LAI Estimation Of Pinus Massoniana Forest Based On Multi Angle UAV Image

Leaf area index（LAI）is the sum of the single-sided plant area per unit surface area,an important parameter that reflects the growth of vegetation,and an important input parameter for many ecological models.Studies have shown that vegetation canopy structure is an important factor that affects the accuracy of LAI estimation,and the hot-dark spot in the bidirectional reflection characteristics of Masson pine forest contain rich canopy structure information.The so-called hot and dark spots refer to the phenomenon that the maximum and minimum reflectance occurs due to the size of the gap in the vegetation canopy and the number of leaves at a particular observation angle.Therefore,how to mine effective information in the hot-dark spot area And utilization is an important basis for improving the accuracy of LAI estimation.Based on this,the study took 72 Pinus massoniana forest plots in Hetian Town,Changting County,a typical soil erosion area in southern red soil,as the research object,and set 9 different observation angles（0°,±15°,±30°）using drones,±45°and±60°;when the observation direction is the same as the light direction,it is"-",otherwise it is"+")to detect the effect of the hot-dark spot area on the estimation of the leaf area index,and then construct the estimation model of the leaf area index,In order to provide a reference for the accurate estimation of the LAI of Pinus massoniana forest canopy.The main research conclusions are as follows:（1）The hot spot and dark spot areas of Pinus massoniana forest were determined by the analysis of its dichroic reflectance spectrum,The results show that under the same solar zenith angle,no matter how the LAI value changes,the reflectance increases with the increase of the observation angle in the negative observation angle,and the reflectance at the observation angle of-45°reaches the strongest point;as the angle increases in the forward observation angle,the reflectivity as a whole tends to decrease,but in many places（+15°,+45°,+60°observation angle）The lowest value of reflectivity appears,so by calculating the average reflectance,it is found that the reflectivity of each band reaches the lowest value at the observation angle of+45°.Therefore,according to the definition of hot spot and dark spot,the observation angle of-45°is the area where the hot spot is located,and the observation angle of+45°is the area where the dark spot is located.（2）The effective information mining in hotspot and dark spot areas is an important basis for LAI estimation.Based on the determined hotspot（-45°）and dark spot（+45°）areas,the hot-dark spots in different wave bands are used for reflection The corresponding hot-dark spot index（HDS）,normalized hot-dark spot index（NDHD）,improved normalized hot-dark spot index（MNDHD）,hot-dark spot ratio index（HDRI）were constructed respectively,and then through According to the correlation analysis of the measured LAI,it is found that the HDRI-r ratio calculated by the reflectance of the red light band is the highest R²of the measured LAI,and the value is 0.5883.Finally,HDRI-r is determined as an important auxiliary index for estimating LAI.（3）The introduction of the hot-dark spot index into the traditional vegetation index can effectively solve the shortcomings of the traditional vegetation index,such as insufficient canopy structure information and easy saturation,and also an important basis for improving the accuracy of LAI estimation.The study combines HDRI-r with selected traditional vegetation indices（NDVI,RVI,GNDVI,NRI,SIPI,VARI,MSAVI）to obtain a variety of hot-dark spot vegetation indices（NHDRI,RHDRI,GNHDRI,NRHDRI,SIPHDRI,VAHDRI,MSAHDRI）,Through linear fitting with the measured LAI,and comparative analysis to obtain the optimal hot-dark spot vegetation index for the estimated LAI.The results show that after the introduction of the optimal hot-dark spot index HDRI-r in the traditional vegetation index,the fitting R2 with the measured LAI has been improved by 66%,114%,494%,304%,503%,150%,361%,in which the fitted R²of the constructed hot-dark spot vegetation index and the measured LAI is from large to small in order of RHDRI>VAHDRI>NRHDRI>NHDRI>GNHDRI>SIPHDRI>MSAHDR I,so the study selects the hot-dark spot vegetation index and The RHDRI with the highest LAI fitting R²is an important parameter for constructing the estimated LAI model.（4）Vegetation spectral information at different observation angles has some help in estimating LAI.Therefore,based on the construction of the hot-dark spot vegetation index RHDRI,combined with other observation angles and vegetation indices with high correlation with LAI,the study adopted Multiple stepwise regression analysis method was used to construct an optimal model for effectively estimating the leaf area index of Masson pine forest in the study area.The specific calculation formula is y=0.047×RHDRI+0.321×RVI_+60°-1.519.The average estimation accuracy is 94.22%and 99.23%respectively.The estimation accuracy of the model introduced with RHDRI is 17.4%higher than that of the model without considering the hot-dark spot information,and the estimation accuracy of the model is improved by 4%relative to the model constructed only with the hot-dark spot vegetation index.（5）The range of LAI of Pinus massoniana forest in the study area obtained from the optimal estimation model is 0.109～4.277m²/m²,with an average value of 2.513m²/m².In typical red soil erosion areas in the south,understory soil erosion is serious,and the soil is relatively poor.As a result,the overall growth of Masson pine in the study area is slower and the plants are shorter.The results of the study are consistent with the actual situation.