Joint Radon Transform for Dense Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking technologies, particularly in vehicular environments, face challenges in accurately tracking dense, extended objects in urban settings due to their reliance on Cartesian distance calculations, which are less effective in such scenarios.
Innovation Solution
The implementation of a joint radon transform association method that uses an energy score to associate detected objects and their movement predictions, replacing traditional distance-based calculations with an energy score-based approach for improved tracking in complex environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Cartesian distance calculations are used for object association, then the tracking system is simple to implement, but tracking accuracy deteriorates in dense urban environments
Solution Approach 1:
The patent transforms the association metric from Cartesian distance to energy score based on Radon transform coefficients. This parameter change enables accurate association of dense extended objects by capturing their spatial-energy characteristics rather than simple Euclidean distances, directly resolving the tracking accuracy issue in urban environments.
Solution Approach 2:
The patent replaces the traditional geometric distance calculation mechanism with a signal processing mechanism using Radon transform. This substitution allows the system to analyze objects in the spatial-frequency domain, providing superior discrimination capability for dense objects while maintaining computational feasibility through efficient transform algorithms.
2Reliability
If traditional distance-based association is used, then computational complexity is low, but tracking reliability deteriorates for extended objects
Solution Approach 1:
The patent segments the association process into multiple stages: initial association using simplified metrics, followed by refined association using Radon transform energy scores for candidate pairs. This segmentation allows the system to achieve high reliability for extended objects while minimizing computational power by applying complex processing only where necessary.
Solution Approach 2:
The patent performs preliminary association filtering before applying the computationally intensive Radon transform. By pre-filtering obvious matches and eliminating clearly incorrect associations, the system reduces the number of pairs requiring full energy score calculation, thereby maintaining tracking reliability while reducing overall computational power requirements.
Data Source
AI summary
An example method for performing a joint radon transform association includes detecting, by a processing device, a target object to track relative to a vehicle. The method further includes performing, by the processing device, the joint radon transform association on the target object to generate association candidates. The method further includes tracking, by the processing device, the target object relative to the vehicle using the association candidates. The method further includes controlling, by the processing device, the vehicle based at least in part on tracking the target object.