基于遥感影像植被指数变化量分析的冬小麦长势动态监测

Monitoring and Evaluation of Winter Wheat Growth Based on Analysis of Vegetation Index Changes on Remote Sensing Images

为进一步探究利用中低分辨率影像监测小麦苗情的机理,丰富小麦长势动态监测的模式,结合2017-2018年定点观测试验,以GF-WFV数据为遥感影像源,研究了孕穗-开花期冬小麦主要长势变化量参数和产量及其与植被指数变化量间的定量关系,以逐步回归方法筛选目标长势变化量参数,分别构建及评价基于GF-WFV影像遥感植被指数变化量的孕穗-开花期叶片含氮量变化量和叶绿素含量变化量监测模型。结果表明,冬小麦叶片含氮量变化量(ΔLNC)和叶绿素含量变化量(ΔCHL)与产量密切相关,而孕穗-开花期的归一化植被指数变化量(ΔNDVI)、比值植被指数变化量(ΔRVI)分别与ΔLNC和ΔCHL相关性最好,因此选择这两个植被指数变化量作为敏感参量构建冬小麦长势监测模型。经验证,基于ΔNDVI和ΔRVI构建的长势线性模型可靠且精度高,其决定系数分别为0.70和0.64,均方根误差分别为0.39%和0.08 mg·L^-1FW。基于预测模型和实测数据分级量化表达冬小麦长势的空间分布状况,能够很好实现了基于GF-WFV时相影像长势不同等级的遥感监测。

In order to explore the mechanism of using low and medium resolution images to monitor wheat seedling situation and enrich the dynamic monitoring mode of wheat growth,combined with the fixed-point observation test in 2017-2018,GF-WFV data was used as the remote sensing image source to study the quantitative relationship between the main growth change parameters and yield of winter wheat from booting to flowering period and the change of vegetation index,and the target length was selected by the stepwise regression method.Based on the change of vegetation index of GF-WFV remote sensing image,the monitoring models of nitrogen content and chlorophyll content in leaves from booting to flowering stage were constructed and evaluated.The results showed that the changes of nitrogen content(ΔLNC)and chlorophyll content(ΔCHL)in winter wheat leaves were closely related to the yield,while the changes of normalized vegetation index(ΔNDVI)and ratio vegetation index(ΔRVI)from booting to flowering stage were closely related toΔLNC andΔCHL,respectively.Thus,these two changes of vegetation index were selected as sensitive parameters to construct the growth potential of winter wheat monitoring model.It is verified that the linear model of growth potential based onΔNDVI andΔRVI is reliable and has high precision.The determination coefficients are 0.70 and 0.64,respectively,and the root mean square errors are 0.39%and 0.08 mg·L^-1FW,respectively.According to the model,the spatial distribution of winter wheat growth conditions were expressed quantitatively.Thus,monitoring growth grading based on GF-WFV temporal images is realized.