一种利用可见光波段无人机遥感的林下植被覆盖识别方法  

A method to recognize vegetation cover of undergrowth based on visible unmanned aerial vehicle data

作　　者：朱雄斌 汪小钦[1] 周小成[1] ZHU Xiongbin;WANG Xiaoqin;ZHOU Xiaocheng(Key Laboratory of Spatial Data Mining&Information Sharing of Ministry of Education, Spatial Information Research Center of Fujian,Fuzhou University,Fuzhou,Fujian 350116,China)

机构地区：[1]福州大学空间数据挖掘和信息共享教育部重点实验室,福建省空间信息工程研究中心,福建福州350116

出　　处：《福州大学学报（自然科学版）》2018年第6期814-820,838,共8页Journal of Fuzhou University(Natural Science Edition)

基　　金：“十二五”国家科技支撑计划资助项目(2013BAC08B01);福建省高校产学合作基金资助项目(2017Y4010);中央引导地方发展专项基金资助项目(2017L3012).

摘　　要：利用可见光无人机遥感数据,通过分析不同地物在像元和对象尺度上的光谱信息,建立一种利用归一化红绿差异指数NGRDI特征并结合面向对象思想的林下植被覆盖识别方法,并对试验区和验证区开展方法应用与适应性分析.分析结果表明:NGRDI在像元尺度可以很好地区分乔木、草地和裸土,在对象尺度也能较好地区分林下有植被覆盖区域和林下无植被覆盖区域,同时该指数能较好地消除地形因素等影响.试验区林下植被识别的总体精度为85.9%,Kappa系数为0.78;验证区林下无植被覆盖区域和林下有植被覆盖区域提取的正确率分别为82.9%和95.1%.故利用可见光无人机影像的NGRDI指数进行林下植被覆盖识别方法是可行的.A method to recognize vegetation cover of undergrowth was proposed based on analysis of spectral characteristics different objects,which was under the use of UAV data.Futhermoer,accuracy analysis and method adaptability analysis was added related to the recognize of undergrowth vegetation cover.The result shows that normalized green-red difference index can not only effectively distinguish different objects,which included trees grassland and bare soil at pixel scale and understory with vegetation cover or not at object scale,but also can eliminate the influence of terrain factors.The overall accuracy and Kappa value was 85.9%and 0.78 respectively in study area,and the accuracy of understory recognize,with vegetation cover or not,was 82.9%and 95.1%respectively in validation area.For the above analysis,we have reasons to believe that it is feasible to recognize vegetation cover of undergrowth based on the method presented in this paper.

关 键 词：无人机遥感 可见光植被指数 林下植被覆盖 面向对象方法 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置] S157[自动化与计算机技术—控制科学与工程]