Natural light field compressive imaging
Cheng Zhang, Jiang Jin-bo, Zhu JinBing · Twelfth International Conference on Information Optics and Photonics · 2021
Compressed imaging of natural scenes utilises the sparsity a priori of most scenes in nature. It can achieve higher-resolution imaging through low-resolution measurements recorded on a smaller focal plane array. This paper proposes an implementation scheme of a compression imaging system for large-scale natural scenes. First, the degree of freedom of the imaging transmission matrix is reduced through separable compressed sensing, thereby reducing the storage space of the measurement matrix and the calibration workload; secondly, for the separable imaging transmission matrix, a separate calibration method is designed to realise the left and Calibration of the separable measurement matrix on the right; finally, high-resolution reconstruction is achieved through the calibrated imaging transmission matrix and the recorded low-resolution measurement. The simulation and experimental results show the effectiveness of the compressed imaging scheme proposed in this paper and successfully achieve the calibration and rapid reconstruction of the natural scene compressed imaging system, providing ideas and references for establishing a new generation of camera design.