Calibrating Moving Obstacle Posture in 3D Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the position and orientation of obstacles in point cloud data simulations for unmanned driving suffer from significant errors, with deviations of meters in spatial directions and angles, leading to inaccurate obstacle positioning.
Innovation Solution
A method that involves obtaining a 3D map with static obstacles, selecting a frame of point cloud data containing both static and moving obstacles, determining posture information, registering the frame with the 3D map, calculating posture offset information, and calibrating the moving obstacles' posture information to reduce errors and improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If obstacle position and orientation are obtained through road data collection and transformation into 3D map, then obstacle positioning can be achieved, but large errors of meters in spatial directions and angles occur
Solution Approach 1:
The patent segments the obstacle positioning process into two independent stages: first obtaining initial positioning information through road data collection and transformation, then performing calibration by matching static obstacles between point cloud data and 3D map. This segmentation allows each stage to focus on specific tasks, with the calibration stage专门 addressing the accuracy issue through static obstacle matching, thereby resolving the contradiction between achieving positioning and maintaining accuracy.
Solution Approach 2:
The patent introduces static obstacles as an intermediary element to bridge the point cloud data and 3D map for calibration purposes. By using static obstacles (which remain unchanged over time) as reference objects, the system can establish accurate coordinate transformations and calibrate moving obstacle positions. This intermediary approach enables precise positioning without directly relying on the error-prone initial road data transformation.
2Productivity
If rough obstacle detection is performed from road data, then obstacle detection can be completed, but deviations of meters or even tens of meters occur in lateral, longitudinal and vertical directions
Solution Approach 1:
The patent performs preliminary action by first detecting obstacles roughly from road data to obtain initial positioning information, then subsequently calibrating these results using static obstacle matching. This preliminary detection stage ensures high productivity by quickly identifying obstacles, while the following calibration stage corrects the meter-level deviations. The two-stage approach resolves the contradiction by handling efficiency and precision at different sequential stages.
3Device complexity
If obstacle positioning is performed without calibration, then the process is simple, but varying degrees of error exist in three angles of pitch, roll, and yaw
Solution Approach 1:
The patent segments the positioning process into a simple initial detection phase and a calibration phase. The segmentation allows the system to maintain simplicity in the overall workflow while introducing targeted calibration operations only when needed. By separating these functions, the system achieves good orientation accuracy without making the entire positioning process overly complex.
Solution Approach 2:
The calibration process uses self-service by automatically matching static obstacles between point cloud data and 3D map to compute correction parameters. The system performs self-calibration without requiring external intervention or complex manual adjustments, thereby improving orientation measurement accuracy while keeping the added complexity minimal and automated.
Data Source
AI summary
A method, an apparatus, a device, and a medium for calibrating a posture of a moving obstacle are provided. The method includes: obtaining a 3D map, the 3D map including first static obstacles; selecting a target frame of data, the target frame of data including second static obstacles and one or more moving obstacles; determining posture information of each of the one or more moving obstacles in a coordinate system of the 3D map; registering the target frame of data with the 3D map; determining posture offset information of the target frame of data in the coordinate system according to a registration result; calibrating the posture information of each of the one or more moving obstacles according to the posture offset information; and adding each of the one or more moving obstacles after the calibrating into the 3D map.