基于数字图像特征的冬小麦渍害监测研究

为探索渍害胁迫下冬小麦灾损程度的可视化监测方法,通过田间试验,分析了麦田16个常用图像特征指数在不同受渍时间下的变化特征及其与冬小麦SPAD值、产量和千粒重的相关关系,并建立了基于图像特征指数衰减量的冬小麦渍害估算模型。结果表明,随渍水时间的增加,红光(R)、红光标准化值(NRI)、超红指数(EXR)、植被颜色指数(CIVE)极显著上升,而绿光标准化值(NGI)、归一化绿红差值指数(NGRDI)、绿-红差值指数(GMR)、超绿指数(EXG)、绿红比值指数(GRVI)则极显著下降;且这9个图像特征指数均与冬小麦SPAD值、产量和千粒重呈极显著相关,相关系数的最大绝对值分别为0.92、0.85和0.91;基于图像指数衰减量所建的SPAD值、产量和千粒重减少量的估算模型均以二次多项式最优,且以CIVE指数衰减量构建的SPAD值、产量和千粒重减少量估算模型的预测精度最高,验证集决定系数分别达到0.98、0.95、0.96。因此,数字图像技术可用于冬小麦渍害监测,且以基于CIVE指数的监测效果最佳。

Monitoring Winter Wheat Waterlogging Based on the Features of Digital Image

In order to explore the visual monitoring method for winter wheat loss suffered from waterlogging disaster, the study analyzed the changes of 16 image feature indices through field experiments, as well as the correlation with SPAD, yield and thousand kernel weight, and then constructed estimation models of winter wheat waterlogging based on the attenuation of image feature indices. The results showed that R, normalized redness index(NRI),excess red index(EXR) and color index of vegetation extraction(CIVE) were all significantly increased with the increase of waterlogging time,while normalized greenness index(NGI),normalized green red difference index(NGRDI),green minus red(GMR),excess green index(EXG) and green-red ratio vegetation index(GRVI) were significantly decreased. The above nine image feature indices were significantly correlated with SPAD, yield and thousand kernel weight of wheat, with the maximum absolute values of 0.92, 0.85 and 0.91,respectively.Moreover, the quadratic polynomial model could be efficiently applied for the modeling of SPAD, yield and thousand kernel weight reduction estimation, and the accuracy of the model built with CIVE was highest,with validation determination coefficient of 0.98, 0.95 and 0.96,respectively. These results indicated that digital image technology could be applied as an effective method for monitoring winter wheat waterlogging, and the CIVE was the best index,thus providing guidance for accurately monitoring of wheat under waterlogging stress by digital image technology.