Unstructured point clouds organize 3D data in a 1D indexed list, where each point’s coordinates and attributes are recorded. Unstructured data allows for flexibility in integrating various data sources, such as unstructured file formats or mobile systems.
Advantages
Simplified Architecture: The straightforward structure of unstructured point clouds reduces project complexity, resulting in smaller file sizes and fewer supporting files.
Flexibility for Integration: Unstructured data supports diverse data sources and is adaptable to various applications, making it ideal for integrating different 3D scanning methods.
Efficient Handling with Octrees: The use of octrees to organize unstructured data enables hierarchical spatial subdivision, improving efficiency for rendering, querying, and filtering large datasets.
Disadvantages
Lack of Structured 2D Indexing: Without a grid-based structure, it is challenging to identify neighbouring points, complicating algorithms that rely on spatial relationships, such as classification and object detection.
Uneditable Links: Unstructured point clouds do not retain editable links or registration between Setups, reducing flexibility for adjusting or modifying the spatial relationship between scans post-processing.
Point-Scanner Relationship Identification: Unstructured data lacks inherent point-to-scanner relationships, making applications such as meshing more challenging.
Loss of Original Scan Data: When unstructured data is created by merging registered structured datasets, it is usually not possible to recover the original scan information, which limits reprocessing and re-registration options.