Publicação

Real-time GPS indoor for USV tracking using lookup table

Ver documento

Detalhes bibliográficos
Resumo:This study introduces an approach that utilizes Lookup Tables (LUT) to enable real-time tracking of an Unmanned Surface Vehicle (USV) in an indoor setting, using a fish-eye camera. The proposed method streamlines image processing and achieves O(1) complexity, significantly reducing application run time. The paper also outlines the process of calibrating the fish-eye camera to correct image distortion, computing the homography matrix for re-projection, and obtaining a virtual top view of the camera’s field of view. The paper provides a detailed explanation of the replacement of the undistortion and re-projection steps with the new LUT method. Experimental results demonstrate a significant enhancement in the process’s run time, making it feasible for real-time tracking, regardless of the image size.
Autores principais:Vieira, Fábio
Outros Autores:Teodoro, Pedro; Jorge, Pedro Mendes
Assunto:Lookup table GPS indoor SIFT Real-time tracking USV pose estimation
Ano:2023
País:Portugal
Tipo de documento:comunicação não publicada nas atas da conferência
Tipo de acesso:acesso restrito
Instituição associada:Escola Superior Náutica Infante D. Henrique
Idioma:inglês
Origem:Escola Superior Náutica Infante D. Henrique
Descrição
Resumo:This study introduces an approach that utilizes Lookup Tables (LUT) to enable real-time tracking of an Unmanned Surface Vehicle (USV) in an indoor setting, using a fish-eye camera. The proposed method streamlines image processing and achieves O(1) complexity, significantly reducing application run time. The paper also outlines the process of calibrating the fish-eye camera to correct image distortion, computing the homography matrix for re-projection, and obtaining a virtual top view of the camera’s field of view. The paper provides a detailed explanation of the replacement of the undistortion and re-projection steps with the new LUT method. Experimental results demonstrate a significant enhancement in the process’s run time, making it feasible for real-time tracking, regardless of the image size.