Název: A real-time method for smoke detection in monitored forest areas
Autoři: Sedlák, Vladimír
Stopjaková, Viera
Brenkuš, Juraj
Citace zdrojového dokumentu: 2017 International Conference on Applied Electronics: Pilsen, 5th – 6th September 2017, Czech Republic, p.205-210.
Datum vydání: 2017
Nakladatel: Západočeská univerzita v Plzni
Typ dokumentu: konferenční příspěvek
conferenceObject
URI: http://hdl.handle.net/11025/35438
ISBN: 978–80–261–0641–8 (Print)
978–80–261–0642–5 (Online)
ISSN: 1803–7232 (Print)
1805–9597 (Online)
Klíčová slova: detekce kouře;zpracování obrazu;odečítání pozadí;pohybové funkce
Klíčová slova v dalším jazyce: smoke detection;image processing;background subtraction;motion features
Abstrakt v dalším jazyce: An improved real-time method for smoke detection based on motion and color analysis is presented. Smoke detection in video surveillance is very important for early fire detection. In general, it is assumed that smoke is a low frequency signal, which may smooth the background. However, these conditions can be met by some pure-color objects, and smoke also produces high frequency signal on the edges. An improved smoke detection method was proposed in order to prevent these problems. The presented technique uses well known background subtraction for identifying of moving regions as well as novel color based algorithm for smoke detection. Detection efficiency is improving using a classifier that provides the identification of smoke-like regions. The main advantage is that it is not required to process each frame of the video signal for sufficient accuracy. Sampling rate was set to a few second in order to find the optimum value between performance and system requirements. Performance of the proposed method was evaluated using different kinds of the input signal, where satisfactory results were achieved.
Práva: © Západočeská univerzita v Plzni
Vyskytuje se v kolekcích:Applied Electronics 2017
Applied Electronics 2017

Soubory připojené k záznamu:
Soubor Popis VelikostFormát 
Sedlák.pdfPlný text288,68 kBAdobe PDFZobrazit/otevřít


Použijte tento identifikátor k citaci nebo jako odkaz na tento záznam: http://hdl.handle.net/11025/35438

Všechny záznamy v DSpace jsou chráněny autorskými právy, všechna práva vyhrazena.