High Definition Map Construction via LiDAR-Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electronic maps used for navigation lack the high definition and detailed road information required for automated driving, with resolutions typically at a meter level, which is insufficient for precise navigation and object placement in autonomous systems.
Innovation Solution
A map construction system on a mobile vehicle that combines a positioning system, image capturing device, and LiDAR to identify and project target objects onto high definition maps, using world coordinates obtained from LiDAR data and image coordinates to accurately place objects in a high definition map.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional electronic maps are used for navigation, then the maps can be used for general travel purposes, but the resolution is only at meter level which is insufficient for automated driving requirements
Solution Approach 1:
The patent combines multiple sensing systems (image capturing device, LiDAR, positioning system) into an integrated map construction system that simultaneously collects visual and spatial data to generate high-definition maps with centimeter-level precision, resolving the contradiction between map resolution and system complexity by merging complementary technologies
Solution Approach 2:
The patent transitions from 2D conventional maps to 3D high-definition maps by incorporating LiDAR point cloud data and spatial coordinates, adding depth and dimensional information to achieve centimeter-level precision in three-dimensional space, which enables automated driving applications
2Manufacturing precision
If high definition maps with centimeter level definition are created for automated driving, then the precision is sufficient for lane lines and shoulders, but the complexity of constructing and maintaining such maps increases significantly
Solution Approach 1:
The map construction system operates autonomously by automatically capturing images, collecting LiDAR data, identifying target objects, and constructing high-definition maps without manual intervention, achieving centimeter-level precision while reducing construction difficulty through self-service automation
Solution Approach 2:
The patent replaces manual map construction methods with automated systems that use image recognition algorithms and LiDAR processing to identify and map target objects, substituting mechanical/manual operations with computational methods to achieve high precision efficiently
3Measurement precision
If target objects are identified and placed in high definition maps using world coordinates, then the accuracy of object placement is improved, but the processing time and computational complexity increase
Solution Approach 1:
The system performs preliminary actions by pre-processing images and LiDAR data to extract target object information and calculate world coordinates in advance, enabling accurate object placement while reducing real-time processing time through preparatory computational steps
Solution Approach 2:
The patent introduces an intermediary coordinate transformation process that converts image coordinates to world coordinates through a structured mathematical model, serving as a mediator between image processing and map construction to achieve accurate placement without excessive computational overhead
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 approach enables the efficient construction of high definition maps by automatically identifying and placing target objects with precise world coordinates, enhancing the accuracy and efficiency of map creation for autonomous driving systems.
Implementation Method 1
The LiDAR is configured to periodically obtain LiDAR data of the surroundings outside the mobile vehicle
Implementation Method 2
The image capturing device is configured to periodically capture a first image of surroundings outside the mobile vehicle
Data Source
AI summary
The disclosure provides a map construction method, and the map construction method includes: periodically detecting a base world coordinate of a mobile vehicle; periodically capturing a first image of surroundings outside the mobile vehicle; periodically obtaining LiDAR data of the surroundings outside the mobile vehicle; identifying a plurality of target objects in the first image and projecting a plurality of LiDAR data points onto the first image to obtain a second image according to the LiDAR data; identifying an image coordinate of each of a plurality of projected target LiDAR data points in a selected target object; obtaining an object world coordinate of each of the target LiDAR data points according to the image coordinates, LiDAR information, and the base world coordinate; and disposing the selected target object into a high definition map according to the object world coordinates.