http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Faster R-CNN 학습데이터 구축과 모델을 이용한 안전모 탐지 연구
김대성(Kim, Daeseong),공정식(Kong, Jungsik),임재훈(Lim, Jaehoon),소병춘(Sho, Byungchoon) 한국방재학회 2020 한국방재학회논문집 Vol.20 No.1
우리나라의 산업재해 중 건설 분야의 경우 전체의 28.55%의 고위험분야에 해당한다. 이러한 산업재해를 저감하기 위한활발하게 진행되고 있지만, 건설 안전분야 적용 연구는 아직 미비한 실정이다. 본 연구에서는 건설 안전분야의 안전 보호구착용 판별에 대한 최신 인공지능 R-CNN 알고리즘을 활용하여 한국형 데이터 학습모델 구축 연구를 수행하였고, 실시간영상 접목이 가능한 객체 탐지 솔루션을 구현하여 검증하였다. 본 연구수행 결과, 신규 한국형 안전모 착용 여부에 대한판별 데이터 학습모델은 평균 정밀도(mean Average Precision) 0.82의 우수한 결과로 건설 안전분야의 산업재해 저감을 이룰것으로 판단된다. Among the industrial accidents occurring in Korea, the construction sector accounts for 28.55% of theaccidents associated with this high-risk sector. Although active progress is being made to reduce such industrial accidents, studies concerning the applications to construction site safety are currently insufficient. In this study, research is conducted on the development of a Korean data learning model using the latest in artificial intelligence algorithms, R-CNN, for the identification of safety protection equipmentworn by workers in the field. As a result of this study, the discriminant data learning model concerning the donning of a new Korean safety hard hat provided excellent results with a mean average precision of 0.82. This advancement is expected to reduce industrial accidents in the field of construction and improve safety.
실리카흄 혼합 시멘트 콘크리트의 성능에 대한 기초적 연구
이승태(Lee, Seungtae),김성수(Kim, Seongsoo),박광필(Park, Kwangpil),김대성(Kim, Daeseong),황재남(Hwang, Jaenam),김환중(Kim, Hwan Joong) 한국구조물진단유지관리학회 2009 한국구조물진단학회 학술발표회논문집 Vol.2009 No.1
This paper is aimed at evaluating the experimental findings including characteristics of both strength and diffusion of cement concrete incorporating silica fume. The main variables are the replacement of silica fume, 0, 5, 10 and 15% of cement by mass. As a results, it was found that the higher replacement of silica fume led to better performance in terms of strength and diffusion.
다단임팩터, SMPS 및 ELPI를 이용한 디젤배출입자의 크기분포 측정
권순박(Soonbark Kwon),김대성(Daeseong Kim),이창(Chang Lee),이규원(Kyoowon Lee),임철수(Cheolsoo Lim),류정호(Jungho Ryu),엄명도(Myungdo Eom),조규백(Gyubaek Cho),김홍석(Hongsuk Kim),정용일(Youngil Jung),이종덕(Jongduk Lee) 한국자동차공학회 2003 한국자동차공학회 춘 추계 학술대회 논문집 Vol.- No.-
The size and mass distribution of diesel exhaust particles were measured using a cascade impactor, SMPS and ELPI with different engine operating conditions in engine speed and ratio. It was found that the total emission was increased with increasing engine speed and load ratio and about 80% of diesel exhaust particles were less than lum. The results of real time measurement of diesel particles by SMPS and ELPI showed that the peak diameter and number concentration increased as engine load ratio increased. It is interesting to note that the increasing rate at the size range of 265~385nm was the highest, while the nuclei mode particles (<50nm) were decreased or maintained constant when the engine load ratio increased. The size distribution measured dy SMPS and ELPI synchronously showed the reasonable correlations, however, the significant differences in absolute number concentration was observed.
임유진(Yujin Lim),조호진(Hojin Cho),김대성(DaeSeong Kim),이진욱(Jinwoog Lee) 한국철도학회 2012 한국철도학회 학술발표대회논문집 Vol.2012 No.10
궤도하부층의 강화노반은 반복적인 열차하중에 의하여 영구변형이 발생한다. 이러한 영구변형이 누적하여 발생하면 궤도틀림을 유도할 수 있다. 궤도구조의 안정성을 확보하기 위해서는 궤도 하부구성층에서의 영구변형을 정확히 예측하는 것이 중요하다. 이에 본 연구에서는 기존 영구변형 모델에 대한 장단점을 비교검토 하였으며 아울러, 대형반복 삼축압축시험기를 이용하여 강화노반 쇄석재료에 대한 반복삼축시험을 실시하여 영구변형 발생 특성을 파악하였다. 구속압, 전단강도특성(전단응력비) 및 반복재하횟수 등이 쇄석 강화노반 재료의 영구변형 모델 수립에 매우 주요한 매개변수가 될 수 있음을 확인하였다. Permanent deformation in trackbed foundation is generated due to repeated track loading. Permanent deformation in the trackbed foundation can cause irregularities of the rail. It is critical to predict magnitude of permanent deformation in design of the trackbed foundation thickness in order to provide stability to the rail structures. A comparative study of permanent deformation prediction models for unbound granular materials was performed in order to check reliability and pros and cons. A large triaxial repetitive apparatus was adapted to perform cyclic triaxial tests to characterize permanent deformation in the track foundation materials such as unbound granular soils. It is verified that confining stress, shear strength parameters (shear stress ratio) and number of load repetition are critical design parameters in construction of prediction model for permanent deformation.