Temperature rising recognition of IR image of electrical equipment based on seeded region growing

A novel automatic seeded region growing method (SRG) for high temperature region extraction of IR image is proposed in this paper. The proposed method efficiently utilizes the RGB values of IR image. The neighborhood average method is selected to reduce image noise. Red component and green component which are more efficient to extract high temperature region because their values are positive correlated with temperature, are extracted to make segmentation using seeded region growing method. Seed points are determined as the connected pixels which exceed a certain threshold, and the grow criteria is based on the gradient of the two component images. Intersection method is used for image fusion, and high temperature region is extracted. Experiment results indicate that the method can accurately recognize the high temperature area, and the outline is clear.