A FULLY AUTOMATIC METHOD FOR ON-ORBIT SHARPNESS ASSESSMENT: A CASE STUDY USING PRISMA HYPERSPECTRAL SATELLITE IMAGES

The recent surge in interest towards hyperspectral imagery has the potential to unlock a new range of applications for the scientific community. However, compared to traditional multi-spectral images, the workload required to process such high-dimensional data is dramatically increased, to the point that new and more flexible strategies must be developed in order to properly monitor the quality of this type of products. In the particular case of sharpness assessment, traditional procedures based on the edge method tend to be extremely time-consuming due to their reliance on visual analysis performed by human operators, and would make proper processing of all bands a daunting task to perform on a large scale. In this paper we propose a flexible and fully automatic approach to edge method-based sharpness assessment that can be applied independently from the number of spectral bands. We then present the results of the application of the methodology on the visible and near-infrared and shortwave infrared spectral cubes of a selection of PRISMA L2D images, which confirm the reliability of the methodology and suggest further improvements.