Aliasing Assessment and Mitigation of PRISM Hyperspectral Sensor Simulated Images

The paper proposes a procedure to assess the aliasing effects that may produce distortions on remote sensing images acquired by hyper-spectral push-broom sensors and that arise because of sampling. Real images recorded over different targets at a resolution that is high for the sensor under investigation are used for the analysis. A model for the system modulation transfer function of PRISM hyper-spectral push-broom sensor is developed by taking into account the different contributions due to optical layout, electronics, detector, satellite motion. By using the sensor model, the high resolution images are pre-filtered and spatially re-sampled in order to obtain simulated images of the sensor. Such images are compared with those obtained by an ideal pre-filtering and re-sampling process in order to evidence possible aliasing effects. Quantitative indexes are adopted to assess the presence of aliasing. Digital filtering is adopted to mitigate aliasing effects; to this aim a multi-resolution filter and a fuzzy filter are evaluated. Quantitative and qualitative results show that the proposed filters are effective in aliasing mitigation with negligible penalties on spatial resolution.