Owl‐Inspired Neuromorphic Vision Sensor with Highly‐Sensitive Ultraweak Deep‐Ultraviolet Light Detection in Dim Environments

作者：Zhaoying Xi, Yu Liu, Sihan Yan, Jia‐Han Zhang, Min Li, Shulin Sha, Cheng Wu, Z. Wu, Shan Li, Xueqiang Ji, Shaohui Zhang, Daoyou Guo, Yong Li, Yingtang Zhou, Mingming Jiang, Zhilai Fang, Weihua Tang, Zeng Liu · 发表于：Advanced Functional Materials · 年份：2026 · DOI：10.1002/adfm.202528620 · 被引用次数：5 · 研究领域：Advanced Memory and Neural Computing、Transition Metal Oxide Nanomaterials、Ga2O3 and related materials

ABSTRACT Neuromorphic vision sensors, which address the shortcomings of traditional artificial vision sensors such as large size and low efficiency, have been widely studied recently. However, the images suffer from low contrast and loss of details when they are applied in dim environments. Inspired by the owl's dual strategy for night vision, a neuromorphic vision sensor for weak‐light applications based on Sn‐doped Ga 2 O 3 polycrystalline thin film was designed and reported here. This sensor demonstrates exceptional UVC detection by using its engineered grain boundaries to achieve both an ultra‐low dark current (88 fA) and high photo‐to‐dark current ratio (7.73 × 10 5 ) under simulated moonlight illumination. The device shows neuromorphic synaptic behaviors, underexposed image retention exceeding 500 s, and long‐term (18 months) stability. Furthermore, the sensor's unique non‐linear photoresponse inherently performs hardware‐level gamma correction, which is leveraged by a new preprocessing system to significantly enhance image contrast, and improving recognition accuracy of underexposed images from 65.20% to 83.44% when integrated with a convolutional neural network. Mimicking the specialized visual adaptations of nocturnal predators, this holistic biomimetic design provides a stabilized device, with low economic cost, and mass production potential, paving the pathway for applications in challenging weak‐light settings.