Point Cloud Coding Device Using Adaptive Time-Difference Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for compressing point cloud data, such as those using predictive coding, face efficiency degradation when dealing with LiDAR systems where point generation clock times are not constant, leading to increased prediction errors and reduced coding efficiency.
Innovation Solution
A coding device that includes a processor or circuit configured to acquire target point attribute data and generate a target point position bit stream by coding target point position data based on attribute data, using a decomposition and synthesis module to improve prediction accuracy and reduce the difference between predicted and actual positions, thereby enhancing coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If predictive coding is used based on constant point generation clock time assumption, then coding efficiency is improved for regular LiDAR, but prediction error increases and coding efficiency degrades for LiDAR with non-constant clock times
Solution Approach 1:
The patent applies dynamics by making the prediction model adaptive to varying clock times. Instead of using a fixed prediction model assuming constant clock intervals, the system dynamically adjusts the prediction based on actual time differences between points. The prediction unit calculates time differences between current and reference points, and uses these variable time intervals to adjust prediction parameters, allowing the system to handle both constant and non-constant clock time scenarios effectively.
2Loss of substance
If position prediction based on past points is used, then data compression is achieved, but prediction error increases when clock times are not constant
Solution Approach 1:
The patent implements feedback by using actual time difference measurements between points to adjust the prediction process. The system measures the real time intervals between reference and current points, feeds this information back to the prediction unit, and uses it to correct prediction parameters. This feedback mechanism ensures that position predictions remain accurate even when clock times vary, maintaining both compression efficiency and prediction accuracy.
3Speed
If simple predictive coding is used for point cloud data, then processing speed is improved, but coding efficiency degrades for irregular point generation patterns
Solution Approach 1:
The patent applies parameter changes by modifying prediction parameters based on actual time differences between points. Instead of using fixed prediction parameters, the system dynamically changes parameters such as prediction weights and reference point selections based on measured time intervals. This allows the coding system to maintain high processing speed while adapting to irregular point generation patterns, thereby preserving coding efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A coding device includes: a target point data acquisition unit configured to acquire target point attribute data indicating an attribute of a target point that is included in point cloud data and is a target of processing of predicting a position; and a target point position data coding unit configured to generate a target point position bit stream indicating the position of the target point by coding target point position data indicating the position of the target point based on the target point attribute data.