Brief Announcement: Optimized GPU-accelerated Feature Extraction for ORB-SLAM Systems
Filippo Muzzini, Nicola Capodieci, Roberto Cavicchioli, Benjamin Rouxel
Abstract
Reducing the execution time of ORB-SLAM algorithm is a crucial aspect of autonomous vehicles since it is computationally intensive for embedded boards. We propose a parallel GPU-based implementation, able to run on embedded boards, of the Tracking part of the ORB-SLAM2/3 algorithm. Our implementation is not simply a GPU port of the tracking phase. Instead, we propose a novel method to accelerate image Pyramid construction on GPUs. Comparison against state-of-the-art CPU and GPU implementations, considering both computational time and trajectory errors shows improvement on execution time in well-known datasets, such as KITTI and EuRoC.
Topics & Concepts
Orb (optics)Computer scienceCUDAPyramid (geometry)Feature extractionImplementationTrajectoryTracking (education)Parallel computingGeneral-purpose computing on graphics processing unitsFeature (linguistics)Artificial intelligenceImage (mathematics)Computer graphics (images)GraphicsLinguisticsPsychologyPedagogyPhilosophyPhysicsProgramming languageOpticsAstronomyRobotics and Sensor-Based LocalizationAdvanced Image and Video Retrieval TechniquesRobotic Path Planning Algorithms