Mobile Image Surveillance With Dynamic Edge Model Deployment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud-based analysis of edge device data for event detection increases response time and consumes bandwidth, while edge-based analysis is advantageous but lacks efficient processing and memory management.
Innovation Solution
A management device identifies a subset of image capture devices within a specified geographic area and time frame, deploying a detection model locally for event analysis, and ceases model use after processing is complete.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If cloud-based analysis is used for event detection, then processing and memory requirements of edge devices are reduced, but response time increases and bandwidth is consumed
Solution Approach 1:
The system dynamically adjusts the location where model processing occurs based on operational needs. Edge devices can execute models locally when rapid response is required, while also capable of uploading data to the cloud for model updates or when local processing is not needed, creating a flexible hybrid architecture that adapts to different operational scenarios
Solution Approach 2:
The system segments the processing architecture into distinct components: edge devices for local inference, cloud infrastructure for model training and updates, and a coordination layer that manages data flow between them. This segmentation allows each component to specialize in specific tasks, optimizing both local response speed and cloud resource utilization
2Device complexity
If cloud-based analysis is used for event detection, then processing and memory requirements of edge devices are reduced, but bandwidth consumption increases
Solution Approach 1:
Instead of continuously uploading all image data to the cloud, the system performs partial processing at the edge by executing detection models locally on edge devices. This partial action at the edge reduces the volume of data that needs to be transmitted over the network, thereby reducing bandwidth consumption while still achieving effective event detection
Solution Approach 2:
The system introduces an intermediary coordination layer that manages the interaction between edge devices and cloud infrastructure. This intermediary optimizes data flow by determining when local processing suffices and when cloud resources are needed, reducing unnecessary bandwidth consumption while maintaining detection effectiveness
3Loss of time
If edge-based analysis is used for event detection, then response time is improved, but processing and memory requirements of edge devices increase
Solution Approach 1:
The system performs preliminary actions by pre-training detection models in the cloud using extensive datasets and computational resources. These pre-trained models are then deployed to edge devices, allowing them to perform inference locally with minimal processing requirements. This preliminary preparation in the cloud enables edge devices to achieve fast local response times without bearing the full burden of model training complexity
Data Source
AI summary
Detection models are distributed as needed to a select subset of image capture devices, based on geographic location and time of the image capture devices. Each image capture device which receives a model processes image data according to the model, and provides a notification of any detected instances of an event the model is trained to detect. Distribution and detection can be based on historical data or live data.