Point Cloud Storage Using Half-Precision for Real-Time Scanning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods to reduce memory requirements for point cloud data either degrade data quality or require significant computational resources, making them unsuitable for high-precision and real-time applications.
Innovation Solution
Convert point cloud data of valid points from single-precision floating-point format to half-precision floating-point format, and store it in this format, while determining valid points based on a preset matching condition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If point cloud data is stored in single-precision floating-point format, then data quality and precision are maintained, but memory requirements increase
Solution Approach 1:
The patent changes the data precision parameter from single-precision (32-bit) to half-precision (16-bit) floating-point format. This parameter change reduces the memory footprint by 50% while maintaining sufficient precision for scanning applications, directly resolving the contradiction between data precision and memory requirements
2Quantity of substance
If data compression is applied to reduce memory requirements, then memory usage decreases, but computational resources increase
Solution Approach 1:
The patent uses a simpler, more efficient data representation format (half-precision floating-point) that requires no complex compression algorithms. This approach achieves memory reduction through format conversion rather than computational compression, eliminating the trade-off between memory savings and computational overhead
3Quantity of substance
If resolution is lowered to reduce memory requirements, then memory requirements decrease, but data quality degrades
Solution Approach 1:
Instead of reducing the number of points (resolution), the patent changes the data format precision parameter from 32-bit to 16-bit floating-point. This maintains the same point cloud resolution and data quality while achieving memory reduction through more efficient data encoding
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to relates to a method, an apparatus, a device, and a medium for storing point cloud data. The method includes: acquiring point cloud data of each point in a single frame image; determining valid points satisfying a preset matching condition from all points in the single frame image based on the point cloud data and a preset matching condition; converting the point cloud data of the valid points from a single-precision floating-point format to a half-precision floating-point format and storing the point cloud data with the half-precision floating-point format. According to the technical solution of the present disclosure, the memory requirements for scanning tasks can be reduced while retaining data details and ensuring real-time performance.