KYOTO, Japan--(BUSINESS WIRE)--

Kyocera Corporation (President: Hideo Tanimoto; hereinafter: Kyocera) today announced a high-resolution AI-based depth sensor for measuring tiny objects that have been difficult to measure using conventional depth-sensing technologies. Kyocera’s new camera delivers record-setting depth measurements with 100μm resolution at a 10cm range, even from reflective or semi-transparent objects.¹ Kyocera’s innovation will support manufacturing, medicine, logistics and a variety of other fields requiring automated identification and precise depth measurement, unlocking the potential of AI and robotics with vision capabilities far exceeding the human eye.

This press release features multimedia. View the full release here:

Kyocera’s AI-Based High-Res Depth Sensor (Graphic: Business Wire)

¹Record-setting performance: AI-based depth sensor capable of measuring extremely small objects with measurement resolution of 100μm at a distance of 10cm from the object, including reflective and semi-transparent objects; source: Kyocera Corporation research, as of November 2024.

Features: Kyocera’s AI-Based High-Res Depth Sensor for Close Imaging

1. Dual-lens sensor measures tiniest objects

A unique configuration incorporating two lenses on a single sensor, Kyocera’s AI-based depth sensor provides the industry’s highest-resolution depth measurements to date among stereo cameras. The exceptionally narrow baseline of Kyocera’s depth sensor allows it to calculate the positional disparities of an object through the left and right lenses at a shorter distance than conventional methods. This precision enables the sensor to make accurate measurements of even the smallest objects.

2. AI stereo vision algorithms measure reflective or semi-transparent objects

Kyocera’s AI stereo vision algorithms enable precise depth measurements of reflective or semi-transparent objects. Traditional stereo vision algorithms involve matching objects between left and right images. However, reflective or semi-transparent objects often lack the necessary contrast, or are challenging to identify as the same object, causing measurement errors. Kyocera’s AI-based methods leverage extensive training data to measure accurately, even with challenging reflective or semi-transparent objects.

Along with their advantages, AI-based approaches often include high annotation costs and extended training times due to the extensive data required for high accuracy. Kyocera therefore developed two key technologies to reduce training costs:

1. Label-free pre-training
2. Data generation using computer graphics for 10x more precise measurements

Kyocera’s distinctive AI solution utilizes pre-training technology without labels, providing equivalent recognition with only 10% of the training data. To address the issue of conventional AI requiring large amounts of training data, Kyocera developed a computer graphic data-generation technology. This technology enables the automatic generation of training data in a CG simulation environment that accurately reproduces the target objects and settings. Additionally, Kyocera improved and accelerated the CG rendering calculation method. This CG simulation for AI training makes it possible to adapt to new objects and environments, allowing for highly accurate 3D distance measurement, even with reflective or semi-transparent objects.

Kyocera’s label-free pre-training technology has been recognized internationally for its novelty and effectiveness, including acceptance at BMVC 2024 (The 35th British Machine Vision Conference), one of the world’s leading conferences in the field of machine vision.

Future Developments

Kyocera's high-resolution AI-based depth sensor has many applications, including integration with robotic arms in manufacturing for precise depth measurement of objects as small as 1 mm. With this level of precision, robotic arms could select individual parts from a batch in a warehouse and recognize high-resolution body measurements or reflective surgical instruments in the medical field. Transport robots in logistics and retail could also more accurately monitor surroundings with this new innovation, helping to address labor shortages and boost societal productivity.

CES 2025 Outlook

This technology will be exhibited at one of the world's largest technology trade shows, CES 2025, held in Las Vegas, Nevada, USA, from Tuesday, January 7 to Friday, January 10, 2025. The Company’s latest innovations in Aerial Display technology, Underwater Wireless Optical Communications, Millimeter-Wave Sensing, Camera-LIDAR Fusion Sensors, Bifocal Mirrors, and more solutions to further safe, autonomous driving will be showcased at the booth.

For more details, please click here

View source version on businesswire.com: