• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.576-579
• /
• 2004

Eventhough the productivity has been improved remarkably by introducing automatic facilities, the 100％ inspection is necessary because the possibility to produce large amount of defective goods is also increased. Since it is extremely unreasonable that workers inspect very large amount of products as 100％ inspection, there has been many researches for the automatic inspection system. In this thesis, we develop an automatic detection system of suspension insulator's minutes cracks System The automatic detection system of suspension insulator's minute cracks： To detect the minute cracks of suspension insulators, images of the insulator are acquired with a progressive scan camera, rotating a suspension insulator on a turning table. And after the shadow and noises are eliminated by preprocessing techniques, we detect minute cracks using the features of them.

• Journal of the Korean Society for Railway
• /
• v.20 no.3
• /
• pp.374-381
• /
• 2017

Concrete sleepers, installed on majority of railroad track in this nation can, if not maintained properly, threaten the safety of running trains. In this paper, an algorithm for automatically identifying cracks in a sleeper image, taken by high-resolution camera, is developed based on Adaboost, known as the strongest adaptive algorithm and most actively utilized algorithm of current days. The developed algorithm is trained using crack characteristics drawn from the analysis results of crack and non-crack images of field-installed sleepers. The applicability of the developed algorithm is verified using 48 images utilized in the training process and 11 images not used in the process. The verification results show that cracks in all the sleeper images can be successfully identified with an identification rate greater than 90%, and that the developed automatic crack identification algorithm therefore has sufficient applicability.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.53-56
• /
• 2021

국내 산업화가 들어선 후 산업화 당시 지었던 낙후된 건물의 증가에 따라 구조물의 손상 조사 및 검사 방법의 수요가 늘어나고 있다. 일반적으로 구조물의 손상은 전문 검사원이 현장에서 직접 측량도구와 시각적인 방식으로 검사한다. 그러나 전문 검사원들이 직접 조사하는 수고에 비해 균열을 검사하는 방식 자체가 단순하고, 일반 사람이 검사하기에는 객관성이 떨어지는 한계가 있어 균열을 자동적으로 검출함으로써 객관성과 편의성을 보장할 기술이 필요하다. 본 연구에서는 이미지 기반으로 다양한 환경에서의 외벽 균열을 검출할 수 있는 딥러닝 모델 개발을 소개한다. 균열 검출을 위해 다양한 외벽 균열 관련 데이터셋을 확보 및 구축하고 각 데이터셋의 검출 정보를 보완할 반자동(semi-auto) 라벨링 작업을 수행하였다. 두 번째로 기존 높은 검출 성능을 보였던 모델들을 선정 및 비교하여 YOLO v5 모델을 최종적으로 선정하였고, 도메인이 각각 다른 데이터셋에 대한 교차 학습을 통해 각 데이터셋의 mAP의 편차가 31%에서 11%로 좁히는 작업을 수행하였다. 이를 통해 실제 상황에서의 균열 영상에서 균열을 검출할 수 있는 측량 시스템을 개발함으로써 실질적인 검사의 도구로 활용될 수 있길 기대한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.6
• /
• pp.1353-1358
• /
• 2010

In this paper, we propose a detection method that automatically detects concrete surface cracks using fuzzy method in the image of concrete surface cracks. First, the proposed method detecting concrete surface cracks detects the candidate crack areas by applying R, G, B channel values of the concrete crack image to fuzzy method. We finally detect cracks by the density information about the detected candidate areas after we remove the detailed noises on the image of the concrete surface cracks. The experiments using real concrete images showed that the proposed method is greatly improved of crack detection compared with the conventional methods.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2007.05a
• /
• pp.365-372
• /
• 2007

