Artificial intelligence for the calibration of mobile spectral analyzers

Abstract:
Mobile spectral analysis has gathered a lot of interest throughout the last years. It enables the determination of object composition, material properties or other application relevant information directly on-site. This could be very interesting in agriculture, food business, environmental monitoring, medical or other fields. Autonomous operated systems could be handheld or mounted to robots as well as unmanned aerial vehicles (UAVs). The components have become compact and affordable so far. The evaluation of spectral data, for example in the near infrared range, requires stable calibration to extract the relevant information to extract concentration or composition information. Complex chemometric models and broad data bases are necessary. These are valid for selected devices only, transfer to other equipment is difficult. Furthermore, changes like aging of the light source must be calibrated from time to time. In close future artificial intelligence (AI) will help to change this situation by providing automated calibration and compensation routines.