Image Fusion and IoT

Abstract

High dynamic range (HDR) imaging can capture an extremely wide range of illumination present in a scene. This document explains the problem of limited dynamic range in the standard imaging pipeline, and then present algorithms for composing ghost-free HDR images from a sequence of dierently exposed LDR images with motion. Firstly, we detect motion regions between LDR images, and then estimate the relative shutter speeds in the static regions. We generate dierent weights in each region to compose HDR images. Finally, we demonstrate the composing experiments using real LDR images, which is shown that the proposed composing algorithm can acquire ghost-free HDR images.