본 발명은 AI 딥러닝을 이용한 자동차 타이어 외관 자동검사방법에 관한 것으로, 더욱 상세하게는 자동차 타이어의 외관을 검사함에 있어서 결함을 찾기 위해서 다수개의 2차원 및 3차원 카메라를 동시에 사용하여 자동검사장치를 구성하고, 불량데이터와 양품데이터를 이용하여 신경망모듈을 선정한 후에 피검사 타이어를 검사함에 있어서 타이어의 종류별, 제조사별로 나누어서 적용되는 신경망모듈로 대비하고 가중치를 줌으로써 빠르게 결함을 찾을 수 있어서 양품과 불량품의 구분을 빠르게 진행할 수 있도록 한 AI 딥러닝을 이용한 자동차 타이어 외관 자동검사방법에 관한 것이다. 본 발명의 바람직한 실시예로 형성된 AI 딥러닝을 이용한 자동차 타이어 외관 자동검사방법에 의하면 타이어의 사이드월, 트레드, 비드 및 인터널까지 연속촬영하고, 데이터를 수집 및 분석할 수 있어서 결함을 쉽게 발견하고, 이미지데이터를 딥러닝을 통해서 쉬운방법으로 디코딩하고 다시 코딩하기 때문에 데이터의 처리속도가 높아지고, 해당 신경망모듈을 먼저 선정한 후에 결함이 발생하기 쉬운 부분을 딥러닝을 통해서 미리 알 수 있어서 결함을 쉽게 찾으며, 촬상하는 카메라가 2차원 라인카메라, RGB카메라 또는 3차원카메라를 복합적으로 사용하여 지나치기 쉬운 결함을 완벽하게 찾아낼 수 있는 등의 효과가 발생한다.

    The present invention relates to a device for automatically inspecting the exterior of a car tire by using AI deep learning and, more specifically, to a device for automatically inspecting the exterior of a car tire by using AI deep learning, which constructs an automatic inspection device by simultaneously using a plurality of 2D and 3D cameras to find defects in inspecting the exterior of a car tire, selects a neural network module by using defective product data and good product data, and can compare images with the defective product data and the good product data by using the neural network module to be applied separately by tire type and manufacturer and can apply weights thereto to quickly find defects in inspecting a tire to be inspected, thereby quickly distinguishing good and bad products. According to a preferred embodiment of the present invention, a device for automatically inspecting the exterior of a car tire throughout AI deep learning can continuously photograph the sidewalls, treads, beads, and internal parts of a tire and can collect and analyze data, thereby easily finding defects, can decode and encode the image data in an easy way through deep learning, thereby increasing the processing speed of the data, can first select the corresponding neural network module and identify parts in which defects are likely to occur, through deep learning, thereby easily finding defects, and can use a combination of the 2D camera, an RGB camera, or the 3D camera, thereby perfectly finding defects that are easy to overlook. The method comprises: a neural network model determination step of determining a neural network model; an application step of applying the determined neural network model; and a determination step of outputting whether a test set of tires is good or bad.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    AI CAR TIRE OUTFIT AUTO TEST METHOD USING AI DEEP LEARNING METHOD


    Weitere Titelangaben:

    AI 딥러닝을 이용한 자동차 타이어 외관 자동검사방법


    Erscheinungsdatum :

    2023-10-18


    Medientyp :

    Patent


    Format :

    Elektronische Ressource


    Sprache :

    Koreanisch


    Klassifikation :

    IPC:    G01M TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES , Prüfen der statischen oder dynamischen Massenverteilung rotierender Teile von Maschinen oder Konstruktionen / B60C VEHICLE TYRES , Fahrzeugreifen / G01N Untersuchen oder Analysieren von Stoffen durch Bestimmen ihrer chemischen oder physikalischen Eigenschaften , INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06T Bilddatenverarbeitung oder Bilddatenerzeugung allgemein , IMAGE DATA PROCESSING OR GENERATION, IN GENERAL



    AI CAR TIRE OUTFIT AUTO TEST METHOD USING AI DEEP LEARNING METHOD

    HUR JO HOON / PARK GEUN SHICK / PARK JONG MOON et al. | Europäisches Patentamt | 2022

    Freier Zugriff

    2 3 AI CAR TIRE OUTFIT AUTO TEST DEVICE THROUGHOUT AI DEEP LEARNING METHOD USING 2D AND 3D IMAGES

    HUR JO HOON / PARK GEUN SHICK / PARK JONG MOON et al. | Europäisches Patentamt | 2022

    Freier Zugriff


    The Helena Tire outfit

    Engineering Index Backfile | 1905


    Method and system for painting-code auto-allocation of vessel outfit

    SONG MUN MIN / JANG GI SAM / KIM HYEONG CHEOL et al. | Europäisches Patentamt | 2016

    Freier Zugriff