Application of Monte Carlo Simulation-Based Photon Migration for Enhanced Understanding of Near-Infrared (NIR) Diffuse Reflectance. Part II: Photon Radial Diffusion in NIR Chemical Images

This is the second of a series of articles applying Monte Carlo simulation-based photon migration as a means of enhancing understanding of near-infrared (NIR) diffuse reflectance in pharmaceutical analysis. This work utilizes Monte Carlo simulation to explore photon movement along the radial direction, as a means of investigating the potential 2-D spatial distortion in NIR chemical images caused by photon radial diffusion. Simulations indicate that the physical sample represented by an observed pixel in an image contributes 17% of the spectral information, while the remaining contribution comes from a radial area around the observed pixel about 500 mm wide. Empirical measurements on a two-medium sample (i.e., Teflon films with different thicknesses laid on the top of a pure theophylline compact) were used to confirm the effect of photon radial movement on a physical feature in an NIR chemical image, that is, edge blurring. As a means of edge clarification, a preliminary Monte Carlo simulation-based deconvolution filter was developed and tested for its efficacy on sharpening the physical feature and maintaining the original chemical information in a measured NIR chemical image. The initial success of the deconvolution filter highlights the potential to develop a filter for practical applications.