Weighted multi-node joint refinement for VSLAM pose-graph robustification

Abstract In visual simultaneous localization and mapping (VSLAM) systems, pose-graph optimization accuracy is often compromised by noise and errors in loop closures, especially in large-scale or complex environments. This paper introduces an intermediate processor to address these challenges and proposes a novel algorithm, Weighted Multi-Node Joint Refinement for VSLAM Pose-Graph Robustification (WMJR-PGO). The WMJR-PGO algorithm incorporates a spatial residual mapping test to enhance loop closure detection and a weighted breadth-first search to refine pose estimation. This approach effectively mitigates front-end errors and ensures back-end consistency. Extensive experiments show that the WMJR-PGO algorithm achieves significant improvements across various noise levels. On the Tours dataset, WMJR-PGO outperforms the RS algorithm by 63 <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mi mathvariant="normal">%</mml:mi> </mml:mrow> </mml:math> in optimization accuracy under a noise standard deviation of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:msub> <mml:mi>σ</mml:mi> <mml:mi>r</mml:mi> </mml:msub> <mml:mo>=</mml:mo> <mml:mn>0.5</mml:mn> </mml:mrow> </mml:math> . Furthermore, the WMJR-PGO algorithm demonstrates superior performance on multiple datasets even with a noise standard deviation of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:msub> <mml:mi>σ</mml:mi> <mml:mi>r</mml:mi> </mml:msub> <mml:mo>=</mml:mo> <mml:mn>0.5</mml:mn> </mml:mrow> </mml:math> , proving its robustness in noisy environments. The proposed intermediate processor serves as an independent optimization module and provides efficient initialization for iterative optimization methods, accelerating convergence and improving overall system performance.