LiDAR Point Cloud Compression Using Height-Based Predictive Trees
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in efficiently processing, transmitting, and compressing large amounts of point cloud data, particularly for applications like virtual reality, augmented reality, and autonomous driving, due to high latency, encoding/decoding complexity, and inefficient compression of data captured by LiDAR equipment.
Innovation Solution
The method involves separating point cloud data into roads and objects based on a base height estimated from laser sensor information, constructing predictive trees for each category, and applying intra- and inter-prediction to compress the data effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If point cloud data is transmitted without classification, then transmission process is simple, but compression performance is poor
Solution Approach 1:
The patent segments point cloud data into multiple categories (road, object, sky, ground) based on height information and spatial distribution. Different prediction modes are applied to each category, enabling specialized compression for each type while maintaining overall system simplicity through automated classification rules.
2Device complexity
If a single predictive tree is used for all point cloud data, then encoding complexity is low, but compression efficiency is insufficient
Solution Approach 1:
The patent divides the point cloud data into multiple segments based on height ranges and spatial characteristics, creating separate predictive trees for each segment. This allows each tree to be optimized for specific data characteristics, improving compression efficiency while keeping individual tree complexities manageable.
Solution Approach 2:
Different prediction modes and parameters are applied to different spatial regions and height ranges. For example, road points in lower height ranges use one prediction strategy while object points in higher ranges use another, optimizing compression for local data characteristics without requiring complex global processing.
3Device complexity
If point cloud data is not classified by height, then processing is simpler, but compression performance deteriorates
Solution Approach 1:
The patent applies height-based classification to create different processing zones with optimized prediction parameters for each zone. This allows the system to achieve better compression performance by adapting to local spatial characteristics while maintaining relatively simple processing through automated height threshold comparisons.
Data Source
AI summary
Disclosed according to embodiments are a point cloud data transmission method, a point cloud data transmission device, a point cloud data reception method, and a point cloud data reception device. The point cloud data transmission method according to embodiments may comprise the steps of: acquiring point cloud data including points through LiDAR equipment provided with laser sensors; encoding the point cloud data; and transmitting signaling data and the encoded point cloud data.