Event Video Sequence Generation with Variable GOP Structures

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing digital video systems face challenges in ensuring a viewable pre-event video sequence when using variable GOP lengths, as decoding may start at an inter-frame, leading to potential loss of previous frames and inadequate buffer space for intra-frames, especially in surveillance applications where event-triggered recording is crucial.

Innovation Solution

A method and system that generate an event video sequence by storing a pre-event sequence with a predetermined GOP structure and combining it with a post-event sequence having a different GOP structure, allowing for variable GOP lengths to control bit rate while ensuring a desired pre-event sequence length, with the pre-event sequence using a shorter or predetermined GOP length to guarantee viewability and the post-event sequence using a longer or dynamic GOP length for bit rate control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If variable GOP length is used to control bit rate, then bit rate control is improved, but pre-event sequence viewability deteriorates

Engineering Contradiction:
Improvebit rate controlVSAvoidpre-event sequence viewability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The video sequence is divided into pre-event and post-event segments with different GOP structures. The pre-event sequence uses a first GOP structure with shorter length to ensure viewability, while the post-event sequence uses a second GOP structure with variable length for bit rate control. This segmentation allows each segment to be optimized for its specific purpose without compromising the other.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different GOP structures are applied to different parts of the video sequence based on local requirements. The pre-event portion uses a conservative GOP structure guaranteed to be viewable, while the post-event portion uses an optimized variable GOP structure. This local differentiation resolves the contradiction by applying appropriate quality levels to each segment.

Inventive Principle:
Principle #3Local quality

2Loss of energy

If longer GOP length is used, then output bit rate is reduced, but pre-event buffer space for intra-frames is insufficient

Engineering Contradiction:
Improveoutput bit rateVSAvoidbuffer space for intra-frames
Core Design Contradiction:
Loss of energyVSVolume of stationary object

Solution Approach 1:

The pre-event sequence is prepared in advance with a predetermined GOP structure that guarantees sufficient intra-frame spacing. By establishing this structure before the event occurs, the system ensures that adequate buffer space is allocated for intra-frames without needing to over-provision the buffer for all possible scenarios. This preliminary action resolves the contradiction by preparing the optimal structure in advance.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If bit rate limit is applied to all cameras, then bandwidth sufficiency is improved, but image quality deteriorates

Engineering Contradiction:
Improvebandwidth sufficiencyVSAvoidimage quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The GOP structure is made dynamic for the post-event sequence, allowing the system to adjust the balance between bit rate and image quality based on actual conditions. While the pre-event sequence maintains a fixed structure for reliability, the post-event sequence can dynamically optimize GOP length to maintain image quality when bandwidth is available and reduce bit rate when bandwidth is constrained. This dynamic approach resolves the contradiction by adapting to actual transmission conditions.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS10779003B2Method and system for generating an event video sequence, and camera comprising such system
Publication Date: 2020.09.15 AXIS
  • US10779003B2 patent drawing
  • US10779003B2 patent drawing
  • US10779003B2 patent drawing

AI summary

A method, system, and non-transitory computer readable storage medium having instructions for generating an event video sequence. The method comprises receiving input image frames encoded into a sequence of image frames arranged in groups of pictures, each group of pictures having a GOP structure. The method further comprises storing a pre-event sequence of image frames in a first-in-first-out buffer, said pre-event sequence having a first predetermined GOP structure, upon receiving an event notification, retrieving the pre-event sequence from the buffer, and generating the event video sequence by combining the pre-event sequence and a post-event sequence of image frames, said post-event sequence having a second GOP structure, which is different from said first GOP structure. The system may be included in a camera.