Résumé: The goal of multi-focus image fusion (MFIF) is to create a single, unified image from multiple images of the same scene with varying depths of focus of foreground and background patterns. To enhance the quality of the fused image, MFIF relies heavily on the precision of the detected focus area. In this paper, we propose a MFIF algorithm using weighted anisotropic diffusion filter (WADF) and a structural gradient. An edge-preserving filter can take the form of either a diffusion of the intensities at the borders or the detection of significantly meaningful edges in an image. WADF is a smart strategy for satisfying such requirements in multi-dimensional space. Image smoothing with an edge-preserving approach is used to create weight map pattern first. After that, the structural gradient-based focus area detection approach is applied to generate the fusion decision map pattern. Finally, a fused image is created by combining the weight map pattern with the fusion decision map pattern via the fusion rule. The performance of the suggested algorithm has been examined using both qualitative and quantitative methods, and it has been demonstrated to be more effective than a select number of other current methods in a number of different tests.

Langue: Anglais