A Photometrically Calibrated Benchmark For Monocular Visual Odometry

Abstract

We present a dataset for evaluating the tracking accuracy of monocular visual odometry and SLAM methods. It contains 50 real-world sequences comprising more than 100 minutes of video, recorded across dozens of different environments -- ranging from narrow indoor corridors to wide outdoor scenes. All sequences contain mostly exploring camera motion, starting and ending at the same position. This allows to evaluate tracking accuracy via the accumulated drift from start to end, without requiring ground truth for the full sequence. In contrast to existing datasets, all sequences are photometrically calibrated. We provide exposure times for each frame as reported by the sensor, the camera response function, and dense lens attenuation factors. We also propose a novel, simple approach to non-parametric vignette calibration, which requires minimal set-up and is easy to reproduce. Finally, we thoroughly evaluate two existing methods (ORB-SLAM and DSO) on the dataset, including an analysis of the effect of image resolution, camera field of view, and the camera motion direction.

* Corrected a bug in the evaluation setup, which caused the real-time results for ORB-SLAM (dashed lines in Figure 8) to be much worse than they should be. * https://vision.in.tum.de/data/datasets/mono-dataset

References in corpus (1)

• Direct Sparse Odometry

Cited by in corpus (28)

• The TUM VI Benchmark for Evaluating Visual-Inertial Odometry
• Keyframe-based monocular SLAM: design, survey, and future directions
• LIFT-SLAM: a deep-learning feature-based monocular visual SLAM method
• BA-Net: Dense Bundle Adjustment Network
• Stereo DSO: Large-Scale Direct Sparse Visual Odometry with Stereo Cameras
• Panoramic annular SLAM with loop closure and global optimization
• LDSO: Direct Sparse Odometry with Loop Closure
• Deployment of Reliable Visual Inertial Odometry Approaches for Unmanned Aerial Vehicles in Real-world Environment
• Project AutoVision: Localization and 3D Scene Perception for an Autonomous Vehicle with a Multi-Camera System
• Simultaneous Localization and Mapping Related Datasets: A Comprehensive Survey
• ORBBuf: A Robust Buffering Method for Remote Visual SLAM
• A Review of Visual Odometry Methods and Its Applications for Autonomous Driving
• Challenges in Monocular Visual Odometry: Photometric Calibration, Motion Bias and Rolling Shutter Effect
• Direct Monocular Odometry Using Points and Lines
• Divide-and-Conquer Adversarial Learning for High-Resolution Image and Video Enhancement
• A Synchronized Stereo and Plenoptic Visual Odometry Dataset
• Characterizing SLAM Benchmarks and Methods for the Robust Perception Age
• Run Your Visual-Inertial Odometry on NVIDIA Jetson: Benchmark Tests on a Micro Aerial Vehicle
• PRGFlow: Benchmarking SWAP-Aware Unified Deep Visual Inertial Odometry
• A Model for Multi-View Residual Covariances based on Perspective Deformation
• Fast Loop Closure Detection via Binary Content
• Photometric Correction for Infrared Sensors
• Virtual Testbed for Monocular Visual Navigation of Small Unmanned Aircraft Systems
• Dual-SLAM: A framework for robust single camera navigation
• GSLAM: A General SLAM Framework and Benchmark
• FlowNorm: A Learning-based Method for Increasing Convergence Range of Direct Alignment
• A Scalable FPGA-based Architecture for Depth Estimation in SLAM
• Advanced Mapping Robot and High-Resolution Dataset