Research on multisource image fusion of power equipment based on LatLRR and multiple filtering

The timely and accurate detection of power facilities is very important to ensure the reliability of energy supply, and a single sensor has certain limitations in the detection of power facilities. Therefore, a fusion method of IR and visible light images for power equipment based on LatLRR and multiple filtering is proposed. First, the multiscale Gaussian filter and bilateral filter are used to decompose the image into basic layers and three detail layers and select different information fusion methods according to different scale levels. Second, multi-source images were decomposed into low-rank parts and significant parts using LatLRR, and different scale hierarchies were summed to compensate for the lack of filter decomposition in terms of image detail and contrast. Finally, the final fusion image is weighted by summing the corresponding scale hierarchy of LatLRR and multiple filtering to minimize the information loss. We conducted subjective and objective experiments on multiple public data sets. This method is better than other methods on multiple evaluation indicators and verify the effectiveness of this method in the field of image fusion of power equipment. This method can effectively capture the rich overall structure and local structure information in infrared and visible light source images.