Spectral reflectance of soil samples with different particle sizes
Description
A total of 449 soil samples were collected from the study area. Each sample was crushed and sieved using soil screens with different mesh sizes (150, 80, 60, 45, 35, and 30 mesh), resulting in soil samples with particle sizes of 0.1, 0.2, 0.3, 0.4, 0.5, and 0.6 mm, respectively, which were labeled as DA1, DA2, DA3, DA4, DA5, and DA6. The reflectance images of the soil samples with varying particle sizes were captured using a Specim IQ hyperspectral imaging system, and the raw data were subsequently converted into reflectance data. The Mahalanobis distance method was employed to analyze the measured total nitrogen (TN) content of all soil samples. Outlier samples exceeding three times the average Mahalanobis distance were excluded. This process resulted in the removal of 144 soil spectral samples from 24 sampling points across the six particle sizes, leaving a total of 2,550 soil spectral samples from 425 sampling points with six particle sizes. The resulting dataset comprises reflectance information for 2,550 soil spectral samples across six particle sizes from 425 retained sampling points.