棉花生长初期灌溉信息遥感提取与校正

发布时间：2018-01-26 05:32

本文关键词： 棉花生长初期灌溉信息植被指数归一化差值植被指数(NDVI) 归一化差值水分指数(NDWI)　出处：《中国生态农业学报》2017年08期 　论文类型：期刊论文

【摘要】：为提高生长初期低覆盖度作物长势的遥感监测精度,需要消除灌溉引起的土壤水分背景变化对归一化差值植被指数(NDVI)的影响。为了实现棉花生长初期灌溉信息提取与校正,提高棉花作物长势监测与产量预判精度,本文以美国加利福尼亚州San Joaquin Valley的2个棉花地块为研究区,选取棉花生长初期灌溉过程中的遥感影像,构建两种灌溉信息提取方法(分阶段阈值法和灌溉线提取法),确定最优灌溉像元提取方法;比较分析灌溉与未灌溉情况下棉花的NDVI与归一化差值水分指数(NDWI)以及土壤调节植被指数的关系,提取含有灌溉信息的像元,并对NDVI进行校正,消除灌溉对NDVI的影响。研究结果表明:在棉花生长初期,灌溉与未灌溉像元NDVI变化率达12%,差异较显著;灌溉与否的棉花NDVI与NDWI间均存在极显著的线性关系,决定系数在0.80以上;利用灌溉线方法提取灌溉信息与分阶段阈值相比精度更高,精度达88%以上;校正后线性回归模型精度达0.95,灌溉校正效果明显,灌溉与未灌溉像元的NDVI差异减小至2%。本研究通过对含有灌溉信息像元NDVI值的校正,去除灌溉对NDVI造成的影响,反映了真实的植被信息,可实现对作物生长初期长势的准确遥感监测,为遥感定量监测提供便利。
[Abstract]:In order to improve the precision of remote sensing monitoring of low coverage crop growth in the early stage of growth. The effect of soil moisture background change caused by irrigation on normalized difference vegetation index (NDVI) should be eliminated in order to extract and correct irrigation information at the early stage of cotton growth. To improve the precision of cotton crop growth monitoring and yield prediction, this paper takes two cotton plots in San Joaquin Valley, California, USA as the research area. Two irrigation information extraction methods (stage threshold method and irrigation line extraction method) were constructed to determine the optimal irrigation pixel extraction method. The relationship between NDVI and normalized difference water index (NDWI) and the vegetation index of soil regulation was compared between irrigated and unirrigated cotton, and the pixel containing irrigation information was extracted. NDVI was corrected to eliminate the effect of irrigation on NDVI. The results showed that the rate of change of NDVI between irrigated and unirrigated pixel reached 12 in the early stage of cotton growth, and the difference was significant. There was a very significant linear relationship between NDVI and NDWI in cotton irrigated or not, and the determining coefficient was above 0.80. The precision of extracting irrigation information by the method of irrigation line is higher than that of stage threshold, and the precision is more than 88%. After correction, the accuracy of the linear regression model reached 0.95.The effect of irrigation correction was obvious, and the NDVI difference between irrigation and unirrigated pixel was reduced to 2.This study was carried out by correcting the NDVI value of the pixel containing irrigation information. Removing the effect of irrigation on NDVI reflects the true vegetation information and can realize accurate remote sensing monitoring of crop growth in the early stage of crop growth and provide convenience for quantitative remote sensing monitoring.
【作者单位】：东北农业大学资源与环境学院;
【基金】：黑龙江省自然科学基金项目(D201404) 黑龙江省普通高等学校新世纪优秀人才培养计划项目资助~~
【分类号】：S562;TP751
【正文快照】： *This study was supported by the Natural Science Foundation of Heilongjiang Province of China(D201404)and the Program for New CenturyExcellent Talents in Heilongjiang Provincial University,China.**Corresponding author,E-mail:zhangxinle@gmail.com(College

【参考文献】