TDI-based continuous window compressed spatio-temporal imaging capable of flexible voxels post-interpretation.

To achieve high frame rates and continuous streaming simultaneously, we propose a compressed spatio-temporal imaging framework implemented by combining time-delay-integration sensors and coded exposure. Without additional optical coding elements and subsequent calibration required, this electronic-domain modulation enables a more compact and robust hardware structure, compared to the existing imaging modalities. By exploiting the intra-line charge transfer mechanism, we achieve a super-resolution in both temporal and spatial domains, thus multiplying the frame rate to millions of frames-per-second. In addition, the forward model with post-tunable coefficients, and two reconstruction strategies proposed therefrom, facilitate a flexible voxels post-interpretation. Finally, the effectiveness of the proposed framework is demonstrated by both numerical simulations and proof-of-concept experiments. With the prominent advantages of prolonged time window and flexible voxels post-interpretation, the proposed system will be suitable for imaging random, non-repetitive, or long-term events.