If you are an autonomous vehicle manufacturer dealing with sensor failure in fog or bright sunlight — this project developed a wide-spectrum image sensor and optical neural network that provides 30x lower power consumption and better pedestrian detection in adverse weather.
Ultra-Low Power All-Weather Computer Vision Sensors and Optical AI Processing
Imagine a camera that can see through thick fog and darkness as clearly as a human eye, but uses almost no battery. Instead of sending images to a power-hungry computer chip, it uses light-based processing to 'think' instantly. It's like replacing a slow, heavy engine with a lightweight, lightning-fast electric motor for how machines see the world.
What needed solving
Current computer vision systems are too power-hungry and fail in adverse weather like fog or extreme light, limiting their use in autonomous transport and wearables.
What was built
A wide-spectrum image sensor using non-toxic quantum dots and an optical neural network (ONN) chip for low-power processing.
Who needs this
Who can put this to work
If you are a headset developer dealing with bulky hardware and short battery life — this project developed a vision system with a form factor 1000x lower than current solutions, allowing for sleeker, lighter wearable devices.
If you are a robot producer dealing with high operational costs and energy drain — this project developed a system that is 1000x lower in cost and 30x lower in power consumption for real-time environment mapping.
Quick answers
How does this affect the cost of computer vision systems?
The project objectives state that the resulting vision systems will be 1000x lower in cost compared to current solutions.
Can this be produced at an industrial scale?
Yes, the project developed waferscale back-end-of-line (BEOL) integration processes that are CMOS-compatible, specifically to enable compatibility with high-volume markets.
What is the IP or licensing status of the technology?
Based on available project data, the project has resulted in new process flows, materials, and device architectures, but specific licensing terms are not provided.
How does it integrate with existing electronics?
The technology uses CMOS-compatible fabrication and includes an FPGA-based control system for real-time image processing.
When will the full system be ready for deployment?
The project runs until 2026-09-30, with full system-level demonstration expected in the next reporting period.
Who built it
The consortium is heavily industry-weighted with a 62% industry ratio, comprising 5 industrial partners (including 2 SMEs) across 4 countries. This strong commercial presence, combined with 1 university and 2 research centers, suggests the project is driven by market application rather than pure academic curiosity.
Contact FUNDACIO INSTITUT DE CIENCIES FOTONIQUES in Spain
Talk to the team behind this work.
Contact us to explore licensing opportunities for 2D-material integrated sensors.