Visual Localization Using Aerial-Ground Models for Large-Scale Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned aerial vehicle/satellite base map-based visual localization methods suffer from low localization success rate and accuracy in large-scale scenarios.
Innovation Solution
A visual localization method that utilizes an aerial-ground model, combining an aerial model with a ground model to determine a second pose when a ground model corresponding to a first pose exists, enhancing localization accuracy and success rate through skyline and semantic information analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If unmanned aerial vehicle/satellite base map-based visual localization method is used, then large-scale scenario localization can be achieved, but localization success rate and accuracy are low
Solution Approach 1:
The patent segments the localization process into two distinct stages: coarse localization using the aerial model for large-scale coverage, and fine localization using the ground model for high accuracy. This segmentation allows each model to operate in its optimal performance range, resolving the contradiction between coverage area and localization accuracy
Solution Approach 2:
The patent implements a nested structure where the ground model is embedded within the aerial model framework. The coarse localization result from the aerial model provides the initial pose estimate, which then feeds into the fine localization using the ground model. This nesting allows the system to maintain both large-scale coverage and high localization accuracy simultaneously
2Productivity
If aerial model is used for localization, then fast and efficient coarse localization can be achieved, but localization accuracy is insufficient
Solution Approach 1:
The patent applies preliminary action by using the aerial model to perform coarse localization first, obtaining an initial pose estimate quickly. This preliminary result then serves as the starting point for the subsequent fine localization using the ground model, enabling the system to achieve both fast initial localization and high final accuracy
3Measurement precision
If ground model is used for localization, then high localization accuracy can be achieved, but computational complexity and resource consumption increase
Solution Approach 1:
The patent implements partial action by selectively applying the computationally intensive ground model only when needed for fine localization, rather than using it for all localization tasks. The system first attempts coarse localization with the simpler aerial model, and only invokes the ground model when higher accuracy is required, thus reducing overall computational complexity while maintaining high accuracy when necessary
Data Source
AI summary
A visual localization method and apparatus are provided. The method includes obtaining a captured first image, determining a first pose based on the first image and an aerial model, determining whether a ground model corresponding to the first pose exists in an aerial-ground model, and when the ground model corresponding to the first pose exists, determining a second pose based on the ground model. The aerial-ground model includes the aerial model and the ground model mapped to the aerial model, a coordinate system of the ground model is the same as a coordinate system of the aerial model, and localization accuracy of the second pose is higher than localization accuracy of the first pose. Performing fine visual localization based on the ground model can improve accuracy and a success rate of localization.