산업용 CT 볼륨데이터에서 2, 3차원 크랙 강화 필터링을 이용한 금속품의 내부 및 표면 크랙 자동 검출 = Automatic Detection of Internal and Surface Cracks of Metal Products using Two and Three-dimensional Crack Enhancement Filtering Methods in Industrial CT Volume Data

국문 초록 (Abstract)

본 논문에서는 산업용 CT 볼륨데이터에서 금속품 내부와 표면 크랙 검출 기법을 제안한다. 제안 방법은 첫째, 다운샘플링 데이터에서 제품 영역을 배경과 분리하여 관심볼륨을 자동으로 설정하고, 밝기값 히스토그램을 이용한 상대누적도수 기반의 임계값 기법을 이용하여 크랙 후보를 추출한다. 둘째, 형상 정보를 고려하기 위하여 다중 스케일 헤시안 행렬 기반의 2차원 크랙 강화 필터링으로 금속품 내부의 크랙을 검출한다. 셋째, 공간상의 크랙 결함 검출 정확성을 높이기 위해 3차원 다중 스케일 헤시안 행렬 기반의 크랙 강화 필터링을 이용하여 3차원 형태의 크랙을 검출한 후, 다단면 씬-슬랩 영역에서 크랙강화 필터링으로 금속품 표면의 크랙을 검출한다. 넷째, 원볼륨데이터에서 크랙 검출율 개선을 위하여 여과기법으로 금속품 내부와 표면의 크랙을 검출한다. 본 제안 방법의 결과를 평가하기 위하여 산업용 CT볼륨데이터와 시뮬레이션 데이터를 사용하여 육안 평가, 정확성 평가와 수행시간 측정을 수행하였다.

다국어 초록 (Multilingual Abstract)

In this paper, we propose a method for detecting internal and surface cracks of metal product in industrial CT volume data. Our method is composed of four main steps. First, product regions are separated from background in down-sampling data and volume of interest is automatically defined. Then crack candidates are extracted using thresholding based on relative cumulative frequency. Second, to consider density and shape information of cracks, internal cracks of metal products are detected using Hessian-based multi-scale 2D crack enhancement filtering. Third, to improve an accuracy of crack detection on three-dimensional space, spatial cracks are detected using Hessian-based multi-scale 3D crack enhancement filtering. For the detection of surface cracks of metal product, the 3D crack enhancement filtering is applied to multiplanar thin-slab areas. Fourth, to improve a crack detection rate in original volume data, internal and surface cracks of metal product are detected using percolation method. The performance of our method is evaluated with the aspects of visual inspection, accuracy assessment and processing time using industrial CT volume data and simulation data. For accuracy assessment, 2D line enhancement filtering, 3D sheet enhancement filtering and finite line integral method are used as comparative methods and Dice"s coefficient is measured.

목차 (Table of Contents)

• 요약
• Abstract
• 1. 서론
• 2. 2, 3차원 크랙 강화 필터링을 이용한 크랙 검출
• 3. 실험 및 결과
• 4. 결론
• References

참고문헌 (Reference)

1 P. Perona, "Scale-Space and Edge Detection Using Anisotropic Diffusion" 2 (2): 629-639, 1990

2 A. F. Frangi, "Multiscale Vessel Enhancement Filtering" 1496 : 130-137, 1998

3 L.C. Chien, "Measuring System for Crack in Concrete using Mutitemporal Images" American Society of Civil Engineers 79-82, 2006

4 T. Yamaguchi, "Image Based Crack Detection for Real Concrete Surface" The Institute of Electrical Engineers of Japan Trans 2008

5 Y. Sato, "Handbook of Biomedical Image Analysis Topics in Biomedical Engineering International Book Series vol.2" Springer 531-589, 2005

6 T. Yamaguchi, "Fast Crack Detection Method for Large-size Concrete Surface Images using Percolation-based Image Processing" 21 : 797-809, 2010

7 F.S. Hiller, "Continuous Sampling Plans Under Destructive Testing" 59 (59): 376-401, 1964

8 K. Ehrig, "Comparision of Crack Detection Methods for Analyzing Damage Processes in Concrete with Computed Tomography" 2011

9 S.K. Sinha, "Automated Detection of Cracks in Buried Concrete Pipe Images" 15 : 58-72, 2006

10 B.R. Suresh, "A real-time automated visual inspection system for hot steel slabs" PAMI-5 : 563-572, 1983

11 Y. Fujita, "A Robust Automatic Crack Detection Method from Noisy Concrete Surfaces" 22 : 245-254, 2011

12 B. Bi, "A Novel Method for 3D Crack Edge Extraction in CT Volume Data" 19 : 429-442, 2011

13 Y. Fujita, "A Method for Crack on a Concrete Structure" 2006

14 N. Tanaka, "A Crack Detection Method in Road Image Using Morphology" 1998

15 P. Ciorau, "A Contibution of Phased Array Ultrasonic Technology to Detection and Sizing Stress Corrosion Cracks" 2008