• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.9-12
• /
• 2019

지난 수 년 동안 계속해서 일반 실상 카메라를 이용한 영상분석기술에 대한 연구가 활발히 진행되고 있다. 최근에는 딥러닝 기술을 적용한 지능형 영상분석기술로 발전해 왔으며 국방기지방호, CCTV, 사용자 얼굴인식, 머신비전, 자동차, 드론 산업이 활성화되면서 많은 시너지를 효과를 일으키고 있다. 그러나 어두운 밤과 안개, 날씨, 연기 등 다양한 여건에서 따라서 카메라의 영상분석 정확성 감소와 오류가 수반될 수 있으며 일반적으로 딥러닝 기술을 활용하기 위해서는 고사양의 GPU를 필요로 하기 때문에 다른 추가적인 시스템이 요구된다. 이에 본 연구에서는 열적외선 영상의 객체 검출에 적용하기 위해 SSD(Single Shot MultiBox Detector) 기반의 경량적인 MobilNet 네트워크로 재구성하여, 모바일 기기 등 낮은 사양의 낮은 임베디드 시스템에서도 활용 할 수 있는 방법을 제안한다. 모의 실험결과 제안된 방식의 모델은 적외선 열화상 카메라에서 객체검출과 학습시간이 줄어든 것을 확인 할 수 있었다.

• Fire Science and Engineering
• /
• v.26 no.6
• /
• pp.51-56
• /
• 2012

The onset of fire on the International Space Station (ISS) is a critical problem that can threaten the life of crew members onboard and thus instantaneous fire detection and extinguishment technology has been considered as one of the most important aspects in the ISS operation. In the present study, a numerical analysis was performed to better understanding of the characteristics of smoke behaviors and detection in a pressurized module of the ISS using the NIST Fire Dynamic Simulator (FDS). Numerical results indicate that the smoke flow patterns under zero-gravity condition are clearly different from those under normal gravity condition. In addition, the results obtained from numerical simulations coupled with the PM internal flows are expected to provide basic and useful information in designing the microgravity fire detection devices and establishing in fire response protocol for astronauts or the crew members.

• Proceedings of the Korea Contents Association Conference
• /
• 2016.05a
• /
• pp.409-410
• /
• 2016

본 논문에서는 승강기 내에서 흡연을 하는 사람을 추출하도록 한다. 흰색 막대를 입에 물거나, 연기를 내품는 사람을 추출하는 것이다. 추출 방법은 장면 전환 검출에서 Average Intensity Measure를 이용하여 추출하도록 한다. 이렇게 추출하여 경찰청이나 법원에 포렌식 증거 자료로 제출하기 위해서이다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.01a
• /
• pp.391-394
• /
• 2023

일반적으로 화재 감지 시스템은 정확하고 빠르게 화재를 감지하는 것은 어려운 문제 중 하나이다. 본 논문에서는 U-net을 활용하여 기존의 화재(불) 영역 추출 기법으로 Segmentation으로 보다 정밀하게 탐지하는 기법을 제안한다. 이 기법은 화재 이미지에서 연기제거 및 색상보정을 통해 이미지를 전처리하여 화염 영역을 추출한 뒤 U-Net으로 학습시켜 이미지를 입력하면 불 영역의 Segmentation을 추출하도록 한다.

• Journal of Korea Society of Industrial Information Systems
• /
• v.17 no.7
• /
• pp.89-94
• /
• 2012

In fact, smoking is prohibited in elevators. It is morally wrong to smoke in elevators. In addition, smoking can be very fatal for our children and for women. In this paper, forensic evidence is submitted to court by people who smoke in elevators. Shots around the face of the person in the elevator extracted partially by scene change detection. Smokers is extracted that the white bar is at the mouth biter. People spouting smoke extraction will proceed in the future. It is extracted by using technology of color histogram, one of the scene change detection method. The extract is a much more accurate extraction ratio than the methods that do not use scene change detection.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.6
• /
• pp.1257-1262
• /
• 2018

In this paper, a deep learning method is proposed to detect the fire effectively by using video of surveillance camera. Based on AlexNet model, classification performance is compared according to kernel size and stride of convolution layer. Dataset for learning and interfering are classified into two classes such as normal and fire. Normal images include clouds, and foggy images, and fire images include smoke and flames images, respectively. As results of simulations, it is shown that the larger kernel size and smaller stride shows better performance.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.37-39
• /
• 2001

