應用電腦視覺於即時火與煙霧之偵測

Fire and Smoke Detection Using Machine Vision Techniques

戴文盛(Wen-ShengTai 戴文盛)

指導教授 ： 李錫捷

元智大學/資訊學院/資訊管理學系/碩士(2012年)

摘要

火災容易造成重大傷亡與損失，目前大部分的偵測火災的裝置主要包含煙霧感測器、溫度感測器等等，然而這些感測器有些會受限於場所的應用，有些則會受限於與火場的距離或是火的大小，溫度感測需要火延燒到一個足以啓動它的程度，故常有誤認火警的情況，如此往往無法達到預期的防火效果，更無法記錄起火原因及演變成火災的過程。因此本論文提出以視訊影像處理的方法偵測火焰與煙霧。
本論文藉由計算影像序列中每一個(Frame)之間飽和度顏色直方圖間的L_1距離，藉由這閥值與時間變化來判別是否有火源與煙霧，再利用運動歷史影像(MHI)移動物件偵測演算法，進行色彩擷取火焰顏色與煙霧顏色特徵，最後透過標記化來分辨火源與煙霧在像素影像中可能出現火源與煙霧的位置。

關鍵字

火災偵測 ； 煙霧偵測 ； 運動歷史影像

並列摘要

Accidental fire can cause considerable damages and heavy casualties. Most of the fire detectors nowadays include smoke detectors and heat detectors, but the equipment is restricted by the place, the distance and the range of fire. When the fire grows and exceeds the limit of the heat detector, the alarm system of the heat detector will start. But sometimes the rate of false fire alarms is too high so the prediction of fire prevention is not as good as effected. Furthermore, the cause and the process of the fire could not be recorded. Therefore, this dissertation will discusses how to detect fire and smoke by using video and image processing.
In this paper, we will calculate the L1 saturation distance of the image sequence in the frame and use the threshold value in time series to differentiate the fire and the smoke. Moreover, we will use motion history images to recognize the moving object and find both the fire and the smoke color features within the moving object as the accident occur. Finally, we will mark the location of the fire and smoke in frame.

並列關鍵字

Fire Detection ； Smoke Detection ； MHI(Motion History Image)

參考文獻

[18] 陳家煌, "Java自由軟體平台之即時火災偵測與查詢監視系統," 碩士, 資訊工程學系, 國立暨南國際大學, 南投縣, 2009.

[15] 陳致宇, "應用機器視覺技術於火災偵測之研究," 碩士, 機電整合研究所, 國立臺北科技大學, 台北市, 2011.

[1] I. O. Sebe, J. Hu, S. You, and U. Neumann, "3d video surveillance with augmented virtual environments," 2003, pp. 107-112.

[2] H. Weiming, T. Tieniu, W. Liang, and S. Maybank, "A survey on visual surveillance of object motion and behaviors," Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, vol. 34, pp. 334-352, 2004.

[4] A. F. Bobick and J. W. Davis, "The recognition of human movement using temporal templates," Pattern Analysis and Machine Intelligence, IEEE Transactions on, vol. 23, pp. 257-267, 2001.