Implementation of LBP based image mosaicing algorithm on FPGA using Nios II softcore

Since a long time, image mosaicing technique was involved in different domains; in order to have much more information about a scene than the individual views of that scene. The basic idea of this technique is to match features between two overlapped images; in order to find a suitable transformation model (Homography matrix), which allows warping images into a single and common reference frame. Different methods were developed for matching points-based features between overlapped images. Recently, researchers focus on using Local Binary Patterns Descriptors (LBPDs) in order to match features between any two successive images. Due to the simplicity of programming the LBP based matching technique; it can be used for hardware implementation. In this paper, some related works will be discussed in the state of the art, then a general framework for implementing LBP based image mosaicing algorithm on FPGA using the embedded processor Nios II will be illustrated.