Abstract Hyperspectral remote sensing images provide rich spatial and spectral information about the Earth’s surface, making them an essential tool for Earth observation. However, existing spaceborne hyperspectral payloads experience slow acquisition speeds and generate large data volumes, posing significant challenges for real-time applications. Moreover, the complex optical design and relatively high cost of traditional hyperspectral payloads hinder their broad-scale in-orbit deployment. In this work, we have proposed and completed the world’s first computational imaging-enabled compact spaceborne snapshot compressive hyperspectral payload, named BUPT-spectra01 , which was successfully launched on November 11, 2024, at the Jiuquan Satellite Launch Center in China. We design a reflective coding structure, which enables BUPT-spectra01 to achieve high compactness (182 mm × 214 mm × 94 mm, 1.535 kg) and low cost. The payload operates in a sun-synchronous orbit at an altitude of 520 km, with a ground imaging swath width of 51 km by 64 km. Through a single exposure (1 ms), the payload enables 47-band hyperspectral imaging with a spectral resolution of 6.5 nm, achieving 47-times data compression simultaneously. To achieve high-accuracy hyperspectral information reconstruction, we design a novel spatial-spectral inference neural network (SSI-Net). Moreover, BUPT-spectra01 can image at a rate of 30 frames per second, which allows video-level hyperspectral observation. In-orbit experiments demonstrate that BUPT-spectra01 achieves accurate classification of ground cover based on hyperspectral features, showing promise in hyperspectral observation applications such as disaster management, environment monitoring, and resource exploration. This breakthrough significantly advances the application of computational imaging in aerospace observation, contributing to the progress of future satellite internet.
Yu et al. (Mon,) studied this question.