ULTRASONIC FLAW DETECTION SYSTEM AND ULTRASONIC FLAW DETECTION METHOD

To provide a detection method capable of minimizing a detection error by ensuring objectivity of a result in an ultrasonic flaw detection and preventing a human error.SOLUTION: The invention relates to an ultrasonic flaw detection system and an ultrasonic flaw detection method utilizing a deep learn...

Full description

Saved in:
Bibliographic Details
Main Authors YOO JUN WOO, LEE JUNG MIN, SEO JUNE SUNG
Format Patent
LanguageEnglish
Japanese
Published 05.01.2024
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:To provide a detection method capable of minimizing a detection error by ensuring objectivity of a result in an ultrasonic flaw detection and preventing a human error.SOLUTION: The invention relates to an ultrasonic flaw detection system and an ultrasonic flaw detection method utilizing a deep learning or machine learning technique. The ultrasonic flaw detection system can include: an ultrasonic flaw detection device for transmitting ultrasonic waves to an inspection target and collecting ultrasonic signals reflected from the inspection target and generating signal data; a pre-processing part of signal data for pre-processing the signal data; a selection part of a defect candidate group for selecting a defect candidate group on the basis of the pre-processed signal data; an image data generation part for generating image data on the basis of signal data included in the defect candidate group; and a defect determination part for finally determining presence or absence of defects of the defect candidate group on the basis of the image data.SELECTED DRAWING: Figure 1 【課題】超音波探傷検査における結果の客観性を確保し、ヒューマンエラーを防止して検査エラーを最小限にできる検査方法を提供する。【解決手段】ディープラーニングまたはマシンラーニング技術を利用した超音波探傷検査システムおよび超音波探傷検査方法に関するもので、超音波探傷検査システムは、検査対象に超音波を送信し、前記検査対象から反射されてくる超音波信号を収集して信号データを生成する超音波探傷装置、前記信号データを前処理する信号データの前処理部、前処理された前記信号データに基づいて欠陥候補群を選別する欠陥候補群の選別部、前記欠陥候補群に含まれる信号データに基づいてイメージデータを生成するイメージデータ生成部、および前記イメージデータに基づいて前記欠陥候補群の欠陥の有無を最終的に判定する欠陥判断部を含むことができる。【選択図】図1
Bibliography:Application Number: JP20230095580