Abstract:Highlights • To enable back-end optimization methods robust to different outliers or noises, the MCMCC-based cost function for VSLAM back-end optimization is designed and the optimization problem is then solved through Levenberg–Marquardt method iteratively. To our knowledge, this is the first time a ITL-based cost function is considered in nonlinear back-end optimization. For its robustness to non-Gaussian outliers, the proposed MCMCC-based back-end optimization method can be applied in VSLAM and 3D reconstruction to further refine camera poses and mappoints. • To promote applications in real scenes, the proposed method is implemented in the visual mode in ORB-SLAM3 and comprehensive analysis is performed by comparing the proposed method with other widely used robust kernel function-based methods in the context of VSLAM on EuRoC, TUM and KITTI datasets. In addition, the method is also tested in real environment, which promises applications in mobile robots and autonomous vehicles when it comes to autonomous localization and environment mapping. • The reprojection error in back-end is analyzed from the prospective of stochastic statistics to provide an insight on the back-end problem in non-Gaussian setting in VSLAM. • The suggestion parameters are provided as guidence for both indoor applications and outdoor applications. Back-end optimization plays a key role in eliminating the accumulated error in Visual Simultaneous Localization And Mapping (VSLAM). Existing back-end optimization methods are usually premised on the Gaussian noise assumption which does not always hold true due to the non-convex nature of the image and the fact that non-Gaussian noises are often encountered in real scenes. In view of this, we propose a back-end optimization method based on Multi-Convex combined Maximum Correntropy Criterion (MCMCC). A MCMCC-based cost function is first tailored for nonlinear back-end optimization in the context of VSLAM and the optimization problem is solved through Levenberg–Marquardt algorithm iteratively. Then, the proposed method is applied to ORB-SLAM3 to test its performance on public indoor and outdoor datasets. The real time performance is also validated using a RaceBot platform in real indoor and outdoor environments. In addition, the reprojection error is statistically analyzed to demonstrate the non-Gaussian characteristics in the back-end optimization process. Finally, the suggestion parameters are also provided through experiments for further study.

New Trend in Back-End Techniques of Visual SLAM: from Local Iterative Solvers to Robust Global Optimization

From Satellite to Ground: Satellite Assisted Visual Localization with Cross-view Semantic Matching

ICE-BA: Incremental, Consistent and Efficient Bundle Adjustment for Visual-Inertial SLAM

Local Optimized and Scalable Frame-to-model SLAM

Nonlinear back-end optimization method for VSLAM with multi-convex combined maximum correntropy criterion

Guaranteed Globally Optimal Planar Pose Graph and Landmark SLAM via Sparse-Bounded Sums-of-Squares Programming

A SLAM Pose Graph Optimization Method Using Dual Visual Odometry

A Review of Common Techniques for Visual Simultaneous Localization and Mapping

PIPO-SLAM: Lightweight Visual-Inertial SLAM With Preintegration Merging Theory and Pose-Only Descriptions of Multiple View Geometry

Submap-Based Pose-Graph Visual SLAM: A Robust Visual Exploration and Localization System

SLAM Back-End Optimization Algorithm Based on Vision Fusion IPS

Simultaneous Localization and Mapping: Through the Lens of Nonlinear Optimization

Evolution of SLAM: Toward the Robust-Perception of Autonomy

Application Of Nonlinear Optimization Method In Stereo Visual Odometry

Visual SLAM Combining Lines and Structural Regularities: Towards Robust Localization

A Survey of Visual SLAM in Dynamic Environment: The Evolution From Geometric to Semantic Approaches

2D LiDAR SLAM Back-End Optimization with Control Network Constraint for Mobile Mapping

A semantic visual SLAM towards object selection and tracking optimization

Real-Time Dynamic SLAM Algorithm Based on Deep Learning

VSLAM Optimization Method in Dynamic Scenes Based on YOLO-Fastest