基于开花期卫星影像的春性中、弱筋小麦籽粒蛋白质含量遥感预测

Remote Sensing Prediction of Grain Protein Content in Medium and Weak Gluten Spring Wheat Using Satellite Image at Anthesie

为了对小麦籽粒品质提前监测与预报,以为小麦生产形势的分析和收购调度提供参考,本研究采用Landsat-5 TM卫星遥感数据,分析长江中下游流域江苏地段春性中、弱筋小麦拔节期和开花期植株性状参数(叶片SPAD值、叶片氮含量和植株氮含量)、卫星遥感植被指数和籽粒蛋白质含量间关系,建立籽粒蛋白质含量卫星遥感预测模型。结果表明:拔节期植株性状参数与籽粒蛋白质含量相关性不显著;开花期植株性状参数与籽粒蛋白质含量均呈极显著正相关(P<0.01),可用于籽粒蛋白质含量提前预测;开花期植株性状参数均可采用NDVI和DSW3等卫星遥感植被指数进行反演,效果最优,并基于NDVI和DSW3分别建立籽粒蛋白质含量预测模型。采用模型预测2007年和2009年籽粒蛋白质含量,与实测值间相关性均达极显著水平(P<0.01),且RMSE值较低,其中NDVI模型预测精度较高。构建了基于开花期数据的小麦籽粒蛋白质含量卫星遥感预测模型,采用NDVI模型制作了2009年江苏省部分区域春性中、弱筋小麦成熟期籽粒蛋白质含量卫星遥感预测专题图。

The present study provides some references for wheat production in situation analysis and acqui- sition schedule through monitoring and predicting grain quality. We measured plant growth quality indexes, remote sensing parameters by Landsat-5 TM satellite data at elongation and anthesis stage and grain protein content at maturity stage in Jiangsu locating in middle and lower reaches of the Yangtse River. The relation- ships between/among these indexes were analyzed to establish remote sensing prediction model of grain pro-tein content （GPC） in medium and weak gluten wheat. The correlations between GPC and plant growth quality indexes, including leaf SPAD reading, leaf nitrogen content and plant nitrogen content, were not significant at elongation. However, plant growth quality indexes at anthesis could be used to predict GPC because these indexes were significantly related with GPC （P 〈0. 01）. Plant growth quality indexes could be better predicted by NDVI and DSW3 than other remote sensing parameters at anthesis. Therefore, two GPC models based on NDVI and DSW3 were established respectively. The correlations between GPC meas- ured values and prediction values by above-mentioned models were very significant in experiments of 2007 and 2009 （P 〈0. 01 ）. NDVI model showed more accurate prediction than DSW3 model evidenced by lower RMSE. Furthermore, we produced thematic maps at anthesis for predicting GPC in medium and weak glu- ten spring wheat growing in Jiangsu some areas in 2009 by using NDVI model.