LIDAR Calibration via Point Cloud Segmentation and NDT Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional SLAM systems face challenges in simultaneous localization and mapping due to high computational complexity, leading to inefficiencies in data collection and processing for creating accurate high-definition road maps for autonomous driving, particularly with synchronization issues between various sensors and large computation requirements.
Innovation Solution
A method for calibrating multiple LIDARs by acquiring and preprocessing point cloud data from both a reference LIDAR and multiple vehicle-mounted LIDARs, using voxelization, outlier detection, ground detection, and normal distribution transform (NDT) matching to achieve accurate calibration and synchronization, facilitated by a computer program executed on a recording medium.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional SLAM systems are used for simultaneous localization and mapping, then localization and mapping can be performed, but computational complexity increases significantly leading to long processing time
Solution Approach 1:
The patent segments the point cloud data into multiple sections along the vehicle's traveling path, processing each section separately through NDT matching. This divides the large-scale computational problem into smaller, more manageable segments that can be processed efficiently while maintaining overall mapping accuracy.
Solution Approach 2:
The patent performs preliminary voxelization of point cloud data before NDT matching, and uses specification values to pre-align coordinate systems. These preliminary actions prepare the data in advance, reducing the computational burden during the actual calibration and mapping processes.
2Quantity of substance
If multiple LIDARs are mounted on a vehicle for data collection, then more comprehensive spatial information can be acquired, but synchronization issues between sensors arise
Solution Approach 1:
The patent introduces a reference LIDAR as an intermediary to establish a reference map, against which vehicle-mounted LIDAR data is calibrated. This reference framework mediates the synchronization between multiple LIDARs by providing a common coordinate system and temporal reference point.
Solution Approach 2:
The patent changes the parameter of coordinate system alignment by using specification values to transform and align the coordinate systems of multiple LIDARs. This parameter transformation enables consistent spatial registration across different sensors mounted at various positions.
3Measurement precision
If high-definition road maps are created with detailed information, then autonomous driving accuracy improves, but data processing complexity increases
Solution Approach 1:
The patent segments the point cloud data into multiple sections along the vehicle's traveling path, processing each section separately through NDT matching. This divides the large-scale computational problem into smaller, more manageable segments that can be processed efficiently while maintaining overall mapping accuracy.
Solution Approach 2:
The patent applies voxelization as a preprocessing step that partially processes the point cloud data before NDT matching, reducing the data complexity while preserving essential spatial information needed for high-definition map creation.
Data Source
AI summary
A calibration method for multiple LIDARS may comprise the steps of acquiring, by a data creation device, first point cloud data from a first LIDAR in order to create a reference map, a plurality of second point cloud data acquired from a plurality of second LIDARs mounted on a vehicle traveling on a path on the reference map and specification values for the first LIDAR and the plurality of second LIDARs; performing, by the data creation device, preprocessing on the first point cloud data and the plurality of second point cloud data; and performing, by the data creation device, calibration on the preprocessed first point cloud data and plurality of second point cloud data.