The application of fractal theory in assessing chlorophyll content using hypersectral data

The variance in leaf chlorophyll concentration can cause the comprehensive variation of spectral curve in geometry and characteristics. Fractal is an appropriate mathematical tool to explain the comprehensive variation. In this paper, the fractal dimension of segmented reflectance spectral curves was used to assess the leaf chlorophyll concentration by the moving window technique. Firstly, the size of moving window was defined in 10nm, 20nm, 30nm, 50nm, 75nm, 100nm and the best window size was chose by comparing the correlation between fractal dimension and chlorophyll content. Secondly, the fractal dimension with the best window was used to build the estimation model of chlorophyll content, and the estimation result was compared with various spectral VIs. The research result showed that (1) the window with the length of 50nm was the best for chlorophyll content assessment. (2) For different chlorophyll content, the variation of fractal dimension was mainly found in 475nm~650nm of visible light spectra and 720nm~770nm of near-infrared spectra, with the trend of changes conforming to the reflectance. (3) Comparing with several spectral VIs, the fractal dimension was better related with chlorophyll content. So, the fractal dimension of segmented reflectance spectral curve can be served as a new comprehensive parameter to estimate the chlorophyll content of vegetation.