Neighbor-Based Point Cloud Filtering for LIDAR Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
LIDAR systems in autonomous vehicles generate noisy point clouds due to dust and particles, leading to false positives in obstacle detection, which can compromise the vehicle's navigation and safety.
Innovation Solution
A method and system for filtering point clouds generated by LIDAR devices in autonomous vehicles, which removes noise points based on intensity measurements and distance thresholds, specifically by identifying points with low intensity and unusual spatial distributions, to improve the accuracy of perception data used for vehicle operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LIDAR device scans through a scene to assemble a point cloud, then distance information to objects is obtained, but noise points caused by dusts or particles are introduced leading to false positives
Solution Approach 1:
The patent extracts and removes noise points from the point cloud by comparing each point's properties (intensity, spatial distribution) against expected characteristics of valid obstacle points. This separation isolates the harmful noise components while preserving the useful obstacle detection data.
Solution Approach 2:
The patent applies different filtering criteria to different points in the point cloud based on their local properties. Each point is evaluated individually using its intensity measurement and spatial relationship with neighboring points, allowing selective removal of noise while preserving valid points with varying characteristics.
2Reliability
If filtering is applied to remove noise points, then false positives are reduced, but processing time and computational complexity increase
Solution Approach 1:
The patent implements a two-stage filtering approach where a first set of points is removed based on basic intensity thresholds, and then a second set is removed using more computationally intensive spatial distribution analysis. This partial application of different filtering depths balances processing time with noise removal effectiveness.
Solution Approach 2:
The filtering process is segmented into distinct stages: first removing points based on intensity measurements, then removing additional points based on spatial distribution relative to neighboring points. This segmentation allows the system to apply computationally lighter operations first, reducing overall processing time while maintaining reliability.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The filtering process effectively reduces false positives in obstacle detection, enhancing the reliability and safety of autonomous vehicle navigation by removing noise points caused by dust and other particles, thereby improving the accuracy of perception data used for vehicle control.
Implementation Method 1
transmitting a laser pulse and detecting a returning pulse, if any, reflected from the object
Implementation Method 2
determining the distance to the object according to the time delay between the transmitted pulse and the reception of the reflected pulse
Data Source
AI summary
A method, apparatus, and system for filtering a point cloud generated by a LIDAR device in an autonomous vehicle is disclosed. A point cloud comprising a plurality of points is generated based on outputs of the LIDAR device. The point cloud is filtered to remove a first set of points in the plurality of points that correspond to noise based on one or more of: point intensity measurements, distances between points, or a combination thereof. Perception data is generated based on the filtered point cloud. Operations of the autonomous vehicle are controlled based on the perception data.