Monocular depth prediction has been well studied recently, while there are few works focused on the depth prediction across multiple environments, e.g. changing illumination and seasons, owing to the lack of such real-world dataset and benchmark. In this work, we derive a new cross-season scaleless monocular depth prediction dataset SeasonDepth 1 from CMU Visual Localization dataset through structure from motion. And then we formulate several metrics to benchmark the performance under different environments using recent stateof-the-art open-source depth prediction pretrained models from KITTI benchmark. Through extensive zero-shot experimental evaluation on the proposed dataset, we show that the long-term monocular depth prediction is far from solved and provide promising solutions in the future work, including geometricbased or scale-invariant training. Moreover, multi-environment synthetic dataset and cross-dataset validataion are beneficial to the robustness to real-world environmental variance.
In the field of large-scale SLAM for autonomous driving and mobile robotics, 3D point cloud based place recognition has aroused significant research interest due to its robustness to changing environments with drastic daytime and weather variance. However, it is time-consuming and effort-costly to obtain high-quality point cloud data and groundtruth for registration and place recognition model training in the real world. To this end, a novel registrationaided 3D domain adaptation network for point cloud based place recognition is proposed. A structure-aware registration network is introduced to help learn feature from geometric properties and a matching rate based triplet loss is involved for metric learning. The model is trained through a new virtual LiDAR dataset through GTA-V with diverse weather and daytime conditions and domain adaptation is implemented to the real-world domain by aligning the local and global features. Extensive experiments have been conducted to validate the effectiveness of the structure-aware registration network and domain adaptation. Our results outperform state-of-the-art 3D place recognition baselines on the real-world Oxford RobotCar dataset with the visualization of large-scale registration on the virtual dataset.
Purpose
This paper aims to study the localization problem for autonomous industrial vehicles in the complex industrial environments. Aiming for practical applications, the pursuit is to build a map-less localization system which can be used in the presence of dynamic obstacles, short-term and long-term environment changes.
Design/methodology/approach
The proposed system contains four main modules, including long-term place graph updating, global localization and re-localization, location tracking and pose registration. The first two modules fully exploit the deep-learning based three-dimensional point cloud learning techniques to achieve the map-less global localization task in large-scale environment. The location tracking module implements the particle filter framework with a newly designed perception model to track the vehicle location during movements. Finally, the pose registration module uses visual information to exclude the influence of dynamic obstacles and short-term changes and further introduces point cloud registration network to estimate the accurate vehicle pose.
Findings
Comprehensive experiments in real industrial environments demonstrate the effectiveness, robustness and practical applicability of the map-less localization approach.
Practical implications
This paper provides comprehensive experiments in real industrial environments.
Originality/value
The system can be used in the practical automated industrial vehicles for long-term localization tasks. The dynamic objects, short-/long-term environment changes and hardware limitations of industrial vehicles are all considered in the system design. Thus, this work moves a big step toward achieving real implementations of the autonomous localization in practical industrial scenarios.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.