Edge AI Video Analytics for Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computing systems for multi-tenant services face challenges in efficiently processing complex image and video data from security cameras, often requiring remote cloud-based processing or specialized cameras with high costs due to the need for advanced image processing capabilities.
Innovation Solution
An edge computing system is deployed at physical locations to perform artificial intelligence image/video processing on received streams, generating metrics that are then analyzed by additional AI modules in a multi-tenant service computing system, reducing bandwidth and computational requirements while allowing for advanced analytics without the need for specialized camera hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image/video streams are provided to remote server environment for complex image processing, then advanced image processing capabilities are achieved, but bandwidth consumption and latency increase
Solution Approach 1:
The system segments the image processing workflow into two parts: preliminary processing (object detection, basic analysis) is performed at the edge device locally, while only extracted metrics and refined analysis are transmitted to the remote server. This segmentation reduces bandwidth consumption while maintaining advanced processing capabilities.
Solution Approach 2:
The edge computing system performs preliminary image processing and extraction of key metrics before transmitting data to the remote server. By conducting preliminary analysis locally, the system reduces the amount of data that needs to be transmitted, thereby reducing bandwidth consumption and latency.
2Measurement precision
If specialized cameras with chip sets are deployed to perform processing, then image processing capability is improved, but device cost increases dramatically
Solution Approach 1:
The system extracts the image processing functionality from specialized camera hardware and relocates it to general-purpose edge computing devices. This extraction allows the use of standard, cost-effective cameras while maintaining advanced processing capabilities through software-based AI modules deployed on edge devices.
Solution Approach 2:
The edge computing system uses general-purpose devices that can perform multiple functions: local processing of image/video streams, metric generation, and communication with remote servers. This multi-functionality eliminates the need for specialized single-purpose cameras, reducing overall system cost.
3Device complexity
If all image/video processing is performed at remote server, then centralized control is maintained, but processing speed and real-time analytics capability decrease
Solution Approach 1:
The system segments processing tasks between edge devices and remote servers. Edge devices handle time-sensitive processing locally to maintain speed, while remote servers perform centralized control and aggregated analysis. This segmentation enables both real-time processing and centralized management.
4Ease of manufacture
If simple security cameras are used that only provide video streams, then device cost is reduced, but advanced analytics capability is lost
Solution Approach 1:
The edge computing system acts as an intermediary between simple security cameras and the remote server. It receives basic video streams from inexpensive cameras, performs advanced analytics locally using AI modules, and transmits processed metrics to the server. This intermediary approach enables advanced analytics without requiring expensive specialized cameras.
Data Source
AI summary
An edge computing system is deployed at a physical location and receives an input from one or more image/video sensing mechanisms. The edge computing system executes artificial intelligence image/video processing modules on the received image/video streams and generates metrics by performing spatial analysis on the images/video stream. The metrics are provided to a multi-tenant service computing system where additional artificial intelligence (AI) modules are executed on the metrics to execute perception analytics. Client applications can then be run on the output of the AI modules in the multi-tenant service computing system.