Exposure Fusion Algorithm for Scenes with an Extremely High Dynamic Range

Nowadays, high dynamic range imaging techniques are often applied for outdoor applications that may encounter scenes with an extremely high dynamic range. In such scenes, a halo can be produced by gradient reversal, and color information is easily damaged. In this paper, we propose an exposure fusion algorithm that does not suffer from any artifacts in extremely high dynamic range scenes. Using the HSV color model, gradient reversal is suppressed in the Value component, and color information is preserved in the other components. Experimental results show that these two artifacts are effectively avoided, and all of the scene's details are preserved.