Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
Permanent URI for this collectionhttps://hdl.handle.net/11147/7148
Browse
2 results
Search Results
Now showing 1 - 2 of 2
Article Novel Methods for Depth-Based Calibration of Multiple RGBD Cameras Using Four Mutually Equidistant Spheres(IEEE-Inst Electrical Electronics Engineers Inc, 2025) CalI, Esra Tuncer; Gumustekin, SevketThis article presents novel calibration methods specifically tailored for multiple depth cameras, utilizing solely depth images. Traditional approaches often rely on infrared (IR) images of checkerboards, which, while feasible, fail to exploit the measured depth values, leading to calibration inaccuracies and 3-D misregistration errors. To overcome this limitation, we designed a 3-D tetrahedron object comprising four spheres placed at each corner. By employing an ellipse-fitting technique, we accurately identified the sphere centers in the depth images. Using these centers, we utilized 3-D reprojection errors and measured depths within a bundle adjustment framework to jointly determine the calibration parameters for four depth cameras. Our proposed methods significantly reduce error values compared with those obtained using IR images of checkerboards. The versatility of our techniques ensures their applicability to various types of depth cameras, independent of their underlying technologies. Here, we demonstrate that by integrating depth information directly into the calibration process, we achieve remarkable improvements. Our first method reduces the average system reconstruction error by 78.98%, while our second method, which introduces a novel cost function tailored to the tetrahedron object, achieves an even more substantial reduction of 82.32%. These results underscore the superiority of our depth-integrated calibration approach, particularly in the context of 3-D reconstruction involving multiple depth cameras.Conference Object Citation - WoS: 7Citation - Scopus: 6Semantic Pose Verification for Outdoor Visual Localization With Self-Supervised Contrastive Learning(IEEE, 2022) Guerrero, Jose J.; Orhan, Semih; Baştanlar, YalınAny city-scale visual localization system has to overcome long-term appearance changes, such as varying illumination conditions or seasonal changes between query and database images. Since semantic content is more robust to such changes, we exploit semantic information to improve visual localization. In our scenario, the database consists of gnomonic views generated from panoramic images (e.g. Google Street View) and query images are collected with a standard field-of-view camera at a different time. To improve localization, we check the semantic similarity between query and database images, which is not trivial since the position and viewpoint of the cameras do not exactly match. To learn similarity, we propose training a CNN in a self-supervised fashion with contrastive learning on a dataset of semantically segmented images. With experiments we showed that this semantic similarity estimation approach works better than measuring the similarity at pixel-level. Finally, we used the semantic similarity scores to verify the retrievals obtained by a state-of-the-art visual localization method and observed that contrastive learning-based pose verification increases top-1 recall value to 0.90 which corresponds to a 2% improvement.