Driven to Distraction: Self-Supervised Distractor Learning for Robust Monocular Visual Odometry in Urban Environments
Authors:
- Dan BarnesDBDan Barnes
- Will MaddernWMWill Maddern
- Geoffrey PascoeGPGeoffrey Pascoe
- Ingmar PosnerIPIngmar Posner
Preprint
Abstract
We present a self-supervised approach to ignoring "distractors" in camera images for the purposes of robustly estimating vehicle motion in cluttered urban environments. We leverage offline multi-session mapping approaches to automatically generate a per-pixel ephemerality mask and depth map for each input image, which we use to train a deep convolutional network. At run-time we use the predicted ephemerality and depth as an input to a monocular visual odometry (VO) pipeline, using either sparse features or dense photometric matching. Our approach yields metric-scale VO using only a single camera and can recover the correct egomotion even when 90% of the image is obscured by dynamic, independently moving objects. We evaluate our robust VO methods on more than 400km of driving from the Oxford RobotCar Dataset and demonstrate reduced odometry drift and significantly improved egomotion estimation in the presence of large moving vehicles in urban traffic.Keywords
All Related Versions
This publication has 12 references indexed in Scilit:
- Multi-task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics2018
- Unsupervised Monocular Depth Estimation with Left-Right Consistency2017
- Unsupervised Learning of Depth and Ego-Motion from Video2017
- CNN-SLAM: Real-Time Dense Monocular SLAM with Learned Depth Prediction2017
- Survey on Background Modeling and Foreground Detection for Real Time Video SurveillanceProcedia Computer Science, 2015
- Object Detection with Discriminatively Trained Part-Based ModelsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2009
- Direct visibility of point setsACM Transactions on Graphics, 2007
- Visual odometry for ground vehicle applicationsJournal of Field Robotics, 2006
- Statistical background subtraction for a mobile observer2003
- Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography1987