. More specifically, it relates to the field of devices, apparatus, systems, and methods for evaluating tire tread wear. Finally, the present invention also relates to the field of computer-implemented arrangements based on machine learning. Introduction

[0002] The present invention relates to a method and system for determining tire condition by estimating and / or indirectly measuring the depth of one or more grooves in the tire, by submitting an image of the tire tread to a machine learning model represented by a trained convolutional neural network. The association of the measured depth of each groove with its position in the tire tread allows, in addition to the identification of any depths below a minimum expected value, information indicative of irregular tire wear and the suggestion of appropriate maintenance or replacement measures.

[0003] The present invention, therefore, also relates to a method for training a neural network to estimate the depth of a tire groove, in which the convolutional neural network is trained with a set of images of tire treads with grooves in different states of conservation and previously known depths, so that this neural network is able to, from an input image, for example, a photograph of a tire tread, identify the grooves individually, estimate their depths and associate them with their position on the tread.

[0004] The invention further relates to an apparatus for performing the method of determining a tire condition and corresponding computer-readable memories. Fundamentals of the invention

[0005] Road transportation plays a fundamental role in transporting cargo and passengers in countries like Brazil. Even maritime transport modes end up relying on road transportation for part of their land journeys, given the infrastructure deficiencies of the rail network. Estimates from the latest ANTT (National Land Transportation Agency) Yearbooks indicate that the road network's influence on cargo is enormous, accounting for 61% to 75% of cargo transported in Brazil, depending on the year.

[0006] Although several factors can be considered critical, as they clearly and profoundly impact a series of other elements that depend on them, e.g., the quality of routine and rest for vehicle drivers, one of the most obvious and evident is the quality of truck components, especially those that suffer the greatest daily wear and tear.

[0007] Of these, tire lifespan is worth highlighting. Its impact is obvious: inadequate tires, due to wear or improper inflation, result in higher fuel consumption, steering instability, balance and alignment problems, and increased susceptibility to adverse road conditions and defects. This can also lead to defects and damage to other vehicle components, such as the suspension and axles. bearings and other parts, unequivocally posing risks to the driver, his cargo, the environment and the general population.

[0008] Technical and logistical solutions that assist in monitoring, rotating, maintaining, and replacing tires—resulting in better tire use, faster tire replacement, and greater safety for everyone involved, from drivers to technicians—have been challenging technicians and engineers worldwide for several years. State of the art

[0009] Various solutions for determining tire wear conditions are known in the prior art, for example, the solution disclosed in Korean patent document number KR102079862. This document describes an invention that acquires tread image information by photographing the exterior of a tire with a camera. Information on the tire tread wear condition can be acquired without a separate measuring device, such as an image detection device and a laser.Said invention relates to a measurement system and method, comprising: a first step of obtaining an image of a tire tread photographed by a camera; a second step of detecting a groove area and a surface area from the tire tread image information; a third step of evaluating tire wear by comparing the image information of the area detected in the step with the existing data constructed in the database of the management server; a fourth step of constructing a database storing the image information in the management server; and a fifth step of providing a service including expert feedback, search results of nearby tire specialty stores, and information on required products to the user's PC or handheld terminal based on the tire wear evaluation result and supplying. of tire image information.

[0010] Essentially, the KR102079862 method and system aim to assess tire wear based on reading camera images through database comparisons, with the aim of improving analysis. However, disadvantages and limitations are readily apparent in this document.

[0011] Its analysis relies on pre-processing of images, requiring more time and less agility, especially for poorly trained technicians. It also focuses on detecting wear, but lacks the ability to quantify tread depth. Therefore, it neither teaches nor suggests how to assess tire condition in relation to a vehicle's systemic variables, for example, failing to correlate tread depth with its position on the tread. This hinders its versatility, as it is limited to comparisons with ideal tread pattern models and suggestions from nearby stores, without providing procedure and maintenance recommendations.

[0012] Therefore, the Korean document KR102079862 does not teach or suggest a detailed measurement and evaluation of the tread grooves, based on a variety of images, in order to allow maintenance, rotation and prevention of tire damage in a practical, routine and everyday manner, both for the driver and the responsible technician.

[0013] Another solution for determining tire wear conditions is revealed by Japanese patent document number JP7020581, which presents and describes a method for estimating uneven tire wear that includes the following steps: acquiring an image including a tread of a target tire; inputting the acquired image into the first trained machine learning model; deriving the output of the first model trained machine learning. Thus, the output of the first trained machine learning model will correspond to the result of estimating the uneven wear of the target tire.

[0014] The aforementioned Japanese document, in short, advantageously intends to use machine learning for its method, including a form of tire rotation suggestion, this being one of the objectives sought by the present invention.

[0015] However, the Japanese document JP7020581 also does not teach or suggest a detailed evaluation of the tread grooves and the comparison of the grooves of the same tire among themselves and in relation to the tread of another tire on the same axle, much less does it teach or suggest the comparison between tires on different axles of the same vehicle.

[0016] Finally, it is worth noting that several prior art documents encounter difficulties such as those mentioned above, including US20200125887, KR20220081236, and US20200098121. Examples of these difficulties include complex image and data conversions, region limitations of the tires analyzed, dependencies on complex structures such as screens, sensors, meters, etc.

[0017] It should be noted that none of the relevant prior art methods allow for a depth measurement and a detailed assessment of tire tread grooves based on a machine-trained model in the form of a trained convolutional neural network. They ultimately fail to enable tire maintenance, rotation, and damage prevention in a practical, routine, and everyday manner, both for the vehicle driver and the technician in charge.

[0018] It is also necessary to emphasize that none of the solutions known from the aforementioned prior art has a sufficiently intuitive to allow a technician to adapt to different situations of trucks and tires analyzed, without requiring great specialization in the subject.

[0019] It's also worth noting that conventional tire tread depth measurement, widely used to check tire condition, faces known problems, as it depends on the quality of the measuring instrument, its condition, calibration, and verification, the dexterity and skill of the person performing the measurement, and mathematical / statistical variations between the results of a series of measurements. Furthermore, conventional measurement is time-consuming, and recording the results for decision-making is uneven and dependent on the attention of the person performing the task.

[0020] As can be inferred from the previous descriptions, there is room for a method and system for determining tire condition, a method for training a neural network to estimate tire tread depth, and corresponding machine-readable equipment and memories that overcome the shortcomings of the current state-of-the-art technology. This solution represents a simple, cost-effective solution that is intuitive for drivers, users, technicians and / or mechanics, and fleet managers. It is easy to install, flexible, adaptable, and does not compromise existing vehicle or road infrastructure. It should also be capable of reducing truck drivers' waiting time during their routes, facilitating tire reorganization, and offering low financial impact while also providing safety and stability, without compromising the health of users and operators involved. Objectives of the invention

[0021] One of the objectives of the present invention is therefore a method for determining a tire condition, according to the characteristics of claim 1 of the attached table of claims.

[0022] Another objective of the present invention is a system for determining a tire condition, according to the characteristics of claim 8 of the attached claim table.

[0023] Yet another object of the present invention is an equipment for determining a tire condition, according to the characteristics of claim 17 of the attached claim table.

[0024] Another objective of the present invention is a method for training a neural network to estimate the depth of a tire groove, according to the features of claim 19 of the attached claim table.

[0025] A further object of the present invention is to provide a computer-readable memory comprising a set of instructions that, when executed, perform the method for determining a tire condition and / or the method for training a neural network to estimate the depth of a tire groove, in accordance with the features of claim 20 of the attached claim table.

[0026] Other characteristics and details of the characteristics are represented by dependent claims. Brief description of the figures

[0027] For better understanding and visualization of the object of the present invention, it will now be described with reference to the attached figures, representing the technical effect obtained by means of exemplary embodiments that do not limit the scope of the present invention, in which, schematically: Figure 1: shows a schematic top view of the set of axles and tires of a vehicle, in a mode of viewing and selecting the tires of interest, according to a system and method of the invention; Figure 2: shows a front view of an image obtained from a region of interest of a tire tread of interest, highlighting the association of the position indicators or codes with the depth indicators; Figure 3: shows a flowchart of a method for determining a tire condition according to the invention; and Figure 4: presents a continuation of the flowchart of the method for determining a tire condition from Figure 3. Detailed description of the invention

[0028] The following detailed description refers to the accompanying drawings, which illustrate, by way of non-limiting illustration, embodiments of the present invention. These embodiments are described to allow a person skilled in the art to reproduce their results. Other embodiments resulting from structural, mechanical, logical, electrical, and electronic changes are possible and may be implemented without departing from the spirit and scope of the present invention. The following detailed description should therefore not be construed as restrictive or limiting.

[0029] The present invention relates to a method and a system for determining the tire condition of a tire of interest. Furthermore, the present invention relates to a method for training a neural network to estimate the depth of a tire tread, a corresponding device, and also to computer-readable memories containing a set of instructions that, when executed, perform the methods of the invention in a system and device of the invention.

