LiDAR Layer Filtering for Static Object Point Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods struggle to accurately distinguish between static objects like guardrails and structures like bushes in multiple layers of LiDAR data, leading to errors in data processing for autonomous vehicles.
Innovation Solution
A method and device that select a reference layer based on the shape similarity to a static object, using regression to identify and remove points in other layers that are not relevant to the static object, enhancing the accuracy of LiDAR data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only points included in each single layer of information are used to determine static objects and structures, then the processing is simple, but the static object and structure points cannot be perfectly distinguished when the structure points form a straight-line shape
Solution Approach 1:
The patent transitions from single-layer 2D point analysis to multi-layer 3D spatial analysis. By utilizing points from multiple LiDAR layers and analyzing their spatial relationships across different elevation planes, the system can distinguish straight-line structure points from static object points even when they appear similar in individual layers. The multi-layer configuration enables detection of depth variations and spatial patterns that resolve the ambiguity.
Solution Approach 2:
The patent segments the multi-layer point cloud data into distinct layers and processes each layer separately before integrating the results. By dividing the complex 3D space into multiple 2D layers and analyzing point distributions within each layer independently, then combining the findings, the system achieves more accurate distinction between static objects and structures while maintaining manageable processing complexity.
2Measurement precision
If multiple layers of LiDAR data are processed to accurately distinguish static objects from structures, then the distinction accuracy improves, but the data processing time and computational load increase
Solution Approach 1:
The patent performs preliminary processing of each LiDAR layer independently before integration, pre-identifying potential static object points and structure points in each layer. This preliminary classification reduces the computational burden during the final integration stage, as the system only needs to reconcile and validate results from already-processed layers rather than analyzing all points simultaneously from scratch.
Solution Approach 2:
The patent applies processing to all multiple layers (excessive action) to ensure complete coverage and accuracy, but uses selective filtering to focus computational resources only on relevant points that contribute to static object vs. structure distinction. Points that can be clearly classified or are irrelevant to the distinction task are processed more efficiently or excluded from intensive analysis.
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
Improves the accuracy of LiDAR data processing by selectively removing irrelevant points, minimizing errors in matching with high-definition maps, thereby enhancing the safety and precision of autonomous vehicle navigation.
Implementation Method 1
A LiDAR sensor constituting a LiDAR system is a device that radiates a laser pulse having a high output power to a surrounding atmosphere and receives a laser pulse reflected from a target object present near the same
Implementation Method 2
receives a laser pulse reflected from a target object present near the same
Data Source
AI summary
A device and method are configured to selectively remove specific points included in multiple layers of data received from a sensor. A vehicle may include a LiDAR sensor as the sensor, and a signal processor configured to select a static object included in the multiple layers, to select target layers from among the multiple layers, to define a shape formed by points included in each of the selected target layers, to select a reference layer based on the shape formed by the points, and to remove, based on a distance between a contour formed by the reference layer and each of points of remaining layers and a reference distance, corresponding points from at least one of the remaining layers.