Reimagining Sight: Perception Beyond Frames with Event-Based Vision

An Event-based Vision Sensor, often referred to as an event camera or neuromorphic camera, is a specialized type of image sensor designed to mimic the behavior of the human retina and provide a fundamentally different way of capturing visual information compared to traditional cameras. Diverging from the fixed-rate frame capture mechanism of traditional cameras, EVS operates based on the concept of "event-driven" data capture. Instead of capturing entire frames of images, EVS only responds to significant changes in the scene, such as alterations in luminosity or motion patterns, effectively reducing power consumption, extending runtime, and data processing load. The remarkable frame rate of EVS empowers it to adeptly capture swift motion and fast-changing scenes with superior efficacy when compared to conventional cameras. Demonstrating exceptional capability, EVS excels in capturing events occurring within mere microseconds.

Event-Based Vision Sensor: Where Vision Meets Versatility

Detection & tracking

Event-based vision sensors excel in human body detection and object tracking by capturing dynamic scene changes and meticulously analyzing the spatial and temporal patterns of these alterations to discern the presence and movements of individuals. Their attributes, including energy efficiency, low latency, and adaptivity to different lighting conditions, render them exceptionally well-suited for a range of applications, including security, robotics, and human-computer interaction.

Anomaly detection in industrial automation

Event-based vision sensors play a pivotal role in industrial automation by swiftly detecting dynamic scene changes, enabling proactive responses to prevent or mitigate issues. These sensors enhance early and efficient anomaly detection while conserving energy and data resources, thereby contributing to heightened productivity, quality control, and safety across manufacturing and industrial processes. Furthermore, their integration with AI/ML algorithms elevates anomaly detection accuracy, allowing seamless adaption to diverse industrial environments over time.

Mobile

The fusion of “images” and “events” introduces the swiftness, efficiency, and excellence of neuromorphic-enabled vision into mobile devices. This advancement significantly augments the existing camera functionalities, particularly in the context of capturing high-speed dynamic scenarios (e.g., sports events) and in challenging low-light environments, courtesy of its pioneering approach in event-based continuous and asynchronous pixel sensing approach.

Automotive

VoxelFlow™ is co-developed by Terranet AB and Mercedes-Benz, equipped with Prophesee Metavision® event-based vision sensors. It enhances the existing radar, LiDAR, and visual systems in accident-prone areas within a range of 30 to 40 meters, enabling fast and accurate understanding and prediction of road conditions for autonomous driving (AD) and advanced driver-assistance systems (ADAS)

AR/VR

Event-based vision sensors exhibit an exceptionally high sampling rate (theoretical values of up to 10 kHz), an expansive dynamic range, low power consumption, and minimal data volume. These attributes render them highly valuable within the realm of virtual reality (VR), encompassing applications such as 3D SLAM, gesture control, and eye tracking, among others.