Λεπτομέρειες βιβλιογραφικής εγγραφής
| Τίτλος: |
Event-based egomotion estimation ; Estimación propioceptiva del movimiento con cámara de eventos ; Estimació propioceptiva del moviment amb càmara d'esdeveniments |
| Συγγραφείς: |
Sanchez Orvay, Sergi |
| Συνεισφορές: |
Universitat Politècnica de Catalunya. Departament de Teoria del Senyal i Comunicacions, Institut de Robòtica i Informàtica Industrial, Andrade-Cetto, Juan, Nájar Martón, Montserrat |
| Στοιχεία εκδότη: |
Universitat Politècnica de Catalunya |
| Έτος έκδοσης: |
2024 |
| Συλλογή: |
Universitat Politècnica de Catalunya, BarcelonaTech: UPCommons - Global access to UPC knowledge |
| Θεματικοί όροι: |
Àrees temàtiques de la UPC::Informàtica::Robòtica, Geometry--Data processing, Robot vision, Mobile robots, Estimation theory, event cameras, egomotion estimation, state estimation, geometry, computer vision, perception, mobile robotics, robotics, cámaras de eventos, estimación de movimiento, estimación de estado, geometría, visión por computador, visión artificial, percepción, robótica móvil, robótica, Geometria--Informàtica, Visió artificial (Robòtica), Robots mòbils, Estimació, Teoria de l' |
| Περιγραφή: |
Event cameras are bio-inspired sensors that detect logarithmic changes in pixel-level illumination, generating events asynchronously with microsecond resolution. This principle provides significant advantages over conventional cameras such as high dynamic range, low latency, and low power consumption, making them ideal for dynamic scenarios and challenging lighting conditions. This work addresses the egomotion estimation of an event camera using a geometric model-based approach and optimization techniques, avoiding the use of learning methods to make it applicable in resource-constrained environments and to deeply understand the problem. A detailed study of the underlying mathematical foundations guides the development of the proposed method. The proposed methodology consists of three main blocks: normal flow estimation, inverse depth estimation, and a robust linear solver. The normal flow estimation method is based on fitting local planes on surfaces of active events, while inverse depth is tackled using optimization techniques to align events at a reference timestamp. Finally, a robust linear solver based on the iterative RANSAC method is implemented, using the inverse depth and normal flow estimates to obtain the camera's linear velocity. The method's performance is evaluated through experiments with synthetic and real data. With synthetic data, the inverse depth optimization algorithm effectively aligns events and converges rapidly to its ground truth value. Additionally, the normal flow estimation recovers both its orientation and magnitude in these examples. Despite certain limitations in synthetic data, this allows the linear solver to obtain the linear velocity. However, poorly conditioned normal flows are observed when tested with real data. To correct these normal flows, a multi-spatial scale maxpooling method is implemented, which improves estimated orientations but fails to recover magnitude. Conversely, the inverse depth estimation algorithm proves highly effective with real data, correctly aligning ... |
| Τύπος εγγράφου: |
bachelor thesis |
| Περιγραφή αρχείου: |
application/pdf |
| Γλώσσα: |
English |
| Relation: |
https://hdl.handle.net/2117/422144 |
| Διαθεσιμότητα: |
https://hdl.handle.net/2117/422144 |
| Rights: |
S'autoritza la difusió de l'obra mitjançant la llicència Creative Commons o similar 'Reconeixement-NoComercial- SenseObraDerivada' ; Open Access |
| Αριθμός Καταχώρησης: |
edsbas.7BD93137 |
| Βάση Δεδομένων: |
BASE |