文章摘要
刘露, 彭致功, 张宝忠, 韩玉国, 魏征, 韩娜娜.基于高光谱的夏玉米冠层SPAD值监测研究[J].水土保持学报,2019,33(1):353~360
基于高光谱的夏玉米冠层SPAD值监测研究
Monitoring of Summer Corn Canopy SPAD Values Based on Hyperspectrum
投稿时间:2018-08-01  
DOI:10.13870/j.cnki.stbcxb.2019.01.055
中文关键词: 夏玉米  高光谱  冠层SPAD值  敏感波段  光谱指数模型
英文关键词: summer corn  high spectrum  SPAD value of canopy  sensitive band  spectral index model
基金项目:国家重点研发计划项目(2018YFC0407703);中国水利水电科学研究院基本科研业务费专项(ID0145B082017,ID0145B742017,ID0145B492017);流域水循环模拟与调控国家重点实验室自主研究项目(2016TS06)
作者单位E-mail
刘露1,2, 彭致功2, 张宝忠2, 韩玉国1, 魏征2, 韩娜娜2 1. 北京林业大学水土保持学院, 北京 100083

2. 中国水利水电科学研究院流域水循环模拟与调控国家重点实验室
, 北京 100038 
pengzhg@iwhr.com 
摘要点击次数: 202
全文下载次数: 574
中文摘要:
      开展夏玉米冠层SPAD值监测技术研究,建立叶绿素含量与敏感波段、光谱指数间的定量关系模型,以促进高光谱技术在玉米快速、无损长势监测及水肥精准管理的应用。以小型蒸渗仪夏玉米光谱反射率与植株冠层SPAD值的监测为基础,研究了夏玉米植株冠层光谱信息与SPAD值的响应关系,并优选出监测夏玉米冠层SPAD值的敏感波段与最优光谱指数。结果表明:夏玉米冠层光谱反射率在可见光波段随玉米冠层SPAD值增加而下降,在近红外波段却与之相反;采用原始光谱反射率、一阶微分光谱监测夏玉米冠层SPAD值的最敏感波段分别为700,690 nm,与SPAD值的相关性分别为-0.498(p<0.05)和-0.538(p<0.01);而根据多元逐步回归分析获得的最优波段组合由405,408,700 nm波段构成;从已报道的73个光谱指数中筛选出与夏玉米冠层SPAD值相关性较高的(SDr-SDb)/(SDr+SDb)、MCARI//OSAVI、TCARI/OSAVI、SDr/SDb和MTCI等5个光谱指数,光谱指数(SDr-SDb)/(SDr+SDb)与SPAD值的相关性在各生育期均达极显著正相关,且在全生育期相关系数高达0.697(p<0.01),进一步优选出监测夏玉米冠层SPAD值最适宜的光谱指数为(SDr-SDb)/(SDr+SDb);在基于敏感波段、光谱指数和最优波段组合建立的夏玉米SPAD值的回归模型中,按照模拟效果由高到低排序依次为最优波段组合、光谱指数、原始光谱反射率、一阶微分光谱,其决定系数分别为0.777,0.539,0.351,0.282;推荐以(SDr-SDb)/(SDr+SDb)指数构建的二次多项式模型与基于405,408,700 nm波段组合建立的线性回归监测模型作为夏玉米植株冠层SPAD值光谱监测适宜模型,且R2大于0.539,RMSE及MAE分别小于6.194和4.702。
英文摘要:
      This study was an exploration into monitoring technology of summer corn canopy SPAD values. A small lysimeter was used in the experiment to monitor spectral reflectivity of summer corn and SPAD values of plant canopy, which provided the basis for investigation of the responsive relationship between canopy spectral information and SPAD values. From this relationship, the sensitive wave band and the optimum spectral index were determined for the SPAD values of the summer corn under investigation. The findings suggested that the canopy spectral reflectivity decreased with the increase of SPAD values in the visible light band, but increased with the increase of SPAD in the near infrared band. The band to which SPAD was most sensitive was 700 nm and 690 nm when monitoring the original spectrum and the first differential spectrum respectively and the correlations with SPAD values were -0.498 (p < 0.05)and -0.538 (p < 0.01). The multivariate stepwise regression analysis found an optimum band combination of 405 nm, 408 nm, and 700 nm. From 73 published spectral indices, five indices, i.e., (SDr-SDb)/(SDr+SDb), MCARI//OSAVI, TCARI/OSAVI, SDr/SDb, and MTCI, were chosen because they produced a higher correlation with SPAD values of summer corn canopy. Among the five indices, the most suitable index was found to be (SDr-SDb)/(SDr+SDb), and the correlation coefficient was as high as 0.697 (p < 0.01) in the whole growth period. A combinational regression model of SPAD, built from sensitive band, spectral index, and optimum band, produced the following simulation performances in descending order:optimum band combination, spectral index, original spectral reflectivity, and first differential spectrum. A quadratic polynomial model built from (SDr-SDb)/(SDr+SDb) index, and a combinational linear regression model built from 405 nm, 408 nm, and 700 nm bands, were recommended as suitable models for spectral monitoring of summer corn canopy SPAD values.
查看全文   查看/发表评论  下载PDF阅读器
关闭