DynaSLAM: Tracking, Mapping, and Inpainting in Dynamic Scenes

By Berta Bescós, José M. Fácil, Javier Civera, José Neira. IEEE Robotics and Automation Letters, 2018

1,214 citations recorded by the source corpus.

The assumption of scene rigidity is typical in SLAM algorithms. Such a strong assumption limits the use of most visual SLAM systems in populated real-world environments, which are the target of several relevant applications like service robotics or autonomous vehicles. In this letter we present DynaSLAM, a visual SLAM system that, building on ORB-SLAM2, adds the capabilities of dynamic object detection and background inpainting. DynaSLAM is robust in dynamic scenarios for monocular, stereo, and RGB-D configurations. We are capable of detecting the moving objects either by multiview geometry, deep learning, or both. Having a static map of the scene allows inpainting the frame background that has been occluded by such dynamic objects. We evaluate our system in public monocular, stereo, and RGB-D datasets. We study the impact of several accuracy/speed trade-offs to assess the limits of the proposed methodology. DynaSLAM outperforms the accuracy of standard visual SLAM baselines in highly dynamic scenarios. And it also estimates a map of the static parts of the scene, which is a must for long-term applications in real-world environments.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/25811285a1c1fd514b315d6a05adc2cb4abe9618

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

DynaSLAM II: Tightly-Coupled Multi-Object Tracking and SLAM (2020, background): “The biggest part of the literature tackles this problem by detecting moving regions within the observed scene and rejecting such areas for the SLAM problem [4]–[7].”
Kimera: From SLAM to spatial perception with 3D dynamic scene graphs (2021, background): “, 2009), masking portions of the scene corresponding to dynamic elements (Bescos et al., 2018; Brasch et al., 2018; Cui and Ma, 2019), or jointly tracking camera and dynamic objects (Bescos et al.”
Aligning Cyber Space With Physical World: A Comprehensive Survey on Embodied AI (2024, neutral): “An agent with autonomous exploration and decision-making capabilities must not only consider which actions to take to explore the environment, but also determine when to stop exploring to answer questions.”
CubeSLAM: Monocular 3-D Object SLAM (2018, background): “Most existing approaches treat dynamic region features as outliers and only utilize static background for pose estimation [4, 30, 31].”
3D Dynamic Scene Graphs: Actionable Spatial Perception with Places, Objects, and Humans (2020, background): “Related work also attempts to gain robustness against dynamic scenes by using IMU motion information [33], or masking portions of the scene corresponding to dynamic elements [9, 13, 19].”
SuMa++: Efficient LiDAR-based Semantic SLAM (2019, methodology): “Motivated by the advances of deep learning and Convolutional Neural Networks (CNNs) for scene understanding, there have been many semantic SLAM techniques exploiting this information using cameras [5], [31], cameras + IMU data [4], stereo cameras [10], [15], [18], [33], [38], or RGB-D sensors [3], [”
MID-Fusion: Octree-based Object-Level Multi-Instance Dynamic SLAM (2018, methodology): “We compare our method with five state-of-theart dynamic SLAM approaches: joint visual odometry and scene flow (VO-SF) [8], StaticFusion (SF) [9], DynaSLAM (DS) [11], Co-Fusion (CF) [12], and MaskFusion (MF) [14].”
ReFusion: 3D Reconstruction in Dynamic Environments for RGB-D Cameras Exploiting Residuals (2019, methodology): “More specifically, we do not employ a deep neural network to detect specific dynamic classes, in contrast to other recent approaches [14], [1].”
Spatial Information Guided Convolution for Real-Time RGBD Semantic Segmentation (2020, background): “As a result, RGBD semantic segmentation for high-level scene understanding becomes extremely important, beneﬁting a wide range of applications such as automatic driving [1], SLAM [2], and robotics.”