Fusion Algorithm Based On Infrared And Visible Images And Its Application In Power Equipment Detection

In on-line monitoring and fault detection of power equipment,multi-sensor images can describe the characteristics of power equipment from different angles.The infrared image is not restricted by the environment and can reflect the temperature information of the device.Visible images can clearly show the details of the device’s appearance.However,the two images are often very different in imaging mechanism,resolution,and field of view.In order to meet the actual application requirements,this paper adopts an effective fusion method for the same scene in different situations.Combined with the advantages of the two kinds of images,the operation fault of power equipment can be recognized in time and corresponding measures can be taken.The main research contents of this paper are as follows:(1)A visible-infrared power image fusion method is proposed.First,infrared and visible source images are decomposed into low and high frequency subband coefficients using NSCT.NSCT is a completely multiscale,displacement invariant,multidirectional transformation method.It achieves translation invariance by subband decomposition similar to contourlet transform without up-down sampling.It is used to represent the contour of the image,which overcomes the pseudo Gibbs phenomenon.(2)Multi-scale decomposition of low-frequency coefficients is carried out by rolling guided filtering and rolling filtering.Low frequency coefficients are further decomposed into base and detail layers to preserve edges and reduce flare.This is because image fusion based on multi-scale decomposition maintains the importance of scale separation by filtering with different fuzzy degrees.Therefore,rolling guided filtering not only considers the scale perception information,but also preserves the edge.(3)The salient fusion rule is adopted to fuse the base layer to retain the salient visual details in the complementary image.The detail layer of maximum absolute value rule is used to extract salient features and edges.The low frequency coefficients of NSCT are reconstructed using the multiscale inverse transformations corresponding to the fusion base and detail layers.The fusion rules based on consistency verification are used to fuse the high frequency coefficients of NSCT.Inverse NSCT is performed on the fused low and high frequency coefficients to obtain the fused image.(4)The above methods are compared and analyzed experimentally.Experiments are carried out from both objective and subjective aspects and used in the image fusion of power equipment to obtain the effect diagram after fusion,which is helpful for relevant personnel to identify and judge the fault points of power equipment.