Assessing a soil-removed semi-empirical model for estimating leaf chlorophyll content