Point Cloud Geometry Encoding Using 2D Occupancy Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud compression technologies fail to effectively combine encoding and decoding simplicity, low latency, and high compression performance, particularly for sparse geometry data sensed by spinning Lidar sensors, which is crucial for real-time applications like autonomous driving.
Innovation Solution
A method of encoding and decoding point cloud geometry data using ordered coarse points in a two-dimensional space, where order index differences and occupancy data are entropy encoded, allowing for efficient compression and decoding by leveraging neighborhood and coordinate differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If existing point cloud compression technologies are used, then compression is achieved, but encoding complexity and latency increase, reducing real-time performance
Solution Approach 1:
The patent segments the point cloud data processing into distinct stages: generating occupancy maps from sensor data, identifying connected components, and encoding only the essential connectivity information. This segmentation allows efficient processing of sparse geometry data by breaking down the complex compression task into manageable steps that can be executed rapidly for real-time applications.
Solution Approach 2:
The patent extracts and encodes only the essential connectivity information between occupied coarse points rather than encoding complete geometric data. By taking out only the necessary connectivity relationships and discarding redundant geometric details, the system achieves high compression ratios while maintaining real-time performance for autonomous driving applications.
2Loss of substance
If existing point cloud compression technologies are used, then compression is achieved, but device complexity increases, making implementation difficult
Solution Approach 1:
The patent applies local quality by generating occupancy maps that represent only the locally relevant spatial information needed for compression. Instead of processing entire point clouds with uniform complexity, the system creates localized occupancy representations that simplify the encoding process while maintaining compression effectiveness for the specific application context.
Solution Approach 2:
The patent changes the representation parameters from detailed 3D point cloud coordinates to simplified 2D occupancy grid positions and connectivity relationships. This parameter transformation reduces the data dimensionality and complexity, making the encoding apparatus more implementable while achieving the required compression ratios for autonomous driving sensor data.
3Reliability
If detailed point cloud geometry data is transmitted, then quality is maintained, but data transmission bandwidth requirements increase
Solution Approach 1:
The patent creates a simplified copy of the point cloud data in the form of occupancy maps and connectivity graphs rather than transmitting the original detailed geometry. This copied representation preserves the essential spatial relationships and connectivity information needed for autonomous driving decisions while reducing the transmission volume by a significant factor compared to raw point cloud data.
Data Source
AI summary
Methods and apparatuses herein encode/decode point cloud geometry data represented by ordered coarse points occupying some discrete positions of a two-dimensional space. At least one first binary data representative of an order index difference representative of a difference between order indices of two consecutive occupied coarse points, is obtained and each of said at least one first binary data is entropy encoded based on a series of at least one second binary data and a coordinate difference between a first coordinate of a current coarse point and a first coordinate of a preceding occupied coarse point in the two-dimensional space. Said current and preceding occupied coarse points having a same second coordinate in the two-dimensional space and, said series is representative of an occupancy data of at least one neighboring coarse point belonging to a causal neighborhood of the current coarse point.