【作者】 隋学艳；

【导师】 李少昆；

【作者基本信息】 石河子大学 ， 作物栽培学与耕作学， 2006， 硕士

【摘要】 遥感技术集空间技术、传感器技术、信息技术和计算机等技术于一身,可以迅速、大面积、多时相或周期性获取地面信息,是监测农业、资源环境及其变化的重要手段之一,也是实现现代化的农业管理方式——精准农业的主要支撑技术。通过遥感技术监测棉花长势,对指导棉花农田管理有重要的实用价值。为了制作基于高光谱的棉花便携式田间快速营养诊断仪,本研究分析了单叶光谱数据、冠层光谱数据与主要栽培生理指标的关系,并建立回归模型,结果如下:1)叶绿素密度与反射率的相关度在可见光波段〉红外波段,一阶微分以490-550nm的蓝边、680-717nm红边的前半部分达到0.01的极显著负相关水平;557-668nm的黄边、717-797nm红边的后半部分达到0.01的极显著正相关水平。用103个样本进行建模,100个样本进行检验所建模型中选定y=1183.5FD754^0.9059为回归模型。2)含水量与光谱反射率的相关度红外波段〉可见光波段,可见光中蓝光、红光波段达到了0.01的极显著正相关水平。用61个样本进行建模,60个样本进行检验所建模型中选定y = -0.0124 Area9802 + 4.9857 Area980+ 2.8937为回归模型。3)叶片纤维素含量的相关性仅在可见光的蓝光、红光部分达到了0.05的显著正相关水平,用37个样本进行建模,16个样本进行检验,所建模型中选定y = 0.39578-0.49776e^{-R429/R743/0.13036}为估算模型。4)对主要生育时期的冠层光谱数据与生物量分析,在整个生育时期内受棉花花、棉絮及棉株内含物的影响整个波段的相关度都有变化。用74个样本进行建模,73个样本进行检验,所建模型中选定y=1261e^{0.3716 P_Area920}为估算模型。5)利用上述4个模型,对干旱胁迫下实时监测的四个小区冠层光谱数据进行分析后发现:受干旱胁迫的棉花整个生育期与正常棉花相比生物量小、叶绿素密度小、单位面积含水量少、纤维素含量多,表现早衰现象,据此可以根据实时采集的光谱数据进行精准灌溉以减免棉花损失,合理利用水资源。更多还原

【Abstract】 Remote sensing technology, congregating special, sensoring, informational and computer technique, can rapidly and periodically obtain land information broadly. It has become one of the advanced methods on monitoring agricultural environment. Using Remote Sensing to monitor cotton grow condition is very useful for cotton field management.To make cotton field nutrition instant diagnoses instrument based on High Remote Sensing, this study analyzes the relationship between leaf spectrum, canopy spectrum and main cultivation indexes, then sets regression models. Results are as follows:1) The correlation degree between Chlorophyll and reflectance is higher in visible light than in infrared light. The first differential coefficient reaches negative significant level at 0.01 in blue edge range from 490 to 550 nm and the former part of red edge from 680 to 717 nm, positive significant level at 0.01 in yellow edge range from 557 to 668 nm and the latter part of red edge from 717 to 797 nm. Set models with 103 modeling samples and 100 test samples, y=1183.5FD754^0.9059 is chosen as regression model.2) The correlation degree between water containment and reflectance is higher in infrared light than in visible light. Set models with 61 modeling samples and 60 test samples, y = -0.0124 Area9802 + 4.9857 Area980+ 2.8937 is chosen as regression model.3) The correlation degree between cotton leaf cellulose containment and reflectance reaches 0.05 significant level only in blue and red range. Set models with 37 modeling samples and 16 test samples, y = 0.39578-0.49776e^{-R429/R743/0.13036} is chosen as regression model.4) After analyzing the relationship between canopy spectrum and biomass of main stages, we find that the relationship of the whole wave band changes affected by flower and batt. Set models with 74 modeling samples and 73 test samples, y=1261e^{0.3716 P_Area920}is chosen as regression model.5) Using these 4 models, we study leaf and canopy spectrum of different water deal cotton. Under draught stress, cotton biomass, Chlorophyll Density and water containment are all lower than normal cotton. While cellulose containment is higher compared with normal cotton and the cotton represents caducity. Thus can we carry out precision irrigation monitor to derate produce loss, use water resource reasonably, relying on real time spectrum.更多还原