Point Cloud Occlusion Culling via Depth Buffer Heuristics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing occlusion culling techniques are inefficient when processing point clouds due to their poor occluding capabilities and the complexity of level-of-detail structures, leading to unnecessary overdraw and memory contention.
Innovation Solution
A computer-implemented method for visualizing point clouds using occlusion culling testing, which involves using a depth buffer from a previous frame to determine occluded points and applying heuristic testing to assess changes in visibility when the observation point changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If traditional occlusion culling techniques are used for point clouds, then overdraw is reduced, but rendering performance deteriorates due to poor occluding capabilities and complex level-of-detail structures
Solution Approach 1:
The patent segments the point cloud into multiple levels of detail (LODs) with different resolution and density characteristics. Each LOD is processed separately through occlusion culling, allowing efficient handling of points at different depths while reducing the complexity of processing the entire point cloud at once. This segmentation enables the system to maintain good occlusion culling effectiveness without being overwhelmed by the complex structure of high-density point clouds.
Solution Approach 2:
The patent implements dynamic occlusion culling that adapts to changing observation points and camera parameters. The occlusion testing is performed dynamically based on the current viewing frustum and depth buffer, allowing the system to adjust which points are culled in real-time. This dynamic approach ensures that occlusion culling remains effective as the camera moves through the scene, preventing both overdraw and unnecessary rendering operations.
2Productivity
If points are grouped into pre-defined volumes for level-of-detail processing, then processing efficiency is improved, but occlusion testing accuracy deteriorates due to large bounding boxes with empty space
Solution Approach 1:
The patent applies different processing qualities to different regions of the point cloud based on their local characteristics. Points are assigned to different LODs according to their spatial distribution and occlusion probability, with higher resolution processing applied to regions that require it and coarser processing applied to regions that can be handled more simply. This local quality approach maintains processing efficiency while improving occlusion testing accuracy in critical regions.
Solution Approach 2:
The patent performs occlusion culling tests on a selective subset of points rather than all points uniformly. By applying occlusion testing to only those points that are potentially visible or require detailed processing, the system achieves both efficiency and accuracy. The heuristic testing mechanism allows the system to perform partial occlusion testing on representative samples to infer occlusion status for groups of points, reducing the total number of tests while maintaining accuracy.
Data Source
AI summary
A method for visualizing a point cloud on a display, wherein the visualizing includes changing an observation point onto the point cloud and efficient handling of changes in points of the point cloud that are differently obscured for different observation points. The method comprises occlusion culling testing using a depth buffer for sampling a first depth from a previous observation point associated with a previous frame towards the tested point to determine whether the tested point was visible or occluded when viewed from the previous observation point. If the tested point is determined to be occluded when viewed from the previous observation point, a heuristic testing is carried out as part of the occlusion culling testing to determine if the change in observation point since the previous frame is likely to have revealed the previously covered point to be visible from the new observation point.