Direct Visual Odometry in Low Light Using Binary Descriptors

Alismail, Hatem; Kaess, Michael; Browning, Brett; Lucey, Simon

doi:10.1109/lra.2016.2635686

Cited by 65 publications

(62 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the main limitation of direct methods is the brightness constancy assumption which is rarely fulfilled in any real-world robotic application [17]. The authors of [18] propose to use binary feature descriptors for direct tracking called Bit-planes. While improving the robustness to bad lighting situations it was also found that Bit-planes have a smaller convergence basin than intensities.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization

Stumberg

Wenzel

Khan

et al. 2020

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Direct SLAM methods have shown exceptional performance on odometry tasks. However, they are susceptible to dynamic lighting and weather changes while also suffering from a bad initialization on large baselines. To overcome this, we propose GN-Net: a network optimized with the novel Gauss-Newton loss for training weather invariant deep features, tailored for direct image alignment. Our network can be trained with pixel correspondences between images taken from different sequences. Experiments on both simulated and real-world datasets demonstrate that our approach is more robust against bad initialization, variations in day-time, and weather changes thereby outperforming state-of-the-art direct and indirect methods. Furthermore, we release an evaluation benchmark for relocalization tracking against different types of weather. Our benchmark is available at https://vision. in.tum.de/gn-net.

show abstract

Section: Related Workmentioning

confidence: 99%

“…To partially overcome this, a pyramid scheme is often used. Usually tracking on multiple channels instead of one can decrease the convergence radius ( [18], [25]). However, in our case, we train the feature maps to in fact have a larger convergence basin than images by enforcing smoothness in the vicinity of the correct correspondence.…”

Section: Deep Direct Slammentioning

confidence: 99%

GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization

Stumberg

Wenzel

Khan

et al. 2020

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

show abstract

“…Crivellaro and Lepetit (2014) relies on the dense computation of a deliberately designed local descriptor to obtain a clear global minimum in energy function while preserving convergence basin by convolving with a low‐pass filter. The method based on the census transform Alismail, Browning, and Lucey (2017) uses a binary descriptor to achieve local illumination invariance during the motion estimation. Based on the analysis from our previous work (Wu & Pradalier, 2019), we found the affine‐model‐based approach shows both good accuracy and large convergence basin, which can be used to formulate our inter‐camera tracking algorithm.…”

Section: Related Workmentioning

confidence: 99%

Robotic weed control using automated weed and crop classification

Aravecchia

Lottes

et al. 2020

Journal of Field Robotics

124

View full text Add to dashboard Cite

Autonomous robotic weeding systems in precision farming have demonstrated their full potential to alleviate the current dependency on agrochemicals such as herbicides and pesticides, thus reducing environmental pollution and improving sustainability. However, most previous works require fast and constant‐time weed detection systems to achieve real‐time treatment, which forecloses the implementation of more capable but time‐consuming algorithms, for example, learning‐based methods. In this paper, a nonoverlapping multicamera system is applied to provide flexibility for the weed control system in dealing with the indeterminate classification delays. The design, implementation, and testing of our proposed modular weed control unit with mechanical and chemical weeding tools are presented. A framework that performs naive Bayes filtering, 3D direct intra‐ and inter‐camera visual tracking, and predictive control, while integrating state‐of‐the‐art crop/weed detection algorithms, is developed to guide the tools to achieve high‐precision weed removal. The experimental results show that our proposed fully operational weed control system is capable of performing selective mechanical as well as chemical in‐row weeding with indeterminate detection delays in different terrain conditions and crop growth stages.

show abstract

“…The designed iteration (42) can directly obtain the attitude matrix on SO(3) and then produce the quaternion. In realworld implementation, the cross product can be easily computed as follows…”

Section: E Real-world Implementationmentioning

confidence: 99%

Rigid 3D Registration: A Simple Method Free of SVD and Eigen-Decomposition

2020

IEEE Trans. Instrum. Meas.

View full text Add to dashboard Cite

A novel solution is obtained to solve the rigid 3D registration problem, motivated by previous eigen-decomposition approaches. Different from existing solvers, the proposed algorithm does not require sophisticated matrix operations e.g. singular value decomposition or eigenvalue decomposition. Instead, the optimal eigenvector of the point cross-covariance matrix can be computed within several iterations. It is also proven that the optimal rotation matrix can be directly computed for cases without need of quaternion. The simple framework provides very easy approach of integer-implementation on embedded platforms. Simulations on noise-corrupted point clouds have verified the robustness and computation speed of the proposed method. The final results indicate that the proposed algorithm is accurate, robust and owns over 60% ∼ 80% less computation time than representatives. It has also been applied to real-world applications for faster relative robotic navigation.

show abstract

Direct Visual Odometry in Low Light Using Binary Descriptors

Cited by 65 publications

References 43 publications

GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization

GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization

Robotic weed control using automated weed and crop classification

Rigid 3D Registration: A Simple Method Free of SVD and Eigen-Decomposition

Contact Info

Product

Resources

About