Fire Detection Algorithms Using Multimodal ... - Bilkent University

CHAPTER 5. FLAME DETECTION USING PIR SENSORS 79Table 5.1: Results with 198 fire, 588 non-fire test sequences. The system triggersan alarm when fire is detected within the viewing range of the PIR sensor.No. of Sequences No. of False Alarms No. of AlarmsFire Test Sequences 198 3 195Non-Fire Test Sequences 588 0 0are issued is presented in Fig. 5.7.5.4 SummaryIn this chapter, a method for flame detection using PIR sensors is proposed.Analog signal from a PIR sensor is sampled with a sampling frequency of 50 Hzand quantized with 8 bits. Single level wavelet coefficients of the output signalare used as feature vectors for flame detection.PIR sensor output recordings containing various human movements andflames of paper and alcohol fire at a range of 5m are used for training the HMMscorresponding to different events. Thresholds for defining the states of HMMsare estimated using an evolutionary algorithm, since the underlying cost functionto be minimized has proved to be multi-modal and non-differentiable. Flame detectionresults of the proposed algorithm show that the single-pixel assumptionfor PIR sensor proves to be a correct one.This thesis demonstrates that low-cost PIR sensors which are commonly usedas indoor and outdoor motion detectors, can be utilized as fire sensors whencoupled with appropriate processing. The main advantage of a PIR based firedetection system over conventional particle sensors is its ability to detect thepresence of fire from a distance which results in a faster response time.