• 한국인터넷방송통신학회논문지
• /
• 제22권4호
• /
• pp.157-163
• /
• 2022

최근 산업 분야 자동화의 발전에 따라 이상 징후 검출에 대한 연구가 활발하게 진행 중이다. 공장 자동화에 사용되는 이상 징후 검출의 응용분야로 카메라를 사용한 결함 검사가 있다. 비전 카메라 검사는 공장 자동화에서 높은 성능과 효율성을 보이지만, 조명과 환경조건의 불안정성을 극복하기가 어렵다. 딥러닝을 이용한 카메라 검사가 훨씬 더 높은 성능을 보이면서 비전 카메라 검사의 문제를 해결할 수 있지만 학습을 위해 엄청난 양의 정상 데이터 및 비정상 데이터를 요구하기 때문에 실제 산업 분야에 적용하기가 어렵다. 따라서 본 연구는 정상 데이터만을 사용한 72개의 기하학적 변환 딥러닝 방법으로 비정상 데이터 수집 문제를 극복하고, 성능 개선을 위한 특이치 노출 방법을 추가한 네트워크를 제안한다. 이를 자동차 부품 데이터 및 이상치 검출용 데이터베이스인 MVTec 데이터 셋에 적용하고 검증함에 의해 실제 산업 현장에서 적용할 수 있음을 보인다.

• 문화기술의 융합
• /
• 제8권6호
• /
• pp.155-159
• /
• 2022

모발의 필수 미네랄 원소 농도 데이터로 인공지능을 학습시킨 후, 피험자의 모발 미네랄 원소 농도로 나이를 예측하도록 하고 그 결과를 피험자의 실제 나이와 비교하여 연관성을 조사하였다. 전체 모발 데이터는 296개이었으며그 중 2/3를 인공지능 학습에 그리고 1/3을 피험자 데이터로 사용하였다. 25세 이하의 성장기 피험자의 실제 나이와 인공지능이 예측한 나이 사이에는 0.678 의 중상 정도의 상관관계가 있었다. 중년의 피험자 그룹에서는 연관성이 거의 없었고 노년의 피험자 그룹에서는 0.522의 약한 상관관계를 보였다. 모발의 미네랄 원소 농도 데이터를 이용한 인공지능의 유용성을 확보하기 위해서는 더 많은 수의 데이터를 제공하여 인공지능을 학습시키는 과정이 필요하다.

• 터널과지하공간
• /
• 제33권6호
• /
• pp.495-507
• /
• 2023

불연속면 거칠기는 암반의 기계적 특성에 큰 영향을 주며 열·수리 역학적 거동에도 많은 영향을 미치는 요소이다. 본 연구에서는 입체사진측량기법을 이용하여 불연속면에 대한 3차원 점군 데이터를 생성시키고 이를 이용하여 불연속면의 거칠기 특성화를 수행하였다. 3차원 점군 데이터로 재생성된 불연속면 프로파일과 프로파일 게이지를 이용하여 수동으로 측정한 프로파일을 비교하여 취득한 점군 데이터가 암반면의 실제 형상을 정확하게 재현하였는지 평가하였다. 또한, 측정 프로파일수가 거칠기 평가에 미치는 영향에 대해 분석하였고, 거칠기의 이방성 평가방법을 제안하고 실제 암반 불연속면에 대한 거칠기 이방성 평가를 수행하였다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1989년도 한국자동제어학술회의논문집; Seoul, Korea; 27-28 Oct. 1989
• /
• pp.131-137
• /
• 1989

The effect of kinematic model choice on robot calibration is examined. This paper presents a complete formulation to identify the actual robot kinematic parameters directly from position data. The method presented in this paper applies to any serial link manipulator with arbitrary order and combination of revolute and prismatic joint.

• 산업공학
• /
• 제4권2호
• /
• pp.35-42
• /
• 1991

Many authors derived MTTF/MTBF for an operating system by analyzing its actual life time data. However, it is difficult to derive MTTF/MTBF when few breakdowns accur throughout a year. In this paper, we address a new approach to solve that problem under a preventive maintenance policy, in which few breakdowns occur, and also introduce a case study using the results obtained.

