Miniature NIR Spectrometers: Designed for Harsh Environments

A new deep-learning framework that was recently developed improves the accuracy of near-infrared spectroscopy for non-destructively measuring internal quality traits.

The bulky bench-top NIR spectrometer is quietly being dismantled and rebuilt as a wafer-scale photonic chip, a self-calibrating algorithm, and a sensor small enough to ride in a shirt pocket. What once demanded a grating, a moving mirror, and a climate-controlled lab now fits inside a handheld module, a bioreactor probe, or a drone payload, and it increasingly figures out what it is looking at on its own.