Remote Sensing Retrieval Of Vegetation Chlorophyll In Typical Road Regions Of South China

In the process of China's rapid development towards a transportation power,the great role of high-grade highway can not be ignored,but the operation of highway not only promotes the rapid economic development,but also has an impact on the surrounding environment and resource utilization.In the process of highway construction,the land occupation and water system change,the pollution caused by vehicles in the process of road operation lead to the impact of road vegetation and other external environment,which makes the monitoring and protection of highway environment put on the agenda.The change of road vegetation environment will lead to the change of chlorophyll content of road vegetation.Using remote sensing monitoring method to study the change of chlorophyll content of road vegetation can provide support for the evaluation of road environment.The chlorophyll content of vegetation will change with the change of vegetation growth status.In this paper,the typical section of Hunan Province-Chang Shao Lou Expressway(both plain and hilly,typical topography;vegetation growth is luxuriant,high vegetation abundance;surrounding vegetation is diverse)is taken as the research object,and the remote sensing quantitative inversion of chlorophyll content in Landsat8 and Sentinel-2A satellite images is carried out by using partial least squares,support vector machine and random forest inversion methods.(1)Reflectance spectral transformation analysis of Landsat8 and Sentinel-2A satellite images.Through the analysis of the correlation between the measured chlorophyll content and the image bands of single band and spectral transformation(first-order differential and second-order differential),the response ability of the spectral transformation band reflectance to the measured chlorophyll content is enhanced,which shows that the spectral transformation is of great significance to the sensitivity analysis of measured chlorophyll content.The correlation between reflectance and chlorophyll content of Sentinel-2A is higher than that of Landsat8.(2)Extraction and analysis of chlorophyll sensitive bands from Landsat8 and Sentinel-2A satellite data.The results show that the original bands with high correlation of chlorophyll sensitive bands of Landsat8 satellite data are B3 and B4,the first-order differential transformed bands are B2fd,the second-order differential transformed bands are B1sd,and NDVI,TVI and MSR4 are more sensitive to vegetation index;the original sensitive bands of chlorophyll sensitive bands of Sentinel-2A satellite data are B2,B3 and B4,the first-order differential transformed bands are B1fd,B2fd,and the second-order differential transformed bands are B2fd After changing,the sensitive band is B2sd,and the vegetation index is more sensitive.(3)Construction and analysis of chlorophyll content inversion model.Partial least squares(PLS),support vector machine(SVM)and random forest chlorophyll content inversion model were constructed,and the inversion results were compared and analyzed.In this study,random forest combined with Sentinel-2A image data had the highest accuracy in retrieving chlorophyll content of vegetation on both sides of typical Expressway in southern China,and its validation set determination coefficient(R2)and root mean square error(RMSE)had the highest accuracy.The results show that:in this study,the accuracy of vegetation chlorophyll content inversion on both sides of typical Expressway in South China is high by using random forest and Sentinel-2A image data,which indicates that it is feasible to quantitatively retrieve vegetation chlorophyll content in hilly road area in South China by using Sentinel-2A image,and this study can provide support for environmental assessment of typical highway in South China.