Vegetation chlorophyll content is a key component in ecosystem function.Study on vegetation chlorophyll content retrieval is carried out in 3 directions in which developing or modifying spectral index to retrieve canopy chlorophyll content is a direction which may be a compromise between multiple stepwise regression and inverting physical models because it has some physical meaning while is easier than the latter.At leaf level

we analyzed the applicability of the spectral indices when applied to chlorophyll content retrieval.The reasons that why some studies found these spectral indices are highly correlated with chlorophyll content of their observed samples and why these relationships can not be applied to other peoples’ observed samples are explained in this paper.A recent study developed a semi-empirical model to retrieve canopy crop chlorophyll content which combines spectral index TCARI and soil adjusted index OSAVI.According to this study

the chlorophyll content is determined by the slope of the intersected isolines whose independent variable y is the value of TCARI and the dependent variable x is the value of OSAVI.So a semi-empirical model was derived which is a logarithmic function of TCARI/OSAVI value and through validation with observed corn canopy reflectance and chlorophyll content

this model gives promising results.In this paper

we give some modifications for this semi-empirical model.First

the intersection point of the isolines is considered which was taken as the origin while actually is not

i.e

the chlorophyll content is determined by（TCARI-a/OSAVI-b） with a and b being the Y-coordinate and X-coordinate of the intersection point respectively in TCARI-OSAVI space.Second

a reciprocal function is thought to be more appropriate than a logarithmic one.Considering these two points

a modified model is given in this paper.With our observed corn canopy reflectance and chlorophyll content

this modified model gives better results.