[0030] In the context of the present invention, the term "tire" refers to any and all inflatable rubber rims, with or without an inner tube, mounted on the wheel joints of various vehicles, comprising, externally, sidewalls, shoulders and a tread with grooves and, eventually, bars and grooves, in addition to comprising, internally, belts, bead, liner, carcass and other elements known from the state of the art of tires of the nature dealt with here, with the external part of the tire being of special interest, more especially the tread and, more specifically, the tread grooves.

[0031] In the context of the present invention, the term "tire of interest" refers to one or more tires that are to be evaluated by the system and methods of the invention, and therefore there is at least one vehicle of interest attached to one or more of the tires of interest.

[0032] In the context of the present invention, the expression "a tire condition" should be understood as being both the individual conservation state of the tire of interest, used and / or in use, in relation to at least one threshold value and / or at least one minimum expected value and / or at least one minimum expected condition, representing an "individualized tire condition", and the conservation state of the tire of interest, used and / or in use, in relation to at least one other tire of interest, representing a "relative tire condition".

[0033] It should be noted that the choice of the tire condition to be evaluated, i.e., the individual tire condition and the relative tire condition, may be decided by a system user or be pre-established by the initial tire verification condition.

[0034] All tire conditions of the invention take into account the state of conservation of the external regions of the tire, in particular, the measurement of the depth of the grooves in the tread, being able to evaluate (i) the depth value of each of the grooves individually and / or (ii) preferably the average of the groove depths of the tire of interest and / or (iii) the groove depths individually between themselves, and, in any of these situations, always in relation to a minimum expected value, representing an individualized tire condition, indicative, for example and without limiting the invention, that either the tire is fit to continue or that the tire needs to be retreaded or that the tire needs to be immediately discarded and / or, further, that the tire needs to have its internal pressure adjusted and / or aligned.Furthermore, a comparison can be made between the measurements of the individual depths of the different grooves or the average groove depths of a tire of interest in relation to the measurements of the individual depths of the different grooves or the average groove depths of one or more of the other tires of interest, representing a relative tire condition, indicative, for example and without limiting the invention, that the tire(s) need to be changed position by means of a rotation.

[0035] In addition to measuring the depth of one or more grooves, the tire conditions may also include, without limiting the scope of the invention, alignment (geometry), internal pressure (calibration or calibration) and / or other parameters indicative of the tire conditions in relation to a normal or preferred condition or minimum ideal expected condition, in particular, a condition that guarantees the safe operation of the vehicle and the longevity of the tires of interest.

[0036] In the context of the present invention, the "depth of a groove" of a tire is the linear measurement of the distance between the surface of the tread and the bottom of the groove, this measurement being taken radially.

[0037] The tire conditions of the invention also comprise the general apparent state of conservation of the tire as a whole, visually inspected by the operator or driver during the steps of the method for determining the tire condition, including, for example and without limitation, the existence of cracks in the bead, punctures, blisters, nails or other objects embedded in the tire, the physical and structural integrity of the sidewalls and tread, etc., where these conditions mentioned here represent, individually or together, an additional individualized tire condition.

[0038] In the context of the present invention, the expression "determining a tire condition" refers to determining the condition of the tire by estimating and / or indirectly measuring the depth of one or more grooves of the tire(s) of interest, by submitting an image of the tread of the tire(s) of interest to a machine learning model, represented by a convolutional neural network trained with a set of images of tire treads with grooves in different states of conservation and whose respective depths were previously known. By associating the depth of each groove with its position in the tire tread (left outer groove, right outer groove, central groove, etc.), the method can be used to determine the condition of the tire.) it is possible, in addition to identifying possible depths or, preferably, identifying possible average groove depths below a minimum expected value, to obtain information indicative of irregular tire wear, resulting from alignment or inflation problems, by comparing the conditions of the different grooves of the same tire of interest. This information helps to accurately determine the condition of each tire of interest and set of tires of interest, so that appropriate maintenance measures (recap, rotation, inflation, etc.) or replacement can be taken.

[0039] More particularly, comparing the average groove depth of a tire of interest to a threshold value allows determining Individual tire conditions that indicate the need for immediate retreading or disposal, or that the tire is fit to continue as is. Furthermore, comparing the tread depths of the same tire allows determining individual tire conditions that indicate under- or over-inflation and / or the need for alignment correction, or that the tire is fit to continue as is. On the other hand, comparing the depths of the different treads individually, or preferably the averages for a tire of interest, with the depths of the different treads individually, or preferably the averages for another tire of interest, allows determining relative tire conditions that indicate the need for tire rotation.

[0040] Still in the context of the present invention, the term "vehicle" must be understood broadly, encompassing motor vehicles in general, extending equally to vehicle implements and accessories, such as trailers, semi-trailers, etc.

[0041] Also in the context of the present invention, the term "neural network" refers to a machine learning model that may be a convolutional neural network, in which the term "train" refers to adjusting the parameters of the machine learning model so that, from a number of reference images associated with previously known depths of tire grooves, it is capable of providing, as output data, the depth of a tire groove of interest from an image of this groove of interest as input data, enabling, as a whole, the evaluation of the grooves in relation to their position on the tread. System according to the invention

[0042] A system for determining a tire condition according to the present invention presupposes the existence of at least one tire of interest, preferably linked to at least one vehicle of interest.

[0043] The system for determining a tire condition of the invention essentially comprises at least one memory that stores information and computer-executable instructions and at least one trained convolutional neural network, wherein the trained convolutional neural network is configured to provide tire groove depth measurements unambiguously associated with the position of each groove in the tire tread, in response to receiving, as input, images of the tire treads, in addition to at least one processor or computer system or processing circuit configured to control the system of the invention, wherein the execution of the instructions performs one or more steps of the method for determining a tire condition of the invention.

[0044] In addition to at least one memory, at least one trained convolutional neural network and at least one processor, the system of the invention further comprises: - At least one processing equipment for acquisition and presentation of information / instructions; - At least one central server; - At least one database; - At least one communication and data network; - One or more sets of information; and - One or more sets of instructions.

[0045] Furthermore, the system for determining a tire condition of the invention may further comprise, eventually and additionally, other devices and / or equipment necessary for the operation of one or more elements of the system of the invention for carrying out one or more methods according to the invention.

[0046] The memory, the trained convolutional neural network, the processors, processing equipment for acquiring and presenting information / instructions, servers, databases, communication and data networks, and other devices and / or equipment eventually and additionally comprised by the system are all interconnected by one or more communication and data networks. Images and data are stored as one or more electrical signals, and these signals are processed by one or more components of the system of the invention.

[0047] In a preferred non-limiting embodiment of the invention, one or more elements of the system of the invention can perform one or more steps of the methods of the invention individually, without a connection to one or more communication and data networks, locally storing the processed information and data, for later sharing as soon as one or more connections to one or more communication and data networks are established or re-established. Memory

[0048] A memory, in the context of the present invention, is any memory or storage device, remote or local, volatile or non-volatile, transient or non-transitory (permanent), which stores information and instructions and, in particular, a computer-readable memory which stores instructions capable of executing a method according to the invention. Trained convolutional neural network

[0049] A trained convolutional neural network, in the context of the present invention, is a machine learning model configured to provide tire tread depth measurements in response to receiving, as input, at least one image of a tire tread, in which the tread depth measurements are unambiguously associated to the respective position of each groove on the tread surface of the tire of interest.

[0050] A trained convolutional neural network according to the invention may comprise interconnected groups of artificial neurons (e.g., neuron models), and may also be a computational device or be represented as a method that will be executed by a computational device.

[0051] A convolutional neural network trained according to the invention is a deep feedforward convolutional neural network trained, with an architecture such that it delivers results quickly and accurately and that it can also and especially be executed on portable processors such as, for example, in a non-limiting manner of the present invention, processors of cell phones, smartphones, tablets and the like, with high processing speed and concomitant precision, and may also comprise layers of neurons that can be configured in a receptive field arranged side by side.

[0052] A non-limiting example of a possible and suitable architecture for a trained convolutional neural network of the invention may be based on the design of a DarkNetl9 network, comprising 19 convolutional layers, particularly suitable due to its good performance in image processing, its simplified architecture, i.e., the way in which layers are connected, and the number of FLOPS (Floating Point Operations Per Second) which, as a unit of measurement capable of measuring a computer's processing capacity, also allows for calculating its computational cost. The number of convolutional layers used was optimized to achieve 95% accuracy while maintaining the lowest possible computational cost. Processor or computer system or processing circuit

[0053] A processor is, in the context of the invention, a central processing unit or CPU or a computer system or even a processing circuit that carries out the instructions of a computer program or application, processing and executing arithmetic, logical operations and data input and output, the computer program being stored in a computer-readable medium with memory for data storage, connection to one or more communication and data networks and to one or more remote databases and / or an information storage and retrieval environment, local and / or centralized and / or decentralized and / or in the cloud, and also equipped with all the usual peripherals of the state of the art, being capable of exchanging information with the electronic and physical medium, interfaces, applications, mobile equipment, other memory devices, etc.

[0054] A processor according to the invention may be, be part of, or be divided into one or more modules. The term "module," according to the invention, refers to an application-specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or group of processors), and a memory that executes one or more software programs or firmware. It also refers to a combinational logic circuit and / or other suitable components capable of providing the functionalities in question.

