Camera-Based High-Precision Map Construction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current high-precision map construction methods rely on costly laser point cloud data, limiting large-scale deployment and timeliness of map updates in autonomous driving applications.
Innovation Solution
A method using camera pose calculation and absolute depth determination to construct three-dimensional point clouds from video data, allowing for the creation of high-precision maps without the need for high-cost laser devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser point cloud data is used for constructing high-precision maps, then map construction accuracy is improved, but device cost increases significantly
Solution Approach 1:
The patent replaces expensive laser devices with inexpensive camera devices for constructing high-precision maps. The camera captures video data that is then processed to extract three-dimensional point cloud information, achieving map construction accuracy without requiring costly laser scanning equipment.
Solution Approach 2:
The patent substitutes the mechanical laser scanning system with an optical camera-based system. By using video data processing and computational algorithms, the system replaces the physical laser point cloud acquisition mechanism with an optical capture and computational reconstruction approach.
2Manufacturing precision
If laser devices are deployed for map construction, then map construction quality is improved, but deployment scale is limited due to high cost
Solution Approach 1:
The patent enables large-scale deployment by replacing expensive laser devices with inexpensive cameras. Multiple low-cost camera units can be deployed across numerous vehicles simultaneously, allowing parallel map construction from multiple sources without the budget constraints that limit laser device deployment.
Solution Approach 2:
The patent makes map construction capability universal by using standard camera devices that are already present in most vehicles rather than specialized laser equipment. This allows any vehicle equipped with a camera to contribute to map construction, greatly expanding the potential deployment scale.
3Measurement precision
If laser point cloud data is used for map construction, then map accuracy is improved, but timeliness of map updates deteriorates
Solution Approach 1:
The patent enables continuous map updates by utilizing video data from vehicles in normal operation. As vehicles continuously capture video data during regular driving, the system can continuously process this data to update maps in real-time, eliminating the need for periodic laser scanning campaigns.
Solution Approach 2:
The system allows vehicles to automatically contribute their video data for map construction and updates during normal operation. The map construction process becomes self-sustaining, using the existing traffic flow and vehicle operations to continuously improve and update maps without requiring dedicated survey missions.
Data Source
AI summary
Provided are a high-precision map construction method, an electronic device, and a storage medium, relating to the field of high-precision map technology and, in particular, to autonomous driving technology. The implementation solution includes: calculating a pose of a camera at each position point according to a pre-acquired video; calculating an absolute depth of each keypoint in the pre-acquired video according to the pose of the camera at each position point; constructing, according to the absolute depth of each keypoint in the video, a corresponding three-dimensional point cloud of each pixel point in the pre-acquired video; and constructing, according to the corresponding three-dimensional point cloud of each pixel point in the pre-acquired video, a high-precision map corresponding to the pre-acquired video.