张雪茹, 冯美臣, 杨武德, 王超, 郭小丽, 史超超. 基于光谱变换的低温胁迫下冬小麦叶绿素含量估测研究[J]. 中国生态农业学报(中英文), 2017, 25(9): 1351-1359. DOI: 10.13930/j.cnki.cjea.170134
引用本文: 张雪茹, 冯美臣, 杨武德, 王超, 郭小丽, 史超超. 基于光谱变换的低温胁迫下冬小麦叶绿素含量估测研究[J]. 中国生态农业学报(中英文), 2017, 25(9): 1351-1359. DOI: 10.13930/j.cnki.cjea.170134
ZHANG Xueru, FENG Meichen, YANG Wude, WANG Chao, GUO Xiaoli, SHI Chaochao. Using spectral transformation processes to estimate chlorophyll content of winter wheat under low temperature stress[J]. Chinese Journal of Eco-Agriculture, 2017, 25(9): 1351-1359. DOI: 10.13930/j.cnki.cjea.170134
Citation: ZHANG Xueru, FENG Meichen, YANG Wude, WANG Chao, GUO Xiaoli, SHI Chaochao. Using spectral transformation processes to estimate chlorophyll content of winter wheat under low temperature stress[J]. Chinese Journal of Eco-Agriculture, 2017, 25(9): 1351-1359. DOI: 10.13930/j.cnki.cjea.170134

基于光谱变换的低温胁迫下冬小麦叶绿素含量估测研究

Using spectral transformation processes to estimate chlorophyll content of winter wheat under low temperature stress

  • 摘要: 近年来,冻害已成为影响我国冬麦区的农业气象灾害之一,及时、快速、准确地获取冬小麦叶绿素含量对于监测冬小麦冻害发生具有极其重要的意义。本研究通过低温胁迫试验,在拔节期对两个冬麦品种进行-6℃,4 h、8 h和12 h的胁迫处理后,测定其冠层光谱反射率,并对原始光谱数据进行15种典型变换处理,分析比较不同光谱变换下冬小麦叶绿素含量的PLSR模型,筛选出能够表征低温胁迫下冬小麦叶绿素含量的最佳光谱变换方式。结果表明,随低温胁迫时间的延长,两个冬小麦品种叶绿素含量呈降低趋势,随着低温胁迫后天数的增加,各处理与对照的差异逐渐减小。胁迫后5 d,近红外区域反射率有较大升高,并随低温胁迫后时间的延长而升高;在可见光区域,短期内差异不明显。胁迫后10 d、20 d、35 d,黄、红波段逐渐趋于水平,同时近红外区域反射率差异逐渐缩小,可见光区域光谱反射率出现不同程度的上升。对原始光谱数据进行15种典型变换处理,发现原始光谱的倒数、对数、幂、平方根等变换难以提高与叶绿素含量的相关性,且建模效果较差。除原始光谱对数的一阶微分(T6)外,其他微分变换处理的叶绿素含量诊断模型都优于原始光谱。综合考虑模型的校正、验证效果、模型最佳因子数以及相对分析误差的大小,二阶微分变换处理(T15)叶绿素含量校正模型的R2和RMSE分别为0.930、0.340,验证模型的R2为0.753,表明基于T15的光谱变换数据可实现低温胁迫下叶绿素含量的准确估算,为最佳光谱变换方式。

     

    Abstract: Chlorophyll content is a vital index of photosynthetic capacity and crop growth status. In recent years, freeze injury had become a main meteorological disaster at jointing stage of winter wheat in the northern region of China. Although global climate had been warming since the 1980s, freeze injury had not weakened. Climate warming had led to a decline in the proportion of wheat varieties planted in winter and to a rise in the proportion of the varieties planted in spring. With increasing warm autumn years, the issue of wheat overgrowth has worsened, decreasing the ability of wheat to resist cold. Climate change has not been stable and extreme weather events have increased, implying that there was still the risk of freeze injury of wheat. Shanxi Province suffered freeze injury and the maximum area affected by freeze injury has been estimated at 0.26 million hectares. This had a severe impact on the growth and development of wheat, and ultimately affected the yield of wheat. The rapid and accurate estimation of chlorophyll content of winter wheat is meaningful in resisting the occurrence of freeze injury. However, the routine methods of measuring chlorophyll content are complex and time-consuming. Therefore, developing a rapid and non-destructive chlorophyll content diagnosis technology can be an effective way to monitor winter wheat freeze injury. Here, two varieties of winter wheat were treated under -6℃ temperature stress for 4, 8 and 12 hours at jointing stage. Moreover, canopy spectra were collected and the raw spectrum transformed with respect to 15 transformation methods and then the spectral transformation processes of chlorophyll content of winter wheat were analyzed in the PLSR model. The aim was to select the optimal spectral transformation of chlorophyll content in winter wheat under low temperature stress, and provide theoretical basis for monitoring freeze injury of winter wheat. The results showed that the chlorophyll content of two winter wheat varieties declined with increasing time of low-temperature stress. With increasing number of days after freeze, the differences between low temperature and control treatments decreased gradually. Moreover, the near-infrared region reflectance increased greatly with strengthening low-temperature stress and the visible region had no significant difference in short-term stress after 5 days. While the yellow and red bands increased, the near-infrared region decreased in differences after 10, 20 and 35 days of stress. Compared with the raw spectrum, the transformation methods under differential treatments (e.g., reciprocal, logarithm, power, and square root transformation methods) failed to improve the relevance of chlorophyll content and therefore the overall performance of model was poor. Other differential transformation processes of chlorophyll content diagnosis models had higher precision than raw spectral analysis, except for T6(lgR)', R is the spectral reflectance. Moreover, the second-order differential for raw spectrum (T15, R") had a higher accuracy (RC2=0.930, RMSEC=0.340; RV2=0.753) respectively for the calibrated and validated models after comprehensive evaluation of predicted performance and complexity level of different models. It showed that the second-order differential for raw spectrum (T15, R") was the most plausible transformation method of spectral reflectance for evaluating chlorophyll content of winter wheat under low temperature stress.

     

/

返回文章
返回