This repository provides a structured, multi-level roadmap designed to guide aspiring developers and researchers from absolute beginner to Visual-SLAM engineer or researcher by 2026. It addresses the perceived high entry barrier of Visual-SLAM by offering a clear learning path, breaking down complex topics into manageable stages, and highlighting essential skills and foundational knowledge.

How It Works

The roadmap is organized into 11 progressive levels, starting with fundamental programming, mathematics, and geometry, and advancing through specialized areas such as Monocular SLAM, RGB-D SLAM, VIO/VINS, LiDAR fusion, event cameras, and cutting-edge World Models and Spatial AI. Each level details specific topics, key concepts, and lists relevant algorithms, systems, and their authors/years, providing a historical and state-of-the-art overview.

Quick Start & Requirements

This is a learning roadmap, not a software project with installation instructions.

• Primary Install/Run: Not applicable.
• Prerequisites: Foundational knowledge in C++ and Python programming, basic probability & statistics, linear algebra, calculus, projective geometry, camera principles, and image processing is recommended for Level 1. Advanced levels require familiarity with libraries like Eigen, Ceres-solver, OpenCV, and deep learning frameworks.
• Links: The primary resource is the repository's README.

Highlighted Details

• Comprehensive 11-level curriculum covering foundational to advanced Visual-SLAM topics.
• Detailed breakdown of concepts, algorithms, and systems, including recent advancements like Transformer-based SLAM, Foundation Model SLAM, 3DGS-based SLAM, and Vision-Language Models (VLMs).
• Inclusion of specific system names, authors, and publication years, offering a historical context and state-of-the-art overview.
• Emphasis on both theoretical underpinnings and practical skill development.

Maintenance & Community

Contributions are welcomed via pull requests and issue discussions. Direct contact is available at hyunggi.chang95[at]gmail.com. Users are encouraged to watch or star the repository for updates.

Licensing & Compatibility

The roadmap is licensed under the MIT License. This permissive license allows for unrestricted use, modification, and distribution, making it compatible with commercial applications and closed-source projects.

Limitations & Caveats

The roadmap acknowledges that SLAM has a "relatively high entry barrier" due to the diverse skill set required. It serves as a guide and does not provide executable code or tutorials, necessitating separate efforts for practical implementation. Some technologies listed are slated for 2025/2026, indicating a forward-looking but potentially evolving landscape for those specific areas.