High dynamic range panoramas using a mobile phone

Abstract

High dynamic range (HDR) imaging is becoming more popular in the field of photographic and computer graphics. It is a technique that uses both image processing and photographic techniques. Even now smart phone camera becomes more and more powerful which is comparable to Digital-Single-lens reflex (DSLR) camera. With a mobile phone, HDR images can be taken easily as the image processing is done via the software. They are made by taking multiple images with different exposures and then combine them into a single image to reproduce greater dynamic range. Consequences, it will take longer time to take a HDR image than a normal image. There are many existing applications for creating 360 degrees spherical panoramas by stitching images taken on a mobile phone. For instance, Google Street View. However, for scenes in which brightness varies dramatically with direction, such panoramas do not retain the original contrast due to auto-gain correction on the cameras. In this project, a different approach is used based on creating a HDR panorama from a sequence of Low Dynamic Range (LDR) or Standard Dynamic Range (SDR) images by additionally taking into account the exposure settings captured in EXIF metadata. HDR panorama can be created by converting each image to HDR and stitch them together instead of using exposure bracketing. All the photos are taken using a mobile phone and then do post processing. The advantage of this method is the possibility of making HDR Panoramas from mobile phone. There are many ways to obtain HDR images. The most common way is to use photograph of the same scene taken with different exposure values. HDR image means that it has a greater dynamic range compared to LDR image. It also stores more information than LDR image. In this paper I will focus on undoing the auto-gain correction by the mobile phone. Various experiments are conducted to validate the correctness of the proposed HDR panorama imaging scheme. Results of implementation of the HDR panorama are discussed and presented.