Point Cloud Azimuth Prediction Using Depth Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for encoding/decoding azimuth information in large-scale point clouds fail to consider the relationship between depth and azimuth information, leading to large prediction residuals and inefficient entropy encoding.
Innovation Solution
Establish a relationship between depth and azimuth information using mathematical derivation or fitting methods, predictively encode the azimuth information based on this relationship, and utilize entropy encoding to improve precision and reduce residuals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional prediction methods (prediction tree or single-chain structure) are used to encode azimuth information, then the encoding process can be completed, but the prediction residuals are large and not centralized, destroying the validity of entropy encoding context model and reducing encoding efficiency
Solution Approach 1:
The patent transitions from one-dimensional azimuth prediction (using only azimuth information) to two-dimensional prediction by incorporating depth information as an additional dimension. This allows the prediction model to consider both depth and azimuth relationships, resulting in more centralized prediction residuals and improved entropy encoding efficiency.
Solution Approach 2:
The patent changes the prediction parameters by introducing depth information alongside azimuth information. Instead of predicting azimuth solely based on previous azimuth values, the method uses a combination of depth and azimuth parameters to generate prediction values, thereby improving prediction precision and residual distribution.
2Device complexity
If only encoded azimuth information is used for prediction, then the prediction process is simple, but the relationship between other information (depth) and azimuth information is not considered, leading to large prediction residuals
Solution Approach 1:
The patent merges depth information and azimuth information into a unified prediction framework. By combining these two types of information, the method achieves more accurate prediction of azimuth values while maintaining a manageable prediction process through established encoding structures.
Data Source
AI summary
An encoding method includes: obtaining original point cloud data; obtaining depth information of a point cloud based on the original point cloud data; establishing a relationship between the depth information and azimuth information of the point cloud; and predictively encoding the azimuth information of the point cloud by using the relationship between the depth information and the azimuth information of the point cloud, to obtain coded stream information.