UAV Asset Localization Using Homography for Identical Landing Targets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned aerial vehicles (UAVs) face challenges in disambiguating between visually identical assets such as autoloaders and charging pads, especially at cruising altitudes, due to the unreliability of visual fiducial markers and the need for precise positioning, which affects the efficiency and accuracy of package delivery operations.
Innovation Solution
A system and method using machine-learned image features and homography transformations to map annotated reference aerial images onto query images, enabling accurate localization and identification of assets without relying on large fiducial markers, by extracting and matching image features across different altitudes and environmental conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual fiducial markers are used for asset identification, then asset localization is possible, but reliability deteriorates at cruising altitudes and visual disambiguation between identical assets becomes unreliable
Solution Approach 1:
The patent creates a digital copy (map) of the geographical area containing visual representations of multiple assets. This map serves as a reference that can be compared against live camera feeds to identify specific assets without relying on visual fiducial markers alone. The system copies the spatial relationships and visual characteristics of assets into a persistent digital representation that enhances identification reliability.
Solution Approach 2:
The patent transitions from two-dimensional visual marker recognition to three-dimensional spatial reasoning by incorporating altitude information and creating volumetric representations of assets. By adding the altitude dimension and using depth from defocus, the system can disambiguate between visually identical assets at different heights and positions in three-dimensional space.
2Difficulty of detecting and measuring
If large fiducial markers are used for asset identification, then asset detection is possible, but device complexity and operational efficiency worsen due to dependency on these markers
Solution Approach 1:
The system uses the asset's own visual characteristics and spatial position to identify itself. Rather than requiring external fiducial markers attached to assets, the methodology enables assets to be identified through their inherent visual features captured by the camera, combined with spatial context from the map and depth information from defocus effects.
Solution Approach 2:
The patent extracts and removes the dependency on fiducial markers from the asset identification system. By taking out the requirement for these markers and replacing them with alternative identification methods using visual features, spatial reasoning, and depth from defocus, the system reduces device complexity and operational constraints.
3Measurement precision
If visual fiducial markers are used for asset identification, then asset localization is possible, but productivity deteriorates due to reduced efficiency in package delivery operations
Solution Approach 1:
The system performs preliminary actions by pre-generating a map of the delivery area with visual representations of assets before the actual delivery operation. This pre-computed spatial and visual information is stored and readily available when the drone needs to identify and interact with assets, eliminating the need for real-time fiducial marker detection and improving delivery efficiency.
Data Source
AI summary
A technique for a UAV includes acquiring a query aerial image with an onboard camera of the UAV and a reference aerial image, the query aerial image including multiple instances of an asset and the reference aerial image including annotated pixels indicating an expected location and an identification for the multiple instances of the asset. The technique further includes identifying a plurality of corresponding pixels between the query aerial image and the reference aerial image, determining a homography transformation describing a relationship between the query aerial image and the reference aerial image, annotating the query aerial image to identify a first instance of the asset included in the multiple instances of the asset within the query aerial image, and instructing the UAV to perform an action associated with the first instance of the asset.