Cloud V2V Virtual View Using Feature Matching for Blind Spots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle-to-vehicle (V2V) communication systems face challenges in providing a comprehensive visual view of a region of interest, especially for vehicles with blocked views, due to limitations in sensor sharing and bandwidth requirements for high-resolution video streaming.
Innovation Solution
A cloud-based virtual view system that requests key points and feature descriptors from surrounding vehicles, matches and combines these data to generate a panoramic visual view, allowing vehicles with blocked views to receive a stitched image of the region of interest, enhancing situational awareness and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution video streaming is used to provide comprehensive visual view of region of interest, then the visual quality and situational awareness are improved, but the bandwidth requirements and system complexity increase significantly
Solution Approach 1:
The patent extracts only the essential visual information (key points and feature descriptors) from the complete video stream rather than transmitting the entire high-resolution video. This selective extraction maintains visual quality for critical regions while dramatically reducing bandwidth requirements by sending only salient features needed for reconstructing the region of interest view.
Solution Approach 2:
The patent segments the visual information into discrete key points and feature descriptors that can be independently processed and transmitted. By dividing the continuous video stream into extractable feature elements, the system enables efficient compression and selective transmission of only the necessary visual data to reconstruct the desired view.
2Area of stationary object
If sensor data from multiple vehicles is collected and processed to generate panoramic view, then the coverage and situational awareness are improved, but the device complexity and processing requirements increase
Solution Approach 1:
The patent introduces a cloud-based server as an intermediary that centralizes the complex tasks of collecting sensor data from multiple vehicles, matching key points across different vehicle perspectives, and generating the panoramic view. This mediator approach distributes the computational complexity from individual vehicles to a centralized system, simplifying on-vehicle device complexity while achieving comprehensive multi-vehicle coverage.
Solution Approach 2:
The patent merges sensor data from multiple vehicles by matching key points and feature descriptors across different vehicle perspectives. This combination of data sources enables the construction of a comprehensive panoramic view that covers areas invisible to any single vehicle, achieving extended coverage through data fusion.
3Reliability
If feature matching and panorama generation are performed in real-time, then the responsiveness and safety are improved, but the processing time and computational load increase
Solution Approach 1:
The patent extracts only the critical key points and feature descriptors needed for panorama generation rather than processing complete video frames. This selective extraction significantly reduces computational load and processing time while maintaining the essential visual information required for safe, real-time operation and accurate region of interest reconstruction.
Data Source
AI summary
Disclosed are systems and techniques for wireless communications. For example, a device can receive, from a first vehicle, a view request for a visual view of a region of interest (ROI). The device can transmit a request for key points and feature descriptors related to a view of the first vehicle and respective view(s) of the other vehicle(s), and can match the key points/feature descriptors related to the other vehicle(s) and the first vehicle. The device can determine, based on the matching, at least one vehicle to provide at least one ROI view of the ROI and determine at least one mapping between the at least one vehicle and the first vehicle, which can be used to combine the at least one ROI view of the at least one vehicle with the view of the first vehicle to generate a combined image having the visual view of the ROI.