[0055] A processing circuit according to the invention is configured to determine a convolutional neural network trained according to the invention.

[0056] This processing circuit may therefore include a processor, such as a central processing unit (CPU), a microcontroller, a microprocessor, a programmable gate array in field (FPGA), a graphics card, or special hardware for convolutional neural networks such as the trained convolutional neural network of the invention.

[0057] Both the central server of the invention's system and the processing equipment for acquiring and presenting information / instructions of the invention each have at least one processor or computer system or processing circuit, which are preferably independent of each other and / or autonomous. Intermediate device

[0058] An intermediate device of the invention is an optional device that can be implemented externally and between the portable equipment and remote server, serving primarily to eventually perform intermediate calculations or processing such as, for example and without limiting the invention, calculating the depth of the grooves of the tire of interest and / or the average depth of the grooves of the tread of the tire of interest. Processing equipment for acquiring and presenting information / instructions

[0059] A processing equipment for acquiring and presenting information / instructions or simply equipment of the invention, is an interface between the system of the invention and the users of the system, and may include any device capable of processing and storing data and / or information and communicating via the communication and data network, and may include, without limiting the scope of the invention, physical, analog, digital, light sensors and combinations thereof, including cameras, lasers, reflectors, light barriers and the like compatible.

[0060] In a preferred embodiment of the invention, the equipment is a portable equipment, preferably a mobile phone or cell phone or smartphone, preferably comprising: (i) A processor or a processing circuit configured to determine a convolutional neural network trained according to the invention and logically connected to at least one memory of the equipment (internal or external); (ii) An interface device with a graphical touch screen for displaying and entering information and / or instructions; (iii) Communication device for communication with one or more data and communication networks; and (iv) Image capture device with a digital camera with a minimum resolution of 224 x 224 pixels.

[0061] Furthermore, the portable equipment of the invention may further comprise devices, peripherals and accessories usual for portable equipment of the nature dealt with here.

[0062] Portable equipment may therefore include any device capable of processing and storing data and / or information and, preferably, communicating via a communication and data network. It may also include personal computers, servers, code readers, telemetry, telematics, biometrics, other cell phones, smartphones, tablets, laptops, smart devices, e.g., smartwatches, and the like, even if they have digital cameras with low image resolution, since resolutions of 224 x 224 pixels, totaling 50,176 pixels, are already sufficient for the trained convolutional neural network to be able to identify a groove and estimate its depth. Each portable device may comprise one or more memories that store instructions, information, and data and can execute one or more programs or applications to perform various functions associated with the system and methods according to the invention.

[0063] Although the portable equipment of the invention is a device capable of interconnection with one or more of the components of the system of the invention, the execution of a method for determining a tire condition, according to the invention, can occur autonomously and, therefore, independently of a connection with any other element of the system, as will be explained in detail later. Central server

[0064] A central server is, in the context of the invention, a computer or computer system or computer circuit or even a set of computer-executable instructions, with one or more centralized computing systems or one or more data processing centers, which makes services and resources available or stores them in and through a communication and data network.It comprises one or more electronic processors capable of executing tasks from a computer program or set of instructions stored on a computer-readable medium, preferably a computer equipped with a processor, memory for data storage, connection to one or more communication and data networks and to one or more remote databases and / or an information storage and retrieval environment, local and / or centralized and / or decentralized and / or in the cloud, and also equipped with all the usual peripherals of the state of the art, being capable of exchanging information with the electronic and physical medium, interfaces, applications, mobile equipment, other memory devices, etc.

[0065] A server can be at least a web server that delivers or serves web pages, an application server that handles web operations, applications between users and applications or databases, a cloud server, a database server, a file server, a service server, a game server implementing games or services for a game, and a media server providing media such as streaming video or audio.

[0066] The system of the invention may include one or more processors for processing commands and information according to instructions from computer programs and one or more memories that store information in one or more data structures, which may be a centralized fleet management system or network, for example, but not limited to a server that may be run privately by a third-party entity or the same entity that is running the server.

[0067] A computer program according to the invention is a program executable on a processor of the invention and thus on a processor of the equipment of the invention, for example in application form.

[0068] The system may also be envisioned as a publicly accessible network system (e.g., a distributed decentralized computing system). Each processor comprises memory that stores information and data and can execute one or more sets of instructions to perform various functions associated with the system and methods according to the invention.

[0069] The central server of the invention is, in particular, the server that hosts and manages the database of one or more vehicles of interest, for example, but not limiting the scope of the present invention, of one or more fleets of vehicles of interest whose tires represent the tires of interest, comprising at least one remote central processor. Communication and data network

[0070] A communication and data network, in the context of the present The invention is a centralized or decentralized network that interconnects one or more active or passive components of the system according to the invention. The servers, processors, data processing center, database, and portable equipment of the invention may be connected to one or more communication and data networks, physical memories, cloud storage, and similar storage devices, the internet, one or more data and / or software clouds, computer terminals, mobile devices, telephones, barcode readers, QR codes, DataMatrix codes, and the like, credit cards, NFC or BLE devices, gas stations, and service stations—in short, to any equipment or interface necessary, directly or indirectly, for the execution of a method according to the invention.

[0071] The communication and data network may comprise any wired or wireless connection, the Internet, or any other form of communication, and may comprise any number of different communication and data networks between any server, device, resource, and system, and / or other servers, devices, resources, and systems described in this document. The communication and data network may enable communication between various computing resources or devices, servers, and systems, and may employ different types of networks, for example, but not limited to, computer networks, telecommunications networks (e.g., cellular), mobile, cable, radio, and similar wireless data networks, and any combination of these and / or other networks. Databases

[0072] A database or database according to the invention is any and all sets of data, files, information, instructions and records that form organized collections of data that are related to each other, hosted in one or more memories or storage devices of the system of the invention and which can be accessed, fed and managed by the data processing centers of the invention.

[0073] Both the servers and the processors and other equipment and devices of the invention's system may comprise one or more databases, independent and / or interconnected. Set of information

[0074] A set of information according to the invention is a set of data acquired and / or transmitted and / or stored by and in one or more of the components of the system of the invention and comprises a plurality of information relevant to the vehicles of interest, the management of a fleet of vehicles, the tires of interest, the operators or users, gas stations, service stations and / or tire shops and accredited tire services.

[0075] Non-limiting examples of such information are: a. Unique identification data; b. Supplementary identification data; c. Data on tires of interest; d. Training image data sets; e. Input image sets; f. Resulting output data sets; g. Threshold values; h. Tread depth indicators or codes; i. Tread position indicators or codes; j. Individualized tire condition indicators or codes; k. Relative tire condition indicators or codes; l. Relative axle condition indicators or codes; m. List of visual check items and their corresponding additional tire condition codes.

[0076] Unique identification data includes, but is not limited to, the vehicle's license plate number, vehicle make, model, year, and color, the name of the operator and / or user and / or driver, including, for example, their login and password for initialization. This data includes the current data at the time of the update (last update), the immediately preceding data, any historical series, events of special interest, and so on.

[0077] Supplementary identification data may include driver data (date of registration, license number and type, identity documentation, incidents, maintenance history) and any other information that may be useful for fleet management, such as vehicle location (geographic coordinates), scheduled route, origin and next destination, current vehicle mileage, average monthly mileage, average monthly wear, weather conditions at the origin, destination, and current location, etc. This data includes current data at the time of the update or the last update, immediately preceding data, any historical series, events of special interest, and so on.

[0078] The tire data of interest includes, but is not limited to, the tire identification code (DOT) of interest or equivalent code that includes the manufacturer's name, place of manufacture, size, model and date or period of manufacture of the tire of interest, its current position on the vehicle, the number of kilometers already driven, the cost per kilometer (CPK), the tire position by axle (steering or first axle, second axle, third axle, etc.), by position on the axle (left steering, second inner axle, etc.), by vehicle type (horse, trailer, first train, second train, etc.), date and mileage of last maintenance and type of maintenance (preventive, predictive, corrective), and other information that may be useful for identifying, characterizing, and classifying the tire of interest. The tire data of interest also includes, in relation to the reference identification, catalog, or manual data (brand, model, etc.), the minimum or reference values ​​expected for tread depth, internal pressure, and other parameters indicative of the tire's condition in relation to an expected normal, preferred, or ideal condition. This data may also include the current data at the time of the update (last update), immediately preceding data, any historical series, events of special interest, and so on.

[0079] The tires of interest are preferably numbered in a specific, previously and arbitrarily defined sequence. Alternatively, the tires of interest can be numbered in a sequence that depends on the desired evaluation method, which may vary according to the interests and demands of the user, operator, and, especially, the fleet manager.

[0080] In a preferred embodiment of the invention, represented especially by Figure 1 as a non-limiting example of the invention, the tires of interest (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) of the first, second and third axles (20, 30, 40), which are, respectively and preferably, the front or steering axle, the drive axle and the free axle, for example, of a vehicle of interest (100), are numbered sequentially according to their current position, starting the numbering from left to right of an axle, for example, from the first axle (20) to the other subsequent axles. Therefore, therefore, still in the non-limiting example of the invention, the first tire (1) is considered to be the front left tire of the first axle (20), the second tire (2) is considered to be the right tire of the first axle (20), the third tire (3) is considered to be the left tire of the second axle (30) or the axle immediately adjacent to the first axle (20), the fourth tire (4) is considered to be the tire immediately to the right of the third tire (3), in the case of axles with dual wheels, or the tire at the right end of the second axle (30) and so on.

