All Collections
Webinar summarries
WEBINAR #2 - The Edge Video Analytics - FLOW solutions for edge computing
WEBINAR #2 - The Edge Video Analytics - FLOW solutions for edge computing

Learn about Edge computing, why is it great and how smart cities can be realized with FLOW solutions.

DataFromSky avatar
Written by DataFromSky
Updated over a week ago

Did you know that video cameras exist for more than 100 years now? Do you know there is 1 camera per 29 people in the world and this ratio is expected to grow by 100% in 5 years? Do you know why is Edge computing important or that cameras can serve as the main driver for the implementation of smart cities and smart traffic? In this webinar, we will look into the issues of building smart cities. We will point out how Edge computing used in our FLOW family products (Camera, Embedded) can solve these problems and discuss the strengths and weaknesses of Edge computing. We will also present to you our real-time video analysis products that run FLOW.

Before diving in. If you are new to FLOW - it might be best you watch the Webinar or read the Webinar summary - Introduction to FLOW first.

In this webinar, we will share with you:

1. Why is Edge computing so important and what are the benefits of moving the processing power to the Edge

2. Find out what types of solutions from FLOW family can be used for Edge computing and help to build smart cities

3. See various LIVE use-cases demonstration (congestion detection, parking detection, pedestrian counting for retail, bike detection in pedestrian zones, and more) with different FLOW products

You can find the webinar timeline in the video description on VIMEO.

Questions and Answers

42:59 Q1: What bandwidth does the TrafficEmbedded unit consume?

A: Depends on the results needed. It can be reduced to IoT device using LoRaWAN network. IoT devices consume only a couple of bytes/hour.

43:32 Q2: What camera types are supported?

A: Any type of IP camera with RTSP stream. Recommended parameters like minimal resolution and FPS depend on the particular use case. Contact us! and we will gladly help you figure out the right camera and camera positioning that would best suit your needs.

44:14 Q3: Where is the data stored?

A: The main purpose of Camera/ Embedded units is to stream insights.

They have their own cache where they store trajectories that are stored there for a limited time (storage for up to 10 000 trajectories) - this data are raw trajectories which can be used for full interactive traffic analysis. The units also save the whole historical data in the basic form of counts of the pre-defined parameters.

45:28 Q4: What type of processor unit is used in Embedded?

A: NVIDIA processors, Embedded nano uses Jetson nano, the Embedded macro uses Jetson Xavier

45:55 Q5: Is recognition of license plates possible?

A: Yes it is with good camera resolution and positioning. It can be activated at any time within the settings of the given video or while uploading new image data. When the function is activated, a special image recognizer for number plate reading for vehicles with a suitable angle is run alongside the standard deep traffic analysis. If there is a reliable number plate detection, it is stored as one of the attributes of the vehicle’s trajectory (similarly as the vehicle’s category or color).


The need for Edge Computing solutions is growing despite some issues in picture quality, power supply, and processing capacity. Its strengths of easy scalability and low latency work well for the smart city and traffic solutions using FLOW video analysis. The existing cameras can become smart with the FLOW AI video analysis software. The cameras can then be used for many purposes such as traffic, security, retail, or parking which opens possibilities for smart city solutions. With FLOW your camera can recognize different traffic participant categories and measure values like speed, acceleration, and timings with high precision.

Apart from Edge Computing, this webinar has also covered 3 types of FLOW products: Camera, Embedded (both utilize Edge Computing), and Enterprise and in which cases are they best used. The webinar also showed you how easy and intuitive it is to set up traffic measuring with the visual programming language dashboard - no need for a PhD in Computer vision as anyone can do it in a minute! The webinar also showed you how to easily set up traffic measuring such as vehicle counts or stationary vehicle detection and congestion detection. You might be interested in reading the article “basics of FLOW Insights” where you can learn all of the basis FLOW analysis tools. Contact us to discuss how we can help you with your project or problem by implementing FLOW. Download the FLOW demo kit here for free!

• Do you have more questions? Contact us!

• Follow us on Linkedin.

• Learn more about FLOW.

• Download the demo version of FLOW.

• Visit our homepage to view products and news.

Thank you for your interest! Have a great day!

Let the traffic flow - with FLOW!

Did this answer your question?