Visual Heading Correction Using Distinctive Feature Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Low-grade inertial measurement units (IMUs) struggle to accurately determine heading angles without significant velocity changes, requiring lengthy S-turn maneuvers for alignment, which distracts from mission goals and is inefficient.
Innovation Solution
A system comprising an image system, GPS, and a coupled processor that captures and matches distinctive features over time, determines unit-vectors, and minimizes errors in the epipolar equation to calculate corrected heading, eliminating the need for S-turns by using visual cues and combining GPS and vision observations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional S-turn alignment procedures are used to align low-grade IMU with GPS, then heading alignment accuracy can be achieved, but the alignment process takes several minutes and distracts from mission goals
Solution Approach 1:
The patent replaces the mechanical S-turn maneuver-based alignment method with a vision-based alignment system. The image system captures images of distinctive features, and the coupled processor uses these visual observations to determine alignment parameters, substituting the need for physical S-turn maneuvers with optical measurement and computational processing.
Solution Approach 2:
The patent introduces distinctive visual features in the field of view as intermediaries for alignment. These features serve as reference points that mediate between the image system and the alignment calculation, enabling the determination of alignment parameters without requiring S-turn maneuvers. The features act as a bridge for transferring spatial information.
2Ease of manufacture
If low-grade IMU is used without velocity changes, then cost is reduced, but heading angle determination accuracy deteriorates to about 0.1 radians
Solution Approach 1:
The patent merges the low-grade IMU with an image system and GPS receiver into an integrated navigation system. The coupled processor combines data from all three sources, fusing the inertial measurements with visual features and GPS position/velocity information to achieve accurate heading determination without requiring high-cost IMU hardware.
Solution Approach 2:
The patent implements feedback by using the image system to continuously observe distinctive features and provide visual measurements back to the coupled processor. This visual feedback loop enables continuous correction and refinement of heading estimates, allowing low-grade IMU to achieve accurate heading determination through ongoing comparison with visual reference data.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A system and a method for determining heading of a traveling vehicle are provided, The system and method employ an image system to capture and match multiple distinctive features of an image area at a first and second point in time and to determine unit-vectors associated with the multiple distinctive features. A global positioning system (GPS) provides a translation vector for the first and second point in time and a coupled processor employs the unit-vectors and the translation vector to provide a corrected heading.