Details-preserving multi-exposure image fusion based on dual-pyramid using improved exposure evaluation

Abstract

Due to a limited dynamic range of widely used image recorders, it is difficult to record complete information on real scenes using a single image, and a restricted range of contrast, brightness and chromaticity can only be recorded. To mitigate this matter, a set of images of the identical scenery could be firstly captured at different exposure situations, and next be merged into an informative image via image fusion. In the paper, we present a well details-preserving image fusion technique via improved exposure evaluation and dual-pyramid model. The proposed method owning to its advantages of cost-effective, computation-stable and adaptive image processing can achieve a high dynamic range image from a set of multiple exposure sequences even for extremely complicated scenes. Experimental results demonstrated that the proposed method has richer details and better visual effects compared with the other commonly used techniques in most cases. Therefore, it could provide useful help and inspiration for image processing and enhancement field such as digital photography, remote sensing and medical imaging.

Keywords

• Image fusion
• Exposure evaluation
• Image quality assessment
• Pyramid algorithm