LiDAR Point Cloud Coding via Hierarchical Block Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional point cloud compression technologies are inefficient as they perform encoding and decoding point by point within a frame, tile, or slice, which limits the overall coding efficiency of LiDAR point clouds.
Innovation Solution
A LiDAR point cloud coding method and device that encode and decode points on a per coding block basis, utilizing a coding block as a point cloud encoding/decoding unit, by decoding a quantized residual block and generating a prediction block using stored reconstructed points, and encoding a quantized residual block with a determined quantization parameter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If point cloud encoding/decoding is performed point by point within a frame, then the processing is simple and straightforward, but the coding efficiency is low
Solution Approach 1:
The point cloud frame is divided into multiple coding blocks, and each coding block is further divided into multiple sub-blocks. This segmentation allows parallel processing of different blocks, improving coding efficiency while maintaining manageable complexity through modular organization of processing tasks.
Solution Approach 2:
The patent introduces a hierarchical block structure that adds a spatial dimension to the processing organization. By organizing points into coding blocks and sub-blocks with specific geometric relationships, the system achieves better compression efficiency through structured processing without proportionally increasing complexity.
2Productivity
If coding blocks are used as encoding/decoding units, then the coding efficiency is improved, but the processing complexity increases
Solution Approach 1:
Coding blocks are divided into smaller sub-blocks that can be processed independently. This segmentation reduces the computational complexity within each block while maintaining the efficiency benefits of block-based processing. The sub-block structure enables finer-grained parallel processing and memory management.
Solution Approach 2:
The patent performs preliminary organization of points into coding blocks and sub-blocks with defined geometric relationships before the actual encoding/decoding process. This preliminary structuring enables more efficient processing during encoding by pre-establishing the processing hierarchy and data organization, reducing runtime complexity.
3Duration of action of moving object
If multiple frames are processed continuously, then the temporal coverage is improved, but the memory requirements and processing load increase
Solution Approach 1:
By dividing frames into coding blocks and sub-blocks, the patent enables selective processing and storage of only relevant portions of point cloud data across multiple frames. This segmentation allows the system to maintain temporal coverage by processing continuous frames while reducing memory requirements through block-level data management and selective retention.
Solution Approach 2:
The block-based processing structure enables the system to discard processed blocks after they have been encoded/decoded, and recover only the necessary blocks when needed. This approach allows continuous frame processing with reduced memory requirements, as the system can manage data in a streaming fashion rather than holding entire frames in memory.
Data Source
AI summary
A LiDAR point cloud coding method includes determining, from LiDAR information, a coding block that is a point cloud encoding/decoding unit, to improve the LiDAR point cloud coding efficiency. The LiDAR point cloud coding method also includes encoding/decoding points on a per coding block basis.