摘要
研究对玉米进行了分块种植和变量施肥控制。氮肥施用采取了4个水平胁迫,并在不同的生长期采集其叶片叶绿素含量和株高数等生理指标数据信息,以及多光谱图像数据信息。对多光谱图像进行冠层分割等处理,并提取了图像的NIR/G的比值特征参数。同时,结合实验室对叶片叶绿素的化学测定,建立了多光谱图像与叶绿素之间的关系模型,相关性系数R2达到了0.637,为基于光谱图像的玉米长势无损监测提供了依据。
Corn was cultivated on different block with nitrogen stressing. Four levels of nitrogen treatment were design, and chlorophyll content, height of corn plant, and multi spectral images were obtained. After corn entropy segmentation based on threshold, the ration value NIR/G was extracted from spectral images. Then the correlation model between chlorophyll content measured by chemical method and NIR/R was built, the coefficient of correlation was 0. 637. It is feasible to predict the nutrient status of corn growth using multi spectral images.
出处
《农机化研究》
北大核心
2012年第11期178-181,共4页
Journal of Agricultural Mechanization Research
基金
国家"863计划"项目(2007AA10Z207)
关键词
玉米
多光谱图像
长势监测
冠层分割
corn
multi-spectral images
growth monitoring
canopy segmentation
