The full case study
AI-powered video surveillance platform, step by step
On this page · 9 sections
About the business
Our client is from a sizeable US-based company that owns different hardware and software solutions for video surveillance. They requested a redesign of their legacy platform and integration of new AI-powered functionality for smart video surveillance.
An updated solution should help big companies monitor and maintain different machinery and equipment effectively. The system should be able to detect objects and abnormal activities, initiate specific scenarios, set up cameras, record, store, and manage videos from different equipment.
Our client wanted to update that system by:
- integrating AI module
- adding new features
- improving UI/UX and usability.
Project tasks
- Complete a discovery and product audit to prepare new product specifications, list of prioritized features, and development roadmap.
- Perform a product redesign, implement all required new features, and create an intuitive and smooth user experience.
- Integrate AI-powered functionality that will detect objects, perform automated analysis, have a base of scenarios and modes.
Project team
- Project Manager
- Business Analyst
- 2 UX/UI designers
- 3 Front-end developers
- 2 Back-end developers
- QA engineer
- Solution Architect
- DevOps specialist
2 month for Discovery & Design
5 months for Development & Testing
Platform features
AI object detection
AI module recognizes and classifies objects
Video recording
System records videos and offers rapid review
AI reporting
AI creates reports with time, pictures and videos
Anomaly detection
AI identifies unusual activity and sends alerts
AI analysis
AI analyzes videos and detects all accidents
Integration with IoT
System easily connects with sensors and alarms
Video recording
We made video recording functionality smarter and easier to use. Platform users can connect their surveillance cameras, stream and store videos, and then review them.
We added more options for more convenient setup and control of cameras. Users can switch between cameras, set multiple camera views, and manage them in a few clicks with new, updated UX.
There is also a smart search that leverages AI to enable more efficient and accurate searching through large volumes of video footage. The system processes the video, activates object recognition, generates and indexes metadata, processes user queries, and provides precise results matching search requests.
AI video analysis
The AI module can work with many custom scenarios depending on the situation. It detects anomalies, automates the understanding of video content, and acts accordingly.
Integrated AI service was capable of detecting and tracking necessary objects (i.e., people, vehicles, or specific items) and analyzing the video to:
- identify the behavior of the object
- define and classify patterns
- detect abnormal activity/behavior
- generate instant alerts
- perform customizable actions.
AI video management
The AI module empowered platform users to analyze, create and retrieve necessary insights within saved video content. AI can dynamically tag video content based on detected objects, actions, or events, making it easier to categorize relevant footage.
It is designed to scale with the growing volume of video data and adapt to changes in the surveillance environment.
Development approach
Here is a tech stack that we used for the updated video surveillance system:
- Nest.js was used to develop most services, and Python – for AI-related ones (since it’s the most powerful and robust stack).
- Amazon AWS was selected as the main cloud service provider since it is the most cost-efficient and scalable option.
- Swagger was used for all API documentation and effective testing.
- TypeScript was selected for both front-end and back-end parts to make source code more stable and cut future platform maintenance costs.
Outcome
We performed a complete redesign of a client’s surveillance system and integrated advanced AI functionality. Once the redesigned solution was set and ready, client’s team started actively testing it. They were satisfied with the end product and the way AI service worked.
As of now this AI-powered system is a bestseller among all client’s solutions. It brought him hundreds of new corporate clients who purchased a license and connected thousands of surveillance cameras.
