Retrieval Of Chlorophyll Content Of Phyllostachys Praecox Using Hyper Spectral Remote Sensing

The bamboo plant has the characteristic of growing rapidly, so it has important ecological significance to evaluate its growth condition by observing the change of biochemical parameters during its growth.Chlorophyll content is a good indicator of nutritional status, photosynthetic capacity and developmental stage, the response of reflectance or transmittance of leaf and canopy to the chlorophyll content can also be used as a effective method to monitor crop conditions, nitrogen and water condition, disease, pollution stress, as well as the reference for the development of precision agriculture. As a good bamboo species used for bamboo shoots, Phyllostachys violascens has the characteristics of high yield, shooting early and good economic benefit, so the planting area of Phyllostachys violascens is increasing. In this study, the data of hyper spectral reflectance and chlorophyll content are observed continuously, and on the basis of which, the correlations between the chlorophyll content and hyper spectral vegetation indexes were analyzed at leaf and canopy scale, respectively. Thus, the estimation model of chlorophyll content are established.The main contents of the research include the following aspects:1. Set and select hyper spectral vegetation indexes and feature-based spectral parameters at leaf scale.Five kinds of vegetation index, including ratio vegetation index, normalized difference vegetation index,difference vegetation index, ratio vegetation index of chlorophyll absorption and vegetation index that reflects spectral feature(e.g. spectral peak or valley) parameters are set to analyze the correlation between them and chlorophyll content of Phyllostachys violascens at leaf scale.2. The chlorophyll content retrieval at leaf scale. Vegetation indexes which have a good correlation with chlorophyll content are selected to build a linear model and the multivariate linear model, respectively.3. Simulate canopy reflectance and calculate vegetation indexes at canopy scale.Canopy reflectance are simulated by PROSPECT-SAIL coupling model,then vegetation indexes including ratio vegetation indexes, normalized difference vegetation indexes, difference vegetation indexes are set on that basis.4. The chlorophyll content retrieval at canopy scale.Correlation between vegetation index and chlorophyll content are analyzed; Then, vegetation index perform excellent would be selected to build chlorophyll content retrieval model that chlorophyll content retrieval model of Phyllostachys violascens at canopy scale are established.Through the above research, this paper draws three conclusions :1.Result at leaf scale shows that: there are better correlations between chlorophyll content and hyper-spectral vegetation index such as GM、Vog3、DD、m ND705、m SR705 and REP during the whole growth period.2. The coefficients of correlation of univariate linear models by these hyper-spectral vegetation indices are all above 0.85 with 99% confidence. The multivariate linear models between the sixHyper-spectral vegetation indices and these chlorophyll content in two strategies both predict with high precision, and the correlation coefficient between predicted values and measured values are all above 0.89.3. Canopy spectral reflectance are simulated by PROSAIL model accuracy, then the difference vegetation index, ratio vegetation index, normalized vegetation index are calculated, after which the correlation analysis of these vegetation index and chlorophyll content are conducted. Indexes can be used to estimate canopy chlorophyll content as follows:R800-R750,R810-R750,(R755-R750)/(R755+R750),(R760-R750)/(R760+R750),(R765-R750)/(R765+R750),(R770-R750)/(R770+R750), R755/R750,R760/R750, R765/R750, R770/R750, R780/R750, R750/R765,R750/R770。4.The R2 of Chlorophyll estimation Univariate linear models at canopy scale are between 0.4 and 0.6.Multivariate linear model are based on(R765-R750)/(R765+R750),R765/R750 and R750/R770,and R2 is 0.5865, RMSE is 5.534 when P is under 0.001, which shows a significant linear correlation. R2 of forecasting canopy chlorophyll calculated by Multivariate linear model and measured canopy chlorophyll is0.3747,and the model can be used in canopy chlorophyll inversion.