基于无人机遥感的云南松林冠参数提取研究OACSTPCD

Taking the typical natural Pinus yunnanensis pure forest in central Yunnan as the research ob-ject,combined with unmanned aerial vehicle(UAV)remote sensing image acquisition and field standard survey,the traditional visual interpretation,multi-scale segmentation and watershed segmentation methods were used to extract the two canopy parameters of single tree crown and canopy density.Based on the ground measured data,the rates of crown extraction accuracy of the three single tree segmentation methods were 91.48％,87.33％and 84.04％,respectively.Taking the visual interpretation results as a reference,the accuracy of multi-scale segmentation method to extract canopy density was 90.24％,and R2 was 0.884 5.The accuracy of canopy density extracted by watershed segmentation method was 87.40％,andR2 was 0.743 7.The results indicate that the multi-scale segmentation method based on UAV remote sensing image can extract the canopy parameter information of P.yunnanensis pure forest well,and the extraction accuracy meets the requirements of forest resource survey,which can effectively improve the efficiency of forest resource survey.