CONTAM - Detection of contaminants in solid matrices and plants using hyperspectral CASI2-SWIR images

Project outcome

Expected scientific results

Two out of four sampled ash roads were correctly identified by the SASI images as unpaved ash roads. The paved ash road (Vlegelstraat) was correctly classified not to be an ash road. The Glaswolstraat was classified as an ash road, although the metal concentrations relatively low. Additional sampling and validation on the ground will provide the necessary information for evaluation of the images. It can be concluded that SASI images can qualitatively detect ash roads. At this stage, the images cannot be used for quantitative analysis (metal concentrations).

 

Correlations between metal concentrations in pine needles and measured reflectances could not be conclusively established due to limited measurements, although a decreasing trend between EGFN and zinc concentrations was observed. Significant negative correlations were found between the stress index EGFN and metal concentrations for the birch leaves, although the variance in the EGFN is large. This may be due to variation in measurement time and light intensity having profound impacts on stomatal aperture and photosynthesis. The maximum fluorescence intensity and the internal cadmium concentrations show a significant positive correlation. Linear regression explains 71 % of the observed variation between individual trees.