A growing need for reliable machine vision in robotics, automotive, and industrial sectors has driven companies to seek improvements in image clarity, especially under challenging lighting. Addressing these requirements, Ubicept has launched the Ubicept Toolkit, which enables both offline and real-time processing of video from CMOS and single-photon avalanche diode (SPAD) sensors. The Toolkit aims to support perception tasks across various industries, offering users new ways to overcome the limitations of traditional imaging technologies. As advanced perception increasingly underpins safety, automation, and data-driven systems, the availability of adaptable and physics-based image processing attracts widespread attention from product developers.
Ubicept’s announcement follows a period where CMOS sensors dominated industrial and commercial imaging, valued for their affordability and ubiquity. However, traditional CMOS technologies have long faced criticism for high noise in low-light and motion-blur in single video frames. Previous solutions usually relied on image post-processing or AI enhancement, which did not always deliver consistently reliable data for sensor-driven applications such as autonomous vehicles or high-speed robotics. The introduction of SPAD sensor technology—and particularly this sort of real-time, hardware-integrated processing approach—marks a significant shift in how imaging challenges are addressed.
What drives Ubicept’s approach to machine vision?
Ubicept’s focus centers on the integration of SPAD sensors with their new image-processing algorithms, allowing for higher contrast and clarity in imaging, including scenarios with low light or high dynamic range. The company leverages the ability of SPAD technology to detect individual photons at sub-nanosecond resolution, producing visual information previously unattainable with CMOS solutions alone. The combination intends to directly improve conventional CMOS output and enables developers to extend existing systems without a complete hardware overhaul.
How does the Ubicept Toolkit support development?
Designed for compatibility, the Toolkit supports any widely available CMOS camera that outputs raw uncompressed data, and also connects seamlessly with SPAD cameras. Through the provided graphical user interface and API, engineers can perform both offline and live processing of camera data, evaluating side-by-side improvements between CMOS and SPAD workflows. Integration with platforms using Qualcomm Snapdragon and NXP chips is planned. Highlighting this flexibility, Ubicept’s CEO Sebastian Bauer stated,
“With the Ubicept Toolkit, we’re now making our advanced single-photon imaging more accessible for a broad range of applications from robotics to automotive to industrial sensing.”
This flexibility opens new avenues for perception model development across robotics, automotive, and other fields.
What sets partnerships and hardware apart?
Ubicept has entered a partnership with Pi Imaging Technology, whose SPAD Alpha camera provides exceptional temporal precision and low-noise operation at up to 73,000 frames per second. By applying Ubicept’s processing to the SPAD Alpha’s raw photon signal, both companies aim to deliver clear, actionable images for analytic and navigation tasks. Michel Antolović, CEO of Pi Imaging Technology, reinforced this synergy, stating,
“By combining our SPAD Alpha’s state-of-the-art hardware with Ubicept’s real-time processing, perception engineers can get the most from what single-photon imaging has to offer.”
Supporting this collaborative vision, Ubicept’s toolkit incorporates algorithms such as Ubicept Photon Fusion (UPF) and firmware like FLARE for efficient photon management, and offers tailored support for early adopters.
Ubicept’s philosophy is not to produce sensors themselves but to develop intellectual property around sensor processing, integrating the image signal pipeline from the sensor level up to object detection and localization. This full-stack approach seeks to bridge the gap between sensor output and actionable perception, enabling developers to build robust computer vision applications without being confined by sensor limitations.
Growing demand for robust imaging systems in automation and robotics puts significant pressure on current camera technologies to deliver precise data in less-than-ideal conditions. Ubicept’s introduction of a real-time, physics-based processing toolkit for both standard CMOS and advanced SPAD cameras addresses this gap by offering developers new options for system accuracy and adaptability. For readers considering machine vision upgrades, understanding the practical advantages of SPAD sensor integration over traditional technologies can be critical for future-proofing hardware investments. With key industry partnerships and a commitment to compatibility, Ubicept’s entry signals a broader shift toward more reliable imaging solutions for complex perception environments.
