Open AccessA novel visible-infrared image fusion method based on visual enhancement and multiscale decomposition
Author(s) -
Lingxiao Li,
Yong Feng,
Zezhong Ma
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2010/1/012141
Subject(s) - artificial intelligence , computer vision , contrast (vision) , image fusion , computer science , fusion , pattern recognition (psychology) , filter (signal processing) , scale (ratio) , image (mathematics) , noise (video) , binary image , fusion rules , image processing , physics , philosophy , linguistics , quantum mechanics
Aiming at the problems of low contrast, low signal-to-noise ratio and scattered energy in the field of optical detection and imaging, a visual enhancement and multi-scale decomposition method for fusion of visible and infrared images is proposed based on the visual characteristics of different bands’ image. Firstly, the infrared image with less texture information and low contrast is preprocessed to weaken the background noise and improve the visual contrast. On this basis, the multi-scale image is decomposed by detail preserving filter, and the saliency map of each scale image is obtained by saliency extraction method. Then, the multi-source images of each scale are fused with saliency map, and binary filtering fusion rules are adopted for the regions with salient details while weighted fusion rules are adopted for other regions. Finally, the images of all scale are reconstructed to get better fusion results. The experimental results show that this method can significantly improve the visual contrast of the fused object, and the objective evaluation indexes are superior to other comparison methods.