Duanhao Huang, Luo Wenjie, Feng Huang, Pengfei Wang
The Risley-prism imaging system (RPIS) exhibits great potential for wide field and variable resolution imaging due to its outstanding capability on beam control. However, the non-linear and non-uniform deflection of light introduced by the prism components, result in geometric distortion, dispersion, and ghosting artifacts. To address these challenges, a robust reconstruction framework is developed based on multi-channel structure-guided fusion. The framework first identifies ghosting regions through theoretical and experimental analysis of ghosting artifacts, enabling precise separation of degraded and smooth areas. Then, distortion and dispersion are corrected while high-quality structural information is transferred across channels, effectively suppressing ghosting and enhancing overall image quality. Besides, an objective evaluation metric termed ghosting-dispersion index (GDI) are proposed to quantitatively measure the severity of ghosting and dispersion. A multi-channel RPIS is fabricated, and experimental results demonstrate that the proposed method consistently achieves the highest GDI scores even when the field of view (FOV) is expanded by 1.15 times, and displays superior correction and reconstruction performance as well as its robustness and adaptability to diverse environments, significantly outperforming existing algorithms for both indoor and outdoor scenarios.