Recent developments in time domain NIRS for non-destructive fruit quality assessment: non-contact measurements and device miniaturization

At present, most near infrared spectroscopy (NIRS) devices for non-destructive assessment of fruit quality exploit continuous wave (CW) light sources (e.g., lamps, LEDs) and detectors (e.g., photodiodes). Unfortunately, CW NIRS is not able to distinguish light scattering (related to microscopic changes of refractive index) from light absorption (related to fruit chromophores) phenomena. Furthermore, when working in reflectance mode, CW NIRS suffers from limited sensitivity to fruit pulp, since source-detector distance and tissue optical properties markedly affect the portion of tissue that is sampled. Potentially, such limitations can be overcome by using multiple source-detector distances, however, the complexity of the setup results increased. As different approach, time domain near infrared spectroscopy (TD NIRS) sensors can acquire the distribution of time-of-flight of the detected photons, by employing picosecond pulsed laser sources, single-photon sensitive detectors, and fast timing electronics. Thus, analysis of TD NIRS data enables to disentangle reduced scattering coefficient from absorption coefficient. Furthermore, fruit pulp is more reliably sampled by photons travelling deeper into the fruit, that is those characterized by later arrival times, on average. TD NIRS devices have not been broadly adopted yet due to the cost of the components, the complexity of the instrumentation, and the requirement for the optical probe to be in contact with the sample. In this work we present the latest advancements in TD NIRS technology concerning feasibility of non-contact measurements and device miniaturization. We present three TD NIRS systems for the non-destructive assessment of fruit internal quality: a multi-wavelength device hosted on a 19” rack, a non-contact prototype for fast measurements suitable for future implementation in industrial sorting lines, and a portable device for in-field measurements. As an example of application, we present the results of measurements performed by the three TD NIRS systems on a set of Mantuan PGI pears.