본 논문에서는 콘크리트 표면 균열 영상에서 균열의 특징들을 추출하기 위하여, 영상 처리 기법을 개선하여 균열의 특징(길이,폭,방향)들을 자동으로 추출 및 분석 할 수 있는 기법을 제안한다. 기존의 영상 처리 기법에서는 비교적 잡음이 적고 균열이 적은 영상을 대상으로 균열을 추출하는 알고리즘을 제시하였기 때문에 많은 잡음과 균열을 가지는 영상에 대해서는 균열 검출 성능이 떨어지는 경향이 있다. 따라서, 본 논문에서 제안한 균열 추출 및 분석 알고리즘은 컬러 영상에서 Histogram Stretching 기법을 적용하여 영상의 콘트라스트 특성을 향상 시킨 후, Robert 연산자를 다시 적용해 균열을 강조하고, 강조된 균열을 Multiple 연산을 이용하여 밝기 차이를 크게 한 후, 개선된 적응 이진화기법을 이용하여 균열의 후보 영역을 추출한다. 추출된 균열 후보 영역을 형상 분석과 위치 및 방향분석을 이용하여 잡음을 제거하고 균열의 특징을 분석한다. 실제 콘크리트 표면 균열 영상을 대상으로 실험한 결과, 균열 검출 성능이 기존의 방법보다 본 논문에서 제안한 방법이 더 우수함을 확인하였다.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.11
• /
• pp.449-456
• /
• 2019

This paper is a study on the construction of a wall-climbing mobile robot using vacuum suction and wheel-type movement, and a comparison of the performance of an automatic wall crack detection algorithm based on machine learning that is suitable for such an embedded environment. In the embedded system environment, we compared performance by applying recently developed learning methods such as YOLO for object learning, and compared performance with existing edge detection algorithms. Finally, in this study, we selected the optimal machine learning method suitable for the embedded environment and good for extracting the crack features, and compared performance with the existing methods and presented its superiority. In addition, intelligent problem - solving function that transmits the image and location information of the detected crack to the manager device is constructed.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.30 no.1
• /
• pp.31-35
• /
• 2010

This paper presents the development of automatic defect detection program using 3D computed tomography image of small solid rocker motor. We applied the neighbor pixel comparison algorithm with beam hardening correction for the recognition of defect. We made the artificial defect specimen in order to decide a standard CT value of defect. The program was tested with 150 small solid rocket motors and it could detect the disbond, crack, foreign material and void. The program showed more reliable and faster results than human inspector's interpretation.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.1034-1037
• /
• 2019

본 논문은 진공을 이용한 흡착방식과 바퀴형 이동방식을 사용하는 벽면 이동로봇의 구성과 로봇 내부에서의 균열검출 및 처리 프로세스에 관한 연구이다. 임베디드 시스템에서 기계학습을 이용한 균열검출을 구현하기 위해 YOLO v3를 수정하여 구동하였으며, 검출된 균열의 영상을 저장하고 위치 정보를 추정하였다. 또한, 균열 정보를 수집하기 위해 고정 IP를 갖는 서버를 구축하고 각 기기 간의 효율적인 통신 네트워크를 구성하였다. 본 기술은 균열검출 작업뿐만 아니라 보수작업에도 활용될 수 있어, 대형 구조물과 건축물 등의 안전진단뿐만 아니라 안전성 향상에 이바지할 수 있을 것으로 예상한다.

• Journal of Korea Society of Industrial Information Systems
• /
• v.24 no.3
• /
• pp.23-29
• /
• 2019

Conventional crack detection methods have a problem of consuming a lot of labor, time and cost. To solve these problems, an automatic detection system is needed to detect cracks in images obtained by using vehicles or UAVs(unmanned aerial vehicles). In this paper, we have studied road crack detection with unmanned aerial photographs. Aerial images are generated through preprocessing and labeling to generate morphological information data sets of cracks. The generated data set was applied to the mask R-CNN model to obtain a new model in which various crack information was learned. Experimental results show that the cracks in the proposed aerial image were detected with an accuracy of 73.5% and some of them were predicted in a certain type of crack region.

• Journal of the Korea Society of Computer and Information
• /
• v.12 no.2 s.46
• /
• pp.77-82
• /
• 2007

In this paper, we proposed the image processing techniques for extracting the cracks in a concrete surface crack image and the enhanced Max-Min neural network for recognizing the directions of the extracted cracks. The image processing techniques used are the closing operation or morphological techniques, the Sobel masking for extracting for edges of the cracks, and the iterated binarization for acquiring the binarized image from the crack image. The cracks are extracted from the concrete surface image after applying two times of noise reduction to the binarized image. We proposed the method for automatically recognizing the directions of the cracks with the enhanced Max-Min neural network. Also, we propose an enhanced Max-Min neural network by auto-tuning of learning rate using delta-bar-delta algorithm. The experiments using real concrete crack images showed that the cracks in the concrete crack images were effectively extracted and the enhanced Max-Min neural network was effective in the recognition of direction of the extracted cracks.