Ego-Vehicle Point Cloud Filtering for Lower Processing Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud representations of an ego vehicle's environment often contain non-relevant features such as moving objects or noise, which are difficult to process and increase processing requirements, affecting the accuracy and efficiency of navigation.
Innovation Solution
A method and system that utilize the ego vehicle as a reference to define a projection area within the point cloud, removing non-relevant points by identifying areas where the vehicle's presence precludes the existence of obstacles or structures, considering both two-dimensional and three-dimensional positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all points in the point cloud are processed to ensure complete environment representation, then measurement precision is improved, but processing power requirements increase
Solution Approach 1:
The patent extracts and removes non-relevant points from the point cloud by identifying the ego vehicle's position and projecting its occupancy area onto the point cloud. Points falling within this projection area are removed as they cannot represent valid environmental features. This extraction principle directly reduces the number of points requiring processing while preserving all relevant environmental information.
2Measurement precision
If moving objects are included in the point cloud to maintain completeness, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent extracts moving objects from the point cloud by comparing current sensor data with historical point cloud data. Points that appear in the current scan but were not present in previous scans (and are not within the ego vehicle's projection area) are identified as moving objects and removed. This extraction simplifies processing by eliminating dynamic elements that would otherwise require complex tracking and prediction algorithms.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The present invention refers to a method for removing non-relevant points of a point cloud (22) corresponding to a point representation of an environment (20) of an ego vehicle (10) with multiple individual points (24), comprising the steps of providing the point cloud (22) covering a position of the ego vehicle (10), recognizing the environment (20) of the ego vehicle (10) using at least one environment sensor (14) of the ego vehicle (10), determining a current position of the ego vehicle (10) within the point cloud (22) based on the recognition of the environment (20) of the ego vehicle (10), determining a projection area (32) with a projection of the ego vehicle (10) into the point cloud (22) based on the determined current position of the ego vehicle (10), and removing individual points (24) from the point cloud (22) in the projection area (32) of the ego vehicle (10) as non-relevant points. The present invention also refers to a driving support system (12) comprising at least one environment sensor (14) for recognizing an environment (20) of an ego vehicle (10) and a processing unit (16), which receives sensor information from the at least one environment sensor (14) via a data connection (18) and processes the sensor information, wherein the driving support system (12) is adapted to perform the above method.