TurboMap: GPU-Accelerated Local Mapping for Visual SLAM

2025-11-03

AI Summary

To address the computational bottleneck of the local mapping module in visual SLAM, this paper proposes an efficient CPU-GPU collaborative optimization framework for local mapping. The method fully migrates key local mapping operations—namely, point triangulation, map point fusion, and local bundle adjustment (Local BA) solving—to the GPU, implementing fine-grained parallel acceleration via CUDA. Concurrently, it optimizes the CPU-side keyframe selection strategy to improve task scheduling efficiency. Integrated end-to-end into ORB-SLAM3, the approach achieves 1.3–1.6× average speedup in local mapping on the EuRoC and TUM-VI datasets. This significantly enhances real-time performance and cross-platform scalability—supporting both desktop and embedded platforms—while preserving original localization and mapping accuracy.

Technology Category

• Search and Optimization: Local Search
• Intelligent Robots: Localization, Mapping, and Navigation
• Computer Vision: Learning & Optimization for CV

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

This paper presents TurboMap, a GPU-accelerated and CPU-optimized local mapping module for visual SLAM systems. We identify key performance bottlenecks in the local mapping process for visual SLAM and address them through targeted GPU and CPU optimizations. Specifically, we offload map point triangulation and fusion to the GPU, accelerate redundant keyframe culling on the CPU, and integrate a GPU-accelerated solver to speed up local bundle adjustment. Our implementation is built on top of ORB-SLAM3 and leverages CUDA for GPU programming. The experimental results show that TurboMap achieves an average speedup of 1.3x in the EuRoC dataset and 1.6x in the TUM-VI dataset in the local mapping module, on both desktop and embedded platforms, while maintaining the accuracy of the original system.

Problem

• Addresses performance bottlenecks in visual SLAM local mapping
• Optimizes map point triangulation and fusion using GPU acceleration
• Accelerates local bundle adjustment and keyframe culling processes

Innovation

• GPU-accelerated map point triangulation and fusion
• CPU-optimized redundant keyframe culling
• GPU-accelerated local bundle adjustment solver