Multi-Camera Trajectory Synchronization for Occlusion Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for predicting the trajectory of moving objects using radar or camera images face challenges in identifying objects with different appearances and are affected by noise due to irregular radio wave reflections, leading to inaccuracies, especially in occlusion states where objects are hidden.
Innovation Solution
An information processing device and method that integrates trajectories from multiple cameras by predicting object trajectories, transforming them into reference coordinates, calculating correlation, and synchronizing images spatially and temporally to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If radar is used for trajectory prediction, then detection range is extended, but measurement precision deteriorates due to irregular reflection of radio waves and difficulty in identifying objects with different appearances
Solution Approach 1:
The system segments the detection task by using multiple cameras positioned at different locations, each capturing trajectories from different viewpoints. This divides the complex radar detection problem into multiple simpler visual detection tasks, improving precision while maintaining extended detection coverage through the network of cameras.
Solution Approach 2:
The system transitions from radar's electromagnetic wave detection to optical dimension detection using cameras. By capturing trajectories in the visual spectrum from multiple angular dimensions, the system achieves better object identification precision while maintaining spatial coverage through multi-camera positioning.
2Measurement precision
If multiple cameras are used for trajectory prediction, then measurement precision improves, but device complexity increases due to the need for coordinate transformation and spatiotemporal synchronization
Solution Approach 1:
The system performs preliminary coordinate transformation by defining a reference coordinate system before trajectory integration. By pre-establishing the transformation framework and synchronization parameters, the system reduces the complexity of real-time processing while maintaining high measurement precision across multiple cameras.
3Measurement precision
If trajectory integration is performed with high correlation threshold, then prediction accuracy improves, but loss of information increases when objects are in occlusion states
Solution Approach 1:
The system merges trajectory information from multiple camera viewpoints into a unified reference coordinate system. By combining detection results from different angles and positions, the system maintains trajectory continuity during occlusion events, reducing information loss while improving overall prediction accuracy through redundant observation paths.
Data Source
AI summary
An information processing device according to the present disclosure includes: a memory configured to store instructions; and one or more processors configured to execute the instructions to: predict, for each of a plurality of cameras, a trajectory of an object included in at least one of a plurality of target images with reference to the plurality of target images captured by the plurality of cameras; transform each of the plurality of predicted trajectories into reference coordinates; calculate a degree of correlation with another trajectory for each of the plurality of trajectories transformed into the reference coordinates; integrate trajectories having the degree of correlation higher than a predetermined value; calculate parameters for spatially and temporally synchronizing images captured by the plurality of cameras with reference to the integrated trajectory; and transform each of the plurality of trajectories into the reference coordinate with reference to the parameter.