Canopy Cover Estimation Combined With UAV Lidar And Sentinel-2B Images

Canopy cover is one of the main standards in the definition of international forests,and it is also a common standard for biological physics and natural resource management.Canopy cover can also be used to establish vegetation dynamic models and ecological processes models and predict vegetation productivity and biomass.Canopy cover is also an important factor in the survey of forest resource.As the main indicators for sub-compartment division,land type division,stand type division,and forest quality evaluation.In recent years,with the increasing attention and protection of the forest,in order to promote the sustainable development of the forest,the canopy cover has received more and more attention.Aiming at the method of current field measurement of canopy cover,it is affected by human factors,which is not stable,and only gets some data for time and effort,which cannot meet the research on the distribution and changes of large space-scale forest.This study has chosen The Wantyon Dianlin Farm(North temperate artificial forest)of Chifeng City,Inner Mongolia Autonomous Region is the test area.It applies to the UAV LIDAR point cloud data and Sentinel-2B multi-spectrum image.Essence The ability to detect remote sensing images and UAV LIDAR data combined with the ability to estimate depression,evaluate model performance,and provides basic support and scientific basis for regional scale stagnation estimation.Mainly obtain the following conclusions:(1)Based on the unmanned LIDAR feature variables,the depression estimation model established by the general addition model(GAM)has good performance(R2 = 0.7231,RMSE= 0.1142,RRMSE = 19.07%).UAV LIDAR depression prediction effect is good,and the effect of estimated when encountering complex scenes will be affected.The height bias and density variables of the drone lidar can refer to the estimation of the aimedital forest lingering of the northern temperate zone.(2)When using GAM to estimate the closedness of depression,the optimal model accuracy of the single variable is R2 = 0.6443(RRMSE = 29.04%),and the optimal model accuracy of the dual variable is R2 = 0.7023(RRMSE = 24.94%).The model accuracy is R2 = 0.7275(RRMSE = 22.92%),and the increase in variables can optimize the fitting effect of the model and reduce the prediction error.(3)Based on Sentinle-2B feature variables,the storm estimation model established by GAM has a good prediction effect.The red-border band of Sentinel-2B is proved to be effective for monitoring the health information of vegetation,estimating the physical physical variables of vegetation,whether in single variables,dual variables or three variable models,the red-border index TCARI2 and red band band REs are feature variables as feature variables Participating in the constructive depression estimation model has a higher advantage of fitting.(4)Using UAV LIDAR-like estimated depression as a training sample,the establishment of the SENTINEL-2B research area depression estimation model is established.The fitting effect of the optimal model of the ground-shaped stagnation-S2 B is used for reference applications in the estimated estimation of artificial forest loses in the North temperate zone.