The quantification of moisture is critical for many industrial and research applications. Quantitative models based on spectroscopy are efficient, non-destructive, and accurate for monitoring the moisture content. Hyperspectral cameras also reveal the spatial distribution of moisture, whereas point spectrometers provide only general distribution. In this study, we monitored the moisture content of a cotton patch along with its drying.
NIR = NEAR INFRARED (900 – 1700 nm)
PLS = PARTIAL LEAST SQUARE
PLS DA = PARTIAL LEAST SQUARE DISCRIMINANT ANALYSIS
WATER ABSORPTION PEAKS IN THE NIR SPECTRAL RANGE
Monitoring the moisture in production is critical, for example, in the food, paper, and wood industries. Near-infrared (NIR) spectroscopy is widely used to monitor moisture content in various applications.
Spectroscopists rely on water absorption peaks being part of the NIR spectral range. As shown below, water strongly absorbs light at 970, 1150, and 1450 nanometers. The Specim FX17 camera’s spectral range covering 900 – 1700 nm is perfect for detecting water absorption peaks. Besides providing relevant data for moisture quantification, the spatial dimension embedded along with the hyperspectral images also reveals the moisture distribution.
In this study, we monitored and quantified the drying time of a cotton patch. We dipped a round cotton pad of ca. 5 cm (typically used to remove makeup) into the water and then applied it on the Specim lab scanner 40×20. We monitored its drying with Specim FX17 hyperspectral camera and took measurements every 4th minute until it dried completely. It required all together sixty-seven (67) measurements over 264 minutes. We analyzed the data with SpecimINSIGHT software.
SPECTRAL ANALYSIS OF THE DRYING COTTON PADS
Each image was first normalized with respect to dark and white references. Then a mosaic was created by combining all 67 images acquired over the time of the experiment into a single file. This mosaic depicts the cotton pad at different drying stages, starting at the top right when it was very wet and dry on the bottom right. The mosaic was filled row after row, from left to right, from top to bottom.
As shown in Fig.1A, a clear gradient from the false RGB mosaic image can be seen after the drying process. Spectra reveal the same trend: those related to the wet samples show the deepest spectral absorption at 970, 1150, and 1420 nm, whereas those peaks vanish along with the drying time. An acceleration of the drying is also noticeable after 3.5 hours (over the last ten measurements).