Smoke detection using image processing

Abstract

In this report, we have proposed a smoke detection method from image. The Smoke Detection using Image Processing (SDIP) is based on four steps. First, Smoke image acquisition and pretreatment is performed. Second, Preprocessing of collected images. Which include converting RGB to HSV, Binarization, Thresholding, Third, Selecting smokes principals. Fourth, Smoke detection based on principals considering Saturation, Intensity and Area of the smoke. The Smoke Detection using Image Processing (SDIP) System is implemented by the MATLAB simulation. We use 46 images for calculating threshold value and 87 images for the identifying smoke. We use some statistical equations to calculate threshold value. We compare our result by trial and error process. We use some graphs to show the result. We also compare some recommended algorithm with our proposed algorithm.