• Journal of the Korean Statistical Society
• /
• 제15권1호
• /
• pp.1-8
• /
• 1986

The k-medians clustering method is considered to partition observations into k clusters. Consistency and advantage of bootstrap confidence sets of k optimal cluster centers are discussed. The k-medians and k-means clustering methods are compared by using actual data sets.

• Journal of the Korean Statistical Society
• /
• 제7권1호
• /
• pp.11-16
• /
• 1978

Bayesian procedures are in vogue to revise the parameter estimates of the forecasting model in the light of actual time series data. In this paper, we study the Bayes forecast for demand and the risk when (a) 'noise' and (b) mean demand rate in a constant process model have moderately non-normal probability distributions.

• 한국농공학회지
• /
• 제8권1호
• /
• pp.1048-1054
• /
• 1966

1. Data of this article come from actual soil survey activity by the UNKUP project personel. 2. Proper national land use must be solved because of the growing population and increasing economic activity. 3. Korea has to be developed for the natural resources of soil and water in the subwate\ulcorner rsheds. 4. This problem. depends On the result of land capability classification which was determind by the soil survey.

• 한국응급구조학회지
• /
• 제15권3호
• /
• pp.81-96
• /
• 2011

Purpose : The purpose of this study was to analyse early childhood teachers' perspectives and actual conditions of earthquakes and tsunami safety education. Methods : A total of 119 teachers were recruited from kindergarten and child-care center located in J city, all of whom gave consent to take part in these research. The data have been collected for 45 days from May to July in 2011. The data was analyzed by SPSS (version 17.0) and chi-square test. Results : The results of this study are summarized as follow. 1. 97.5% of the subjects considered earthquakes and tsunami safety education necessary, however, only 34.6% of them were given the education related to earthquakes and tsunami. 2. 63.9% of the early childhood teachers provided children in the kindergarten or childhood center with earthquakes and tsunami safety education, however, 30.3% of them were found to be without any experiences of earthquakes and tsunami safety education. 3. 68.5% of the teachers were informed earthquakes and tsunami safety knowledge. 4. 35.0% of respondents answered that the single session of earthquake and tsunami education was not efficient, thus expecting to receive continuing education. Conclusions : The study suggested that more teachers necessarily provide earthquake and tsunami safety education to reduce children's mortality rate in earthquakes and tsunami. The education program should be developed and systematized to improve teacher's knowledge. Moreover, the program should be continuously provided to teachers and children in the kindergarten or childhood center.

• 한국조경학회지
• /
• 제26권1호
• /
• pp.1-11
• /
• 1998

The purpose of this paper is twofold : to identify loyalty applicable to segmentation of theme park users and to find characteristics of the segments. Thetheme park was regarded as a product and Lotte World was regarded as a brand. One hundred thirty five college students were selected by nonprobability sampling for two waves thirty of data collection. Both behavioral and attitudinal dimension of loyalty were measured in the first wave by the proportion of visit of the Lotte World to 3 major theme parks for one year, including the Lotte World, and by calculating the mean score of selected 7 attitudinal items, respectively. After 14 weeks, the same respondents were asked the number of actual visits of the Lotte World. Medians of two dimensions and cluster analyses were utilized to classify the respondents into 4 categories : high, spurious, latent, and low loyalty. Then ANOVA and $$\chi$^2$ test of independence were conducted to find the difference in intention to visit the Lotte World and actual visitation of it among groups. Only intention was significantly different by the group and the mean score of intention was highest in the high loyalty group. Although no statistical difference was found in actual visitation among groups, the theory of planned behavior provided a theoretical support to conclude that the loyalty is a useful variable for segmentation of theme park users because intention is an antecedent variable to the behavior. Discriminant analyses showed that characteristics of each loyalty group can be differentiated by motivations and constraints. When median was a group classification criterion, 73.2 percent of high loyalty group was correctly classified. A few comments were suggested on data collection, and inclusion of new discriminant variables was discussed for the future research.