[0081] According to the preferred embodiment of the invention, in the example of Figure 1 it is also possible to verify that the first tire (1) is represented with a different color from the other tires (2, 3, 4, 5, 6, 7, 8, 9, 10), as a result of one of the steps of the method for determining a tire condition of the invention, in which the system suggests to the user or operator the sequence of selection of the tires of interest (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) and that, therefore, the selection begins with the first tire (1).

[0082] It should be noted that the numbering above is for reference only and may be done differently, depending on the interests and demands of the user, operator and, in particular, the fleet manager, and also on the type and characteristics of the vehicle, and may also include, in addition to a main number, other additional numbers to reference, for example, vehicle implements and accessories, such as trailers, semi-trailers, etc.

[0083] It should also be noted that the schematic graphic representation of the set of axles and tires of a vehicle, in a mode of viewing and selecting the tires of interest as represented in Figure 1, according to a system and method of the invention, takes into account the current position of each of the tires of interest, that is, the last position recorded in the system, stored in one or more databases, in which the validation of this position is carried out in one of the steps of the method of determining a condition of tire of the invention.

[0084] The training image datasets comprise images used to train a convolutional neural network in a machine learning model. These images contain images of tire tread grooves with known groove depths, each labeled with the linear measurement in millimeters of the corresponding grooves. The images were acquired using image capture devices and captured in a variety of device tilt positions (phone, camera, etc.), lighting conditions, shadows, and tire dirt levels. In short, the aim was to reproduce as closely as possible the real-life conditions of vehicles and their tires.

[0085] The sets of input images include images of the regions of interest of the treads of the tires of interest, these images being obtained by the portable processing equipment for acquisition and presentation of information / instructions of the invention, and which will serve as input data for the trained convolutional neural network, which will apply the machine learning model to these input images.

[0086] The resulting sets of output data comprise at least one, preferably at least three, values ​​resulting from linear depth measurements taken in a radial direction assigned to at least one, preferably, and respectively, to at least three grooves of the region of interest of the tread of the first tire, directly associated with their respective positions in the tread (left outer groove, right outer groove, central groove, etc.). These resulting output data will serve to determine the individualized tire condition and relative.

[0087] The limit values ​​refer, essentially, to the minimum (or maximum, depending on the adopted reference) reference values ​​that must be respected to guarantee the safety of the vehicle's operation and the longevity of the tires of interest, including, without limiting the scope of the invention, the values ​​of the minimum tolerated depth of a groove, the values ​​related to the identification of calibration problems, that is, of the internal pressure, as well as in relation to the alignment, geometry and other useful parameters indicative of the tire's conditions in relation to a normal, preferential or ideal minimum (or maximum, depending on the adopted reference) expected condition.

[0088] ID onwards.The groove depth codes represent the condition of the groove and may, for example, without limiting the scope of the invention, be numbers, letters, alphanumeric sequences, geometric figures, colors, image patterns, the measured value itself (in millimeters, rounded or not), sounds, lights, expressions such as "irregular" or "ok", and these variants may be used individually or together and may or may not be accompanied by colors, etc.

[0089] Groove position identifiers or codes are the codes assigned to each unique groove position identifier of interest according to its position or location on the tread of interest of the tire of interest, being unambiguously linked to the depth codes, which may be at least one code, for example, for the left outer groove, or at least two codes, one for the left outer groove and one for the right outer groove, or, even, at least three codes, two for the outer grooves (left and right) and the third for the central groove. It should be noted that, if there is more than one intermediate groove between the extreme right and left grooves, the third code may represent, for example, the arithmetic mean of the intermediate grooves.The groove position codes can, for example, without limiting the scope of the invention, be configured in a similar way to the groove depth codes, but can also be used together or in combination with the depth codes, indicating, for example, a left outer groove with an irregular groove depth (equal to or less than the pre-established threshold value).

[0090] Individual tire condition identifiers or codes are codes representing the condition of the external regions of the tire, in particular, the depth of the grooves in the tread. Both the depth of each groove individually and, preferably, the average depth of the grooves of the tire of interest can be evaluated, and, in both situations, always in relation to a minimum expected value, representing an individual tire condition. In addition to the depth measurement of one or more grooves, tire conditions can also include, without limiting the scope of the invention, alignment (geometry), internal pressure (calibration or calibration) and / or other parameters indicative of the tire's condition. in relation to a normal or preferred condition or an ideal minimum expected condition, in particular, a condition that ensures the safe operation of the vehicle and the longevity of the tires of interest.

[0091] As previously described, determining a tire's condition involves determining its condition by estimating and / or indirectly measuring the depth of one or more grooves in the tire(s) of interest. By associating the depth of each groove with its position in the tire's tread (left outer groove, right outer groove, center groove, etc.), it is possible, in addition to identifying potential depths or, preferably, identifying potential average groove depths below a minimum expected value, to obtain information indicative of irregular tire wear, resulting from alignment or inflation problems, by comparing the conditions of the different grooves of the same tire of interest.This information helps to accurately determine the condition of each tire of interest and set of tires of interest, so that appropriate maintenance measures (recap, rotation, calibration, etc.) or replacement can be taken.

[0092] More specifically, comparing the average tread depth of a tire of interest to a threshold value allows determining individual tire conditions that indicate the need for immediate retreading or disposal, or that the tire is fit to continue as is. Furthermore, comparing the tread depths of the same tire of interest with each other allows determining individual tire conditions that indicate under- or over-inflation and / or the need for alignment correction, or that the tire is fit to continue as is. On the other hand, comparing the depths of different treads, individually or, preferably, the averages of a tire of interest with the depth of the different grooves, individually or, preferably, the averages of another tire of interest, allows to determine relative tire conditions indicative of the need to promote tire rotation.

[0093] Differences between the tread depth codes of the same tire may indicate irregular tire wear, and it is up to the user, operator and / or, in particular, the fleet manager, to determine the threshold values ​​that will trigger a specific alert for maintenance, replacement or tire rotation.

[0094] In addition to the difference between the depths of at least two grooves on the same tire being considered indicative of irregularities in tire wear, the invention also considers the indication of this difference in terms of the position of one groove relative to the other—that is, the absolute value of the depth and the difference between the depths of all grooves, for example, indicating the need for alignment. Furthermore, differences between the depths of both outer grooves relative to the depth of the center groove(s) may result from uneven tire inflation (pressure), indicating the need to increase or decrease tire pressure. Center grooves with a shallower depth (by a configurable delta) than the depth of both outer grooves indicate, for example, overinflation and, therefore, increased wear in the center strip of the tread.On the other hand, a central groove with a greater depth (in a configurable delta) than the depth of both outer grooves indicates, for example, a lack of pressure, and thus greater wear on the tread edges. Individual tire condition codes may, for example, without limiting the scope of the invention, be numbers, letters, alphanumeric sequences, geometric figures, colors, or patterns. of image, the value itself measured in millimeters, rounded or not, sounds, lights, expressions such as "perform geometry" and / or "perform balancing" and / or "perform rotation" and / or "attention" and so on, these variants being able to be used individually or together and even accompanied or not by colors etc.

[0095] Relative tire condition codes or identifiers are codes representing the condition of the tire of interest, used and / or in use, in relation to another tire of interest on the same vehicle of interest, for example, by comparing the individual tread depth measurements or the average tread depth of a tire of interest with the individual tread depth measurements or the average tread depth of one or more other tires of interest. Differences between tire tread depth identifiers may indicate uneven tire wear, and it is up to the user, operator, and / or, especially, the fleet manager, to determine the threshold values ​​that will trigger a specific tire maintenance or rotation alert.The relative tire condition codes may, for example, without limiting the scope of the invention, be numbers, letters, alphanumeric sequences, geometric figures, colors, image patterns, the value itself measured in millimeters, rounded or not, sounds, lights, expressions such as "perform rotation" or "attention" and so on, and these variants may be used individually or together and may or may not be accompanied by colors, etc.

[0096] The list of visual check items and their respective additional tire condition codes comprises the additional items that represent the overall apparent condition of the tire as a whole, visually inspected by the operator or driver during the steps of the tire condition method. determination of the condition of the tire, including, for example and without limitation, the existence of cracks in the bead, punctures, bubbles, nails or other objects embedded in the tire, the physical and structural integrity of the sidewalls and tread, etc., in which these conditions mentioned here represent, individually or together, an additional individualized tire condition. Instruction set

[0097] An instruction set, according to the invention, comprises one or more instructions, sequential and / or non-sequential, unique and / or repeated, the instruction set being acquired and / or transmitted and / or stored by and in one or more of the components of the system of the invention, being, essentially, instructions executable on machines or processors.

[0098] The instructions may be executed and / or stored by and on a server, in a processing center, in processing equipment for acquiring and presenting information / instructions of the invention, and may also be stored in one or more databases or other computer-readable storage medium, volatile or non-volatile.

