Detect Flame Fire Using Fractal Geometry in Color Digital Images

Abstract

Pattern recognition, which takes features extraction as a basis for decision making, is considered as of the cutting-edge technologies. It is used in various useful applications such as tracking and monitoring objects. In this research an algorithm for detecting fire flame in the colored digital images is built. The algorithm basically depends an extracting the features of flame spots with all its spectra with reference to fractal dimension. The input image is cut into a set of equal dimension squares, then fractal dimension for each square is calculated using two-dimension variation algorithm, which is are of the algorithms used in calculating fractal dimension. Fractal dimension values in the output fractal dimension array are analyzed to detect flame spots using the computer through determining the common features and characteristics of the flame with all its spectra.

Keywords

• pattern recognition
• fire flame
• fractal geometry