Why Not Replace? Sustaining Long-Term Visual Localization via Handcrafted-Learned Feature Collaboration on CPU

TL;DR

This paper introduces a CPU-efficient hierarchical visual localization framework that combines handcrafted and learned features, significantly improving long-term accuracy and robustness in complex industrial environments.

Contribution

It proposes a novel integration of handcrafted and learned features for long-term visual localization, optimizing for CPU efficiency and environmental robustness.

Findings

Achieved 47% average error reduction in localization accuracy.

Demonstrated improved robustness under seasonal and diurnal variations.

Validated the approach through systematic experiments and latency profiling.

Abstract

Robust long-term visual localization in complex industrial environments is critical for mobile robotic systems. Existing approaches face limitations: handcrafted features are illumination-sensitive, learned features are computationally intensive, and semantic- or marker-based methods are environmentally constrained. Handcrafted and learned features share similar representations but differ functionally. Handcrafted features are optimized for continuous tracking, while learned features excel in wide-baseline matching. Their complementarity calls for integration rather than replacement. Building on this, we propose a hierarchical localization framework. It leverages real-time handcrafted feature extraction for relative pose estimation. In parallel, it employs selective learned keypoint detection on optimized keyframes for absolute positioning. This design enables CPU-efficient, long-term visual localization. Experiments systematically progress through three validation phases: Initially establishing feature complementarity through comparative analysis, followed by computational latency profiling across algorithm stages on CPU platforms. Final evaluation under photometric variations (including seasonal transitions and diurnal cycles) demonstrates 47% average error reduction with significantly improved localization consistency. The code implementation is publicly available at https://github.com/linyicheng1/ORB_SLAM3_localization.