Traffic Event 3D Reconstruction for Blind-Spot Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to provide an objective overview of traffic events, such as collisions or near misses, as they often rely on single-direction video observations, missing blind spots and making it difficult to grasp the full situation.
Innovation Solution
A system that combines first and second three-dimensional models generated from multiple event videos to create a comprehensive three-dimensional model, allowing users to view the event from any specified viewpoint.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a pseudo fixed point monitoring image is generated by arranging frame images from multiple vehicles at the same point, then the positional relationship between feature points can be corrected, but the observation is limited to a certain direction and blind spots from fixed point areas cannot be observed
Solution Approach 1:
The patent transforms the two-dimensional fixed-point monitoring image into a three-dimensional virtual observation space. By generating a three-dimensional model that combines multiple vehicle camera viewpoints, the system enables observation from any direction and height, not limited to a single fixed point. This dimensional transition resolves the contradiction by maintaining positional accuracy while adding observation flexibility.
Solution Approach 2:
The three-dimensional model serves multiple observation functions simultaneously. It can display the traffic event from various viewpoints (front, rear, left, right, and elevated perspectives) within a single integrated model, making the system universally applicable for comprehensive situation analysis rather than requiring multiple separate fixed-point observations.
2Loss of information
If video images are observed from the viewpoint of each vehicle separately, then the situation can be captured from multiple perspectives, but it makes it difficult to grasp the overview of the traffic event
Solution Approach 1:
The patent merges multiple separate vehicle video perspectives into a single integrated three-dimensional model. This combination allows the system to preserve the rich multi-perspective information from different vehicles while presenting it as a unified overview that can be easily navigated and understood, resolving the contradiction between information completeness and operational ease.
Solution Approach 2:
The three-dimensional model acts as an intermediary that translates multiple separate vehicle camera views into a cohesive spatial representation. This intermediary structure enables users to access both detailed vehicle-specific perspectives and the overall event overview without confusion, mediating between the complexity of multiple sources and the need for simple comprehension.
3Measurement precision
If a three-dimensional model combining multiple event videos is generated, then an objective overview of the traffic event can be grasped, but the system complexity increases
Solution Approach 1:
The patent segments the complex task of creating a comprehensive three-dimensional model into manageable steps: extracting feature points from individual vehicle videos, matching these features across multiple sources, and then assembling them into the final model. This segmentation reduces the apparent complexity by breaking down the complex process into discrete, manageable operations.
Solution Approach 2:
The system creates a virtual three-dimensional copy of the traffic event scene based on multiple vehicle camera observations. This digital replica preserves the spatial relationships and details of the actual event without requiring physical presence at the scene, simplifying the system's operational complexity while maintaining high measurement precision for event reconstruction.
Data Source
AI summary
A traffic event reproduction system according to the present procedure includes: a server; a first transmission apparatus configured to transmit first event information including a first event video to the server; a second transmission apparatus configured to transmit second event information including a second event video to the server. A server includes: a reception unit configured to receive first and second event information; a model generation unit configured to generate first and second three-dimensional models including a party to a traffic event extracted from the first and second event videos as an object; a combining unit configured to combine the first and second three-dimensional models to generate a combined three-dimensional model in a three-dimensional space; and a display unit configured to display a video viewed from a user-specified viewpoint using the combined three-dimensional model.