Handcrafted Local Feature Descriptor-Based Point Cloud Registration and Its Applications: A Review.

Area of Science:

• Computer Vision
• Computer Graphics
• Remote Sensing

Background:

• Local feature descriptors (LFDs) are crucial for point cloud registration.
• Recent years have seen significant advancements in LFD-based registration methods.
• A systematic review is lacking to consolidate these developments.

Purpose of the Study:

• To provide a comprehensive review of state-of-the-art and widely referenced LFD-based point cloud registration methods.
• To analyze the strengths and limitations of existing methodologies.
• To offer guidance for future research in this domain.

Main Methods:

• Systematic survey of existing LFD-based registration methodologies.
• In-depth analysis of method strengths and limitations.
• Summary of relevant applications and datasets.

Main Results:

• Identification of key advancements in LFD-based registration.
• Critical examination of various approaches across registration subtasks.
• Provision of practical recommendations for researchers.

Conclusions:

• The review addresses the gap in systematic consolidation of LFD-based registration research.
• It offers valuable insights and guidance for researchers in computer vision, graphics, and remote sensing.
• The work highlights the importance of LFDs in point cloud registration and future research directions.