Effects of leaf dust retention on spectral characteristics of Euonymus japonicus

  Objective  Leaf dust retention will affect the spectral characteristics of vegetation, which will weaken the response ability of vegetation index to vegetation and affect the accuracy of inversion evaluation. In order to explore the influence of leaf dust retention on vegetation spectral response characteristics and prediction models, this study took the common greening tree species Euonymus japonicus in Beijing as the research object.

  Method  Leaf samples were collected from closed area, semi closed area and open area, and environmental dust was collected. Hyperspectral data from different dust-retaining leaves were measured using an ASD FildSpec Handheld spectrometer through indoor control experiments. Five characteristic bands and spectral angles were used to study the influence of leaf dust retention on the spectral characteristics of leaves, and the influence of dust retention on the accuracy and stability of the prediction model for leaf dust retention was also studied.

  Result  The characteristics of vegetation spectral curves were gradually weakened and the characteristics of dust were gradually enhanced with the increase of leaf dust retention, but the overall trend of spectral curves was basically the same. However, when the leaf dust retention was greater than 120 g/m², the spectral curve basically showed the spectral characteristics of dust. When the leaf dust retention was less, the simulation accuracy of the prediction model was relatively high, but with the increase of dust retention, the determinant coefficients of all prediction models decreased; when the leaf dust retention was greater than 120 g/m², the prediction ability of all prediction models for leaf dust retention will be worse. And the root mean square error (RMSE) increased with the increase of leaf dust retention per unit area, and the stability and prediction accuracy of the prediction model were gradually reduced. The spectral angle was very sensitive to the spectral changes of dust-retaining leaves in the range of 350−1 770 nm. It is not necessary to discuss the degree of dust-retaining using the spectral angle of leaves in different regions, while simple comparison with the threshold was need to be done.

  Conclusion  In this study, the response characteristics of leaf dust retention to vegetation spectrum were studied through indoor control experiments, which could provide basic theory and data support for establishing physical model of spectral reflection of dust retention vegetation.