• 한국멀티미디어학회논문지
• /
• 제18권10호
• /
• pp.1197-1204
• /
• 2015

Smoke detection plays an important role for the early detection of fire. In this paper, we suggest a newly developed method that generated LBPV(Local Binary Pattern Variance)s as special feature vectors from RGB contrast images can be applied to detect smoke using SVM(Support Vector Machine). The proposed method rearranges mean value of the block from each R, G, B channel and its intensity of the mean value. Additionally, it generates RGB contrast image which indicates each RGB channel’s contrast via smoke’s achromatic color. Uniform LBPV, Rotation-Invariance LBPV, Rotation-Invariance Uniform LBPV are applied to RGB Contrast images so that it could generate feature vector from the form of LBP. It helps to distinguish between smoke and non smoke area through SVM. Experimental results show that true positive detection rate is similar but false positive detection rate has been improved, although the proposed method reduced numbers of feature vector in half comparing with the existing method with LBP and LBPV.

• 한국산학기술학회논문지
• /
• 제14권9호
• /
• pp.4472-4477
• /
• 2013

본 논문은 기존 논문들에서 사용되었던 다양한 색상모델의 연기색상을 비교분석하여, 화재 영상감시 시스템의 연기 검출에 최적인 컬러모델을 제시하기 위한 컬러영상의 연기색상 분석에 대하여 기술한다. 각 표준 색상 모델에서의 연기색상과 비연기 색상간의 분리도 특성을 비교하기 위하여 히스토그램 교차 분석 기법을 사용하였다. 표준색상모델로는 RGB, YCbCr, CIE-Lab, HSV 컬러모델을 사용하였으며, 계산된 히스토그램 교차(Histogram Intersection)값이 작으면 연기와 비연기 영역분할 특성이 우수한 컬러모델이며 큰 값을 가지는 컬러모델에서는 연기분할 특성이 좋지 않다. 4개의 표준 컬러모델을 분석한 결과, RGB 색상모델과 HSV 색상모델이 각각 평균 히스토그램 교차 값이 0.14, 0.156 으로서 연기와 비연기 색상 분리도가 매우 우수하여 컬러영상의 색상기반 연기검출에 가장 최적이며 실용적인 컬러모델로 확인되었다.

• 한국화재소방학회논문지
• /
• 제17권4호
• /
• pp.86-91
• /
• 2003

공기조화설비가 설치된 감시공간에서의 화재 시 열$.$연기기류의 유동형상은 정상 유동해석과 다른 기류 유동형상을 나타내어 화재감지기의 응답특성지연 해결 및 연기감지농도를 향상, 화재초기에 경보를 발생하여 인명피해 및 재산피해를 최소화하기 위한 성능위주의 화재감지장치 개발을 위한 필요성에 의해 연구를 수행하였다. 본 연구는 높은 연기응답특성을 가지며 공기순환에 의한 응답특성지연에 영향을 받지 않는 능동형태의 연기감지장치로서 감시공간의 공기를 공기흡입관을 통하여 연기농도 분석장치로 흡입하여 연기를 감지하는 공기흡입형 광전식 연기감지장치 개발에 필요한 연기농도 분석기술 및 공기흡입관을 통한 균등 공기흡입기술에 대하여 수행하였다. 연구결과, 공기흡입배관에 설치된 흡입구를 통한 공기흡입이 균등하게 이루어져 균일한 감도특성을 나타내어 공기순환에 의한 연기감지의 지연에 영향을 받지 않으며 연기감지성능은 수동형태의 연기감지기보다 우수한 응답특성을 나타내었다.

• 방재기술
• /
• 통권19호
• /
• pp.17-21
• /
• 1995

This is to study installation standards of aspirating devices and detectors, which are components of ADS(aspirating fire detection system). The BFPSA(British Fire Protection System Association) code was mainly referred is studying. ADS aspirated air and smoke through the pipe and then checks if there is fire or not. It is now in the limelight because it can early alarm in case of fire and prevent false-alarming.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.1149-1151
• /
• 2003

We performed a validation study of the MODIS active fire detection algorithm using high resolution SPOT image as the reference data set. Fire with visible smoke plumes are detected in the SPOT scenes, while the hotspots in MODIS data are detected using NASA's new version 4 fire detection algorithm. The detection performance is characterized by the commission error rate (false alarms) and the omission error rate (undetected fires). In the Sumatra and Kalimantan study area, the commission rate and the omission rate are 27% and 34% respectively. False alarms are probably due to recently burnt areas with warm surfaces. False negative detection occur where there are long smoke plumes and where fires occur in densely vegetated areas.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권3호
• /
• pp.1301-1314
• /
• 2018

In this paper, we present a low-cost fire detection system using a thermal camera and a smartphone. The developed system collects thermal and RGB videos from the developed camera. To detect fire, candidate fire regions are extracted from videos obtained using a thermal camera. The block mean of variation of adjacent frames is measured to analyze the dynamic characteristics of the candidate fire regions. After analyzing the dynamic characteristics of regions of interest, a fire is determined by the candidate fire regions. In order to evaluate the performance of our system, we compared with a smoke detector, a heat detector, and a flame detector. In the experiments, our fire detection system showed the excellent performance in detecting fire with an overall accuracy rate of 97.8 %.

