• 정보처리학회논문지A
• /
• 제17A권2호
• /
• pp.81-92
• /
• 2010

진동 정보를 통해 기계 설비의 상태나 고장 유무를 판단하는 연구들이 다수 진행 중에 있는데, 대부분의 연구에서는 설비에 대한 진동을 모니터링하거나 고장 유무를 판별하여 사용자에게 알리는 수준이다. 본 논문에서는 진동에 의한 고장 진단과 판별을 보다 정교하게 수행하는 선박 엔진의 고장 감지 기법과 시스템을 제안하였다. 일차적으로 이중적 진동 정보 판별 기법을 적용하여 진동 정보를 확인한 다음에 고장 유무를 검사한다. 만일 고장이 발생한 경우에는 진동 정보의 오류 부분만을 이용하여 고장 진동 파형에 대한 오차 범위를 기준으로 어떤 유형의 고장인지를 판별할 수 있는 기법을 적용하였다. 또한 선박의 진동 경향 분석과 엔진 안전 보존을 목적으로 진동 정보를 데이터베이스에 저장하고 추적할 수 있도록 시스템을 구현하였다. 제안 시스템을 선박 엔진의 고장 판별 유무와 고장 진동 파형 감별 인자에 대해 실험을 수행한 결과 고장 유무 판별은 약 100% 정확성을 가졌고 고장 진동 파형의 유형 인식에서는 약 96% 정확성을 가졌다.

• Journal of Information Processing Systems
• /
• 제16권5호
• /
• pp.1008-1014
• /
• 2020

Aiming at the defect detection quality of denim fabric, this paper designs an improved algorithm based on the optimized Gabor filter. Firstly, we propose an improved defect detection algorithm of jean fabric based on the maximum two-dimensional image entropy and the loss evaluation function. Secondly, 24 Gabor filter banks with 4 scales and 6 directions are created and the optimal filter is selected from the filter banks by the one-dimensional image entropy algorithm and the two-dimensional image entropy algorithm respectively. Thirdly, these two optimized Gabor filters are compared to realize the common defect detection of denim fabric, such as normal texture, miss of weft, hole and oil stain. The results show that the improved algorithm has better detection effect on common defects of denim fabrics and the average detection rate is more than 91.25%.

• 반도체디스플레이기술학회지
• /
• 제20권1호
• /
• pp.99-103
• /
• 2021

Warpage defect detecting of scaffold is very important in biosensor production. Because warpaged scaffold cause problem in cell culture. Currently, there is no detection equipment to warpaged scaffold. In this paper, we produced detection model for shape warpage detection using deep learning based CNN. We confirmed the shape of the scaffold that is widely used in cell culture. We produced scaffold specimens, which are widely used in biosensor fabrications. Then, the scaffold specimens were photographed to collect image data necessary for model manufacturing. We produced the detecting model of scaffold warpage defect using Densenet among CNN models. We evaluated the accuracy of the defect detection model with mAP, which evaluates the detection accuracy of deep learning. As a result of model evaluating, it was confirmed that the defect detection accuracy of the scaffold was more than 95%.

• International Journal of Internet, Broadcasting and Communication
• /
• 제14권3호
• /
• pp.149-154
• /
• 2022

In recent years, the application of deep learning method to computer vision has shown to achieve great performances. Thus, many research projects have also applied deep learning technology to railroad defect detection. In this paper, we have reviewed the researches that applied computer vision based deep learning method to railroad defect detection and inspection, and have discussed the current trend and the direction of those researches. Many research projects were targeted to operate automatically without visual inspection of human and to work in real-time. Therefore, methods to speed up the computation were also investigated. The reduction of the number of learning parameters was considered important to improve computation efficiency. In addition to computation speed issue, the problem of annotation was also discussed in some research projects. To alleviate the problem of time consuming annotation, some kinds of automatic segmentation of the railroad defect or self-supervised methods have been suggested.

• 전자공학회논문지B
• /
• 제33B권4호
• /
• pp.141-147
• /
• 1996

In this paper an improved morphological algorithm for directional defect detection is proposed, where the defect is parallel to the texture image. The algorithm is based on obtaining the background image while removing the defect by comparing every directional morphological result with max or min except that of defect. The defect can of defect and the background image. For a computer simulation, it is shown that the proposed method has better performance than the conventional algorithm.

