Study On Remote Sensing Extraction Of Vegetation Coverage Based On LAI

Soil erosion is one of the main eco-environmental problems in China Vegetation coverage is the key factor affecting soil erosion.How to obtain vegetation coverage accurately becomes an important problem in quantitative estimation of soil erosion.In this paper,the extraction technology of vegetation coverage is studied.Under the condition of soil erosion,the vegetation coverage in space is extracted based on remote sensing data Studies have found that vegetation coverage has a good correlation with Leaf Area Index(LAI).Many scholars have studied and compared the difference between vegetation index and LAI quantitative characterization of real vegetation coverage by runoff plot experiments.Therefore,from the perspective of LAI representation of vegetation coverage,the extraction technology of vegetation coverage under the influence of soil erosion is realized.Based on this,vegetation coverage is achieved by LAI.The research area is designated as Pingshan County,Shijiazhuang,Hebei.Selection the same period of 16 meters of GFI WFV and a resolution of 30 m Landsat 8 OLI remote sensing image.Through pretreatment,select NDVI,RVI,SAVI.,MSAVI,ARVI and PVI.The vegetation types are divided into farmland,broad-leaved forest and mixed forest,coniferous forest and shrub grass.Regression analysis was made between the extracted vegetation index and the measured vegetation LJi[.Obtaining the optimal inversion model.The accuracy is verified by the measured data and image sampling method.Finally,the relationship between the retrieved LAI and the actual coverage of UAV is verified.The study results show that in the 248 LAI-Vl models,fitting accuracy of the model by using the GF 1 WFV image,the single variable optimal model of different vegetation types are as follows:1)When the vegetation is farmland,logarithm model correlation coefficient NDVI and the highest was 0.5792.2)When the vegetation is broad-leaved forest and mixed forest when the correlation coefficient of polynomial model of MS AVI is up to 0.5747.3)When vegetation and coniferous forest,the polynomial inversion model of MSAVI and NDVI has higher correlation coefficient.,which is 0.5887 and 0.5833,respectively.4)When the vegetation is irrigated,the correlation coefficient of NDVI polynomial model is the highest,reaching 0.6167.5)In four vegetation types,the accuracy of multivariate linear models was higher than that of single variable models.The correlation coefficients of LAI models for different vegetation types were 0.7485,0.7419,0.6787 and 0.7213,respectively.Using this model,we estimate the LAI of 4-9 months in Pingshan County.Using UAV to take photos to estimate sample coverage,the correlation coefficients between the LAI and the inversion obtained are greater than 70%.It is proved that this method i5 feasible and the vegetation coverage is extractedIn this study,remote sensing technology is used to replace the ground survey,and we can extract vegetation information from near ground surface quickly and in large scale.It provides new methods and ideas for vegetation cover extraction technology.The feasibility of this method is verified by experiments.Further research on LAI and vegetation coverage should be strengthened in the fiuture.To provide technical support for grasping the change of vegetation coverage and the spatial distribution of soil eros ion in the whole country,it will lay a good foundation for the development of our ecological environment.