[0099] The set of instructions may also be the set of steps and possible substeps of a method of the invention executed by the system of the invention or in one or more components of the system of the invention. Computer-readable medium

[0100] A computer-readable medium according to the invention is a medium comprising instructions that, when executed on a system or on equipment of the invention, performs one or more methods according to the invention. Other devices and / or equipment necessary for operation

[0101] The system of the invention may further comprise, if necessary, any and all additional devices and / or equipment that are necessary for the operation of the system of the invention to perform one or more methods of the invention. Operation of the system for determining a tire condition

[0102] The system according to the invention, fed with information and instructions and with access to information, instructions and updates of these information and instructions, comprises memories that store information and instructions, a trained convolutional neural network, a remote and / or local central server, remote and / or local databases, computer-readable media and at least one portable processing equipment, which execute a method of the invention for determining a tire condition, resulting from the interaction of one or more of its components that perform one or more steps and possible subtasks that may or may not be sequential, possibly repeating themselves, in isolation or in a grouped manner, if necessary and depending on the conditions determined for the method.

[0103] The main operating steps of the system comprise, in a non-limiting manner, at least one set of preparation steps for determining the tire condition, at least one set of steps for determining the tire condition itself and at least one set of completion steps with suggested actions by means of procedural instructions based on the determined tire condition.

[0104] The preparation steps include initialization, in the application executed by the equipment processor, through the equipment interface device, informing at least one unique identification data, including vehicle and / or driver data, in which this information is validated by the system, which returns, from the information in a database, with the data of the tire of interest and suggesting to the operator a first tire of interest. The operator selects (or indicates) the first tire of interest to validate, in particular, the expected position for the respective tire and, optionally, receives additional information about the tire, such as brand, model, life (if still original or if retreaded or remolded, etc.) and instructions for capturing the image of the tread.

[0105] The steps for determining the tire condition itself include obtaining, using the image capture device, e.g., a camera, of the portable equipment, a first image of the region of interest. This image is stored in the equipment's memory and, preferably, converted in the equipment's processor into a pre-processed image, to which the machine learning model is applied as input data in the equipment's processor to obtain, in the equipment's processor, a resulting set of output data comprising one or more estimated values ​​of the depth of one or more of the tread grooves, these depth measurements being associated, in an attributive manner, with their respective positions on the tread. These depth measurements are used for various comparisons in the equipment's processor. These comparisons will be detailed in detail below.This information can be presented to the user on the equipment's graphical interface device by displaying the tread depth codes as numbers, letters, or alphanumeric codes superimposed on the image, identifying each tread groove with its respective depth code and / or position code. The system then presents the user, through the equipment's interface device, with a list of visual check items, inviting the user to select at least one additional tire condition code. This condition code information. tire depth values ​​associated with tire tread position codes are processed in the equipment processor to obtain at least one condition code for the first tire and suggest one or more procedural instructions according to the condition of the first tire.

[0106] The finalization steps with suggested procedure instructions based on the determined tire condition include informing the fleet manager or person responsible for maintenance, through the remote central processor or the portable equipment processor, one or more tire condition codes for the first tire of interest and receiving from the fleet manager or person responsible for maintenance, through the remote central processor, at least one procedure instruction equivalent to: (a) follow (see prediction, maintenance interval, estimated lifespan); or (b) discard; or (c) recap; and / or (d) calibrate; and / or (e) check alignment; and / or (f) change positions / rotate.

[0107] Ideally, the steps can be repeated in full for the remaining tires of interest or partially, with the decision-making process being left until the end, accumulating decisions for a group of tires, axles, etc., in addition to considering the potential for autonomous execution of the method using portable equipment, without connection to a central server, database, or any communication and data network. Finally, the process is completed.

[0108] It should be noted that the steps, tasks and subtasks are executed using directly or indirectly one or more of the active components or liabilities of the system of the invention, based on one or more of the information of the system of the invention, according to a method of the invention and, always, according to one or more of the instructions of the system of the invention.

[0109] That said, it is clear that, by executing the steps described above, the system of the invention determines the state of conservation of a tire of interest based on the estimate or, in other words, indirect measurement of the depth of the grooves of the tire of interest, by submitting an image of the tread of the tire of interest to a machine learning model represented by a convolutional neural network trained with a set of images of treads with grooves in different states of conservation and whose respective depths were previously known.

[0110] By associating the depth of each groove with its position on the tire tread (left outer groove, central groove, right outer groove, etc.), for the same tire of interest, the system is capable of, in addition to identifying any depths lower than a minimum expected value, for retread or disposal instructions, gathering information indicative of irregular tire wear, resulting from alignment or calibration problems.

[0111] Furthermore, by comparing the conditions of different tires of interest, the system can suggest tire rotation. It's important to note that, due to safety concerns, a previously retreaded tire should not be suggested to replace a tire on the steer axle of the vehicle of interest.

[0112] With this information, the system helps determine the status of each tire and set of tires, and suggests to the operator and, in particular, to a fleet manager, the appropriate and necessary maintenance measures. (recap, rotation, calibration, alignment, etc.) or replacement of one or more tires using procedural instructions. Method for training a neural network to estimate the depth of a tire groove

[0113] The method for training a neural network to estimate the depth of a tire groove, according to the invention, essentially comprises generating an input image dataset and training the neural network to map the input image dataset relative to the reference image.

[0114] Training serves to configure the neural network to provide tire tread depth measurements unambiguously associated with the position of each tread in the tire, in response to receiving, as input, images of tire treads.

[0115] The preferred neural network, without limiting the scope of the invention, is a convolutional neural network, but may be any other suitable similar architecture, such as, for example, (alternatively, a Vision Transformer (ViT) model), provided that this model is trained for this purpose, such as the convolutional neural network mentioned herein. Furthermore, the trained neural network of the invention may comprise the combination of two or more architectures.

[0116] More broadly, the method comprises training one or more integrated neural networks with information that allows the unambiguous association of features of an image (photo) of a tire to one or more conditions of this tire, especially the unambiguous association of an image (photo) of a region of interest of this tire to one or more conditions of this tire, individually in relation to the tire of interest itself or in relation to one or more other tires of interest of the same vehicle.

[0117] For training the invention's convolutional neural network, it was A database was created with approximately eleven thousand images photographed on two thousand tire samples. The images were captured of tire tread grooves with known groove depths, with each training image labeled with the linear measurement, in millimeters, of the corresponding grooves. The images were obtained using image capture devices (phone, camera, etc.) and captured in a wide range of device tilt positions, lighting conditions, shadows, and tire dirt levels. In short, the aim was to reproduce as closely as possible the real-life conditions of vehicles and their tires. This contrasts with images obtained, for example, in a laboratory, which could result in inadequate training, unrelated to, or distant from, the application's reality, and the resulting network would have difficulty identifying the characteristics and even identifying the grooves of interest.

[0118] With the collected data, a deep feedforward convolutional neural network was created, which can include layers of neurons that can be configured in a receptive field arranged side by side, in which the input data are tire images and the output data are the depths of the tire grooves.

[0119] The input image is fed with a size of 224 rows by 224 columns and only one color channel, which means that the image is fed in grayscale or grayscale instead of RGB (or BGR), which should not, in fact, limit the scope of the present invention. This input image undergoes preprocessing, normalizing the values ​​that would originally range from 0 to 255 (corresponding to the pixel colors) to range from 0 to 1. This normalization is done to reduce the magnitude of the values ​​calculated by the neurons in the chosen convolutional neural network architecture and, thus, avoid overflow.

[0120] The input image dataset is provided as input to the convolutional neural network which, during its training, adjusts its convolution kernels to extract information from the image relevant to the grooves of the tire in question.

[0121] The developed machine training model outputs a depth value generated by each respective neuron. Ideally, the model outputs at least three depth values ​​generated by at least three neurons, each representing the depth of the tire's respective grooves. The reading is preferably made from one side to the other (from one end of the tread to the other), for example, from left to right, with the first value corresponding to the first groove on the left, the second value corresponding to the middle groove, and the third value corresponding to the groove on the right. Therefore, the outer values ​​are reserved for the extreme left and right grooves, respectively, and the middle value is reserved for the middle groove, located between the extreme grooves.In the case of tires that have more than one intermediate groove, the intermediate value will be calculated as the average of the groove depths in the middle of the tire tread.

[0122] For regression, MSE (mean squared error) was empirically chosen as the error metric. Furthermore, the ADAM optimizer is used to drive training. For a given error in a given training iteration, the optimizer algorithm calculates new neural network weight values ​​that tend to reduce the error in the next iteration.

[0123] In this way, it was possible to extract unique characteristics from the captured images, in which the unique characteristics are directly linked to a specific groove depth, that is, a depth reference directly linked to the characteristics of the image groove, with each of the references being stored within the mathematical model of the convolutional neural network, as a result of training.

