Quantitative Inversionand Dynamic Monitoring Of The Vegetation Canopy Water Content In The Road Area Based On Remote Sensing

Due to the gradual development of expressway networks and the increasing frequency of human transportation activities,they pose a serious threat to the ecological environment along the expressway line.For protecting the limited vegetation resources and the deteriorating ecological environment,it is necessary to analyze the road environment of high-grade highways and periodically monitor the road environment.Timely and accurate monitoring of vegetation water content can obtain information on the health status of the vegetation.In this paper,the leaf equivalent water thickness,which can characterize the vegetation water content,is selected as the research objective to measure the physiological status,morphological structure and health of the vegetation.This paper takes the roadway of Chang Yi Expressway in Hunan Province as the research area,selects multi-spectral remote sensing image of Landsat 8 OLI as the data source,combines the basis of physical radiation transmission model,constructs two types of vegetation equivalent water depth inversion models based on the reflectance of water feature band,moisture index and complex ratio index of the image water,evaluate the accuracy of the model,and estimates the equivalent water thickness of the vegetation canopy.The main research contents and conclusions are as follows:(1)Based on PR04SAIL radiation transmission model,analysis and obtain the equivalent water thickness of vegetation of Landsat 8 OLI image,which are bands of B6 and B7(2)By establishing a statistical model of the water index and the measured equivalent water thickness,select the most relevant Normalized Multi-band Drought Index(NMDI)to participate in the inversion of the equivalent water thickness physical model;(3)Construct six complex ratio indices and analyze their correlation with the equivalent water thickness,and select the significant NDII/NDVI complex ratio index as the equivalent water thickness inversion model factor;(4)Using B6,B7,NMDI,and ND11/NDVI as retrieving factors to model vegetation equivalent water thickness inversion based on PR04SAIL Model,Successfully establish a combined model of PR04SAIL+Support Vector Machine Regression(SVR)and PR04SAIL+K-Neighbor Regression Algorithm(KNN),Successfully invert the equivalent water thickness of vegetation in the road area;(5)Comparing the inversion results of the PR04SAIL+SVR model and the PR04SAIL+KNN model,it can be seen that the inversion results of the two models are in the same trend and the distribution is similar,which verifies the feasibility of using the multi-spectral remote sensing image to invert the equivalent water thickness of the vegetation.The results show that it is fast to invert vegetation equivalent water by the Support Vector Machine regression model and physical model on multi-spectral remote sensing images,which is suitable for estimating the equivalent water thickness of vegetation in a large area;K-Nearest Neighbor algorithm combined with radiation transmission model has high precision for inversion of vegetation equivalent water thickness and can accurately reflect the vegetation moisture content of the road.The research results provide a strong support for the quantitative inversion of the equivalent water thickness of vegetation for the combination of radiation transmission model and artificial intelligence algorithm,which lays the foundation for the quantitative inversion of vegetation parameters in the highway area of the southern hilly region using multi-spectral data.