Visual Target Tracking Using Spatial Feature Changes in UAV Flight
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing visual tracking methods for moving targets are inadequate when the aerial vehicle or imaging device is at different spatial dispositions relative to the target, such as varying heights, distances, or orientations, leading to reduced tracking accuracy and precision, especially in environments without reliable GPS signals.
Innovation Solution
A system and method that determine changes in features between image frames captured at different times to adjust the movement of a movable object, such as an unmanned aerial vehicle (UAV), to accurately track a target object, using image analyzers and motion controllers to adjust the position and orientation of the imaging device based on changes in bounding box size and position.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual visual tracking methods are used where operators manually select and control the aerial vehicle to track the target, then tracking can be performed with simple equipment, but tracking precision and response speed deteriorate due to human reaction time and manual operation limitations
Solution Approach 1:
The system implements autonomous tracking where the aerial vehicle automatically selects, tracks, and adjusts its position relative to the target without manual intervention. The vehicle independently processes image data, identifies targets, and controls its motion to maintain optimal tracking conditions, making the system self-sufficient in the tracking task.
Solution Approach 2:
The patent replaces manual mechanical control with automated vision-based control systems. Image processing algorithms and computer vision techniques substitute for human operators, enabling automatic target detection, feature extraction, and trajectory calculation to achieve precise tracking without manual intervention.
2Measurement precision
If existing visual tracking methods are used that do not account for spatial dispositions, then the system structure remains simple, but tracking accuracy deteriorates when the aerial vehicle and imaging device are at different heights, distances, and orientations relative to the target
Solution Approach 1:
The system transitions from 2D image plane tracking to 3D spatial tracking by incorporating depth information and spatial disposition calculations. It accounts for variations in height, distance, and orientation by computing three-dimensional positions and transformations, enabling accurate tracking regardless of the aerial vehicle's spatial configuration relative to the target.
Solution Approach 2:
The patent introduces spatial disposition parameters as intermediary variables that mediate between the aerial vehicle's position and the target's position. These intermediate calculations of relative height, distance, and orientation enable the system to compensate for spatial variations and maintain tracking accuracy across different configurations.
3Extent of automation
If automatic visual tracking methods using tracking algorithms are used, then autonomous tracking is achieved, but tracking precision deteriorates when the spatial relations between features change or when the model has errors
Solution Approach 1:
The system implements dynamic tracking that adapts to changing spatial conditions. It continuously updates the spatial disposition model based on real-time image data and adjusts tracking parameters dynamically. The system can handle variations in target scale, orientation, and position by adapting its feature extraction and matching algorithms to current spatial conditions.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system continuously monitors tracking performance and spatial disposition accuracy. Based on feedback from image analysis and position estimation, the system adjusts its tracking parameters and spatial model to compensate for errors and maintain precision under varying spatial conditions.
4Loss of information
If tracking indicators are used that only provide 2D information on the image plane, then the system remains simple, but sufficient spatial information about the aerial vehicle's disposition relative to the target is lost
Solution Approach 1:
The system enhances traditional 2D bounding box indicators by adding three-dimensional spatial information. It computes and displays depth distance, height differential, and orientation angles alongside the 2D tracking box, providing a comprehensive view of the aerial vehicle's spatial disposition relative to the target that preserves all relevant spatial information.
Data Source
AI summary
A method for controlling a movable object to track a target object includes determining a change in one or more features between a first image frame and a second image frame, and adjusting a movement of the movable object based on the change in the one or more features between the first image frame and the second image frame. The one or more features are associated with the target object, and the first image frame and the second image frame are captured at different points in time using an imaging device on the movable object.