Study On Leaves' Polarized Hyperspectral Characteristics And The Estimation Model Of Its Chlorophyll Content

Plants exist widely on the earth's surface,and leaves are important organs for photosynthesis,nutrient transformation,respiration and transpiration in plants.Chlorophyll content in plant leaves is an important physiological index of plant photosynthesis ability and nutrient supply,as well as an indicator of crop growth and disaster status.Because of its large amount of information,short cycle,objective,and its advantages in continuous,dynamic tracking and monitoring,remote sensing technology is widely used in the estimation of crop biochemical parameters.Compared to traditional optical and radiological remote sensing methods,because of the multi-dimensional characteristics of polarization information,polarized remote sensing has more unique advantages in crop automatic observation.In this paper,it is studied the relationship between the polarized spectrum of different plant leaves and the chlorophyll content by the experiments of polarized spectrum observation on a variety of plants and leaves.So as to find out the correlation of them,and then use the polarized reflectance spectrum to reckon the chlorophyll content of the plant,so to understand the growth conditions and the disaster situation of crops.It can provide scientific basis and technical support for the automatic observation of crops,the recognition,classification and application study of polarized remote sensing.The main works are as follows:(1)Using the polarimetric imaging ground experiment platform,polarization observation experiments were carried out from different orientations.By comparing the forward,lateral and backward directions of DOP images,it is found that the degree of polarization of the target in the forward image is the largest,the information is the most abundant,the lateral direction is the second,and the information contained in the backward image is the least,Therefore,the polarization detection for the target characteristics of the plant should be based on the forward orientation.(2)The polarized spectra of plant leaves with different surface features such as burr,fluff and wax layer were studied by using indoor multi-angle observation platform and ASD spectrometer with polarizing plate.Through the comparison and analysis of the experimental results,it is found that the more burrs on the leaf surface,the more the villi,the thinner the wax layer,the lower the polarization reflection;the most influential of the three characteristics on thedegree of polarization is wax layer,burr is the second,the impact of the villi is relatively small;the influence of wax layer on the polarization reflection and the degree of polarization are both very obvious.(3)The correlation between the chlorophyll content and the polarization hyperspectral spectrum of a variety of smooth leaves was studied by using the indoor multi-angle observation platform,ASD spectrometer and chlorophyll meter with polarizing plate.By analyzing the relationship between polarized hyperspectral and chlorophyll content,a chlorophyll content estimation model based on the green light polarization was established and the accuracy was evaluated.The results show that for the smooth blade,in the 420 ~ 720 nm measurement range,The relationship between the degree of polarization and chlorophyll content is the best,followed by the polarization reflection,again for the maximum reflection and total reflection,the least reflective relationship is the minimum reflectance.The R2 and RMSE of the estimation model of chlorophyll content based on exponential form of polarization degree were 0.7527 and 9.5759,respectively.And the model has passed the significance test of the reliability of 0.01,meaning that it can be used to estimate the chlorophyll content.