[0124] The unique features extracted from the images during convolutional neural network training comprise a reference data package. This reference, or template, is a mapping of image subdivisions and / or each of its pixels, quantifying, for example, the intensity of colors and / or shadows in each of these subdivisions and the position of this subdivision within the image. This makes it possible to identify a groove in the tire tread surface, quantify its depth, and determine its position within the tread. These features are extracted by the convolutional layers and determined through training iterations. These iterations define which features should be extracted to feed the subsequent layer, which extracts new features based on the previous layer.

[0125] It should be noted that it is possible to record more than one reference or template for one or more grooves or depths in the convolutional neural network database during its training.

[0126] The comparison between the obtained image and the recorded images (stored templates) is performed by the system of the invention, in which, each time a tire image is read / captured, a comparison is made with the template(s) stored in the integrated networks (levels) of the trained convolutional neural network. The analysis of each of the obtained images is performed iteratively, and with each analysis, the stored template is compared to a newly generated template.

[0127] As neural networks, by definition, usually provide the result of the comparison between templates in the form of a probability of a positive result for the proximity / similarity correlation and a negative result for the disparity / difference correlation, and knowing that only the probability of a positive result provided does not guarantee the accuracy of the comparison result, since mathematical models are subject to margins of error intrinsic to the model itself, a cross-reference is used to define the limits for the positive and negative results that should be considered for the comparison, and thus increase the accuracy of the result.

[0128] Tests have shown that the invention's convolutional neural network is capable of providing a response with an average error between ± 1 mm and ± 2.0 mm, within a range of groove depths that varies from any value greater than 0 mm, with the presence of a groove, to values ​​equal to or greater than 25 mm, remembering that a value equal to 0 mm represents either a part of the band that does not have grooves or a grooved region of an extremely worn tire, with no grooves, popularly called a "bald tire".

[0129] The method for training a neural network to estimate the depth of a tire groove comprises the following method steps: (i) Generate at least one training image dataset; (ii) Configure a neural network to receive, as input, the training image dataset, in which each training image corresponds to the image of a tire tread; (iii) Receive training images of tire treads with grooves of previously known depth and position; (iv) Configure the neural network to locate one or more grooves in an image of a tire tread; (v) Configure the neural network to identify each of the grooves individually, for example, starting from one of the edges towards the other edge of the tread, and more particularly, from the left edge towards the right edge of the tread, identifying a left groove, a central groove and a right groove; (vi) Configure the neural network to assign each of the training grooves its known depth, assigning it a unique depth identifier, unambiguously associating the value of the known depth with its known position, e.g., left, center, right, on the tread, also assigning it a unique position identifier; (vii) Configure the neural network to generate, in response to an input image of a tire tread, at least one, preferably at least three, depth values ​​uniquely associated with at least one, preferably at least three corresponding groove positions in the tread, e.g., from left to right, as previously mentioned; and (viii) Produce a neural network by repeatedly training the trained convolutional neural network with each of the training data sets.

[0130] Once the convolutional neural network has been trained and the system, as a whole, has been fed with the information and instructions described herein, the processors will be ready to manage the system and its trained convolutional neural network, which includes, in particular, the management of the entire system according to the invention for executing the method for determining a tire condition according to the invention. Method for determining a tire condition

[0131] The method for determining a tire condition according to the invention is a computer-implemented method that preferably comprises the following method steps: A. Obtain, by means of the image capture device of a portable equipment, a first image of a region of interest of the tread of a first tire of interest and store the first image obtained in the transient memory and / or in the non-transient memory of the portable equipment; B. Apply to the image, in the processor of the portable equipment or of an intermediate device or of a remote server, a machine learning model, preferably, but without limiting the invention, in the form of a neural network, preferably a trained convolutional neural network; C. Obtain, in the processor, as a result of the application of the machine learning model, a resulting set of output data comprising at least one, preferably at least three values ​​resulting from depth measurements, in the radial direction of the tire of interest, assigned, in the form of a groove depth code, to the at least one, preferably, and respectively, to the at least three grooves of the region of interest of the tread of the first tire of interest, starting from one of the edges towards the other edge of the tread, and, more particularly, from the left edge to the right edge, in the form of a groove position code; and calculate the average of the groove depth values; D. Assign at least one of the groove depth codes unambiguously to at least one of the corresponding groove position codes; and compare, in the processor, the average of the groove depth values ​​with a pre-established threshold value, the result of the comparison being able to be less than, equal to, or greater than the threshold value; E. Compare the groove depth codes of the same tire of interest with each other, and identify any differences between the depths in association with the position codes of the corresponding grooves; F. Process, in the processor, the results of the comparisons made in steps D and E and indicate at least one individual tire condition code, based on the comparisons made in steps D and E; G. Repeat steps A through F for a second tire of interest or up to a last or nth tire of interest; and / or H. Compare the tread depth codes of the first tire of interest to those of a second tire of interest or even a last or nth tire of interest and process, in the processor the results of the comparisons made and indicate at least one relative tire condition code, based on the comparisons made; I. Determine and report, based on at least one individual tire condition code and at least one relative tire condition code, at least one suggested procedural instruction equivalent to: a. Follow; or b. Discard; or c. Recap; and / or d. Inflate; and / or e. Check alignment; and / or f. Change position / rotation; and J. Close the process.

[0132] Below, the steps mentioned above will be analyzed and described individually, for the purposes of non-limiting examples of preferred embodiments of the invention. Step A:

[0133] Step A comprises obtaining, by means of the image capture device of the portable equipment, a first image of a region of interest of the tread of the first tire of interest and storing this first image obtained in the transient memory of the equipment and / or in the non-transient memory of the portable equipment.

[0134] Step A may or may not be preceded or complemented by tasks such as, for example and in a non-limiting manner of the invention, pre-processing, in the equipment processor, of the image obtained, to obtain a pre-processed image suitable for submission to the neural network. trained convolutional.

[0135] Furthermore, step A may be preceded by one or more steps comprising receiving from a local or remote database one or more data from the tire of interest.

[0136] Thus, step A could perform, in a non-limiting manner, the initialization of the system of the invention and the preparation of the procedure for determining a tire condition, which may or may not be preceded or complemented by tasks such as, for example and in a non-limiting manner of the invention, informing or "entering data" into an application executed by the processor of the portable equipment or the intermediate device or the remote server, through the interface device of the portable equipment, at least one unambiguous identification data.

[0137] At this stage, a connection is established between the equipment and a remote central processor, sending, through the equipment's communication device, at least one unique identification data to a remote central processor, for verification and validation of the at least one unique identification data and access to the corresponding data stored in the remote central processor's database. This stored data may include other unique identification data, in addition to supplementary identification data and, in particular, data on the tires of interest related to the identification data, both unique and supplementary.

[0138] In a preferred embodiment of the invention, the system database is stored in the memory of the portable equipment itself and, thus, although the portable equipment of the invention is a device capable of interconnection with one or more of the components of the system of the invention, the execution of the method for determining a tire condition occurs in autonomously, independent of a connection with any other element of the system.

[0139] The unique identification data includes, but is not limited to, the vehicle identification plate data of the vehicle of interest, make, model, year and color of the vehicle, the name of the operator and / or user and / or driver, including their login and password for initialization.

[0140] This same step A may or may not be preceded or complemented by tasks such as, for example and in a non-limiting manner of the invention, indicating to the operator, through the equipment interface device, among one or more tires of interest, a first tire of interest, physically attached or not to a vehicle and / or an implement of interest.

[0141] In a preferred embodiment of the invention, this step comprises the schematic graphic representation of the axles and their respective tires of interest of the vehicle(s) of interest, as represented, for example, in Figure 1, in which the tire of interest at the time, in the case of the example the first tire of interest, appears highlighted, e.g., with a different color, flashing, in high relief, etc.

[0142] At this stage, the system can also return to the user with information relevant to the vehicle, driver and tire of interest in question, comprising receiving from the remote or local central processor information from the central or local database relating to at least one unique identification data item and, eventually, one or more additional identification data items, but, in particular, comprising at least one, preferably more than one, of the data items of the tire of interest.

[0143] Step A may or may not be preceded or complemented by tasks such as, for example and in a non-limiting manner of the invention, the selection of a first tire of interest to an operator, who highlights it by pressing a finger on this tire on the interface device / screen of the portable equipment.

[0144] Finally, still in relation to step A, this may or may not be preceded or complemented, for example and in a non-limiting manner of the invention, by a task in which the system will return, through the system's database, local and / or remote, one or more data on the tire of interest, including, for example, but not limited to, the number of kilometers already driven, the current cost per kilometer "CPK", date and mileage of the last maintenance and type of maintenance, i.e., preventive, predictive, corrective, number of tire lives, i.e., if original, if already retreaded or remolded, etc. Step B:

[0145] In step B, a machine learning model in the form of a convolutional neural network is applied to the image, on the processor of the portable equipment or an intermediate device or a remote server, trained with a method for training a convolutional neural network according to the invention, that is, trained with a first set of data of training images of tires associated with values ​​of linear measurements, taken in a radial direction, of tire grooves, assigning each groove a position code. Step C:

[0146] Step C comprises obtaining, in the processor, as a result of the application of the machine learning model, a resulting set of output data comprising at least one, preferably at least three values ​​resulting from depth measurements, in the radial direction of the tire of interest, assigned, in the form of a groove depth code, to at least one, preferably, and respectively, to at least three grooves of the region of interest of the tread of the first tire of interest, starting from one of the edges towards the other edge of the tread, and, more particularly, from the left edge to the right edge, in the form of a groove position code; and calculate the average of the groove depth values.

