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.