Improved method “grid translation” for mapping environmental pollutants using a two-dimensional CAT scanning system

This research reports on a new method that improves two-dimensional tomographic mapping of air pollutants. In traditional reconstruction techniques, a single grid of cells was used to reconstruct a two-dimensional map from measured line integrated open-path Fourier transform infrared (OP-FTIR) spectrometer measurements. Typically, in environmental two-dimensional imaging, the area of interest is sparsely sampled with rays. As a result, the reconstruction grid resolution has to be coarse in order to avoid grid cells that are hit by very few rays. Unfortunately, successful reconstruction of a peak on a coarse grid depends on its position with respect to a grid cell. If it lies mostly within one grid cell, the reconstructed peak value will be fairly accurate; if it overlaps two or more cells, the predicted concentrations are lower. On average, this leads to an underestimation of the peak value, as well as a strong variability when a peak changes location or when a different reconstruction resolution is selected. This paper presents a recently developed “grid translation” method that allows the choice of the reconstruction resolution to be less critical than using previous single grid methods. In addition, this method substantially improves the quantitative and qualitative reconstruction accuracy of concentration maps under the configuration constraints of OP-FTIR CAT scanning systems.