DH-PTAM: A Deep Hybrid Stereo Events-Frames Parallel Tracking And Map** System
Authors:
Abanob Soliman,
Fabien Bonardi,
Désiré Sidibé,
Samia Bouchafa
Abstract:
This paper presents a robust approach for a visual parallel tracking and map** (PTAM) system that excels in challenging environments. Our proposed method combines the strengths of heterogeneous multi-modal visual sensors, including stereo event-based and frame-based sensors, in a unified reference frame through a novel spatio-temporal synchronization of stereo visual frames and stereo event stre…
▽ More
This paper presents a robust approach for a visual parallel tracking and map** (PTAM) system that excels in challenging environments. Our proposed method combines the strengths of heterogeneous multi-modal visual sensors, including stereo event-based and frame-based sensors, in a unified reference frame through a novel spatio-temporal synchronization of stereo visual frames and stereo event streams. We employ deep learning-based feature extraction and description for estimation to enhance robustness further. We also introduce an end-to-end parallel tracking and map** optimization layer complemented by a simple loop-closure algorithm for efficient SLAM behavior. Through comprehensive experiments on both small-scale and large-scale real-world sequences of VECtor and TUM-VIE benchmarks, our proposed method (DH-PTAM) demonstrates superior performance in terms of robustness and accuracy in adverse conditions, especially in large-scale HDR scenarios. Our implementation's research-based Python API is publicly available on GitHub for further research and development: https://github.com/AbanobSoliman/DH-PTAM.
△ Less
Submitted 9 June, 2024; v1 submitted 2 June, 2023;
originally announced June 2023.
IBISCape: A Simulated Benchmark for multi-modal SLAM Systems Evaluation in Large-scale Dynamic Environments
Authors:
Abanob Soliman,
Fabien Bonardi,
Désiré Sidibé,
Samia Bouchafa
Abstract:
The development process of high-fidelity SLAM systems depends on their validation upon reliable datasets. Towards this goal, we propose IBISCape, a simulated benchmark that includes data synchronization and acquisition APIs for telemetry from heterogeneous sensors: stereo-RGB/DVS, Depth, IMU, and GPS, along with the ground truth scene segmentation and vehicle ego-motion. Our benchmark is built upo…
▽ More
The development process of high-fidelity SLAM systems depends on their validation upon reliable datasets. Towards this goal, we propose IBISCape, a simulated benchmark that includes data synchronization and acquisition APIs for telemetry from heterogeneous sensors: stereo-RGB/DVS, Depth, IMU, and GPS, along with the ground truth scene segmentation and vehicle ego-motion. Our benchmark is built upon the CARLA simulator, whose back-end is the Unreal Engine rendering a high dynamic scenery simulating the real world. Moreover, we offer 34 multi-modal datasets suitable for autonomous vehicles navigation, including scenarios for scene understanding evaluation like accidents, along with a wide range of frame quality based on a dynamic weather simulation class integrated with our APIs. We also introduce the first calibration targets to CARLA maps to solve the unknown distortion parameters problem of CARLA simulated DVS and RGB cameras. Finally, using IBISCape sequences, we evaluate four ORB-SLAM3 systems (monocular RGB, stereo RGB, Stereo Visual Inertial (SVI), and RGB-D) performance and BASALT Visual-Inertial Odometry (VIO) system on various sequences collected in simulated large-scale dynamic environments.
Keywords: benchmark, multi-modal, datasets, Odometry, Calibration, DVS, SLAM
△ Less
Submitted 20 October, 2022; v1 submitted 27 June, 2022;
originally announced June 2022.