[0147] Step C may further comprise, for example and without limiting the invention, informing, in the graphical interface device of the portable equipment, at least one, preferably at least three, depth and / or position codes, graphically superimposed on the image obtained, identifying each tread groove with its respective depth code and / or position code. Thus, each tread groove of interest in the field of the image obtained by the operator and projected on the application screen will appear with a code equivalent to its groove depth, as can be seen, for example, in Figure 2, in which each groove has an indication of its depth. Step D:

[0148] Step D comprises assigning at least one of the groove depth codes unambiguously to at least one of the corresponding groove position codes; and comparing, in the processor, the average of the groove depth values ​​with a pre-established threshold value, the result of the comparison being: a. Less than the threshold value; or b. Equal to the threshold value; or c. Greater than the threshold value.

[0149] If the result is lower than the limiting value, it is an indication that the tire needs to be retreaded or discarded immediately; if it is equal to the limiting value, it is an indication that the tire needs to be retreaded (provided that this tire is still suitable for retreading based on the number of retreads previously determined by the fleet manager); and if it is greater than the limit value, it is an indication that the tire is ready to continue.

[0150] In this way, it is possible to determine at least an initial individual tire condition, with regard to whether the tire is suitable, whether it needs to be retreaded or whether it needs to be discarded immediately. Step E:

[0151] Step E involves comparing the groove depth codes of the same tire of interest with each other, and identifying any differences between the depths in association with the position codes of the corresponding grooves.

[0152] In a preferred embodiment of the invention, comparisons between the grooves are made two by two, for example, but without limiting the invention, comparing (i) the left groove with the central groove; and (ii) the right groove with the central groove; and (iii) the left groove with the right groove.

[0153] If both side grooves are deeper (in a configurable delta) than the central groove, this indicates that the tire is over-inflated; if both side grooves are deeper (in a configurable delta) than the central groove, this indicates that the tire is under-inflated.

[0154] If the left and central grooves have a greater depth (in a configurable delta) than the right groove, or if the right and central grooves have a greater depth (in a configurable delta) than the left groove, this is an indication that the tire has an alignment problem.

[0155] In this way, it is possible to determine at least a second individual tire condition with regard to the need for adjustment. calibration and / or alignment.

[0156] It should be noted that the terms "greater" and "lesser" refer to differences between the depths compared, either more or less than a predetermined differential, for example, by the fleet manager, in the form of a configurable delta. Step F:

[0157] Step F comprises processing, in the processor, the results of the comparisons performed in steps D and E and indicating at least one individual tire condition code, based on the comparisons performed in step D - at least one individual tire condition relating to a suitable tire, need for retreading or immediate disposal - and in step E - at least one individual tire condition relating to the need for adjustment of the pressure and / or alignment. Step G:

[0158] Step G involves repeating steps A through F for a second tire of interest or until a final or nth tire of interest, in which case this step may be skipped, continuing the method execution for one or more subsequent steps and, eventually, until the last step of the method. This configuration can be determined according to the needs and preferences, for example, of the fleet manager. Step H:

[0159] Step H involves comparing the groove depth codes of the first tire of interest to those of a second tire of interest or even a last or nth tire of interest and processing, in the processor, the results of the comparisons made and indicating at least one relative tire condition code, based on the comparisons made.

[0160] When the comparison indicates that the difference between the average of the groove depths of a first tire of interest in relation to one or more other tires of interest is greater or less than a predetermined tolerance or differential (in the form of a configurable delta), there is an indication of the need to rotate or exchange the positions of one or more tires of interest.

[0161] This allows for at least a first relative tire condition to be determined, with regard to the need to replace or rotate one or more of the tires of interest. It should be noted that the indicative conditions for rotating one or more of the tires of interest must be established and predetermined according to the fleet manager's needs and preferences. For safety reasons, it is essential not to use retreaded tires on steer axles.

[0162] It should be noted that the terms "greater" and "lesser" here also refer to differences between the depths compared, either more or less than a tolerance or a predetermined differential, for example, by the fleet manager, in the form of a configurable delta. Stage I:

[0163] Step I comprises determining and reporting, based on at least one individual tire condition code and at least one relative tire condition code, at least one suggested procedural instruction equivalent to: a. Follow; or b. Discard; or c. Recap; and / or d. Inflate; and / or e. Check alignment; and / or f. Change position / rotation.

[0164] It should be noted that step I can be carried out either for each tire of interest or from the second tire of interest or even at the end of the verification of all tires of interest or even in a mixed manner, according to the needs and preferences, for example, of the fleet manager.

[0165] Step I may further comprise, for example and in a non-limiting manner of the invention, sending, by means of the communication device of the portable equipment, in order to inform the fleet manager (or other user involved in the process), at least one condition code of the first tire and respective tire depth condition codes of interest and tire groove position codes to the remote and / or local central processor, in addition to suggesting one or more procedural instructions from a group comprising, without limiting the scope of the invention, continuing with the tire as is, also checking the maintenance interval, estimated remaining life before retreading / remolding, etc., and / or calibrating and / or checking alignment and / or retreading / remolding the tire and / or changing its position, i.e., rotating it, indicating the new preferred position of the tire, or replacing the tire with a new one or even with a retreaded one.It's important to note that any retread recommendation must be preceded by an assessment to ensure the tire in question has not exceeded the maximum number of permitted or recommended retreads. This maximum number of retreads depends essentially on the tire brand and model, but especially on the axle on which the tire is located. Steer, drive, and free axles place different stresses on tires and, therefore, impose different limits on the maximum number of permitted or recommended retreads.

[0166] In a preferred embodiment of the invention, the system database is stored in the memory of the portable equipment itself and, thus, the method execution occurs autonomously, independent of a connection with any other element of the system. Step J:

[0167] Step J represents the completion of the method, which may be automatic or triggered by one or more system elements, or triggered by one or more operators, users, fleet managers, or a combination of these. It should be noted that other steps may be incorporated before or after Step J. Thus, Step J can also be skipped, continuing the method execution to one or more additional steps and, eventually, to a new, final step of the method. Considerations on the steps of the method for determining a tire condition

[0168] It should be noted that the aforementioned method may also, optionally and without limiting the invention, comprise one or more visual verification steps, presenting to the operator or user, through the interface device of the portable equipment, a list of visual verification items and receiving, through the operator or user, from the interface device of the equipment, at least one additional tire condition code, for visual verification.

[0169] Visual inspection items include, for example and without limitation, the existence of cracks in the bead, punctures, bubbles, nails or other objects embedded in the tire, the physical and structural integrity of the sidewalls and tread, etc., in which these conditions mentioned here represent, individually or together, an additional individualized tire condition.

[0170] The steps of the method of determining a tire condition of the invention may be performed in the suggested sequences or be repeated individually or in groups, if necessary and depending on the condition to be investigated.

[0171] All steps and possible substeps of the method of the invention are performed using directly or indirectly one or more of the active or passive components of the system of the invention, using one or more of the information of the system of the invention and always in accordance with one or more of the instructions of the system of the invention.

[0172] Interaction between the system components during execution of the method of the invention occurs through communication between them via one or more communication and data networks. Images and data are stored as one or more electrical signals, and these signals are processed by one or more components of the system of the invention.

[0173] In the context of the invention, the term "inform" means the transmission / communication or execution of information or instructions by one or more components of the system to one or more components of the system, which may be related to one or more of the activities of one or more of the components of the system and may also include instruction information.

[0174] In the context of the invention, the term "send" means the transmission / communication of an information or instruction or selection by one or more system components to one or more system components.

[0175] In the context of the invention, the term "receive" means the receipt of information or an instruction, which may be a relationship between one or more components of the system or a relationship between one or more components of the system, a return or response to a sending and similar activities.

[0176] In the context of the invention, the term "select" means the choice, as a decision-making process, or selection of data options, information and instructions made available and / or stored on the server, processor, database of data and interface devices, carried out by one or more system components and / or by one or more operators or users, and may even include the fleet manager.

[0177] In the context of the invention, the term "obtain" essentially means to fetch or capture an image or information or instruction that may be internal or external to the system, through one or more elements of the system.

[0178] In the context of the invention, the term "storing" means storing data in storage devices or memories so that they can be retrieved later.

[0179] In the context of the invention, the terms "process", "apply" and "compare" refer to the processing or subjection to processing of one or more information and / or instructions and / or images and / or data by one or more components of the system, including the determination "determine" and the decision "decide" of sequences of steps of the method of the invention and operation of the system, executing one or more applications or computer programs, and the processing may also involve interaction with one or more operators or users or fleet managers.

[0180] In the context of the invention, the term "consult" means the consultation of data, images, information and instructions made available / stored on the server, processor, database and equipment and its devices, carried out by one or more components of the system and / or by one or more operators or users or fleet managers.

