Nano-scale particle classification using image histogram maximum value index of rayleigh scattered images

Considering the Rayleigh light scattering behaviour by small particles, current study adopts a new technique to classify nano-scale particles with noise intensity histograms features. Noise was generated using the scattered light by five different sized particles with a continuous spectrum of light. Each captured video frame was divided into its red, green and blue (RGB) planes (single channel arrays) and noise was isolated using a modified frame difference method. Mean and the standard deviation of the maximum value index of intensity histograms over predefined number of frames (N) were used to discriminate the particles. Results shows that the classifier was able to distinguish four types of particles (Polyurethane smoke, kerosene smoke, water steam and cooking oil smoke) with the 100% accuracy when N≥100. The classifier failed to distinguish wood smoke particles from the other particles since the maximum value index of the intensity histograms of wood smoke images varies widely due to its complex composition (wood smoke contains different sized particles such as water vapour, resin smoke etc.).