• International Journal of Advanced Culture Technology
• /
• 제9권4호
• /
• pp.260-267
• /
• 2021

In this paper, we intend to use a new mixed model of YoloV5 and DeepSort. For fire detection, we want to increase the accuracy by automatically extracting the characteristics of the flame in the image from the training data and using it. In addition, the high false alarm rate, which is a problem of fire detection, is to be solved by using this new mixed model. To confirm the results of this paper, we tested indoors and outdoors, respectively. Looking at the indoor test results, the accuracy of YoloV5 was 75% at 253Frame and 77% at 527Frame, and the YoloV5+DeepSort model showed the same accuracy at 75% at 253 frames and 77% at 527 frames. However, it was confirmed that the smoke and fire detection errors that appeared in YoloV5 disappeared. In addition, as a result of outdoor testing, the YoloV5 model had an accuracy of 75% in detecting fire, but an error in detecting a human face as smoke appeared. However, as a result of applying the YoloV5+DeepSort model, it appeared the same as YoloV5 with an accuracy of 75%, but it was confirmed that the false positive phenomenon disappeared.

• 한국대기환경학회지
• /
• 제27권6호
• /
• pp.681-691
• /
• 2011

The size distributions of smoke particles from fire are prerequisite for the studies on fire detection and adverse health effects. Above the flame of the fire, coagulation dominates and the smoke particles grow from 1 to 50 nm up to 100 to 3,000 nm, sizes ranging from the free-molecular regime to the continuum regime. The characteristics of the agglomeration of the smoke particles are well known, independently for each of the free-molecular and continuum regimes. However, there are not many systematic studies in the entire regime by the complexity of the mechanisms. The purpose of this work is to find the characteristics of the development of the size distribution of smoke particles by agglomeration in the entire size range covering the free-molecular regime, via transition regime, to the near-continuum and continuum regime for each variation of parameters such as fractal dimension, primary particle size and dimensionless coagulation time. In this work, the dynamic equation for the discrete-size spectrum of the particles was solved using a nodal method based on the modification of a sectional method. In the calculation, the collision frequency function for the entire regime, which is derived by using the concept of collision volume and general enhancement function, was applied. The self-preserving size distribution for the entire regime is compared with the ones for the free-molecular or continuum regimes for each variation of the parameters.

• 한국재난정보학회 논문집
• /
• 제18권2호
• /
• pp.364-373
• /
• 2022

연구목적: 본 논문은 지하공동구의 초기 화재 감지를 위해 CCTV를 활용한 AI 연기 객체 감지 모델을 개발하는데 목적이 있다. 연구방법:비정형성이 높은 연기 객체의 감지 성능을 제고하기 위해 화재 감지에 특화된 딥러닝 객체 감지 모델을 지하공동구 연기 감지에 특화되도록 학습시켰고, 학습데이터셋의 정제 및 학습 중 Gradient explosion 완화 등 감지 성능 개선을 위한 방법들을 적용해 모델 결과를 비교하였다. 연구결과: 결과는 제안된 방법을 통해 모델 성능을 향상시켰고 mAP 등의 지표를 평가를 통해 개발 모델이 우수한 성능을 보유하고 있음을 보여준다. 최종 모델은 지하공동구 환경의 연기에 대해 미탐이 낮은 반면 오탐이 다수 발견되는 성능을 보였다. 결론: 본 논문의 모델은 지하공동구 관리시스템과 연계를 통해 보완함으로써 지하공동구의 연기 객체 감지에 활용할 수 있을 것으로 판단된다.

• 한국안전학회지
• /
• 제37권3호
• /
• pp.16-23
• /
• 2022

This study aims to provide a fundamental material for identifying fire and no-fire signals using the detection signal characteristics of IoT-based fire-detection systems. Unlike analog automatic fire-detection equipment, IoT-based fire-detection systems employ wireless digital communication and are connected to a server. If a detection signal exceeds a threshold value, the measured values are saved to a server within seconds. This study was conducted with the detection data saved from seven fire accidents that took place in traditional markets from 2020 to 2021, in addition to 233 fire alarm data that have been saved in the K institute from 2016 to 2020. The saved values demonstrated variable and continuous VC-Signals. Additionally, we discovered that the detection signals of two fire accidents in the K institution had a VC-Signal. In the 233 fire alarms that took place over the span of 5 years, 31% of smoke alarms and 30% of temperature alarms demonstrated a VC-Signal. Therefore, if we selectively recognize VC-Signals as fire signals, we can reduce about 70% of false alarms.