Recently, researchers from Peking University Shenzhen Graduate School and Xidian University have innovatively proposed a multispectral fusion strategy inspired by bionic vision. This strategy leverages the team’s independently developed air-liquid-solid spraying technology (ALS) to successfully fabricate a seven-channel narrowband perovskite photodetector array covering the ultraviolet-visible spectrum. Combined with a deep learning-based color fusion algorithm (MSCF-DNN), the system achieves the capability to capture full-spectrum information and accurately reproduce RGB colors in a single shot, offering a new technological pathway for applications in remote sensing and medical imaging. The research, titled “Biovision-Inspired Perovskite Intelligent Camera for Panchromatic and Metameric Sensing,” was published in the international academic journal Advanced Materials.
Figure 1. Research Published in Adv. Mater.
This multispectral fusion strategy effectively addresses the “metamerism” phenomenon—where colors with different spectral distributions appear identical to the human eye—by enabling precise identification and differentiation, significantly improving the color confusion issues common in traditional cameras. Experimental results show that the seven-channel fusion images reduce color difference (ΔE) in the red-green region by approximately 50% compared to conventional RGB schemes, markedly enhancing color reproduction accuracy. The study further integrates material characterization and optoelectronic performance testing to systematically validate the response uniformity and long-term stability of the perovskite photodetector across different spectral channels. By leveraging artificial intelligence algorithms, the research achieves efficient analysis of complex spectral information, providing new insights for high-precision, multispectral intelligent sensing and vision applications.
Figure 2. Bioinspired Perovskite Intelligent Vision System
Author Information Professor Shihe Yang from the SAM at Peking University Shenzhen Graduate School, along with Professor Xueli Chen, Dr. Zhong Ji, and Dr. Yujin Liu from Xidian University, are the co-corresponding authors. Doctoral student Yu Li from the SAM, master’s student Zikun Jin from the AI4S program at the SAM and the School of Information Engineering, and Dr. Yujin Liu from Xidian University are the co-first authors. The work was supported by the National Natural Science Foundation of China, the Shenzhen Peacock Plan, the Shenzhen Science and Technology Program, the Interdisciplinary Exploration Fund of Xidian University, and the Guangdong Key Laboratory of Nano-Micro Materials Research.
Link to the paper: https://doi.org/10.1002/adma.202508984
