Real-time Smoke Detection Research with False Positive Reduction using Spatial and Temporal Features based on Faster R-CNN
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Fire must be extinguished as quickly as possible because they cause a lot of economic loss and take awayprecious human lives. Especially, the detection of smoke, which tends to be found first in fire, is of greatimportance. Smoke detection based on image has many difficulties in algorithm research due to the irregular shapeof smoke. In this study, we introduce a new real-time smoke detection algorithm that reduces the detection offalse positives generated by irregular smoke shape based on faster r-cnn of factory-installed surveillancecameras. First, we compute the global frame similarity and mean squared error (MSE) to detect the movement ofsmoke from the input surveillance camera. Second, we use deep learning algorithm (Faster r-cnn) to extractdeferred candidate regions. Third, the extracted candidate areas for acting are finally determined using space andtemporal features as smoke area. In this study, we proposed a new algorithm using the space and temporalfeatures of global and local frames, which are well-proposed object information, to reduce false positives based ondeep learning techniques. The experimental results confirmed that the proposed algorithm has excellentperformance by reducing false positives of about 99.0% while maintaining smoke detection performance.
Abstract
Fire must be extinguished as quickly as possible because they cause a lot of economic loss and take awayprecious human lives. Especially, the detection of smoke, which tends to be found first in fire, is of greatimportance. Smoke detection based on image has many difficulties in algorithm research due to the irregular shapeof smoke. In this study, we introduce a new real-time smoke detection algorithm that reduces the detection offalse positives generated by irregular smoke shape based on faster r-cnn of factory-installed surveillancecameras. First, we compute the global frame similarity and mean squared error (MSE) to detect the movement ofsmoke from the input surveillance camera. Second, we use deep learning algorithm (Faster r-cnn) to extractdeferred candidate regions. Third, the extracted candidate areas for acting are finally determined using space andtemporal features as smoke area. In this study, we proposed a new algorithm using the space and temporalfeatures of global and local frames, which are well-proposed object information, to reduce false positives based ondeep learning techniques. The experimental results confirmed that the proposed algorithm has excellentperformance by reducing false positives of about 99.0% while maintaining smoke detection performance.
참고문헌 (10)
- B. C. Ko, "Modeling and Formalization of Fuzzy Finite Automata for Detection of Irregular fire Flames", IEEE Transaction on Circuits and System Video Technology, 21(12), pp.1903~1912 (2011)[DOI]
- J. Gubbi, "Smoke detection in video using wavelets and support vector machines", Fire Safety Journal, 44(8), pp.1110~1115 (2009)[DOI]
- B. C. Ko, "Wildfire smoke detection using temporospatial features and random forest classifiers", Optical Engineering, 51 (2012)[DOI]
- T. Chen, "An Early FireDetection Method Based on Image processing", 2004 International Conference on Image Processing, pp.1707~ (2004)
- T. B. CelikT, "Wavelet Based Real-time Smoke Detection in Video", EUSIPO 2005 (2005)
- C. Yu, "A Real-time Video Fire Flame and Smoke Detection Algorithm", Procedia Engineering, 62, pp.891~898 (2013)[DOI]
- S. Frizzi, "Convolutional Neural Network for Video Fire and Smoke Detection", IECON 2016 (2016)
- Y. LeCun, "Gradient-based Learning Applied to Document Recognition", Proceeding of IEEE, 86(11), pp.2278~2324 (1998)[DOI]
- A. Krizhevsky, "ImageNet Classification with Deep convolutional neural Network: Advanced in Neural Information Processing System 25", 26th Annual Conference on Neural Information Processing Systems, pp.1106~1114 (2012)
- G. E. Hinton, "Improving neural networks by preventing co-adaptation of feature detectors", Clinical Orthopedics and Related Research, 2012, pp.1~18 (2012)