基於卷積神經網路及色彩影像技術之火焰辨識;Flame Identification Based on Convolutional Neural Network and Color Image Technology

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Electrical Engineering > Electronic Thesis & Dissertation > Item 987654321/77691

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/77691

Title:	基於卷積神經網路及色彩影像技術之火焰辨識;Flame Identification Based on Convolutional Neural Network and Color Image Technology
Authors:	周哲伍;Chou, Che-Wu
Contributors:	電機工程學系
Keywords:	火焰辨識;視訊監控;卷積神經網路;flame recognition;video surveillance;convolutional neural network
Date:	2018-07-25
Issue Date:	2018-08-31 14:52:52 (UTC+8)
Publisher:	國立中央大學
Abstract:	中文摘要本篇論文提出一種深度學習的演算法藉由卷積神經網路演算法以及色彩空間分析，同時結合了背景相減法、形態學、角點偵測以及感興趣區域，來實現連續影像火焰追蹤辨識。近幾年，智慧型家庭監控系統的崛起並且基於即時影像的技術更加成熟，所以對於火焰的防範，有別於過去較傳統的方式來偵測，例如煙霧或溫度感測器，但相較於即時影像偵測，大多為時已晚。就如近期有一場火災奪走七條消防員的生命，如果有此監控影像的使用，也許就不會有這個悲劇發生，因此本文所探討的系統是希望在火焰開始擴展以前，藉由即時影像進行火焰辨識，進而告知人們，以至於達到降低人員及財物的損失。 ;Abstract This paper proposes a deep learning algorithm by using “convolutional neural network algorithm” and “color space analysis” for real-time recognition, which combines background substraction, morphology, corner detection and regions of interest to achieve the flame tracking recognition of continuous image. In recent years, the rise of smart home monitoring systems and the technology based on real-time image have become more and more mature, so the preventive measure of fire is very different from traditional methods, such as smoke sensors and temperature sensors. Therefore, compared to the real-time image detection, the traditional methods are too slow to measure the prevention. As it happened recently, there are seven firefighters died on the fire disaster. If the surveillance image is used, there may not be this tragedy. As a consequence, the system discussed in this article can notify people by using real-time image to flame recognition before the flame begin to expand, so as to reduce the loss of personnel and property.
Appears in Collections:	[Graduate Institute of Electrical Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	196	View/Open

社群 sharing

Loading...