See more projects
nda

The customer chose to protect their confidential information

Computer Vision Integration to an AD System

Automotive
ADAS/AD
Customer rating
4.9
This ranking reflects our expertise and success
Computer Vision Integration to an AD System

About the client

As a team of Automotive software developers, we were approached by a Germany-based original equipment manufacturer.

Location:Germany
[object Object] on the map

Business Context

This IT project focused on integrating intelligent ride support, based on computer vision technology. Its successful completion allowed the customer to accelerate the introduction of the L2 AD system in their vehicles and laid the foundation for L3 system development.

Challenges

Andersen's customer, an automotive company, faced challenges in independently designing and developing intelligent car-driver support mechanisms aimed at enhancing safety by continuously monitoring ride parameters and aiding decision-making processes to reduce the risk of accidents.

While deploying their AI computer vision system for ride support capable of learning from amber and red flags and reacting accordingly – potentially issuing warnings or taking autonomous actions – the company encountered a significant hurdle.

The primary challenge involved the deployment of an embedded AI computer vision solution, as decisions needed to be made in milliseconds, along with the integration of this system into the car's network infrastructure.

Project overview

The automotive computer vision technology project aimed to achieve three objectives: continuously monitoring ride and driver parameters, alerting the driver, and recommending appropriate actions.

Duration15 months
Technologies
C++
Вlib
OpenCV
Solution architecture
Solution architecture

Solution

The resulting computer vision system acts as an independent "observer" making decisions through the analysis of data received from cameras. Here's how the process works:

  1. Front and interior cameras stream signals to the image capture component, which converts the video signal into .jpeg files at a predefined frequency;
  2. A slicer component extracts the relevant portion of the image (image cropping) required for the computer vision algorithm;
  3. The computer vision technology compares the image with training data and categorizes it.

Based on the categorized data, the computer vision automotive system may exhibit various behaviors:

  • The system triggers a sound signal and suggests taking a break (displaying the nearest possible stop on the map on the screen) when the driver's eyes are closed for an extended period;
  • Additionally, the resulting computer vision system triggers a sound signal, prompting the driver to apply the brake pedal when the car in front is too close (based on distance measured by a proximity sensor) and has its brake lights on (detected by computer vision);
  • The solution activates a sound signal and sends a message to the steering system to take corrective action when the driver is crossing road markings without using their turn signal.

Project results

The devised and executed computer vision solution effectively met the customer's needs through the following measures:

  • Continuous monitoring of ride and driver parameters, including driver eye open/closed status, proximity to the car in front, and crossing road markings;
  • Alerting the driver to detected warning signals;
  • Recommending appropriate actions based on the identified warning types.

Let's talk about your IT needs

What happens next?

An expert contacts you after having analyzed your requirements;

If needed, we sign an NDA to ensure the highest privacy level;

We submit a comprehensive project proposal with estimates, timelines, CVs, etc.

Customers who trust us

SamsungVerivoxTUI

Let's talk about your IT needs