1 August 2018 - Big Data & Data Science

RealCarBrand — an IoT and Big Data-based project

The solution consists of the CCTV cameras installed at the key roads of the city, paired with Raspberry Pi 2 or 3, and 3 or 4G modems. The cameras take photos of each passing car, the IoT node identifies the brand and the model and this information is sent to the cloud processing server. This way, a huge dataset of actually used car brands and models is formed real-time, which can be leveraged to deliver the predictive analysis for the market, like producing certain amounts of spare parts, opening new service centers and certifying the technicians for them.

Project requirements

In order for the system to work we had to ensure the following functionality:

  • IoT side had to be able to take the photos of the cars, identify the car model using OCR (Optical Character Recognition), send the report to the cloud server and delete the photo afterward to comply with GDPR.
  • Back-end side had to be able to accept the reports, filter and aggregate them and store the data in a  NoSQL database
  • Front-end side had to be able to represent the charts and reports and provide the end user APIs to interact with any third-party analytical system

Project results

The solution works as follows:

  1. CCTV cameras are installed at the key roads of the city, paired with Raspberry Pi 2 or 3, and 3 or 4G modems
  2. The cameras take photos of each passing car, the IoT node identifies the brand and the model, and this information is sent to the cloud-based processing server
  3. A huge dataset of actually used car brands and models is formed real-time
  4. This data can be leveraged to deliver the predictive analysis for the market, like producing certain amounts of spare parts, opening new service centers and certifying the technicians for them.

The system we developed has shown good results during the alpha-testing phase. One IoT node is able to identify about 150 cars per minute (rush hour on a three-lane road), which makes the system quite feasible from the analytical perspective.

 

Location: Vienna, Austria

Partnership period: May 2018

Team size: 3 people

Team location: Kharkiv, Ukraine
Services: Cloud infrastructure design and development, Big Data solutions, Web Development, Python development, data visualization, data normalization, IoT

Expertise delivered: AWS cloud administration, DevOps services, Big Data architecture, Python development, data science, data visualization, IoT architecture

Technologies: Cassandra, Python, Django, Chart.js

 

Product Overview

Client’s goals

The customer wanted to provide their users with the clean and simple tools for real-time analysis of the car brands and models currently in circulation in any city of the globe, where this system is installed. The main objectives were the following:

  • Design and develop the structure of an IoT node to be deployed on the streets
  • Ensure the secure and reliable aggregation of the data in the cloud
  • Provide the means for data visualization
  • Give the customers access to the historical data
  • Provide a clean API to enable interaction with third-party systems

Implementation and challenges resolved

IT Svit team has once more proved our expertise and accomplished all the tasks:

  • Using the OCR algorithm enables the solution to identify the car brands and models seen on the photos
  • The photos are deleted on spot, in order not to waste the resources and comply with the GDPR
  • The data on the car brand and model is sent to the central server in AWS or GCP cloud
  • These messages are queued and processed by the aggregate workers, and are stored in a Cassandra NoSQL database in the back-end
  • The customers can interact with the frontend via a web portal, where they can view the charts and reports, or get the data directly to their third-party software via API
Contact Us



[dynamichidden your-post "CF7_ADD_POST_ID"]
[dynamichidden your-country "CF7_ADD_COUNTRY"]
[dynamichidden your-link "CF7_ADD_LINK"]

Our website uses cookies to personalise content and to analyse our traffic. Check our privacy policy and cookie policy to learn more on how we process your personal data. By pressing Accept you agree with these terms.