• 동력기계공학회지
• /
• 제15권4호
• /
• pp.37-41
• /
• 2011

This paper introduces the special eddy current sensor and its characteristic for bolt hole defect evaluation in gas turbine rotor. In the past, Fluorescent penetration inspection method was used for qualitative defect evaluation in gas turbine rotor bolt hole. This method can defect the bolt hole defect but can not evaluate the defect size. Nowadays, eddy current method is used quantitative defect evaluation due to advanced sensor design technology. And eddy current method is more time and cost saving than the old method. We developed bolt shape eddy current sensor for the rotor bolt hole defect detection and evaluation. The eddy current sensor moves to the bolt hole guided by screw nut and detects the defect on the bolt hole. The bolt hole mock-up and artificial defects were made and used for the signal detection & resolution analysis of eddy current sensor. The results show that signal detection capability is enough to detect 0.2 mm depth defect. And the resolution capability is enough to differentiate 02, 0.5, 1.0 and 2.0 mm depth defect.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 춘계학술대회
• /
• pp.108-111
• /
• 2022

제품의 결함 탐지를 위한 머신 비전 시스템에 딥러닝을 적용하기 위해서는 다양한 결함 사례에 대한 방대한 학습 데이터가 필요하다. 하지만 실제 제조 산업에서는 결함의 종류에 따른 데이터 불균형이 생기기 때문에 결함 사례를 일반화할 수 있을 만큼의 제품 이미지를 수집하기 위해서는 많은 시간이 소요된다. 본 논문에서는 적은 데이터로도 학습이 가능한 샴 신경망을 제품 결함 탐지에 적용하고, 제품 결함 이미지 데이터의 속성을 고려하여 이미지 쌍 구성법과 대조 손실 함수를 수정하였다. AUC-ROC로 샴 신경망의 임베딩 성능을 간접적으로 확인한 결과, 같은 제품끼리만 쌍을 구성하고 결함이 있는 제품 간에는 쌍을 구성하였을 때, 그리고 지수 대조 손실로 학습하였을 때 좋은 임베딩 성능을 보였다.

• International Journal of Internet, Broadcasting and Communication
• /
• 제14권2호
• /
• pp.136-146
• /
• 2022

Some surface defects make the weak quality of steel materials. To limit these defects, we advocate a one-stage detector model RetinaNet among diverse detection algorithms in deep learning. There are several backbones in the RetinaNet model. We acknowledged two backbones, which are ResNet50 and VGG19. To validate our model, we compared and analyzed several traditional models, one-stage models like YOLO and SSD models and two-stage models like Faster-RCNN, EDDN, and Xception models, with simulations based on steel individual classes. We also performed the correlation of the time factor between one-stage and two-stage models. Comparative analysis shows that the proposed model achieves excellent results on the dataset of the Northeastern University surface defect detection dataset. We would like to work on different backbones to check the efficiency of the model for real world, increasing the datasets through augmentation and focus on improving our limitation.

• Smart Structures and Systems
• /
• 제32권4호
• /
• pp.253-266
• /
• 2023

The time-reversal method is employed to improve the ability of pipeline defect detection, and a new approach of identifying the pipeline defect depth is proposed in this research. When the L(0,2) mode ultrasonic guided wave excited through a lead zirconate titinate (PZT) transduce array propagates along the pipeline with a defect, it will interact with the defect and be partially converted to flexural F(n, m) modes and longitudinal L(0,1) mode. Using a receiving PZT array attached axisymmetrically around the pipeline, the L(0,2) reflection signal as well as the mode conversion signals at the defect are obtained. An appropriate rectangle window is used to intercept the L(0,2) reflection signal and the mode conversion signals from the obtained direct detection signals. The intercepted signals are time reversed and re-excited in the pipeline again, result in the guided wave energy focusing on the pipeline defect, the L(0,2) reflection and the L(0,1) mode conversion signals being enhanced to a higher level, especially for the small defect in the early crack stage. Besides the L(0,2) reflection signal, the L(0,1) mode conversion signal also contains useful pipeline defect information. It is possible to identify the pipeline defect depth by monitoring the variation trend of L(0,2) and L(0,1) reflection coefficients. The finite element method (FEM) simulation and experiment results are given in the paper, the enhancement of pipeline defect reflection signals by time-reversal method is obvious, and the way to identify pipeline defect depth is demonstrated to be effective.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제17권1호
• /
• pp.86-93
• /
• 2017

In this paper, we propose a novel defect detection method using component tree representations of scanning electron microscopy (SEM) images. The component tree contains rich information about the topological structure of images such as the stiffness of intensity changes, area, and volume of the lobes. This information can be used effectively in detecting suspicious defect areas. A quasi-linear algorithm is available for constructing the component tree and computing these attributes. In this paper, we modify the original component tree algorithm to be suitable for our defect detection application. First, we exclude pixels that are near the ground level during the initial stage of component tree construction. Next, we detect significant lobes based on multiple attributes and edge information. Our experiments performed with actual SEM wafer images show promising results. For a $1000{\times}1000$ image, the proposed algorithm performed the whole process in 1.36 seconds.