Object Position Estimation via Radio Tag and Camera Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object position estimation systems face challenges in accurately identifying and tracking objects when an observation device, such as a camera, is unable to perform ID identification, particularly in environments with varying clothing or multiple individuals with similar features, leading to limited application range and high costs for multi-camera systems.
Innovation Solution
A system that calculates object position likelihoods and ID likelihoods using data from multiple observation units, including those incapable of ID identification, by determining association values based on tracking state information and feature matching across different observation points in time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a camera is used for object identification, then position measuring precision is improved, but ID identification precision deteriorates because objects with similar image-based features cannot be identified with high precision
Solution Approach 1:
The patent combines multiple types of observation devices (radio tags and cameras) to integrate their respective advantages. Radio tags provide reliable ID identification while cameras provide precise position measurement. By merging these devices and using a data association filter to process information from both sources, the system achieves both reliable ID identification and precise position measurement simultaneously, resolving the contradiction between the two performance aspects.
2Reliability
If radio tags are used for object identification, then ID identification reliability is improved, but position measuring precision deteriorates
Solution Approach 1:
The system merges radio tags and cameras into a unified observation system. Radio tags provide reliable ID identification through unique ID information, while cameras provide precise position measurement through image-based features. The data association filter processes information from both devices, combining their strengths to achieve both reliable identification and precise positioning, thereby resolving the contradiction between ID reliability and position precision.
3Adaptability or versatility
If multiple cameras are arranged all over the place to improve detection coverage, then detection capability is improved, but system cost increases extremely high
Solution Approach 1:
Instead of deploying multiple expensive cameras throughout the entire space, the patent combines a limited number of cameras with radio tags. This hybrid approach provides comprehensive detection coverage by leveraging the complementary strengths of both device types, significantly reducing system cost while maintaining or improving detection capability compared to using only multiple cameras.
Solution Approach 2:
The observation devices are designed to perform multiple functions: radio tags provide both ID identification and position information, while cameras provide position measurement and visual confirmation. This multi-functionality reduces the need for separate dedicated devices for each function, thereby reducing overall system cost while maintaining comprehensive detection capability.
Data Source
AI summary
ID likelihoods and position likelihoods of an object are detected by a first observation device, and position likelihoods of the object and tracking states of the object are detected by a second observation device; thus, the object detected by the second observation device and the object ID are associated with each other by an association unit, and based upon information from the second observation device and the association unit, the ID likelihoods of the object detected by the second observation device are determined by a second object ID likelihood determination unit so that the object position is estimated by an object position estimation unit based upon the ID likelihoods and position likelihoods.