Bioinspired in-sensor orthogonal adaptation for robust all-weather multispectral vision

Reliable machine vision under diverse illumination and weather conditions remains a major challenge for autonomous systems. Biological vision systems offer inspiration, utilizing distinct mechanisms for adapting to brightness and spectral composition. In this work, we report a bioinspired organic vision sensor with electrically switchable spectral response, enabling in-sensor orthogonal adaptation for robust multispectral imaging in all environments. The organic vision sensor transitions rapidly between photovoltaic mode for high-fidelity visible detection and photomultiplication mode for sensitive near-infrared imaging. This dual-mode operation supports robust vision performance in both bright and low-light scenarios, as well as in low-visibility conditions such as fog or haze. Experimental demonstrations and system-level simulations reveal substantial improvements in object detection under varying illumination and visibility. This approach establishes a foundation for adaptive multispectral imaging, advancing practical deployment of robotics, autonomous vehicles, and other electronic vision systems that require reliable performance in all-weather conditions.