가속도계 설치 개수 및 위치에 따른 딥러닝 기반 교량 손상 추정 영향 분석

이강혁(Lee, Kanghyeok),심성한(Sim, Sung-Han),신도형(Shin, Do Hyoung) 한국방재학회 2021 한국방재학회논문집 Vol.21 No.5

최근 노후화 시설물의 급증과 함께 노후시설물로 인한 재난이 자주 발생하고 있다. 노후시설물로 인한 재난의 방재대책으로 모니터링 시스템을 기반한 시설물 유지관리가 활용될 수 있다. 본 연구진은 선행연구를 통해 모니터링 시스템에서 활용할 수 있는 딥러닝 기반의 교량 손상 추정 기술 기반 유지관리 기술을 제안하였다. 하지만, 선행 연구를 통해 제안된 기술은 기술의 사용에 있어 고가의 가속도계가 다수 필요하므로 높은 비용이 든다는 문제가 존재하였다. 이를 해결하기 위해, 본 연구에서는 손상 추정에 있어 가속도계의 위치 및 개수를 줄여 교량 손상 추정의 활용성을 높이기 위한 연구를 수행하였다. 그 결과, 기존 교량 손상 추정의 효과적인 위한 최적의 가속도계 설치 위치를 제시할 수 있었다. In recent times, there has been an increase in the occurrence of disasters caused by a rapid increase in the aging of infrastructure. A structural health monitoring system can be utilized to detect severe damages in infrastructure and prevent accidents. In a previous study, the authors developed a method for damage detection in bridges using monitoring systems. However, the operational cost of the proposed method was high due to the requirement of multiple expensive accelerometers. The aim of this study is to reduce the number of accelerometers used in the aforementioned method by determining the optimal installation locations of the accelerometers. As a result of this study, optimal accelerometer installation locations for bridge damage detection, have been proposed.

한국리듬운동학회지 저자 네트워크 분석

박성건 ( Sung-geon Park ),박주영 ( Joo Young Park ) 한국리듬운동학회 2016 한국리듬운동학회지 Vol.9 No.1

This study aims to find out the network structure, influence and progress of the network, by analyzing the network of Korean Society for Rhythmic Exercise(KSRE) using the social network analysis structure. For this study, we collected the text data of the titles and authors(affiliations) of 65 papers in 2010-2015 published on the KSRE online journal web site. The data analysis and visualization were performed by social network analysis tool, Node XL, by Microsoft. Also, the influences of authors and their groups were investigated by applying the weight factors depending on the importance of authors, and the network size was analyzed by Eigenvector Centrality, Betweenness Centrality and Closeness Centrality. The results show that the influence under the application of the weight factors is bigger in the ‘Kim, Eunhye" and "Kangwon National University" than the other authors and their affiliations. In terms of the network size, "Choi, Young-jun" and "Dongnam Health College" was the biggest. "Seo, Han-kyo" and "Shinhan University" was doing the role of the midium connecting authors. In conclusion, since the research performance of KSRE is low compared to other academic societies, the efforts to expand the authors and their affiliations are required in the future.

스포츠 데이터 분석에 적용된 통계 및 데이터마이닝 기법에 대한 연구동향

박성건 ( Park Sung-geon ),박주영 ( Park Jooyoung ) 한국리듬운동학회 2016 한국리듬운동학회지 Vol.9 No.2

The purpose of this paper is to analyze and compare statistical analysis and data mining techniques used for sports data analysis by investigating previous studies, and to provide the analyzed results to researchers who study sports data analysis. The analysis was performed based on 84 literatures collected from ACM Portal, IEEE xplore, Pubmed, and RISS for Higher Education. Statistics are used to standardize the records for performance evaluation of players and teams and to identify related factors of win and loss based on the statistical methods such as t-test, ANOVA, correlation, regression and so on, while data mining and machine learning algorithms are used to analyze unstructured data such as image and text, and to predict the outcome of games. In conclusion, sports data analysis has become one of the promising research fields as the technical convergence of sports and other studies progresses quickly. We predict that research and development related to sports data analysis will increase continually in the future.

2014 소치 동계올림픽 여자 컬링 경기분석

박성건(Park, Sung-Geon),윤형기(Yoon, Hyoung-Ki),이수원(Lee, Soo-Won) 한국체육과학회 2016 한국체육과학회지 Vol.25 No.2

The purpose of this study is to provide the curling stakeholder with a better method for the performance enhancement. This paper is composed as follows: 1) difference analysis of the team average score according to the possession of the first and the last stone per end, 2) difference analysis of the number of the blank ends by the team, 3) shot type analysis according to the shot orders, 4) introduction of the developed Curling Performance Analysis System. For the analysis, data was collected from Results and Statistics of the curling game in 2014 Sochi Winter Olympics. Chi‐square test in S PSS 2 3 .0 was applied for the analysis with the statistical significance p<0.05. As a result, it was found that the average score of KOR and JPN with the last stone in opening end was revealed as relatively higher than the other teams. It was also found that the number of the average blank ends per game for the team which shows outstanding performance was relatively higher than the women team"s average. It appears that the Lead makes a shot by exploring the condition of ice in stadium and the strategy of the opposite team. Also, the Second and the Third try to hit the opponent’s stones out of the house. Skip with possession of the first stone tries to "Steal" if possible, while Skip with the possession of the last stone tries to put up her own’s stone on the middle of house. In conclusion, it is needed to develop the Performance Analysis System which can offer the advanced analysis information to the curling stakeholder.

