Hyperspectral Images Acquisition: an Efficient Capture and Processing Stitching Procedure for Medical Environments

Abstract

The complexity of the Hyperspectral (HS) imaging-based applications demands faster and more efficient acquisition and processing systems. Moreover, HSI technology is being more and more used within the medical imaging field, increasing and making more restrictive the requirements of the implementations. In this work, the authors present an efficient methodology for the acquisition and stitching processes when capturing line-scan based HS images. In order to verify the proposed methodology, a full hardware and software setup has been implemented. The proposed method has been tested scanning a reference polymer at different working distances and scroll speeds. The obtained results in laboratory show that the delta in PSNR keeps below 5% for all cases. Also, the correlation between comparable spectral signatures are not affected neither by the working distance nor by the acquisition speed.