Hybrid Vision RTLS Tracking for Occluded Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tracking systems, both vision-based and RTLS, face challenges in accurately locating and identifying objects that are occluded or in close proximity, especially in scenarios like team games where objects have similar appearances and move unpredictably, leading to positional data uncertainty and incorrect identity swapping.
Innovation Solution
A synergetic tracking system that combines vision-based and RTLS technologies to fuse positional data from cameras and telemetry signals from transponders, using methods like trilateration and triangulation to resolve occlusion and identity issues, even with limited lines of sight, by integrating RTLS and vision-based systems to provide accurate and continuous tracking of multiple objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vision-based tracking is used to locate objects, then positional data fidelity is high, but tracking fails when objects are occluded or in clusters
Solution Approach 1:
The patent combines vision-based tracking and RTLS into a unified hybrid system. The vision system provides high-precision positional data when objects are visible, while the RTLS provides continuous tracking data when objects are occluded. The system merges these two data sources through data fusion algorithms to maintain continuous, accurate tracking throughout all conditions.
Solution Approach 2:
The patent introduces transponders attached to tracked objects as intermediary elements. These transponders enable the RTLS to detect and track objects even when they are occluded from camera view. The transponders act as mediators that bridge the gap between vision-based and RTLS tracking, allowing seamless transition between the two systems.
2Reliability
If RTLS is used to track objects, then tracking continues under occlusion, but positional accuracy decreases due to signal attenuation and multipath fading
Solution Approach 1:
The patent merges RTLS and vision-based tracking systems where each compensates for the other's weaknesses. The vision system corrects RTLS positional inaccuracies when objects are visible, while RTLS maintains tracking continuity when vision is blocked. This combination achieves both continuous tracking and high positional accuracy.
Solution Approach 2:
The patent implements feedback mechanisms where the hybrid system continuously evaluates the quality and availability of data from both vision and RTLS sources. When vision data is available and reliable, it overrides or corrects RTLS data. When vision data is unavailable or unreliable, the system relies on RTLS data. This feedback loop ensures optimal positional accuracy under all conditions.
3Measurement precision
If objects have similar appearances and move in clusters, then vision-based tracking cannot distinguish individual objects, but this is a common scenario in team games
Solution Approach 1:
The patent uses transponders as unique identifiers attached to each object. These transponders serve as intermediaries that provide unambiguous identification of individual objects even when they have similar appearances or are in close proximity. The RTLS reads transponder IDs to distinguish and track each object uniquely, solving the identification problem in team game scenarios.
Solution Approach 2:
The patent segments the identification function between vision-based appearance recognition and RTLS-based transponder identification. While vision can struggle with similar appearances, the transponder system provides unique identification for each object. This segmentation of identification methods ensures accurate object distinction even in challenging scenarios.
4Reliability
If multiple tracking systems are combined, then tracking reliability under occlusion improves, but system complexity increases
Solution Approach 1:
The patent designs the hybrid system with multi-functional components. The transponders serve multiple purposes: providing unique identification, enabling RTLS tracking, and potentially storing additional data. The fusion algorithm handles multiple data sources and conditions through a unified mathematical framework. This universality reduces overall system complexity despite combining multiple tracking approaches.
Solution Approach 2:
The patent dynamically changes operational parameters based on conditions. The system adjusts the weight and trust given to vision versus RTLS data based on visibility conditions, signal quality, and object density. This adaptive parameter adjustment simplifies decision-making in the fusion algorithm and optimizes performance without requiring complex rule-based systems.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
A method for tracking objects in a scene may include receiving visual-based information of the scene with a vision-based tracking system and telemetry-based information of the scene with a RTLS-based tracking system. The method may also include determining a location and identity of a first object in the scene using a combination of the visual-based information and the telemetry-based information. Another method for tracking objects in a scene may include detecting a location and identity of a first object and determining a telemetry-based measurement between the first object and a second object using a real time locating system (RTLS)-based tracking system. The method may further include determining a location and identity of the second object based on the detected location of the first object and the determined measurement. A system for tracking objects in a scene may include visual -based and telemetry-based information receivers and an object tracker.