Machine-Learning Video Frame Prioritization for Mobile Entities

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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

Engineering Contradiction:
Improvevideo data completenessVSAvoidtransmission time
Core Design Contradiction:
Loss of informationVSLoss of time

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.

Inventive Principle:
Principle #1Segmentation

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

Engineering Contradiction:
Improvevideo data transmission rateVSAvoidavailable bandwidth
Core Design Contradiction:
ProductivityVSQuantity of substance

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.

Inventive Principle:
Principle #2Taking out (Extraction)

3Loss of time

If frame prioritization is implemented to reduce transmission time, then critical frames are transmitted faster, but video data processing complexity increases

Engineering Contradiction:
Improvetransmission timeVSAvoidvideo data processing complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

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.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4625358A1Video data provision
Publication Date: 2025.10.01 BRITISH TELECOM PLC
  • EP4625358A1 patent drawingFigure 1
  • EP4625358A1 patent drawingFigure 2
  • EP4625358A1 patent drawingFigure 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.