Point Cloud Compression Using Ground-Object Height Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current point cloud compression technologies are inefficient in encoding and decoding, particularly in distinguishing between static ground points and dynamic object points, leading to increased signaling overhead and processing requirements.
Innovation Solution
A Geometry Point Cloud Compression (G-PCC) encoder and decoder classify points as ground or object points based on height thresholds, using a global motion vector for object points and separate encoding techniques for ground points, reducing the need for local motion vectors and improving operational efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If point cloud compression uses traditional encoding methods without classification, then all points are processed uniformly, but this leads to increased signaling overhead and reduced encoding efficiency
Solution Approach 1:
The patent segments the point cloud into two distinct sets: ground points and object points. This segmentation allows different encoding strategies to be applied to each set, improving overall encoding efficiency. Ground points are encoded using one method while object points use another, reducing the need for complex per-point signaling.
Solution Approach 2:
The patent applies different encoding qualities and methods to different regions of the point cloud. Ground points (static background) are encoded with simpler methods, while object points (dynamic foreground) receive more detailed encoding. This local differentiation optimizes the balance between compression efficiency and quality where it matters most.
2Productivity
If point cloud compression processes all points with the same method, then implementation is simple, but this increases processing operations and reduces operational efficiency
Solution Approach 1:
The patent divides the point cloud processing into two separate pipelines: one for ground points and one for object points. This segmentation enables specialized processing for each type, improving operational efficiency by applying appropriate algorithms to each set rather than using a one-size-fits-all approach.
Solution Approach 2:
The patent changes the processing parameters based on point classification. Ground points use one set of encoding parameters while object points use different parameters. This parameter differentiation allows optimized processing for each category, improving operational efficiency despite the added classification step.
3Productivity
If point cloud encoding uses height threshold classification to separate ground and object points, then encoding efficiency improves, but this requires additional data structures and threshold coding
Solution Approach 1:
The patent uses height threshold values as a simple parameter change to classify points into ground or object sets. This parameter-based classification is more efficient than complex geometric or semantic methods. The thresholds are coded into the data structure, adding minimal overhead while enabling efficient compression through differentiated encoding of the two point sets.
Data Source
AI summary
A device to code a point cloud data that 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; code a data structure including data that represents a top threshold and a bottom threshold; classify points having height values between the top threshold and the bottom threshold into the set of ground points; classify points having height values above the top threshold or below the bottom threshold into the set of object points. The one or more processors code the ground points and the object points according to the classifications. The one or more processors code a geometry data unit header that includes data that overrides or refines the data of the data structure for the at least one of the top threshold or the bottom threshold.