Spinning Sensor Point Cloud Geometry Coding With Bounded Azimuth Angles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud compression methods, particularly for sparse geometry data captured by spinning sensor heads, face challenges in achieving efficient, low-latency encoding and decoding while maintaining high compression performance, especially in applications requiring real-time data transmission for autonomous driving and immersive worlds.
Innovation Solution
The method involves encoding and decoding azimuthal angle data based on sensor characteristics, bounding the absolute value of azimuthal angle data by a boundary value dependent on sensor characteristics, and using quantization and entropy coding to improve coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing point cloud compression methods are used for sparse geometry data captured by spinning sensor heads, then compression performance can be achieved, but encoding and decoding latency is high and coding efficiency is insufficient
Solution Approach 1:
The patent changes the parameter representation by bounding azimuthal angle data using a boundary value dependent on sensor characteristics. This transforms the angle data into a constrained range, enabling more efficient entropy coding and reducing the computational complexity of encoding and decoding operations, thereby improving coding efficiency while reducing latency
Solution Approach 2:
The patent performs preliminary bounding of azimuthal angle data before compression. By pre-processing the angle data to ensure it falls within a defined boundary, the system prepares the data in an optimized format that reduces subsequent encoding/decoding latency and improves overall coding efficiency
2Loss of information
If azimuthal angle data is not bounded, then full angular range is represented, but dynamic range is large and compression performance deteriorates
Solution Approach 1:
The patent applies parameter transformation by introducing a boundary value that constrains the azimuthal angle data. This change in parameter representation reduces the dynamic range from the full angular spectrum to a bounded interval, significantly improving compression performance while preserving the essential angular information needed for point cloud reconstruction
3Productivity
If quantization step is increased to improve compression, then coding efficiency improves, but measurement precision of azimuthal angle decreases
Solution Approach 1:
The patent changes the parameter space by bounding the azimuthal angle data, which allows for more efficient quantization. The bounding transforms the angle representation in a way that improves the trade-off between quantization coarseness and precision loss, enabling better coding efficiency while maintaining acceptable azimuthal angle precision for the application
Data Source
AI summary
A method of encoding a point cloud into a bitstream of encoded point cloud data representing a physical object is provided. The method includes: encoding, into the bitstream, azimuthal angle data (ϕcoded,n) representative of an azimuthal angle (ϕ(Pn)) responsive to a capture angle of a sensor that captured a point of the point cloud, wherein the encoding is based on a boundary value (B) that depends on sensor characteristics, and an absolute value of the azimuthal angle data (ϕcoded) is bounded by the boundary value.