Point Cloud Preprocessing via Cell Segmentation for LOD Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud rendering technologies face limitations in presenting objects at desired levels of detail, particularly for elongated objects, and are inefficient in managing level of detail (LOD) creation, which is tangled with octree node choice, leading to suboptimal rendering.
Innovation Solution
A computer-implemented preprocessing method that subdivides a large point cloud into cells using slice planes, allowing for independent levels of detail creation for each cell, separating the determination of visible cells from LOD choice, and enabling parallel processing for faster rendering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If octree hierarchy is used to organize point cloud data, then spatial organization and navigation are improved, but the level of detail creation becomes tangled with node choice, reducing rendering flexibility
Solution Approach 1:
The patent segments the point cloud data into multiple independent cell structures, where each cell can have its own level of detail independently controlled. This separates the spatial organization function from the LOD creation function, allowing flexible LOD selection without being constrained by octree node hierarchy.
Solution Approach 2:
The patent extracts the LOD creation process from the octree node structure by implementing independent LOD generation for each cell. This extraction allows LOD parameters to be controlled separately from spatial partitioning, resolving the tangling between node choice and LOD selection.
2Device complexity
If entire point cloud is treated as one object for rendering, then simplicity of processing is maintained, but rendering efficiency and image quality for specific regions are reduced
Solution Approach 1:
The patent divides the entire point cloud into multiple cells, each processed independently for LOD creation and rendering. This segmentation enables parallel processing of different regions, significantly improving rendering efficiency while maintaining manageable complexity through modular cell-based processing.
Solution Approach 2:
The patent applies different levels of detail to different cells based on their specific characteristics and importance. This local quality approach allows high-detail rendering for important regions while using lower detail for less important areas, optimizing overall rendering efficiency and image quality.
3Stability of the object's composition
If octree subdivision is used for LOD creation, then hierarchical organization is achieved, but elongated objects are not optimally handled and rendering versatility is limited
Solution Approach 1:
The patent employs asymmetric cell subdivision strategies that adapt to the geometry of the point cloud data. For elongated objects, cells can be subdivided along the appropriate axes to create optimal representations, rather than being constrained by fixed octree directional subdivision. This asymmetric approach maintains hierarchical organization while greatly improving versatility for different object types.
4Speed
If coarse LOD is selected in octree hierarchy, then larger nodes are automatically selected, but this limits the ability to present objects at desired distinct levels of detail
Solution Approach 1:
The patent segments LOD control into independent cell-level units rather than node-level units. This allows different cells to have different LOD selections simultaneously, enabling fine-grained control over which regions are rendered at what detail level, independent of other regions.
Solution Approach 2:
The patent implements dynamic LOD selection where each cell can independently choose its level of detail based on viewing conditions, importance, and performance requirements. This dynamic approach replaces the static hierarchical LOD selection of octree structures, allowing flexible adaptation to different rendering scenarios.
Data Source
AI summary
The invention concerns a computer-implemented preprocessing method for dividing a point cloud in cells, whereby for each cell a set of levels of detail is created, each level of detail comprising a set of points and a minimal distance; a data structure for storing a preprocessed point cloud; a computer-implemented rendering method for creating a pixelated image based on a stored preprocessed point cloud and a view frustum; and a viewing method involving a computing system to select a view frustum with a user interface and to present a pixelated image of a point cloud on a screen.