Multilayer LiDAR Road Detection With Threshold-Based Road Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in efficiently detecting and isolating road markings in autonomous vehicles due to insufficient precision with commercial GNSS systems and the complexity of processing large datasets from LIDAR sensors, which require prior identification of the road and are hindered by the loss of 3D information with single-camera systems.
Innovation Solution
A road detection method using a rotating multilayer LIDAR sensor that filters data based on polar coordinates, determines standard deviations, and applies Gaussian filters to segment road data, allowing for efficient identification and tracking of road markings by deeming data as part of the road if standard deviations are below a threshold, and prioritizes data processing to reduce computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LIDAR data is processed to detect road markings, then road detection accuracy is improved, but data processing time increases due to large dataset volume
Solution Approach 1:
The patent segments the LIDAR point cloud data into multiple horizontal slices at different height levels. Each slice contains points within a specific vertical range, allowing parallel processing of smaller subsets of data rather than processing the entire large dataset at once. This reduces computational complexity and processing time while maintaining detection accuracy.
Solution Approach 2:
The patent extracts and removes points that do not belong to the road surface by comparing height values against expected road elevation ranges. By filtering out irrelevant points (vegetation, buildings, sky points) and keeping only road-related points, the processing load is reduced while preserving the essential road marking detection capability.
2Reliability
If all LIDAR data is processed, then complete road coverage is achieved, but computational load increases
Solution Approach 1:
The patent applies different processing strategies to different spatial regions. Road areas receive detailed processing with multiple height slice analysis, while non-road areas are quickly identified and excluded. This localized approach ensures complete road coverage where needed while reducing computational effort in irrelevant regions.
Solution Approach 2:
The patent performs preliminary filtering of points based on height criteria before detailed road marking detection. By pre-identifying and removing obviously non-road points in an initial pass, the system reduces the dataset size beforehand, making subsequent detailed processing more efficient without compromising road coverage.
3Productivity
If road data is segmented using height filtering, then processing speed is improved, but measurement precision may be affected
Solution Approach 1:
The patent segments road data into multiple horizontal slices at different heights rather than using a single threshold. This multi-level slicing approach captures road markings that may appear at different vertical positions due to perspective, curvature, or varying road surfaces, maintaining precision while enabling efficient slice-by-slice processing.
Solution Approach 2:
The patent dynamically adjusts height thresholds and slice parameters based on the specific data characteristics and detection requirements. By adapting the filtering parameters rather than using fixed thresholds, the system maintains high detection precision across varying road conditions while optimizing processing speed through appropriate parameter selection.
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
This method effectively reduces data processing time and enhances road detection accuracy by isolating road markings from LIDAR data, enabling efficient road segmentation and tracking, even under partial occlusion, while maintaining road detection capacity with degraded sensor resolution.
Implementation Method 1
LIDAR data of the environment of the motor vehicle are acquired in polar coordinates with the LIDAR sensor
Data Source
Figure 1
Figure 2
AI summary
Method of road detection for an automotive vehicle fitted with at least one rotary multilayer LIDAR sensor (1) and with at least one means for processing data, wherein LIDAR data of the environment of the automotive vehicle are acquired, from the LIDAR data acquired are deleted the data for which the angle of the beam with respect to the plane of the vehicle is greater than or equal to zero, for each angle of the beam with respect to the plane of the vehicle, and for each LIDAR datum, theoretical coordinates of a point are determined, corresponding to the LIDAR datum, dependent notably on the angle of the beam with respect to the plane of the vehicle and on the angle of the beam in the plane of the vehicle, the standard deviation between the theoretical coordinates and the coordinates of the LIDAR datum is determined, and it is determined that the LIDAR datum forms part of the road if the standard deviation is less than a predefined threshold.