컬링 경기력 지표를 이용한 컬링 경기의 순위 예측

박성건(Park, Sung-Geon),윤형기(Yoon, Hyoung-Ki),이수원(Lee, Soo-Won) 한국체육과학회 2016 한국체육과학회지 Vol.25 No.2

The purpose of this study is to develop the Performance Indicator in Curling(PIC) to predict the ranking of women’s curling games. For the research, data was collected from Results and Statistics of the World Curling Federation and 2014 Sochi Winter Olympics. This paper selects highly influential variables in curling games from the previous studies and logistic regression. Chi‐square, binary logistic regression, and spearman’s correlation in SPSS 23.0 was applied for the analysis with the statistical significance p<.05. As a result, it was found that the highly influential variables in the outcome of curling game is the normalized olympics medal point, the average shot accuracy of Lead, Third and Skip, and the average score with the possession of the last stone in 10th end. It was also found that correlation between PIC and ranking of woman’s curling game is .702 in 2014 Sochi Winter Olympics. In conclusion, the PIC developed in this study is a suitable indicator for predicting the ranking in women"s curling games.

국민안전을 위한 강력범죄 수배차량 검거시스템

지문세,기혜정,기창민,문범섭,박성건,Ji, Moon-Se,Ki, Heajeong,Ki, Chang-Min,Moon, Beom-Seob,Park, Sung-Geon 한국정보통신학회 2021 한국정보통신학회논문지 Vol.25 No.12

본 연구의 최종 목표는 블랙박스, 스마트폰, CCTV 등으로부터 수집된 영상에서 수배차량이 범죄차량인지 여부를 분석할 수 있는 시스템을 개발하는 것이다. 데이터 수집은 자체 개발된 블랙박스를 이용하였다. 본 연구에 활용된 데이터는 차량 유형 8개(트럭, RV, 승용차, 승합차, SUV, 버스, 스포츠카, 전기차)와 차량 모델 434개 등 총 83,753건의 데이터를 사용하였다. YOLO v5를 이용한 차량인식 결과, mAP가 80%로 나타났다. 자체 개발한 블랙박스를 이용하여 ReXNet으로 차량 모델을 식별한 결과, 정확도는 99%로 나타났다. 이러한 결과는 데이터 라벨링의 정확도를 개선하는 것이 차량인식 성능 향상에 도움이 된다는 것을 의미한다. The final goal of this study is to develop a system that can analyze whether a wanted vehicle is a criminal vehicle from images collected from black boxes, smartphones, CCTVs, and so on. Data collection was collected using a self-developed black box. The used data in this study has used a total of 83,753 cases such as the eight vehicle types(truck, RV, passenger car, van, SUV, bus, sports car, electric vehicle) and 434 vehicle models. As a result of vehicle recognition using YOLO v5, mAP was found to be 80%. As a result of identifying the vehicle model with ReXNet using the self-developed black box, the accuracy was found to be 99%. The result was verified by surveying field police officers. These results suggest that improving the accuracy of data labeling helps to improve vehicle recognition performance.

효율적인 국가재난대응을 위한 재난안전통신망(PS-LTE) 표준운영절차(SOP) 교육훈련 모델 개발 및 활용 방안

이상직(Lee,Sangjik),박성건(Park,Sung-Geon),기창민(Ki,Chang-Min),박지홍(Park,Jihong),조현정(Jo,Hyoun-Jung) 한국방재학회 2021 한국방재학회논문집 Vol.21 No.6

본 연구는 PS-LTE 기반 재난안전통신망을 체계적으로 활용할 수 있는 표준운영절차(SOP) 교육 훈련 모델 개발 프로세스를 제시함으로써;국내 SOP의 수준을 CMMI 3단계 이상으로 높이기 위한 기초자료를 제공하는 것이다. 본 연구팀이 개발한 표준운영절차의 타당성 검증을 위해 국내 SOP 전문가 그룹 129명에게 설문조사 및 인터뷰를 실시하였으며;설문은 2020년 9월 1일~9월 30일까지 진행되었다. 연구 결과;국가적인 재난재해 상황에서 국민들의 생명과 재산의 피해를 최소화하고 국내 SOP 수준을 CMMI 3단계 이상으로 발전시키기 위해서는 네 가지 전략이 필요하며;구체적으로는 1) 재난안전통신망 표준운영절차에 대한 거버넌스 구축;2) 재난안전통신망 표준운영절차에 대한 필수교육 훈련은 연 1회;3) 재난안전통신망 표준운영절차는 가이드라인에 따라 구축;4) 기술 분야에 대한 개선이다. 국내 SOP가 CMMI 3단계 이상의 발전은 재난안전에 반드시 필요하다. This study provides information for developing the domestic standard operating procedure (SOP) to CMMI level 3 or higher by presenting the SOP education and training model development process that systematically utilize the PS-LTE-based disaster safety communication network. The survey was conducted with 113 domestic SOP experts. Results revealed that four strategies can minimize the damage to people s lives and property in a national disaster and develop the domestic SOP level to CMMI level 3 or higher-establishment of governance for the SOPs for disaster safety communication networks; training on SOP once a year; establishment of SOP according to the guidelines; and improvement in the technical field. In the future;if SOP develops to CMMI level 3 or higher;it will contribute to the protection of public safety and property from disasters.

단일 이미지 인식으로 피트니스 분야 디지털 휴먼 구현에 필요한 데이터셋 구축에 관한 연구

강수형 ( Soo-hyuong Kang ),박성건 ( Sung-geon Park ),박광영 ( Kwang-young Park ) 한국정보처리학회 2023 한국정보처리학회 학술대회논문집 Vol.30 No.1