ADAS Driving Assistance & Automated Driving, Parking Assistance System

Computer Vision

Last update: 12 Jul. 2024

What is Computer Vision?

Computer Vision, as a field, encompasses the development and implementation of systems that can understand and interpret visual data from the real world to replicate human vision capabilities. Its technology serves as the backbone for multiple industries, with Valeo leading the charge in the automotive domain. With expertise in surround view camera perception, Valeo excels in making sophisticated algorithms function optimally even on lower-end computing hardware.

The versatility of Valeo’s Computer Vision system is exemplary, as it can be tailored to meet varying perception needs depending on project specifications. Its core capabilities include camera calibration, camera soiling detection, and identifying vulnerable road users. The system’s other advanced features are adaptable, ranging from emergency vehicle classification, lane and park slot detection, to object identification and traffic sign recognition. All these capabilities aim to support tasks such as trailer coupling and maneuvering.

The Computer Vision algorithms leverage camera images and corresponding lens distortion calibrations. These algorithms analyze the visual data, detecting and classifying different entities within the scene. This is then merged with data from other sensors and tracked over time, contributing to a complete understanding of the vehicle’s environment. Thus, Computer Vision acts as an indispensable tool in ensuring safe and efficient navigation for automated parking and driving systems.

0/0

What are our Computer Vision system benefits?

Use software viewing cameras to detect vulnerable road users, objects, lanes, park slots, freespace, kerbs, etc.
Additional sensing system combines with other systems increasing overall safety
Offers increased range for 360° perception allowing for more accurate and natural trajectory planning for autonomous maneuvers
Offers robust classification compared to other sensor options e.g classification of pedestrians