현재 이동 클라이언트의 발전과 무선 이동 데이터 통신의 발달에 의해 보다 정학한 공간 데이터 수정을 현장에서 수행 할 수 있게 되었다. 이러한 환경을 고려하여, 이 논문에서는 무선 이동 클라언트에서의 공간데이터 변경을 위해서 2-tier 트랜잭션 모델(2)을 이용한다. 이동 트랜잭션은 완료 즉시 서버에 재 접속할 필요가 없기 때문에 이동 트랜잭션의 완료 시점과 재 접속 후 베이스 트랜잭션으로 재 수행하는 시점 사이에 간격이 존재하게 된다. 그리고 고안 데이터 변경 트랜잭션은 교환가능한 트랜잭션이 아니며, 완전히 직렬가능(fully serializerability)해야 한다. 이러한 이유로 갱신 손실 문제(lost update problem)가 발생한다. 이 논문에서는 갱신 손실 문제를 해결하기 위하여 영역 잠금의 타입스탬프 값과 영역 잠금의 영역의 겹침을 이용하여 갱신 손실 가능한 공간객체 집합을 검출해내는 방법을 제시한다. 검출된 갱신 손실 가능한 공간 객체 집합의 완료 시점을 뒤로 연기(postpone)하는 프로토콜도 함께 제시한다.

• Journal of the Korean Society for Railway
• /
• v.12 no.1
• /
• pp.9-15
• /
• 2009

In this study, the l/35 reduced-scale model experiment were conducted to investigate designed ventilation system performance at rescue station in tunnel fire. A model tunnel with 2 mm thick of steel, 10 m long, 0.19 m high and 0.26m was made by using Froude number scaling law. The cross-passages installing escape door at the center. were connected between accident tunnel and rescue tunnel. The n-heptane pool fire, $4cm\times4cm$, with heat release rate 698.97W were used as fire source. The fire source was located in the center and portal of accident tunnel as Worst case.. An operating ventilation system extracted smoke amount of 0.015 cms. The smoke temperature and carbon monoxide. concentration in cross-passage were measured to verify designed ventilation system. The results showed that, in center fire case without ventilation in accident tunnel, smoke did not propagated to rescue station. In portal tire case, smoke spreaded to rescue station without ventilation. But smoke did not propagated to rescue station with designed ventilation.

• Fire Science and Engineering
• /
• v.24 no.2
• /
• pp.82-88
• /
• 2010

In this study, we have performed several tests for composite plastic materials to be applied on interior materials of a vehicle to identify their combustion characteristics using cone calorimeter, smoke density chamber and toxicity index chamber. We have prepared a total of 12 samples for 4 major parts of a vehicle wherein each major part has 3 different materials. The results of cone calorimeter test showed ignition time of PVC sheet and PVC leather were 2s. The 8 samples showed under less than 10s of ignition time. The sample comprising Nylon and PE had the biggest maximum heat release rate of 635 $kW/m^2$. The sample comprising Rubber showed the smallest maximum heat release rate but with the biggest total heat release. The results of smoke density chamber test showed the sample that is made up with Rubber had the biggest specific optical smoke density. The sample comprising PVC leather and PUR showed the biggest VOF4 which enables the initial smoke production. The results of toxicity index test showed that all samples contained carbon dioxide content exceeding its lethal concentration. The sample comprising PVC showed high content of hydrogen chloride and hydrogen bromide. The PVC sheet showed the biggest toxicity index calculated by using lethal concentration and test results. Toxicity index of all sample wes over 1.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.11
• /
• pp.1371-1376
• /
• 2019

In case of fire in a structure, it is difficult to suppress fire because it can not accurately grasp the location of fire in case of fire. In this paper, we propose a system algorithm that can guide the optimal evacuation route in case of deep learning-based (RNN) structure disaster. The present invention provides a service to transmit data detected by sensors to a server in real time by using installed sensor, to transmit and analyze information such as temperature, heat, smoke, toxic gas around the sensor, to identify the safest moving path within a set threshold, to transmit information to LED guide lights and direction indicators in a structure in real time to avoid risk factors. This is because the information of temperature, heat, smoke, and toxic gas in each area of the structure can be grasped, and it is considered that the optimal evacuation route can be guided in case of structure disaster.