Cloud-enabled Video Analytics for Smart Surveillance. : Trenser

Overview

As a technology partner to a leading video analytics solution provider in the Retail and Logistics domains, our objective was to modernize the software across their product portfolio, including cameras and video gateways, by enabling seamless cloud connectivity. The project involved integrating advanced AI-driven analytics, connecting devices to cloud services, and incorporating both edge and cloud-based analytics for real-time monitoring, incident detection, and actionable insights.

Software Modernization: Upgrade legacy video gateway software to seamlessly integrate with cloud services, enabling scalable and centralized device management.
Enhanced Real-Time Monitoring: Implement advanced video analytics to improve asset security, optimize store traffic, and streamline warehouse operations.
AI-Driven Intelligent Insights: Leverage cutting-edge AI models for object detection, pattern recognition, and incident analysis at both the edge and cloud, facilitating faster and more informed decision-making.
Robust Data Security: Implement end-to-end encryption and other data protection measures to ensure compliance with privacy regulations and safeguard sensitive information across edge and cloud environments.

Designed and developed an AWS cloud-based solution combining both Edge and Cloud AI, with secure data transmission using a multi-layer approach:

Edge Layer (On-Premises): Redesigned the NVR/Gateway software leveraging AWS IoT Greengrass to enable local execution of AI models for real-time processing. Key features include object detection, facial recognition, and activity recognition, dynamically optimized to maximize in-camera AI capabilities. The solution significantly reduced bandwidth consumption by transmitting only AI-generated insights to the cloud. This design ensures compliance with industry security standards while delivering robust and efficient performance.
Ingestion and Processing Layer: Integrated AWS Kinesis Video Streams to process on-demand video streams from edge devices and AWS IoT Core for secure, reliable connectivity. Enhanced the processing pipeline by leveraging AWS Rekognition for AI-driven image and video analysis. Utilized Lambda functions for centralized processing, enabling deeper insights through the correlation of Edge AI and server-side AI outputs.
Storage Layer: Used Amazon S3 for scalable storage of video archives and Amazon DynamoDB for fast, scalable storage of metadata and configuration details related to video streams.
Delivery Layer: Integrated Amazon CloudFront to provide secure, low-latency access to live and archived streams. Developed a React.js web interface to facilitate interactive viewing by fetching streams directly from CloudFront.
Security Layer: Leveraged AWS Cognito to manage user authentication, ensuring secure access for both Edge-to-Cloud connections and live video streaming.

Time & Material Contract: Multi-year offshore Time & Material contract with flexibility in team composition changes, resource re-allocation, and release timelines.
Agile Development: An offshore team of 8 developers and 2 testers adopted Agile/Scrum methodologies, delivering incremental features in 2-week sprints.
Requirements Management: A dedicated engineer collaborated closely with the customer to align technical deliverables with business objectives.

Project Outcome

Market Differentiation: The cloud-connected solution provided the customer with a distinct competitive edge, enabling them to offer advanced analytics services to their end users, driving customer satisfaction, and unlocking new revenue streams.
Enhanced Security: Real-time AI-powered video analytics allowed end users to proactively mitigate security threats, safeguard assets, minimize theft, and create safer environments.
Operational Efficiency: Retailers and logistics providers leveraged actionable insights to refine store layouts, optimize warehouse traffic flow, and enhance overall supply chain efficiency.
Timely Product Releases: A highly skilled, co-located engineering team with expertise in diverse technologies ensured efficient development cycles and on-time delivery, facilitating consistent market releases.