A breakthrough in short-wave infrared sensing could bring military-grade vision systems into mass-market devices, from self-driving cars to smartphones
New chips designed for safer autonomous driving
Cameras capable of seeing through smoke, fog, and low-light conditions may soon become a reality for autonomous vehicles. Researchers at Xidian University have developed a new generation of imaging chips based on short-wave infrared, SWIR, technology, a spectrum of invisible light that can penetrate atmospheric obstacles more effectively than visible light.
Combined with SPAD, single-photon avalanche diode sensors, these chips can detect extremely low levels of light, down to individual photons, enabling high-resolution imaging even in near-total darkness.
Until now, such systems have largely been limited to military applications, including drones and advanced surveillance platforms, due to their high cost and complex manufacturing requirements. The new approach aims to bring these capabilities into commercial markets using scalable production methods.
A cost breakthrough, but industrial scale remains the real test
Traditional SWIR sensors rely on indium gallium arsenide, InGaAs, a material that is both expensive and difficult to integrate into standard semiconductor manufacturing.
The research team instead developed a silicon-germanium-based solution, which could allow SWIR sensors to be produced using existing chip fabrication processes. Early estimates suggest a dramatic reduction in production costs, potentially enabling large-scale adoption.
However, the transition from laboratory innovation to industrial deployment will depend on factors such as manufacturing yield, long-term reliability, and integration with existing automotive and consumer electronics systems.
Autonomous vehicles: improved visibility in extreme conditions
One of the main limitations of current autonomous driving systems is reduced performance in adverse weather. Standard cameras rely on visible light, while LiDAR systems can be affected by scattering from fog, rain, or airborne particles.
SWIR-based sensors offer a complementary approach. By operating at longer wavelengths, they can reduce scattering effects and improve detection of objects at greater distances, even in low-visibility environments.
If successfully integrated, these chips could enhance both fully autonomous systems and advanced driver-assistance systems, ADAS, improving safety in challenging driving conditions.
Smartphones: beyond night mode photograph
The same technology could also redefine mobile imaging.
SPAD sensors enable ultra-low-light photography, potentially surpassing current software-based night modes by capturing real optical signals rather than relying heavily on image processing.
In addition, SWIR sensing can interact differently with certain materials, opening the door to new imaging applications in areas such as health monitoring and secure biometric authentication.
For example, facial recognition systems could become more robust by analyzing subsurface features, such as blood vessel patterns, making them harder to spoof compared to traditional visible-light systems.
