Kilometer-range compressive hyperspectral single-pixel imaging

Hyperspectral imaging is widely used in remote sensing, agricultural monitoring and other fields. However, long-range hyperspectral imaging typically employs a scanning structure, which results in a low light throughput per measurement. Here, we demonstrate a compressive and compact hyperspectral single-pixel imaging system specifically designed for high-precision spectral measurements at kilometer-range distances. To reduce the influence of sunlight fluctuations on the measurement acquisition process, we also propose an optimized method for modulation pattern ordering. First, we demonstrated the superior image quality of the reconstructed results with our method to the results with the traditional method by numerical simulations. Next, we verified the reconstructed spectral accuracy of the system on an optical table in the laboratory. Then, we performed passive hyperspectral imaging experiments over a range of 4.4 km, with a spectral window of 500–750 nm. We verified the superiority of the optimized modulation patterns by comparing the recovered images obtained using both the traditional and optimized patterns. Experimental results show that the system achieved compressive hyperspectral imaging with a spectral resolution of 1.8 nm and a spatial resolution of 0.18 mrad, simultaneously. Our system offers an efficient and compact measurement solution for remote hyperspectral imaging applications, such as environmental monitoring, precision agriculture, and long-range target detection.