Point Cloud Compression Using Height-Based Ground Object Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current point cloud compression technologies are inefficient in encoding and decoding, particularly in distinguishing and processing ground and object points within point clouds, leading to increased signaling overhead and processing complexity.
Innovation Solution
The proposed solution involves a Geometry Point Cloud Compression (G-PCC) encoder and decoder that classify points as ground or object points based on height values, using global motion vectors for object points and distinct encoding techniques for ground points, such as zero motion vectors or local motion vectors, to reduce signaling overhead and processing operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If current point cloud compression technologies are used to encode and decode point clouds without distinguishing ground and object points, then the encoding and decoding processes are simpler, but the signaling overhead increases and processing efficiency decreases
Solution Approach 1:
The patent segments the point cloud into two distinct sets: ground points and object points. This segmentation allows each set to be processed using optimized encoding techniques tailored to their specific characteristics, thereby improving overall compression efficiency without excessively increasing complexity
Solution Approach 2:
The patent applies different encoding strategies to different parts of the point cloud based on their local characteristics. Ground points use one encoding approach while object points use another, allowing each region to be processed with the most appropriate method for its specific properties
2Loss of information
If global motion vectors are used for all points in the point cloud, then the encoding process is simpler, but the number of bits required for encoding increases
Solution Approach 1:
The patent divides the point cloud into ground points and object points, allowing different motion vector strategies to be applied to each segment. This reduces the overall signaling overhead by optimizing the motion compensation approach for each specific point set
Solution Approach 2:
The patent changes the motion vector parameters differently for ground points versus object points. By adapting the motion compensation parameters to the specific characteristics of each point set, the patent reduces the number of bits required while maintaining compression quality
Data Source
AI summary
An example device for coding point cloud data includes a memory configured to store data representing points of a point cloud, and one or more processors implemented in circuitry and configured to: determine height values of points in a point cloud; classify the points into a set of ground points or a set of object points according to the height values; and code the ground points and the object points according to the classifications. The one or more processors may determine top and bottom thresholds and classify the ground and object points according to the top and bottom thresholds. The one or more processors may further code a data structure, such as a geometry parameter set (GPS), including data representing the top and bottom thresholds.