
3D depth sensing technology has gained significant attention in the era of spatial computing due to its important applications in machine vision and artificial intelligence. Among the mainstream non-contact optical 3D depth measurement methods, structured light projection, time-of-flight, and stereo vision are the most common. Structured light-based 3D sensing technology, in particular, offers high spatial resolution, compact size, wide field of view, and low power consumption, making it widely used in machine vision, facial recognition, human-computer interaction, motion-sensing games, and biomimetic robotics.
Traditional structured light depth sensing system typically consists of a vertical-cavity surface-emitting laser (VCSEL) array of ~300 randomly positioned emitters, which is multiplicated by a diffractive optical element (DOE) with a fixed repeated pattern. However, such system faces significant limitations. The distance between speckles is fixed and known, with no additional scales for calibration or adjustment, limiting the system flexibility. Depth sensing accuracy is constrained, typically achieving only 1 mm accuracy at 60 cm, and requiring complex internal calibration to address errors caused by temperature drift and distance variations. Traditional speckle patterns are static and non-programmable, making them unsuitable for dynamic sensing scenarios.
To address these challenges, Professor Connie Chang-Hasnain, Professor Shaohua Yu and their team developed an addressable structured light (ASL) 3D sensing system. This innovative system integrates a metasurface-based diffractive optical element (Metasurface-DOE, MDOE) with an individually addressable VCSEL array (IA-VCSEL), achieving unprecedented control over structured light patterns. A defining feature of this ASL-3D system is the introduction of a "built-in caliper" mechanism, which enables dynamic calibration and precise control over speckle pattern positioning. This advancement has led to sub-100-micron depth sensing accuracy, surpassing the capabilities of conventional structured light systems. Relevant research results were recently published in Photonics Research, Volume 12, Issue 6, 2024. [Chenyang Wu, Xuanlun Huang, Yipeng Ji, Tingyu Cheng, Jiaxing Wang, Nan Chi, Shaohua Yu, and Connie J. Chang-Hasnain. Addressable structured light system using metasurface optics and an individually addressable VCSEL array[J]. Photonics Research, 2024, 11(6): 2327]
At the core of the ASL-3D system lies the IA-VCSEL array, which consists of an 8×8 grid of VCSELs with uniform 100 μm spacing. All VCSELs share a common cathode, while their anodes are independently addressable, allowing for real-time dynamic switching of individual or multiple VCSELs with response times in the microsecond range (Fig. 1a). The MDOE component of the system is fabricated using a reflective SOI (Silicon-On-Insulator) metasurface, featuring a thin film with nanorod structures of single-crystal silicon. By carefully designing the dimensions and rotation angles of silicon nanorods, the metasurface achieves beam splitting with a pre-designed yet random pattern, crucial for forming unique and adjustable speckle patterns.
Figure 1 (a) Prototype of the ASL-3D system based on the reflective MDOE and the IA-VCSEL array, (b) dynamic control of the speckle pattern density and distribution.
One of the key differentiators of the ASL-3D system is its capacity to encode speckle patterns dynamically. Unlike conventional systems that generate fixed patterns, the ASL-3D system offers exceptional flexibility in controlling speckle pattern density and distribution. Through dynamic on/off control of individual VCSELs, the system can generate speckle patterns in real-time, adapting to different scenarios and improving sensing accuracy. The far-field results have demonstrated that the speckle patterns produced align precisely with the IA-VCSEL array's programmed coding configurations, underscoring the accuracy and flexibility of the system (Fig. 1b).
Another groundbreaking feature of the ASL-3D system is the introduction of adjustable calibrated unique distances (CUDs), made possible by the IA-VCSEL's encoding capability. By leveraging these CUDs, the system can dynamically recalibrate speckle patterns, compensating for environmental factors such as temperature drift. This "built-in caliper" functionality allows the ASL-3D system to adjust and correct for measurement errors in real-time, reducing error accumulation and enhancing the stability of the sensing process. As a result, in depth sensing experiments, the ASL-3D system consistently achieves sub-100-micron depth measurement accuracy, far surpassing traditional structured light methods with millimeter-level precision.
The implications of this technology extend far beyond basic 3D sensing. The ASL-3D system opens new possibilities for a wide range of applications, including spatial computing, facial recognition, wearable devices, and machine vision systems. In AR/VR environments, the system can enhance depth perception, improve gesture recognition, and enable precise eye-tracking mechanisms (Fig. 1a). In industrial and robotics contexts, the system's adaptability and precision facilitate more accurate quality control, object recognition, and robotic manipulation. Additionally, by integrating the ASL-3D technology with AI-driven algorithms, the system can further enhance environmental perception, contributing to smarter automated systems capable of real-time, adaptive decision-making.
Comments:
Professor Connie Chang-Hasnain stated that, "This research integrates addressable VCSEL array and metasurface technology, achieving significant breakthroughs in 3D depth sensing. The innovative design of the 'built-in caliper' significantly enhances depth sensing accuracy and system flexibility. The experiments validate the system's superior performance in dynamic speckle control and depth measurement, offering important support for applications in AR/VR devices, smart interaction, and consumer electronics." Dr. Xuanlun Huang remarked that, "This study proposes a novel addressable structured light system by combining the individually addressable VCSEL array and metasurface technology. It significantly improves the accuracy and adaptability of 3D depth sensing. The 'built-in caliper' design provides a new approach for dynamic calibration and high-precision depth measurement, demonstrating broad potential applications in 3D sensing technology."