Robust aerial image mosaicing algorithm based on fuzzy outliers rejection

The use of unmanned aerial vehicles (UAVs) imagery for acquiring data is constantly evolving, due to the ease of use and low data acquisition costs. All of these made UAVs very popular with end customers and data acquisition companies. For most cases, the acquired UAV images need further more processing before being analyzed, and that because of changing of illumination condition in aerial environment and fog generated because of UAV flying speed. Image mosaicing is a practical solution for these problems; in which overlapped views of the same scene are combined to form a large image with high quality. The common problem associated with image mosaicing algorithms is the false associations (outliers) produced when defining the overlapping region between every two successive views. This article presents an image mosaicing algorithm based on efficient fuzzy technique for outliers rejection. Our proposed technique is based on using RANdom SAmpling Consensus (RANSAC) and bidirectional approaches with a fuzzy inference system in order to separate between inliers and outliers. The experimental results prove that the proposed method has good performance for aerial images and gives better results when compared with other techniques. Thus our approach is insensitive to the ordering, orientation, scale and illumination of the images.