Fire detection based on flame color and area

• Published in: 2012 IEEE International Conference on Computer Science and Automation Engineering Vol. 3; pp. 222 - 226
• Main Authors: Wenhao Wang, Hong Zhou
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2012
• Subjects: Accuracy; connected regions; contour extraction; Feature extraction; fire detection; Fires; flame color; flame recognition; Image color analysis; Image edge detection; Light sources; Noise
• ISBN: 1467300888; 9781467300889
• DOI: 10.1109/CSAE.2012.6272943

Contour Extraction of Flame Object is a key problem in fire recognition Based on Video Sequences applications, which has a direct impact on improving fire recognition accuracy. To extract flame object precisely, this paper introduces a kind of the new method of extraction flame object based on the threshold of the area. Firstly, adaptive threshold generated by iterative method is used to segment the image, Secondly, we use knowledge of set theory to extract object contour, Finally, it can judge whether fire occurs from the fire flame color, fire spreading etc characteristic information. Experiments proved that the method is effective and robust, and that it can extract flame object effectively from a sequence of image, and that it is significant for improving accuracy and reducing false alarms. So it is good method of fire detection in the complex large outdoors situation.