Vehicle Driving Environment Display Using Temporal Point Cloud Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The limited LIDAR angle and high cost of point cloud data in intelligent driving systems, such as autonomous driving and ADAS, result in poor visual expression of the vehicle's driving environment, limiting the effective display of the surrounding environment.
Innovation Solution
A method and apparatus that classify point cloud data to determine semantic types of points, combining data from multiple moments to enhance visual expression by displaying the driving environment, including sparsification to reduce data volume and improve processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If point cloud data is used for visual expression of driving environment, then the driving environment can be displayed, but the visual expression effect is poor due to limited LIDAR angle and enormous amount of data
Solution Approach 1:
The patent extracts and classifies point cloud data by semantic types (vegetation, building, vehicle, pedestrian, etc.), separating useful semantic information from the enormous amount of raw point cloud data. This extraction process filters out redundant data while preserving visually important information, thereby improving visual expression effect without requiring processing of all raw data points.
Solution Approach 2:
The patent segments point cloud data into different semantic categories (vegetation points, building points, vehicle points, pedestrian points, etc.). By dividing the enormous point cloud dataset into meaningful segments based on semantic types, the system can selectively process and display only the relevant segments, reducing the effective data volume while maintaining or improving visual expression quality.
2Area of stationary object
If point cloud data from multiple moments is combined to extend environment range, then the expressible range is extended, but the data processing complexity increases
Solution Approach 1:
The patent performs preliminary classification of point cloud data by semantic types before combining data from multiple moments. By pre-classifying data into categories (vegetation, building, vehicle, etc.), the system establishes a consistent framework that simplifies the subsequent fusion process. This preliminary action prevents the need for complex real-time classification during data fusion, thereby reducing processing complexity while extending the expressible environment range.
3Ease of manufacture
If LIDAR angle is limited, then the system cost is reduced, but the visual expression coverage is insufficient
Solution Approach 1:
The patent compensates for the limited LIDAR angle by utilizing the time dimension - combining point cloud data from multiple moments (temporal dimension) to extend the spatial coverage. By accumulating and fusing data across different time points, the system effectively expands the visual expression coverage beyond the instantaneous field of view of the limited-angle LIDAR, without requiring additional hardware investment.
Data Source
AI summary
Embodiments of this disclosure disclose a method and apparatus for displaying a vehicle driving environment, medium, and device, wherein the method includes: acquiring first point cloud data at a first moment; classifying the first point cloud data, to obtain second point cloud data containing a type attribute of a point; determining, based on the second point cloud data and point cloud data at a second moment, point cloud data of a driving environment of a vehicle at the first moment, wherein the second moment is a moment before the first moment; and displaying, based on the point cloud data of the driving environment, the driving environment at the first moment. Embodiments of this disclosure enable a user to effectively perceive morphology of an object in an environment around a vehicle, improving an effect of visual expression by a point cloud.