%A He Yi,Zhou Xiaocheng,Huang Hongyu,Xu Xueqin %T Counting Tree Number in Subtropical Forest Districts based on UAV Remote Sensing Images %0 Journal Article %D 2018 %J Remote Sensing Technology and Application %R 10.11873/j.issn.1004-0323.2018.1.0168 %P 168-176 %V 33 %N 1 %U {http://www.rsta.ac.cn/CN/abstract/article_2906.shtml} %8 2018-02-20 %X Tree number is the index that describes the stand density,and extracting tree number in districts is important for researching forest reserve information.This essay investigates Jiangle forest farm in Fujian to study the applicability of maximum algorithm and multiscale segmentation algorithm in extracting tree number.The first step in this dissertation is to use ebee unmanned aerial vehicle remote sensing with fixed wings to get image whose resolution is more than 10cm,and gets orthoimage after processing.Based on it,this essay uses 20 area samples including coniferous forest and broad-leaved forest.The second,the tree number of samples were extracted from maximum algorithm and multiscale segmentation algorithm;Lastly,this essay uses the tree number abstracted fromthe two algorithm and the tree number from visual interpretation statistics to precision analysis.The results show that the tree number of samples was extracted by the two algorithm in the overall accuracy of about 90%.In the coniferous forest plots the tree number of extraction accuracy by local maximum algorithms is better than the broad-leaved forest plots the tree number of extraction accuracy.In the broad-leaved forest plots the tree number of extraction accuracy by multiscale segmentation algorithm is better than the tree number of extraction accuracy by local maximum algorithms.Therefore,this research argues that local maximum value algorithm is more suitable for unmanned aerial vehicle (uav) remote sensing image on coniferous sample tree number of rapid extraction,the multiscale segmentation algorithm is more suitable for unmanned aerial vehicle (uav) remote sensing image on broad-leaved sample tree number of accurate extraction.