Aerial Target Tracking with Feature Re-Acquisition After Occlusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object tracking systems for aerial vehicles face challenges such as target deformation, occlusion, and long-term tracking, especially with limited computing resources, making it difficult to maintain accurate tracking over time.
Innovation Solution
A system and method that utilize a feature model to represent imagery characteristics of a target, allowing a controller to extract features from images captured by a movable object, determine similarity, and generate control signals to track the target, including re-targeting capabilities when the target is lost.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If object tracking is performed using conventional methods, then tracking can be achieved in the short term, but tracking accuracy deteriorates over long periods due to target deformation, occlusion, and limited computing resources
Solution Approach 1:
The system performs preliminary actions by proactively detecting potential tracking failures before they occur and preparing recovery strategies. The controller continuously monitors tracking quality metrics and preemptively switches to alternative tracking methods or re-acquires target features when degradation is detected, ensuring continuous accurate tracking over extended periods.
Solution Approach 2:
The system implements feedback mechanisms by continuously evaluating tracking performance and using this information to adjust tracking parameters and strategies. When tracking accuracy deteriorates due to deformation or occlusion, the feedback loop triggers re-acquisition of target features or adjustment of feature model parameters, maintaining reliable tracking throughout the duration.
2Reliability
If advanced tracking algorithms are used to maintain accuracy, then tracking reliability improves, but computational resource consumption increases
Solution Approach 1:
The system dynamically adjusts the complexity of tracking algorithms based on current operational conditions and available resources. When resources are abundant and tracking is stable, simpler methods are used. When tracking challenges arise or resources are limited, the system adaptively switches to more sophisticated algorithms only when necessary, optimizing the balance between reliability and resource consumption.
Solution Approach 2:
The system changes parameters of the tracking algorithm based on target characteristics and environmental conditions. Feature extraction parameters, matching thresholds, and model update frequencies are dynamically adjusted to match the complexity required for the current tracking scenario, reducing unnecessary computational overhead while maintaining reliability.
3Reliability
If the target is re-acquired after loss, then tracking continuity is restored, but time is lost during the re-acquisition process
Solution Approach 1:
The system performs preliminary actions by maintaining feature models and candidate target data in advance during normal tracking operation. When tracking failure is detected, these pre-prepared features enable immediate re-acquisition without extensive searching, significantly reducing the time penalty while restoring tracking continuity.
Solution Approach 2:
The system prepares compensatory measures in advance by maintaining alternative feature representations and backup tracking hypotheses during stable tracking periods. When tracking is lost, these pre-prepared alternatives serve as cushions that enable rapid recovery, minimizing the disruption and time loss associated with re-acquisition.
Data Source
AI summary
A method for supporting target tracking includes obtaining a feature model for a target that represents imagery characteristics of the target, extracting one or more features from one or more images captured by an imaging device carried by a movable object, and applying the feature model on the one or more features to determine similarity between the one or more features and the feature model.