Rancang Bangun Pendeteksi Api Berbasis Vision untuk Pengendalian Kamera Pengawas

Abstract

Fires is a disaster caused by flame or uncontrolled burning that can endanger human life, buildings and ecology. Fires are unpredictable when it will happens. For examples forest fire, Fires in housing or building can occur during the day or night. The use of CCTV/webcam which can monitor the location for 24 hours is very effective to prevent the occurrence of fire. Therefore, we built a system that can detect early fire based on image processing by using the method of HSV color segmentation . The system processes the data image which captured by the webcam in real time through the LabVIEW software. The fire detection process is done by using Image enhancement method and combined with result of HSV color segmentation to produce the segment object in the form of blob so that it can know the number of detected object and the center point of the fire object. Based on the results of testing and analysis, it is concluded that user control in terms of color sampling and color tolerance of fire objects have an important role in the segmentation process. The accuracy of the system that has been made is 98% with error of 0.25% for the detection conditions at night. The system is able to detect the presence of fire in real time around 30 fps at a resolution of 1024 x 768. Error detection is mainly caused by light reflections from lights or light sources on the surface of shiny objects.