[0181] In the context of the invention, the term "trigger" means the actuation and / or instruction of the server, processor, database and equipment and their presentation and / or acquisition devices, carried out by one or more components of the system, e.g., the triggering of a task of storing information in the database by the processing center or the triggering of a presentation device by an acquisition device, passing through the server and then storing the respective information in the database, and so on, and / or by one or more operators, users, or fleet managers. The term "trigger," as an instruction, can be a command to either start or end one or more operations of one or more of the system components.

[0182] In the context of the invention, the term "present" means the presentation of information or instruction or image or data or set of data by one or more components of the system, in particular, by the graphical interface devices of one or more components of the system, which may be related to one or more of the activities or decisions of one or more of the components of the system and may also include information in instructions, lists, images, etc. Conclusion

[0183] It should be noted that the system and methods according to the invention, despite being ingenious, are intuitive and easy to understand for the user, vehicle driver, fleet manager and maintenance professional, dispensing with any and all additional tools for measuring, controlling and recording one or more tires of one or more vehicles simultaneously and, very importantly, it is capable of functioning without requiring the connection of the user's equipment to any other elements of the system or devices or communication and data networks.

[0184] The system and methods of the invention inform the user of the condition of a vehicle's tires with a very high level of detail, as it individually evaluates each of the grooves of each of the vehicle's tires. By directly associating the depth of a groove with its relative position on the tread, by associating the condition of a tire in in relation to another tire on the same axle and, finally, by associating the condition of the tires on the same axle with the condition of the tires on another axle of the same vehicle, the system and methods of the invention reach an unprecedented level of determining the conditions of the running elements of a vehicle.

[0185] It should also be noted that both methods of the invention, in particular the method for determining a tire condition, can be performed independently of the system of the invention as a whole, for example, without the use of a central processor and even without a communication and data network.

[0186] The present invention therefore provides a system, methods, equipment and computer-readable memory or computer-readable memory that improve in a new and inventive way the solutions to problems of the state of the art of supervision, maintenance and logistics of tires in a vehicle or in fleets of several vehicles. Final considerations

[0187] It will be readily understood by those skilled in the art that modifications may be made to the present invention without thereby departing from the concepts set forth in the description above. Such modifications should be considered as falling within the scope of the present invention. Consequently, the particular embodiments described in detail above are merely illustrative and exemplary and are not limitative of the scope of the present invention, which should be given the full scope of the appended claims and any and all equivalents thereof.

Claims

CLAIMS 1. Method for determining a tire condition, implemented on a computer, characterized by the fact that it comprises the following method steps: A. Obtain, by means of the image capture device of a portable equipment, a first image of a region of interest of the tread of a first tire of interest and store the first image obtained in the transient memory and / or in the non-transient memory of the portable equipment; B. Apply a machine learning model to the image, on the processor of the portable equipment or an intermediate device or a remote server; C. Obtain, in the processor, as a result of the application of the machine learning model, a resulting set of output data comprising at least one, preferably at least three values ​​resulting from depth measurements, in the radial direction of the tire of interest, assigned, in the form of a groove depth code, to at least one, preferably, and respectively, to at least three grooves of the region of interest of the tread of the first tire of interest, starting from one of the edges towards the other edge of the tread, and, more particularly, from the left edge to the right edge, in the form of a groove position code; and calculate the average of the groove depth values; D. Assign at least one of the groove depth codes unambiguously to at least one of the corresponding groove position codes; and compare, in the processor, the average of the groove depth values ​​with a pre-established threshold value; E. Compare the groove depth codes of the same tire of interest with each other, and identify any differences between the depths in association with the position codes of the corresponding grooves; F. Process, in the processor, the results of the comparisons made in steps D and E and indicate at least one individual tire condition code, based on the comparisons made in steps D and E; G. Repeat steps A through F for a second tire of interest or up to a last or nth tire of interest; and / or H. Compare the groove depth codes of the first tire of interest to those of a second tire of interest or even a last or nth tire of interest and process, in the processor, the results of the comparisons made and indicate at least one relative tire condition code, based on the comparisons made; I. Determine and report, based on at least one individual tire condition code and at least one relative tire condition code, at least one suggested procedural instruction equivalent to: a. Follow; or b. Discard; or c. Recap; and / or d. Calibrate; and / or e. Check alignment; and / or f. Change position / rotation.

2. Method, according to claim 1, characterized by the fact that in step D at least a first individualized tire condition is determined; in which the result of the comparison being lower than the threshold value indicates that the tire needs to be retreaded or discarded immediately; in which the result of the comparison being equal to the threshold value indicates that the tire needs to be retreaded; and in which the result of the comparison being higher than the threshold value indicates that the tire is fit to continue.

3. Method, according to claim 1, characterized by the fact that in step E at least a second individualized tire condition is determined; in which the result of the comparison indicates that the tire is over-inflated when the lateral grooves have a greater depth than that of the central groove; in which the result of the comparison indicates that the tire is under-inflated when the lateral grooves have a shallower depth than that of the central groove; in which the result of the comparison indicates that the tire has an alignment problem when the left and central grooves have a greater depth than that of the right groove, or in the case of the right and central grooves having a greater depth than the left groove.

4. Method according to claim 1, characterized in that in step H at least a first tire condition is determined. relative; and in which the need for rotation or change of position of one or more of the tires of interest is determined by the difference between the average depth of the grooves of a first tire of interest in relation to one or more other tires of interest.

5. Method, according to claim 1, characterized by the fact that the method further comprises one or more steps of visually verifying the external conditions of one or more tires of interest, determining at least one more individualized tire condition.

6. Method, according to claim 1, characterized by the fact that the image of a region of interest of the tread comprises a minimum definition of 224 x 224 pixels.

7. The method of claim 1, wherein the machine learning model comprises a feedforward deep convolutional neural network.

8. System for determining a tire condition, characterized by the fact that it comprises: At least one memory that stores information and computer-executable instructions; At least one neural network trained to provide tire tread depth measurements uniquely associated with the position of each tread in the tire tread, in response to receiving, as input, images of tire treads; and At least one processor or computer system or processing circuit configured to control the system of the invention, wherein the execution of the instructions performs one or more more steps of a method of determining a tire condition.

9. System, according to claim 8, characterized by the fact that it additionally comprises: At least one processing equipment for acquiring and presenting information / instructions; At least one database; One or more sets of information; and One or more sets of instructions.

10. System, according to claim 9, characterized by the fact that it additionally comprises: - at least one central server; and - at least one communication and data network.

11. The system of claim 8, wherein the trained convolutional neural network is a feedforward deep convolutional neural network.

12. The system of claim 8, wherein the processor or computing system or processing circuit is configured to determine the trained convolutional neural network.

13. System, according to claim 10, characterized by the fact that the central server comprises at least one remote central processor, in which the central server hosts and manages the database of the tires of interest of one or more vehicles of interest of one or more fleets of vehicles of interest.

14. System, according to claim 9 or 10, characterized by the fact that both the central server and the processing equipment each have at least one processor or computer system or processing circuits that are independent of each other and / or autonomous.

15. System according to any one of claims 8 to 14, characterized in that the set of information comprises: a. Unique identification data; b. Supplementary identification data; c. Data on the tires of interest; d. Training image data sets; e. Input image sets; f. Resulting output data sets; g. Threshold values; h. Tread depth indicators or codes; i. Tread position indicators or codes; j. Individualized tire condition indicators or codes; k. Relative tire condition indicators or codes; l. Relative axle condition indicators or codes; and m. Listing of visual check items and their additional tire condition codes.

16. System characterized by the fact that it performs at least one, preferably all, of the steps of the method defined in any one of claims 1 to 7.

17. Equipment, characterized by the fact that it performs at least one, preferably all, of the steps of the method defined in any one of claims 1 to 7, in which the equipment is portable and comprises at least: (i) A processor or processing circuit configured to determine a trained neural network and logically connected to at least one memory of the equipment (internal or external); (ii) An interface device for displaying and entering information and / or instructions; (iii) Image capture device with a digital camera with a minimum resolution of 224 x 224 pixels.

18. Equipment according to claim 17, characterized in that it additionally comprises a communication device for communicating with one or more data and communication networks.

19. Method for training a neural network to estimate the depth of a tire groove, characterized by the fact that it comprises the following steps: (i) Generate at least one training image dataset; (ii) Configure a neural network to receive, as input, the training image dataset, in which each training image corresponds to the image of a tire tread; (iii) Receive training images of tire treads with grooves of previously known depth and position; (iv) Configure the neural network to locate one or more grooves in an image of a tire tread; (v) Configure the neural network to identify each of the tread grooves individually; (vi) Configure the neural network to assign each of the training grooves its known depth, assigning it a unambiguous depth identifier, unequivocally associating the known depth value with its known position on the tread, also assigning it a unambiguous position identifier; (vii) Configure the neural network to generate, in response to an input image of a tire tread, at least one, preferably at least three, depth values ​​uniquely associated with at least one, preferably at least three corresponding groove positions in the tread; and (viii) Produce a neural network by repeatedly training the trained neural network with each of the training data sets.

20. Computer-readable memory comprising a set of instructions that, when executed, perform the method defined in any one of claims 1 to 7 and / or the method defined in claim 19.

Images