The region of interest is captured using CCD camera and identified by smoke sensor in the wireless sensor node. The physical and dynamic characteristics and colour information of interesting region can be obtained with an application of the digital image processing algorithm. The fire source is identified according to the acquired characteristics and smoke level. The system is based on the continuous image sampling. The experimental results show that the system can accurately identify and confirm the fire. The video sensor node is designed with the sensors such as MQ2 sensor for smoke sensing, SHT75 sensor for temperature and humidity sensing, OPT101 sensor for light sensing and CCD camera. By combining sensor output with image output, the false alarm rate is zero and improves the stability. The amount of data processing can be reduced because of the use of proposed algorithm and thus shorten the execution time and storage. Alarm is activated only for fire image and fire incidents.

Keywords: fire detection; fire characteristics; image process; image segmentation; smoke sensor; video sensor node;