Research on video stabilization technology of underwater robot based on SIFT and LBP fusion algorithm

In this paper, an algorithm for fusion of SIFT (scale-invariant feature transform) and LBP (local binary pattern) is proposed to address the challenges of video stabilization technology for underwater robots in complex underwater environments. The algorithm makes full use of the scale and rotation invariance of SIFT and the robustness of LBP to illumination changes to effectively improve the accuracy and robustness of underwater video stabilization. Through this fusion approach, image features can be extracted and matched more accurately in various complex underwater environments, thus realizing a more precise estimation of underwater robot motion and providing a reliable data base for subsequent video stabilization processing. Experimental validation on a variety of underwater datasets shows that this fusion algorithm outperforms the traditional single-feature algorithm in terms of feature matching accuracy, image stabilization quality and computational efficiency.