Random forest microplastic classification using spectral subsamples of FT-IR hyperspectral images.

In this work, a random decision forest model is built for fast identification of Fourier-transform infrared spectra of the eleven most common types of microplastics in the environment. The random decision forest input data is reduced to a combination of highly discriminative single wavenumbers selected using a machine learning classifier. This dimension reduction allows input from systems with individual wavenumber measurements, and decreases prediction time. The training and testing spectra are extracted from Fourier-transform infrared hyperspectral images of pure-type microplastic samples, automatizing the process with reference spectra and a fast background correction and identification algorithm. Random decision forest classification results are validated using procedurally generated ground truth. The classification accuracy achieved on said ground truths are not expected to carry over to environmental samples as those usually contain a broader variety of materials.

Short-range remote detection of liquid surface contamination by active imaging Fourier transform spectrometry.

An imaging Fourier transform spectrometer developed at TUHH was used for short-range remote detection and identification of liquids on surfaces. The method is based on the measurement of infrared radiation emitted and reflected by the surface and the liquid. A radiative transfer model that takes both the real and imaginary parts of the refractive index of the materials into account has been developed. The model is applied for the detection and identification of potentially hazardous liquids. Measurements of various liquids on diverse surfaces were performed. The measured spectra depend on the optical properties of the background surface. However, using the radiative transfer model, automatic remote detection and identification of the liquids is possible. The agreement between measured spectra and spectra calculated using the radiative transfer model is excellent.

Passive remote sensing of pollutant clouds by Fourier-transform infrared spectrometry: signal-to-noise ratio as a function of spectral resolution.

In a passive infrared remote sensing measurement, the spectral radiance difference caused by the presence of a pollutant cloud is proportional to the difference between the temperature of the cloud and the brightness temperature of the background (first-order approximation). In many cases, this difference is of the order of a few kelvins. Thus the measured signals are small, and the signal-to-noise ratio (SNR) is one of the most important quantities to be optimized in passive remote sensing. A model for the SNR resulting from passive remote sensing measurements with a Fourier-transform infrared spectrometer is presented. Analytical expressions for the SNR of a single Lorentzian line for the limiting cases of high and low spectral resolutions are derived. For constant measurement time the SNR increases with decreasing spectral resolution, i.e., low spectral resolutions yield the highest SNRs. For a single scan of the interferometer, a spectral resolution that maximizes the SNR exists. The calculated SNRs are in good agreement with the measured SNRs.