Remote Sensing Quantitative Inversion Study Of Vegetation Chlorophyll Content Sentinel 2 In Hilly Road Region In South China

The construction and operation of expressways cause serious pollution to the surrounding environment,so it is very important to protect and maintain the highway environment for the sustainable development of human beings.The chlorophyll content of vegetation in the road area can effectively reflect the health status of vegetation and evaluate the environmental pollution degree of expressway,so it is of great significance to use remote sensing technology to obtain the chlorophyll content in real time to monitor the environment of expressway.In this study,the expressway Changchang Expressway located in the Southern Hills is selected as the experimental object.Combined with the radiation transmission model,the feature band is extracted by constructing the correlation between the spectral band of Sentinel-2A image and the measured chlorophyll content of vegetation.Based on the feature band,the linear regression and BP neural network inversion model of vegetation chlorophyll content are constructed,and the optimal model is applied to the whole On the sentinel-2A images of the study areas,the large-scale inversion and distribution map of chlorophyll content of vegetation in the road area were realized.The main research contents and conclusions are as follows:(1)Based on the radiative transfer model,through the analysis of the sensitivity of chlorophyll content and vegetation spectrum,it shows that the sensitive band of chlorophyll content is mainly located in the visible and near-infrared band,and with the increase of vegetation chlorophyll content,vegetation reflectivity decreases.(2)The correlation of single band,vegetation index,first-order differential,second-order differential and measured chlorophyll content was constructed,and the high correlation bands B2,B3,B5,green edge chlorophyll index cigreen,green band normalized vegetation index gndvi,B1,B2 first-order differential and B1 second-order differential were selected as the characteristic bands of inversion model input.(3)Based on the characteristic band of radiative transfer model,a linear regression model and BP neural network inversion model of chlorophyll content were constructed,and the model was verified by sentinel 2 image data.(4)Compared with the accuracy of the model,the results show that the accuracy of the model constructed by univariate linear regression is higher than that of BP neural network on the whole;the model constructed by BP neural network is over fitted,and part of the measured data is added to the model to reduce the fitting degree of the model and improve the prediction accuracy of the model;the model constructed by univariate linear regression is the best one Based on the model,it is applied to the sentinel-2A image of the whole study area to get the distribution map of the chlorophyll content of the vegetation in the road area.Compared with the actual situation of the vegetation in the road area,the overall space of the chlorophyll content is similar.The results show that the accuracy of single linear regression combined with radiative transfer model used in sentinel-2A image inversion of the vegetation chlorophyll content in the hilly area of South China is high,which meets the monitoring requirements.The spatial distribution map of the vegetation chlorophyll content in the road area is generally similar to the actual situation.Using BP neural network to construct the inversion model of the chlorophyll content,adding some measured data to participate in the modeling can solve the over fitting problem The results show that it is feasible to use sentinel-2A image to quantitatively retrieve the chlorophyll content of the vegetation in the southern hilly region,and enrich the research of remote sensing image to retrieve the vegetation parameters.