UAV-Borne Thermal Images Registration Using Optimal Gradient Filter

In this study, a novel method for Unmanned Aerial Vehicle (UAV)-borne thermal image registration is proposed. In the proposed method, an optimum gradient filter (OGF) is used to extract details from images and the filtered images are used for image reconstruction. This approach can enhance image texture in some areas within the image. The gradient filter coefficients are optimized using particle swarm optimization and for the image matching process, scale invariant feature transformation (SIFT) algorithm has been implemented. The proposed Methods (OGF-SIFT) is tested on eight paired stereo thermal high and low-resolution images acquired from two thermal sensors mounted on a UAV-borne thermal platform. After all, this method is compared to other thermal image matching methods. For the performance evaluation, random sample consensus and 2D-projective transformation are adopted. The results demonstrate that the proposed method outperforms other methods in terms of the number of true matches.