Wavelength Estimation of Light Source via Machine Learning Techniques using Low Cost 2D Layered Nano-material Filters

We created a vector of low-cost filters for estimating the wavelength of a light source in the range of 300nm-1100nm using gradually differing mixtures of nano-flakes of semiconducting materials, Molybdenum-Disulfide (MoS₂) and Tungsten-Disulfide (WS₂). The nano-flakes were produced by method of Liquid-Phase Exfoliation and Sonication. We studied the incident and transmitted intensities of light passing through these filters and derived a statistical model for the behavior pattern of the filters for incident light. By employing machine learning techniques we estimated the wavelength distribution of incident light with accuracy of ≤ %1 using the incident and transmitted sensor readings for intensity of light.