| 基于优化光谱指数的新疆春小麦冠层叶绿素含量估算 |
| |
| 引用本文： | 亚森江·喀哈尔,尼加提·卡斯木,尼格拉·塔什甫拉提,张 飞,茹克亚·萨吾提,阿不都艾尼·阿不里,师庆东,苏比努尔·居来提.基于优化光谱指数的新疆春小麦冠层叶绿素含量估算[J].麦类作物学报,2019(2):225-232. |
| |
| 作者姓名： | 亚森江·喀哈尔 尼加提·卡斯木 尼格拉·塔什甫拉提 张 飞 茹克亚·萨吾提 阿不都艾尼·阿不里 师庆东 苏比努尔·居来提 |
| |
| 作者单位： | （1.新疆大学资源与环境科学学院，新疆乌鲁木齐 830046; 2.新疆大学绿洲生态教育部重点实验室，新疆乌鲁木齐 830046; 3.新疆大学干旱生态环境研究所，新疆乌鲁木齐 830046） |
| |
| 基金项目： | 国家自然科学基金项目（41761077，41671348） |
| |
| 摘 要： | 为筛选可用于干旱半干旱区春小麦冠层叶绿素含量估算的高光谱植被指数,2017年通过测定春小麦关键生育时期冠层的田间高光谱与叶绿素含量,利用光谱指数波段优化算法分别计算400～1 300 nm光谱波段中不同波段两两组合的比值光谱指数(ration spectral index,RSI)、归一化光谱指数(normalized difference spectral index,NDSI)、叶绿素指数(chlorophyll index,CI)、简化光谱指数(CI/NDSI,NPDI),并将这些参数及其他17个不同高光谱植被指数分别与实测冠层叶绿素含量进行Pearson相关分析,通过变量重要性准则筛选最优光谱参数,使用偏最小二乘回归法建立冠层叶绿素含量的预测模型。结果表明:(1)RSIs、NDSIs、CIs和NPDIs与冠层叶绿素含量的相关性都优于前人研究中定义的17种高光谱植被指数,并且冠层叶绿素含量与NDSI(R_(849),R_(850))、RSI(R_(849),R_(850)),CI(R_(849),R_(850))和NPDI(R_(849),R_(850))表现出强相关性。(2)用此4个优化光谱指数分别建模时,以CI(R_(849),R_(850))、 CI(R_(539),R_(553))、 CI(R_(540),R_(553))、 CI(R_(536),R_(553))为自变量的X-3模型预测精度最高(r~2=0.74,RMSE=0.272 mg·g~(-1))。(3)结合4个优化光谱指数构建的组合模型预测精度,其r~2=0.83,RMSE=0.187 mg·g~(-1)。
|
| 关 键 词： | 春小麦 冠层叶绿素含量 优化光谱指数 组合模型 |
Estimation of Spring Wheat Canopy Chlorophyll Content in Xinjiang Based on Optimized Spectral Indices |
| |
| YASENJIANG Kahaer,NIJAT Kasim,NIGARA Tashpolat,ZHANG Fei,RUKEYA Sawut,ABDUGHENI Abliz,SHI Qingdong,SUBINUER Julaiti.Estimation of Spring Wheat Canopy Chlorophyll Content in Xinjiang Based on Optimized Spectral Indices[J].Journal of Triticeae Crops,2019(2):225-232. |
| |
| Authors: | YASENJIANG Kahaer NIJAT Kasim NIGARA Tashpolat ZHANG Fei RUKEYA Sawut ABDUGHENI Abliz SHI Qingdong SUBINUER Julaiti |
| |
| Abstract: | In order to screen the hyperspectral vegetation index for estimating the chlorophyll content of spring wheat in arid and semi-arid regions,the field hyperspectral and chlorophyll contents of the canopy were determined during the key growth period of spring wheat in 2017.Initially,various two-band combinations in the 400-1 300 nm band were used to optimize ration spectral index(RSI),normalized difference spectral index(NDSI),chlorophyll index(CI),and spectral index(CI/NDSI,NPDI).Then,the relationship between the indices(RSIs,NDSIs,CIs,NPDIs and 17 different hyperspectral vegetation indices from the literature) and measured canopy chlorophyll content were examined,and the optimal spectral parameters were demonstrated by variable importance in projection(VIP).Ultimately,the partial least squares regression(PLSR) method was utilized to develop a predictive model of canopy chlorophyll content.The results revealed that the newly identified NDSIs,RSIs,CIs and NPDIs always performed better than the hyperspectral vegetation indices from the former studies,and canopy chlorophyll content exhibited strong correlation,with NDSI(R849,R850),RSI(R849,R850),CI(R849,R850) and NPDI(R849,R850).According to the prediction models established respectively by the four optimized spectral indices,the X-3 model revealed that the highest rPre(0.74) and lowest RMSEPre(0.272 mg·g-1) were identified with four optimized chlorophyll indices(CI(R849,R850),CI(R539,R553),CI(R540,R553),and CI(R536,R553).The combination model with four newly identified spectral indices had the best ability for estimating canopy chlorophyll content in spring wheat,and the accuracy of model was r=0.83 and RMSE=0.187 mg·g-1. |
| |
| Keywords: | Spring wheat Canopy chlorophyll content Optimized spectral index Combination model |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《麦类作物学报》浏览原始摘要信息 |
| 点击此处可从《麦类作物学报》下载免费的PDF全文 |
|
