• 한국정밀공학회지
• /
• 제23권8호
• /
• pp.80-88
• /
• 2006

This paper proposes a new steel strip surface inspection system. The system acquires bright and dark field images of defects by using a stroboscopic IR LED illuminator and area camera system and the defect images are preprocessed and segmented in real time for feature extraction. 4113 defect samples of hot rolled steel strip are used to develop KNN (k- Nearest Neighbor) classifier which classifies the defects into 8 different types. The developed KNN classifier demonstrates about 85% classifying performance which is considered very plausible result.

• 한국해양공학회지
• /
• 제22권2호
• /
• pp.72-77
• /
• 2008

The application of infrared thermography for detecting defects under the surface of a material was studied. Defects in a specimen were made by back-drilled circular holes. To get alarge temperature difference at the surface, a halogen lamp was used for surface heating. We confirmed that the defect location had a good relationship with the maximum temperature difference. The sizes of the defects could be calculated by means of the FWHM. The value of the FWHM of a temperature difference decreased with time. Therefore in an extremely short time after the heating, the true defect size could be measured.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권12호
• /
• pp.4763-4775
• /
• 2020

This research aims to develop a deep learning-based method that automatically detects and segments the defects on railroad surfaces to reduce the cost of visual inspection of the railroad. We developed our segmentation model by modifying a fully convolutional network model [1], a well-known segmentation model used for machine learning, to detect and segment railroad surface defects. The data used in this research are images of the railroad surface with one or more defect regions. Railroad images were cropped to a suitable size, considering the long height and relatively narrow width of the images. They were also normalized based on the variance and mean of the data images. Using these images, the suggested model was trained to segment the defect regions. The proposed method showed promising results in the segmentation of defects. We consider that the proposed method can facilitate decision-making about railroad maintenance, and potentially be applied for other analyses.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1993년도 학술논문발표회 논문집 제12권 1호
• /
• pp.169-173
• /
• 1993

In this paper, the features extraction of reflected signals from various type of defects existing in the solid has been studied by Wiener filtering technique. In this experiment, three types of the defect have been considered; a flat cut, a angular cut and a circular hole. All of the defects have the same size, 20mm in diameter and have been located at 45mm in depth from the aluminum surface. In the result of the experiment, it has been found that the wiener filtering technique used for features extraction from the reflected signal corresponding to each defect have been very effective for defect classification.

• 한국전기전자재료학회:학술대회논문집
• /
• 한국전기전자재료학회 2001년도 하계학술대회 논문집
• /
• pp.578-581
• /
• 2001

Silicon wafers using DRAM devices required for high cleaning technology and this cleaning technology was evaluated by defect level or electron life time. This paper examined the correlation of SPV(Surface Photo Voltaic Analyzer) which analyzes diffusion length of minority carriers and DLTS(Deep level Transient Spectroscope) which analyzes defect level.

• 한국항해항만학회지
• /
• 제28권1호
• /
• pp.43-49
• /
• 2004

결함검사는 생산공정에 있어서 상품의 디자인과 함께 매우 중요한 부분으로서, 상품의 경쟁력을 높이는데 필수 불가결한 것이다. 만약, 실시간 결함검출이 상품에 대한 어떤 손상도 없이 할 수 있다면, 품질 및 공정의 효율적 관리와 고비용 인력의 절감을 통하여 생산원가를 줄일 수 있다. 본 논문에서는 철판과 같은 표면에 결함이 있는 경우 필요한 정보만을 추출할 수 있는 3가지 공간필터법에 대하여 제안하였고, 공간필터의 특성을 통하여 결함검출 시스템을 구성하였다. 그리고, 최적의 표면결함 계측용 공간필터법을 개발하기 위하여 결함의 크기와 형태, 광도의 크기 및 외부 광간섭 그리고 슬리트의 개수와 같은 파라메타의 변화에 따른 측정 성능을 비교 및 분석하였다.

