Camera Buffering for Event Video Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video monitoring systems face inefficiencies in bandwidth usage and storage due to continuous streaming of full video feeds, even when only specific events are of interest, and lack an integrated platform for utilizing recorded videos for various use cases, especially in cloud-based setups without local NVRs.
Innovation Solution
A cloud-based video monitoring system where cameras store a predetermined temporal segment of the video stream and generate a pre-event video file upon detecting a triggering event, sending it to a network video recorder (NVR) for storage, allowing for efficient event-triggered recording without continuous full recording, and enabling integrated video utilization through metadata and persistent connections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If continuous full video recording is maintained, then complete video coverage is achieved, but network bandwidth consumption increases significantly
Solution Approach 1:
The camera performs preliminary recording by storing video frames in a local buffer memory before an event is detected. This preliminary action ensures that pre-event footage is captured and ready for transmission, eliminating the need for continuous full-video streaming while maintaining complete event coverage.
Solution Approach 2:
The system extracts only the relevant video segments (pre-event, during-event, and post-event footage) from the continuous video stream and transmits them to the NVR. This extraction approach maintains complete event coverage while dramatically reducing network bandwidth consumption by excluding irrelevant video portions.
2Adaptability or versatility
If both full recording and event-triggered recording are maintained, then all user requirements are met, but storage and bandwidth efficiency decrease
Solution Approach 1:
The system dynamically adjusts the recording mode based on detected events. During normal operation, only essential metadata is transmitted. When an event is detected, the system dynamically switches to transmitting pre-event buffered footage, during-event footage, and post-event footage. This dynamic approach provides recording flexibility while optimizing storage and bandwidth usage.
Solution Approach 2:
The single recording system serves multiple functions: it performs continuous local buffering, event detection, pre-event recording, during-event recording, and post-event recording. This multi-functional approach eliminates the need for separate full-recording and event-triggered-recording systems, reducing storage and bandwidth requirements while maintaining adaptability to different user needs.
3Ease of manufacture
If cloud-based NVR is used without local NVR, then deployment cost is reduced, but recording reliability may be compromised
Solution Approach 1:
The camera performs preliminary video buffering in local memory before cloud transmission. This preliminary local storage acts as a safety buffer, ensuring that video footage is preserved even if cloud connectivity is interrupted. This approach reduces deployment cost by eliminating the local NVR hardware while maintaining recording reliability through the camera's built-in buffer.
4Loss of information
If pre-event video buffering is implemented, then event context is improved, but camera memory requirements increase
Solution Approach 1:
The system implements partial buffering by storing only a limited temporal window of pre-event video frames in the camera's buffer, rather than buffering extensive periods. This partial action approach provides sufficient event context for most security applications while keeping the buffer memory size manageable and cost-effective.
Data Source
AI summary
One embodiment of the present invention provides a cloud-based video monitoring system comprising a camera and a network video recorder (NVR). During operation, the camera stores in a buffer last predetermined temporal segment of a video. The camera also stores a first video file in a local storage corresponding to an event detected by the camera. This first video file is non-overlapping with a second video file corresponding to the event and this second video file is stored in the NVR. The camera generates a third video file by pre-pending last predetermined temporal segment of a video prior to the event to the first video file and sends the third video file to the NVR.