Multi-Camera Geolocation Using Markers for Cross-View Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing monitoring systems using two-dimensional cameras struggle to accurately track and identify objects across different camera views due to perspective changes and lack of geospatial coordination, leading to inefficiencies in object tracking and identification.
Innovation Solution
A system utilizing multiple imaging devices with markers having known geospatial coordinates, a controller to determine transforms between camera views, and multi-valued logic to confirm object identity across images, enabling accurate geolocation and velocity estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple imaging devices are deployed to improve monitoring coverage, then the area covered is increased, but the complexity of tracking objects across different camera views increases
Solution Approach 1:
The patent introduces markers with known geospatial coordinates as intermediary reference points between multiple camera views. These markers serve as a common reference framework that mediates the transformation between different camera coordinate systems, enabling consistent object tracking across cameras without requiring complex direct coordination between each camera pair.
Solution Approach 2:
The patent transforms the tracking problem from two-dimensional image coordinates to three-dimensional geospatial coordinates by introducing the vertical dimension of elevation angles. This dimensional transformation allows objects to be consistently identified across camera views through their unique geospatial positions rather than relying solely on 2D image overlap, which is ambiguous in multi-camera systems.
2Measurement precision
If geospatial coordinates are introduced to improve object identification accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent performs preliminary action by pre-establishing the geospatial coordinates of markers before the actual object tracking begins. These pre-computed reference coordinates are stored and used during object identification, eliminating the need for real-time complex calculations and reducing computational burden during dynamic object tracking operations.
Solution Approach 2:
The patent creates a virtual copy of the physical world by mapping camera observations to a standardized geospatial coordinate system. This coordinate system acts as a simplified representation that preserves essential location information while filtering out camera-specific complexities, enabling precise object identification without requiring direct manipulation of complex camera parameters.
3Measurement precision
If transforms between camera views are calculated to enable object tracking, then object tracking accuracy is improved, but processing time increases
Solution Approach 1:
The patent calculates camera transforms and marker coordinates in advance before object tracking begins. By pre-computing the relationship between camera views and geospatial coordinates, the system avoids performing these complex transformations in real-time during object tracking, significantly reducing processing time while maintaining high tracking accuracy.
Solution Approach 2:
The patent uses detected marker positions in images as feedback to verify and refine the camera transforms. By continuously comparing observed marker locations with expected locations based on pre-calculated transforms, the system can correct for drift and maintain accuracy without requiring continuous re-computation of complex transformation matrices.
Data Source
AI summary
An example system includes a first imaging device located at a first location and a second imaging device located at a second location; a plurality of markers, wherein each of the plurality of markers corresponds to a known geospatial coordinate; and a controller operably coupled to the plurality of imaging devices, wherein the controller is configured to: receive a first image from a first imaging device, receive a second image from a second imaging device, wherein the second image captures a second marker of the plurality of markers and the object, estimate, based on a first transform, a first object location of an object in the first image, estimate, based on a second transform, a second object location of an object in the second image, and determine that the object in the first image and the second image is the same object.