Despite the fact, that dense SLAM systems are extensively developed and are getting popular, feature based ones still have many advantages over them. One of the most important matters in sparse systems are features. The performance and robustness of a system depends strictly on the quality of constraints imposed by feature observations and reliable matching between measurements and features. To improve those two aspects, higher-level features can be used, and planes are a natural choice. We tackle the problem of plugging planes into the g2o optimization framework with two distinct plane representations: one based on a properly stated SE(3) parametrization and one based on a minimal parametrization analogous to quaternions. Proposed solutions were implemented as extensions to the g2o framework and experiments that verify them were conducted using simulation. We provide a comparison of performance under various conditions that emphasized differences.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.