퍼지 스트레칭을 이용한 세라믹 X-선 영상에서의 비파괴 결함 검출 = Defect Detection of Ceramic X-ray Images using Fuzzy Stretching

국문 초록 (Abstract)

본 논문에서는 X-선 영상을 통해 용접한 세라믹 소재 부품의 가공 불량, 균열 등과 같은 결함 영역을 효과적으로검출하는 방법을 제안한다. 제안된 방법은 세라믹 X-선 영상에서 결함 영역과 배경 간의 명암 대비를 강조하기 위해퍼지 스트레칭 기술을 적용한다. 또한, 결함 영역의 위치 특성을 분석하여 영상 내의 불필요한 잡음을 제거하고 관심영역 (ROI: Region of Interest)을 추출한다. 추출된 ROI 영역에 Region-Growing 알고리즘을 적용하여 결함 영역의픽셀 위치를 식별하고, 해당 영역을 원본 X-선 영상에 표시하여 최종적으로 결함을 검출한다. 실험 결과, 제안된 방법은 세라믹 비파괴 검사 영상에서 높은 검출 성능을 보여 70개의 영상에서 평균 95.75%의 검출 성과를 달성하였다.
이 방법은 세라믹 부품 검사의 효율성을 높이고 결함 검출 과정에서의 자원과 시간을 절감하는 데 유용한 기술로서의 잠재력을 제시한다.

다국어 초록 (Multilingual Abstract)

In this paper, we propose an efficient method for detecting defects in ceramic components, such as processingirregularities or cracks, welded onto parts in X-ray images. Our approach utilizes fuzzy stretching techniques to enhancethe contrast between defect areas and background regions in ceramic X-ray images. Subsequently, by analyzing the spatialcharacteristics of defect areas, unnecessary noise is removed from the images and extract the ROI. Region-Growingalgorithm then detect the pixel locations of defect areas within the extracted ROI and mark these areas on the originalX-ray images, ultimately achieving precise defect detection. Experimental results demonstrate that the proposed methodachieves a high detection performance in ceramic non-destructive testing images, with an average detection rate of 95.75%across 70 images. This approach offers the potential to enhance the efficiency of ceramic component inspections andreduce resources and time requirements during the defect detection process.

참고문헌 (Reference)

• 1 "Korea Society for Nondestructive Testing"
• 2 Tanzila Saba, "IoMT Enabled Melanoma Detection Using Improved Region Growing Lesion Boundary Extraction" Computers, Materials and Continua (Tech Science Press) 71 (71): 6219-6237, 2022
• 3 K. B. Kim, "Flaw detection in ceramics using sigma fuzzy binarization and gaussian filtering method" 9 (9): 403-414, 2014
• 4 K. B. Kim, "Automatic segmentation of ceramic materials with relaxed possibilistic c-means clustering for defect detection" Institute of Advanced Engineering and Science 19 (19): 1505-1511, 2020
• 5 Neeraj Shrivastava, "Automatic Seeded Region Growing Image Segmentation for Medical Image Segmentation: A Brief Review" World Scientific Pub Co Pte Lt 20 (20): 1-21, 2020
• 6 Joonsung Park, "Automatic Extraction of Soft Tissue Tumor from Ultrasonography Using ART2 Based Intelligent Image Analysis" Bentham Science Publishers Ltd. 13 (13): 447-453, 2017
• 7 Mythili Thiruganam, "Automatic Defect Detection and Counting In Radiographic Weldment Images" Foundation of Computer Science 10 (10): 1-5, 2010
• 8 Ye TIAN, "Adaptive Binarization for Vehicle State Images Based on Contrast Preserving Decolorization and Major Cluster Estimation" Institute of Electronics, Information and Communications Engineers (IEICE) E105.D (E105.D): 679-688, 2022