• Smart Structures and Systems
• /
• 제33권1호
• /
• pp.55-64
• /
• 2024

This paper proposes a fusion imaging-based coating-defect classification method for steel structures that uses zero-shot learning. In the proposed method, a halogen lamp generates heat energy on the coating surface of a steel structure, and the resulting heat responses are measured by an infrared (IR) camera, while photos of the coating surface are captured by a charge-coupled device (CCD) camera. The measured heat responses and visual images are then analyzed using zero-shot learning to classify the coating defects, and the estimated coating defects are visualized throughout the inspection surface of the steel structure. In contrast to older approaches to coating-defect classification that relied on visual inspection and were limited to surface defects, and older artificial neural network (ANN)-based methods that required large amounts of data for training and validation, the proposed method accurately classifies both internal and external defects and can classify coating defects for unobserved classes that are not included in the training. Additionally, the proposed model easily learns about additional classifying conditions, making it simple to add classes for problems of interest and field application. Based on the results of validation via field testing, the defect-type classification performance is improved 22.7% of accuracy by fusing visual and thermal imaging compared to using only a visual dataset. Furthermore, the classification accuracy of the proposed method on a test dataset with only trained classes is validated to be 100%. With word-embedding vectors for the labels of untrained classes, the classification accuracy of the proposed method is 86.4%.

• 청정기술
• /
• 제23권1호
• /
• pp.90-94
• /
• 2017

현재 풍력발전은 우리나라에서 가장 주목 받고 있는 신재생에너지 분야로 많은 연구가 진행 중이다. 풍력발전을 구성하는 요소 중 핵심 요소인 블레이드에 손상이 발생할 경우에 발전 효율에 직접적인 영향을 미치므로 효율적인 유지보수를 위해 초기에 결함을 측정하는 기술이 매우 중요한 상태이다. 그러나 기존의 초음파 비파괴 검사 및 열화상 비파괴 검사는 소요시간이 길고 실시간 모니터링이 어려우므로 초기 결함 측정이 불가능하다. 기존의 문제를 보완하고자 본 논문에서는 표면파를 이용한 블레이드 표면상태 실시간 모니터링에 관한 연구를 수행하였다. 압전센서 기반의 시스템을 구성하여 블레이드 표면샘플에 대해 공정 변수 별 기초 성능 실험을 하였고, 소형 블레이드에 대해 공정변수 별 실험을 통해 블레이드의 크랙, 벗겨짐, 장애물 등을 실시간으로 모니터링이 가능한지 연구 하였다.

• 대한기계학회논문집B
• /
• 제30권5호
• /
• pp.405-412
• /
• 2006

In order to realize a stably propagating flame in a narrow channel, flame instabilities resulting from flame-wall interaction should be avoided. In particular flame quenching is a significant issue in micro combustion devices; quenching is caused either by excessive heat loss or by active radical adsorptions at the wall. In this paper, the relative significance of thermal and chemical effects on flame quenching is examined by means of quenching distance measurement. Emphasis is placed on the effects of surface defect density on flame quenching. To investigate chemical quenching phenomenon, thermally grown silicon oxide plates with well-defined defect distribution were prepared. ion implantation technique was used to control defect density, i.e. the number of oxygen vacancies. It has been found that when the surface temperature is under $300^{\circ}C$, the quenching distance is decreased on account of reduced heat loss; as the surface temperature is increased over $300^{\circ}C$, however, quenching distance is increased despite reduced heat loss effect. Such abberant behavior is caused by heterogeneous surface reactions between active radicals and surface defects. The higher defect density, the larger quenching distance. This result means that chemical quenching is governed by radical adsorption that can be parameterized by oxygen vacancy density on the surface.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2011년도 제41회 하계 정기 학술대회 초록집
• /
• pp.203-203
• /
• 2011

The importance of stepped single-crystal surfaces as model catalysts has been well recognized [1]. We re-investigated the adsorption properties of $H_2$ and CO, most important species in platinum-based catalysts, on nearly defect-free and highly stepped surfaces of one and the same Pt(111) crystal. While both being symmetric and single-peaked from the nearly defect-free surface, temperature-programmed desorption (TPD) spectra from the highly stepped surface saturated at 90 K with H and CO were triply- and doubly-peaked, respectively. Once pre-adsorbed, CO preempted step and then terrace sites, inhibiting the dissociative $H_2$ adsorption completely. Pre-adsorbed H inhibited the CO adsorption on terrace sites only, leaving defect sites intact for CO adsorption even at the saturation H precoverage. On defect-free Pt(111), while pre-adsorbed CO inhibited the dissociative $H_2$ adsorption completely, pre-adsorbed H could not inhibit the CO adsorption completely. These intriguing, but interesting results are discussed in terms of energetics/kinetics and the role of surface step sites in the dissociative adsorption of $H_2$ on Pt(111) [2].