Chlorophyll content retrieval from hyperspectral remote sensing imagery

Chlorophyll content is the essential parameter in the photosynthetic process determining leaf spectral variation in visible bands. Therefore, the accurate estimation of the forest canopy chlorophyll content is a significant foundation in assessing forest growth and stress affected by diseases. Hyperspectral remote sensing with high spatial resolution can be used for estimating chlorophyll content. In this study, the chlorophyll content was retrieved step by step using Hyperion imagery. Firstly, the spectral curve of the leaf was analyzed, 25 spectral characteristic parameters were identified through the correlation coefficient matrix, and a leaf chlorophyll content inversion model was established using a stepwise regression method. Secondly, the pixel reflectance was converted into leaf reflectance by a geometrical-optical model (4-scale). The three most important parameters of reflectance conversion, including the multiple scattering factor (M0), and the probability of viewing the sunlit tree crown (PT) and the background (PG), were estimated by leaf area index (LAI), respectively. The results indicated that M0, PT, and PG could be described as a logarithmic function of LAI, with all R2 values above 0.9. Finally, leaf chlorophyll content was retrieved with RMSE = 7.3574 μg/cm2, and canopy chlorophyll content per unit ground surface area was estimated based on leaf chlorophyll content and LAI. Chlorophyll content mapping can be useful for the assessment of forest growth stage and diseases.