Machine-Learning Video Frame Prioritization for Mobile Entities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmanned vehicles often encounter environments with poor connectivity, limiting the bandwidth available for transmission of telemetry and video data, which can hinder effective operation and mission completion.
Innovation Solution
A computer-implemented method prioritizes the transmission of significant video frames using a machine learning model to identify operational and mission-specific events, ensuring critical frames are transmitted first, and employs progressive transmission techniques to ensure timely delivery of video data to operators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all video frames are transmitted with equal priority, then complete video data is provided to the operator, but transmission time and bandwidth consumption increase significantly in poor connectivity environments
Solution Approach 1:
The video stream is segmented into individual frames, and the frame selection system divides these frames into different priority categories (high, medium, low) based on their informational value. This segmentation allows the system to transmit only the most important frames first, reducing transmission time while preserving critical video information for the operator.
2Productivity
If video transmission bandwidth is increased to maintain constant quality, then video data transmission improves, but available bandwidth is insufficient in poor connectivity environments
Solution Approach 1:
The frame selection system extracts and identifies only the most significant frames from the complete video stream based on motion detection, event recognition, and informational value. By transmitting only these extracted significant frames rather than the entire video stream, the system achieves effective video data transmission at reduced bandwidth requirements, making it suitable for poor connectivity environments.
3Loss of time
If frame prioritization is implemented to reduce transmission time, then critical frames are transmitted faster, but video data processing complexity increases
Solution Approach 1:
The frame selection system performs preliminary analysis and prioritization of video frames before transmission begins. By pre-processing the video stream to identify and categorize frames by importance (using motion detection, event recognition, and informational value assessment), the system prepares the transmission queue in advance. This preliminary action reduces real-time processing complexity during transmission while still achieving fast delivery of critical frames.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented method of providing video data from one or more mobile entities to an operator of the one or more mobile entities is provided. The method obtains video data comprising one or more video streams, each video stream being captured by a respective video camera attached to a respective mobile entity and comprising a plurality of sequentially ordered frames. The method processes each of the video streams using a machine learning model trained to identify one or more significant frames. The method initiates provision of at least some of the video data to the operator, the provision of the significant frames being prioritised over the provision of other frames of the video data.