基于无人机多光谱数据的玉米苗株估算  

Estimation of maize seedling number based on UAV multispectral data

作　　者：赵晓伟 黄杨[1] 汪永强[1] 储鼎 ZHAO Xiaowei;HUANG Yang;WANG Yongqiang;CHU Ding(Heilongjiang Provincial Research Institute of Surveying and Mapping,Harbin 150081,China;Northeast Institute of Geography and Agroecology,Chinese Academy of Sciences,Changchun 130102,China)

机构地区：[1]黑龙江省测绘科学研究所,哈尔滨150081 [2]中国科学院东北地理与农业生态研究所,长春130102

出　　处：《自然资源遥感》2022年第1期106-114,共9页Remote Sensing for Natural Resources

基　　金：国家自然科学基金项目“土壤水分与表面粗糙度的光学与雷达遥感协同反演算法研究”(编号:41971323)资助。

摘　　要：为能及时监测和评估东北大面积的玉米出苗情况,估算苗株数,依据低空无人机(unmanned aerial vehicle,UAV)遥感影像为玉米苗株数的快速估算提供有效支持。研究基于UAV多光谱数据,通过对比ExG,GBDI,ExG-ExR,NGRDI,GLI等颜色指数分割玉米与土壤背景,借助OTSU算法确定最佳阈值,选定最佳颜色指数ExG。优化出最佳形态学特征参数的组合:面积A、周长B、矩形长D、矩形周长G、椭圆长轴长度H、形状因子Q。借助支持向量机回归(support vector regression,SVR)模型,预测出玉米苗株数,评价精度,并估算和绘制了局地玉米苗株数的空间分布图。该SVR模型测试的精度达到96.54%,统计误差为0.6%。研究成果能够在短时间内迅速、快捷、准确地预测玉米苗株数和长势趋势。To monitor and evaluate maize seedlings in Northeast China and estimate their number in time,this study provided effective support for the rapid estimation of the maize seedling number using unmanned aerial vehicle(UAV)remote sensing images.Using the multispectral UAV data,the color indexes ExG,GBDI,ExG-ExR,NGRDI,and GLI were compared to segment maize seedlings from the soil background.Then,the optimal threshold was determined using the Otsu algorithm,and ExG was selected as the optimal color index.According to optimization,the best combination of morphological parameters consists of area(A),perimeter(B),rectangle length(D),rectangle perimeter(G),ellipse long axis length(H),and shape factor(Q).Then,the number of maize seedlings was predicted using the support vector regression(SVR)model and the prediction accuracy was assessed.Finally,the spatial distribution map of the local maize seedling number was developed.Tests revealed that the accuracy and the statistical error of the SVR model were 96.54%and 0.6%,respectively.These results allow the number and growth trends of maize seedlings to be predicted quickly and accurately in a short time.

关 键 词：UAV 苗株数 支持向量机回归 颜色指数 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]