Inversion Model Of Crop LAI Based On Lidar And Multispectral Images

Biological resources are one of the most important resources on earth.The destruction of habitats poses a great threat to species extinction.Due to the over-exploitation of land resources and diseases and insect pests,many vegetation has become extinct in fragile habitats.50% of the wild habitats in the tropics have been destroyed,over 80% in the subtropics,70% of China’s natural forests have been harvested,50% of grasslands have been destroyed,and a quarter are threatened by deserts.In order to guarantee the sustainable use of resources and conserve biodiversity,it is the most important to guarantee the quality of biological resources.In the agricultural production process,macro-observation of vegetation growth and growth change is an important basis that fully reflects the information on the agricultural situation and provides farmers with an objective basis of agricultural production and estimation of crop production.The leaf area index(LAI)is a parameter of one commo n production,and traditional leaf area index measurement methods are not only destructive,but also time,labor and efficiency are low.In recent years,unmanned aerial vehicle remote sensing has developed vigorously.Low-altitude remote sensing based on unmanned aerial vehicles has been widely used.The technique of LAI inversion using multispectral data is relatively mature.However,the research on retrieving LAI by using laser radar with new sensors mounted on unmanned aerial vehicles is still few,whi ch is the current research hotspot and the general trend of future agricultural monitoring,and has research value.In this paper,40 sorghum LAI samples were collected from different regions of Helen cooperative experimental field in July 2018,taking plots with obvious differences in crop growth in Helen cooperative experimental field in Heilongjiang province as the research area.Based on the analysis of laboratory measured LAI data,LAI inversion models based on single multispectral data,single lidar data and multi-source fusion are established respectively.Meanwhile,laboratory data are used to analyze and evaluate the input of different models,explain the influence mechanism of interference factors on remote sensing inversion model,and finally dete rmine the optimal LAI inversion model.The results show tha t:(1)LAI,as an important indicator of crop growth characteristics,is an important physiological parameter that represents the canopy information structure and growth trend of plants.Research o n LAI has extremely important influence on monitoring the growth trend of sorghum and improving the yield of sorghum per unit area.(2)Flight planning and data acquisition schemes based on different flight platforms.According to the characteristics of high speed and efficiency of unmanned aerial vehicles,appropriate research areas are selected to acquire lidar and multispectral data.The method solves the problems of coverage calculation,flight route design,ground target laying,data matching and the like in the data acquisition process,and lays a foundation for acquiring effective data.(3)Based on UAV multispectral data,the vegetation index with high correlation coefficient with measured LAI is selected by linear stepwise regression method:(RVI,N DVI,SAVI,DVI,TM,EVI,HJVI,SIR,PSRI,NDVIedge,CI,MSR).The correlation analysis with measured LAI is established,and the vegetation index with correlation greater than 0.7 is selected and introduced as the input of the model,so that the determination coefficient R of the remote sensing inversion model reaches 0.7351 and RMS is 0.1422.(4)Based on the high density point cloud data acquired by laser radar,set reference parameters to separate ground points from non-ground points.The sorghum land digital surface model(DSM)and digital elevation model(DEM)in the study area are obtained by using the sorghum encrypted point cloud data processed by LIDAR360.The canopy height model(CHM)is obtained by making the difference between the digital surface model and the digital elevation model.After that,the effective canopy structure parameters,namely HCV,CRR,standard deviation,mean value,vegetation height index,echo density change measurement and skewness,were extracted.The correlation analysis between canopy structure parameters and LAI was established.Parameters with correlation greater than 0.6 were selected to construct sorghum LAI inversion model,and the accuracy was verified and evaluated with measured LAI.The experimental results show that the determination coefficient R2 between the LAI estimated by the model and the measured LAI is 0.7824 and RMS is 0.1434.(5)Koryan LAI inverse deduction model based on multi-source remote sensing data.Compared with the single sensor LAI inverse deduction model,the accuracy and stability of the model are obviously improved.Using radar and multi-spectral laser sensor Koryang vertical and level structure parameters,along with the reliance spectrum parameter measurement altitude relevance,through the method of reverting,through the red derivation gap index ndviedge,the red edge chlorophyll index Ciel,The relationship between the four parameters of vskewn,the Retirement Research Center exceeded 0.8,and the involvement with the lai repercussions parameter was reliant and the measured value was 0.8347.The RMS is 0.142704.The optimization was performed for the single sensor inverse deduction model l.