Multi-Camera Person Tracking with Trajectory-Based Time Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing biometric authentication systems using multiple cameras with different shooting intervals face reduced accuracy due to unused images and inconsistent frame rates, leading to degraded person tracking and judgment accuracy.
Innovation Solution
A method that synchronizes image processing across cameras with different frame rates by determining a common execution interval for image processing, allowing accurate tracking and judgment of individuals across multiple cameras.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a predetermined interval is set for authentication processing using images captured by multiple cameras, then authentication can be executed at consistent time points, but images taken at intervals shorter than the set interval are not used, causing deterioration of person tracking accuracy
Solution Approach 1:
The patent divides the image processing task into two separate processing paths: one for person tracking that processes all images at their original capture intervals, and another for authentication that operates at predetermined time points. This segmentation allows each path to optimize for its specific purpose without compromising the other.
Solution Approach 2:
The system performs person tracking processing on all captured images before authentication processing. By preliminarily calculating trajectories from all available images, the system accumulates comprehensive tracking data that improves authentication accuracy when authentication is performed at predetermined intervals.
2Productivity
If multiple cameras with different shooting intervals are used, then more frequent monitoring is achieved, but inconsistent frame rates cause degraded judgment accuracy
Solution Approach 1:
The patent introduces trajectory information as an intermediary that bridges images from cameras with different frame rates. By calculating and matching trajectories across images from multiple cameras taken at different times, the system can accurately determine whether captured images show the same person despite inconsistent frame rates.
Solution Approach 2:
The system changes the evaluation parameter from direct image comparison to trajectory-based matching. Instead of comparing images at uniform intervals, it calculates trajectories from images at their native intervals and uses trajectory matching to determine person identity, thereby accommodating different frame rates while maintaining accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In a determination method, a computer executes a process including, when a plurality of first images in a time series photographed by a first camera every first time and a plurality of second images in a time series photographed by a second camera every second time are acquired, calculating a trajectory of a person that is in each of the first images and a trajectory of a person that is in each of the second images, determining a third time according to a time for acquiring each of the first images from the first camera and a time for acquiring each of the second images from the second camera, and judging whether a person whose trajectory is calculated from an image of the plurality of first images corresponding to the third time after determined and a person whose trajectory is calculated from an image of the plurality of second images corresponding to the third time after determined are the same.