Real-time monocular SLAM: Why filter?

By Hauke Strasdat, José M. M. Montiel, Andrew J. Davison. IEEE International Conference on Robotics and Automation, 2010

580 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/9ee001cfe71341c509159f5c881b3f25a2fe5cfc

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

In Electrical and Electronic Engineering

Citations in the corpus

On-Manifold Preintegration for Real-Time Visual--Inertial Odometry (2015, background): “The front-end also selects a subset of images, called keyframes [32], for which we want to compute a pose estimate.”
D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry (2020, background): “On the other hand, since SLAM/VO is essentially a state estimation problem where uncertainty plays an important role [19, 63, 69] and mean-while many learning based methods have started estimating uncertainties, the next question is, how can we incorporate such uncertainty-predictions into…”
KinectFusion: Real-time dense surface mapping and tracking (2011, result): “Compared to filters, this architecture means that an order of magnitude more scene features could be packed into the map [25], and the result was realtime accuracy now comparable to results in off-line reconstruction.”
Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving (2017, unknown): “However, the authors conclude that “filter-based SLAM frameworks might be beneficial if a small processing budget is available, but that BA optimization is superior elsewhere that with limited resources” [83].”
A constant-time SLAM back-end in the continuum between global mapping and submapping: application to visual stereo SLAM (2016, methodology): “As already introduced in previous sections, one can establish two main variations of BA techniques, also known as smoothing methods (Strasdat et al., 2010), according to the representation chosen for KFs and LMs: GBA and RBA.”
High-precision, consistent EKF-based visual-inertial odometry (2013, background): “When many features are visible (the common situation in images of natural scenes), EKF-SLAM’s runtime can be unacceptably high (in fact, higher than that of iterative minimization in certain cases (Strasdat et al., 2010)).”
Extending Monocular Visual Odometry to Stereo Camera Systems by Scale optimization (2019, background): “Bundle adjustment of camera poses and 3D points are conducted only for keyframes [16].”
Keyframe-based visual–inertial odometry using nonlinear optimization (2015, background): “Subsequent research (Strasdat et al., 2010), however, has shown that nonlinear optimization based approaches, as commonly used for offline SfM, can provide better accuracy for a similar computational work when compared to filtering approaches, given that the structural sparsity of the problem is…”
Keyframe-Based Visual-Inertial SLAM using Nonlinear Optimization (2013, methodology): “In [18] it was shown that in purely visual SLAM optimization-based approaches provide better accuracy for the same computational work when compared to filtering approaches.”
Visual simultaneous localization and mapping: a survey (2012, background): “Whenlocalizationandmappingis being done with a single camera, the map will suffer from a scale ambiguity problem (Nistér 2004; Strasdat et al. 2010a).”
Vision based navigation for micro helicopters (2012, methodology): “In (Strasdat et al. (2010)) the authors illustrate the requirement of many features in order to do robust pose estimation and thus favor a key-frame based SLAM approach in most scenarios.”
Visual Odometry : Part II: Matching, Robustness, Optimization, and Applications (2012, background): “Number of Features and Distribution The distribution of the features in the image has been found to affect the VO results remarkably [1], [9], [29].”
Advanced closed loop visual navigation for micro aerial vehicles (2014, background): “Thrun and Liu (2003) studied the problem of fusing laser point clouds exploiting the advantageous properties of sparse extended information filters, while Bahr et al.”
Iterated extended Kalman filter based visual-inertial odometry using direct photometric feedback (2017, background): “While for vision-only systems it has been shown that the number of landmarks is a crucial parameter (Strasdat et al. (2010)), it seems to be different for visual-inertial systems.”