UbiCept | STEX

Elevator Pitch

Ubicept revolutionizes how machines perceive the world by combining next-generation image sensors with advanced computer vision algorithms.

Description

Cameras often blur in challenging lighting environments when under motion, translating to poor visual data for computer vision algorithms. Scenarios like solar glare, low light, and wide dynamic range, often result in poor video quality for optical computer vision. Despite the advances in LiDAR to offset these shortcomings, visual imaging remains desirable for scenarios such as characterization of obstructions, optical character recognition, and quality control. Our technology overcomes these shortcomings.

Ubicept leverages single photon perception. Single photon perception refers to utilizing individual, low-level photon data to achieve consistent and crisp imaging streams, particularly useful for computer vision applications.

Technology Description

For both human and machine vision, Ubicept enables perception in traditionally challenging imaging scenarios, including objects under fast motion with challenging lighting. Important specs include the following:

- Fast Motion and Tracking: The zero read noise of the sensor is well-suited for burst-mode exposure for optical flow and image formation, practically eliminating motion blur while simultaneously reducing noise.

- High-Dynamic Range: 150 dB+ as sensor technology improves

- Operation in All Lighting Conditions: 0.01 lux to direct sunlight

- Low Latency: The effective framerate of the camera is 100k fps, and Ubicept’s embedded processing approach can provide sparse feature tracking at 10 kHz.

- Color: Color imaging is possible and will be demonstrated in FY23.

- No cooling requirements at standard temperatures.

- Size and Weight: Sensors are of a comparable size as conventional CMOS sensors, thus can be packaged with similar size and weight requirements.

- Reduced Calibration: The same imaging optics can be used for passive fast motion optical imaging, passive slow-motion imaging, and (optional) active LiDAR.

- Cost, Scalability, and Manufacturability: Sensors can be produced similarly to traditional silicon CMOS fabrication. Thus, with scale, cost will be competitive with CMOS cameras.