A new approach for flame image edges detection

In this paper, a simple and robust approach for flame image analysis is presented. It is based on the Local Binary Patterns and thresholding techniques. The main features of an image are obtained by Local Binary patterns, and two level thresholding is used to make the edges visible and clear. The simplicity and robustness of the proposed approach in noisy environment makes it suitable for the subsequent analysis of the flame features. Various experimentations are carried out on synthetic as well as real images. The results show the proposed approach gives good localization and effective edge detection as compared to the existing methods.