| Based on the TLS(TLS,Terrestrial Laser Scanning)point cloud data,the three-dimensional scene of the stand could be reconstructed,the parameters of the single tree could be accurately extracted,and the spatial overview of the stand after thinning could be simulated,which had important guiding significance for the determination of technical parameters of forest tending and thinning.This study took Larix principis-rupprechtii plantation in Saihanba Mechanical Forest Farm as the research object and obtained the TLS point cloud data of the sample plot,the high spatial resolution remote sensing image of the unmanned aerial vehicle,and the field survey data of the sample plot.On the basis of cropping,resampling,denoising,ground point classification,and normalization processing of TLS point cloud data,the single tree position,DBH,tree height,crown width,and crown boundary were extracted,and the extracted values were compared with the actual measured values and analyzed to test their reliability;For the scanning plots,the interval is 10%thinning intensity,up to 70%thinning intensity,and different thinning schemes were set.The unit circle analysis method of microenvironment and forest competition determined the corresponding harvested trees;analyzed the response of stand spatial structure parameters to different thinning schemes by means of zero-element distribution and uni-element distribution;based on tree size factor,forest competition factor and site conditions,the growth model of single tree diameter at breast height of Larix principis-rupprechtii predicted the growth trend of a single tree and the change of stand productivity after implementing different thinning schemes,and comprehensively analyzed the impact of different thinning schemes on stand spatial structure and stand productivity,so as to optimize the thinning scheme.The main research results were as follows:(1)The detection accuracy of TLS point cloud single tree segmentation was 94.58%,the over-detection error was 4.26%,and the missed detection error was 1.16%,and its Bias and RMSE were divided into 0.44m and 0.27m.The Bias,rBias,RMSE and rRMSE of the extracted DBH are divided into-0.34cm,-1.68%,0.89cm and 4.39%;the Bias,rBias,RMSE and rRMSE of the extracted tree height were divided into-0.09m,-0.43%,0.74m and 3.71%;the Bias,rBias,RMSE and rRMSE of the extracted crown width were0.24m,6.44%,0.49m and 13.05%,respectively;the hexagonal algorithm could capture the exact shape of the canopy point cloud data and effectively express the geometric features of the tree canopy profile.The results showed that the TLS point cloud data had high reliability in the application of extracting single wood parameters.(2)The unit circle analysis method based on tree size,tree microenvironment and forest competition could comprehensively evaluate the quality of individual trees,quantitatively represent the status of individual trees in the forest stand,and objectively determine the harvested trees of different thinning schemes.Based on the cloud data of the sample site,the simulated harvesting of different thinning schemes was carried out,and the results of different thinning schemes were displayed visually.(3)From the zero-element distribution and uni-element distribution of stand spatial structure parameters,it could be seen that when the thinning plan was40%of the thinning intensity,the stands were in a random distribution state,the spatial distribution pattern of stands was reasonable,and the individual size differences and stand density were lower.(4)The R~2,MSE,ME,MRE,RMSE and TRE of the single-tree diameter growth model of Larix principis-rupprechtii were 0.8153,0.2491cm,0.300cm,-5.39%,0.4991cm,and 1.34%,respectively.The overall accuracy reached the prediction standard,and it had a good prediction effect for the diameter at breast height of Larix principis-rupprechtii.From the five-year growth of stand sectional area and stand volume,it could be seen that when the thinning plan was 30%of the thinning intensity,the increase in stand sectional area and stand volume was the highest,and the increase in stand productivity was the most significant.Therefore,considering the stand spatial structure and stand productivity after different thinning schemes,the optimum thinning intensity for the scanned plots in this study was 30%-40%,and the reserved wood density was 302 plants/ha-402 plants/ha. |