Point Cloud Geometry Encoding Using Sensor Sensing Number
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud codecs struggle to combine encoding and decoding simplicity, low latency, and high compression performance, particularly when dealing with sparse geometry data sensed by spinning sensors or other types of sensors.
Innovation Solution
The method involves encoding and decoding point cloud geometry data by representing it as ordered coarse points in a two-dimensional space, using entropy encoding and decoding based on a sensing number that represents the average number of consecutive senses needed by a sensor to detect another coarse point.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If existing point cloud codecs are used to compress sparse geometry data, then compression performance is improved, but encoding complexity and latency increase
Solution Approach 1:
The patent changes the parameter representation by introducing a sensing number parameter that characterizes the sparsity of point cloud data. This parameter is used to adaptively select encoding strategies and optimize the compression process, allowing the system to handle sparse data efficiently without excessive complexity
Solution Approach 2:
The patent segments the point cloud data processing into distinct stages: obtaining sensing numbers, determining occupancy patterns, and applying differential encoding. This segmentation allows each stage to be optimized independently, reducing overall encoding complexity while maintaining compression performance
2Loss of information
If existing point cloud codecs are used to compress sparse geometry data, then compression performance is improved, but encoding and decoding simplicity is reduced
Solution Approach 1:
The patent implements self-service by having the encoder generate sensing numbers and occupancy patterns that inherently guide the decoding process. The decoder uses these same parameters to reconstruct the point cloud without requiring complex external information, simplifying both encoding and decoding operations
Solution Approach 2:
The patent performs preliminary action by calculating sensing numbers and determining occupancy patterns before the main compression process. This preparation enables the subsequent encoding steps to proceed with simpler operations, as the difficult analytical work has already been completed
3Loss of information
If existing point cloud codecs are used to compress sparse geometry data, then compression performance is improved, but latency increases
Solution Approach 1:
The patent applies periodic action by processing point cloud data in regular patterns based on sensor rotation periods. This allows the system to predict when points will be sensed and pre-calculate encoding parameters, reducing processing latency while maintaining compression efficiency
Data Source
AI summary
A method of encoding, into a bitstream, point cloud geometry data represented by ordered coarse points occupying some discrete positions of a set of discrete positions of a two-dimensional space, includes obtaining at least one binary data (fj) representative of an order index difference representative of a difference between order indices of two consecutive occupied coarse points and encoding each of the at least one binary data (fj) by: obtaining coordinates (sj,λj) of a current coarse point (Pj) within the two-dimensional space; obtaining a sensing number (Nsens(sj,λj)) from the current coarse point (Pj), the sensing number (Nsens(sj,λj)) being representative of an average number of consecutive sensing needed by a sensor (λj) associated with the current coarse point (Pj) to sense another coarse point; and entropy encoding, into the bitstream, the binary data (fj) based on the sensing number (Nsens(sj,λj)).