High dynamic range image matching using radial feature descriptors

Obtaining a top match for a given query image from a set of images forms an important part of the scene identification or scene reconstruction process. The query image might not be exactly similar to the images in the data set, with possible variations including change in scale, change in viewing angle and change in the lighting conditions. Designing an image matching algorithm which is invariant to all these transforms forms an important part of this research. When the information obtained from images captured from normal image capturing devices are used for matching, certain amount of detail in the scene is lost due to the encoding of the images at the capture time. Using high dynamic range images with the purpose of utilizing all the details obtained at the time of capture for image matching is proposed in this thesis. Once the high dynamic range images are obtained through the fusion of low dynamic range images, feature detection is performed on the query images as well as on the images in the database. A junction detector algorithm is used for detecting the features in the image. The features are described using the wedge descriptor which is modified to adapt to high dynamic range images. Once the features are described, a voting algorithm is used to identify a set of top matches for the query image. The images are subjected to color correction to enhance the accuracy of the match.