耕地土壤有机质与速效氮磷钾含量高光谱遥感反演研究
[Abstract]:The content of soil organic matter and soil nutrients is of great significance to the growth of supplying crops. The development of hyperspectral remote sensing provides an effective technical approach for the monitoring of regional organic matter and soil nutrients. In order to explore the relationship between soil organic matter and nutrient status and satellite image spectra, Hyperion hyperspectral satellite imagery and farmland fertility survey were used to determine the soil properties (organic matter, alkali-hydrolyzed nitrogen, available potassium). The spectral characteristics of available phosphorus), the spectral inversion model of soil attribute is constructed, and the optimal inversion model of each soil attribute is selected, and the accuracy of the model is evaluated by comparing with the measured value. The main results are as follows: the results show that vegetation in remote sensing images has a large degree of interference with the precision of soil attribute inversion model. The precision of the soil attribute inversion model based on the soil pixel in the image is higher than that of the soil attribute inversion model based on the vegetation pixel. The results of soil pixel spectral sensitivity analysis showed that the content of organic matter had a good response to the Hyperion range of 782.95-813.48 nm. The model with the first derivative of reflectivity has the best fitting accuracy (R2 = 0.777RMSE = 5.31), and the correlation between the model inversion results and the measured values (R2 = 0.809 ~ RMSE = 5.19), which can be used for the rapid determination of the distribution of regional organic matter content. The available phosphorus content has a good response ability to the 1467.33 nm ~ 1 00.29 nm band of Hyperion. The fitting accuracy of the model based on the first derivative of reflectivity is the best (R2 = 0.767RMSE = 19.55), and the correlation between the model inversion results and the measured values (R2 = 0.783rMSE = 9.04), and the fitting accuracy of the model is R2 as follows: The RMSE of 0.314 and 0.405 are 38.06 and 52.47 respectively. It can not be used for the rapid determination of alkali-hydrolyzed nitrogen and available potassium in hyperspectral. For vegetation pixels, only available phosphorus content has a good response to the 1457.23 nm band of Hyperion. The fitting effect of the model established by the ratio index is the best (R2 = 0.304 RMSE is 38.96), and the correlation between the model inversion results and the measured values (R2 = 0.740? RMSE = 16.77), organic matter, The fitting accuracy of the optimal inversion model for alkali-hydrolyzed nitrogen and rapidly available potassium is 0.171 0.196 and 0.163 RMSE respectively, which is 9.1U 39.63 and 79.79, which can not be used for the rapid determination of organic matter, alkali-hydrolyzed nitrogen and available potassium in vegetation covered areas in hyperspectral remote sensing images.
【学位授予单位】:福建农林大学
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:S158;S127
【参考文献】
相关期刊论文 前10条
1 刘焕军;张小康;张新乐;武洪峰;金慧凝;于胜男;邱政超;;面向土壤分类的高光谱反射特征参数模型[J];遥感学报;2017年01期
2 刘凡;马玲;杨光;陈建华;马雪莲;王海江;;灰漠土土壤全氮含量的高光谱特征分析及估测[J];新疆农业科学;2017年01期
3 Chang TIAN;Xuan ZHOU;Qiang LIU;Jian-wei PENG;Wen-ming WANG;Zhen-hua ZHANG;Yong YANG;Hai-xing SONG;Chun-yun GUAN;;控释肥料对早熟油菜产量、养分吸收和肥料利用率的影响(英文)[J];Journal of Zhejiang University-Science B(Biomedicine & Biotechnology);2016年10期
4 南锋;朱洪芬;毕如田;;黄土高原煤矿区复垦农田土壤有机质含量的高光谱预测[J];中国农业科学;2016年11期
5 金慧凝;张新乐;刘焕军;康苒;付强;宁东浩;;基于光谱吸收特征的土壤含水量预测模型研究[J];土壤学报;2016年03期
6 王莉雯;卫亚星;;湿地土壤全氮和全磷含量高光谱模型研究[J];生态学报;2016年16期
7 白由路;;植物营养与肥料研究的回顾与展望[J];中国农业科学;2015年17期
8 许丽丽;李宝林;袁烨城;高锡章;刘海江;董贵华;;2000-2010年中国耕地变化与耕地占补平衡政策效果分析[J];资源科学;2015年08期
9 栾福明;熊黑钢;王芳;时卉;王昭国;张芳;王晶晶;;基于小波分析的土壤速效K含量高光谱反演[J];干旱区地理;2015年02期
10 吴才武;夏建新;段峥嵘;;土壤有机质预测性制图方法研究进展[J];土壤通报;2015年01期
相关博士学位论文 前5条
1 陈红艳;土壤主要养分含量的高光谱估测研究[D];山东农业大学;2012年
2 张婷婷;基于PLS模型的农业土壤成分高光谱遥感反演研究[D];吉林大学;2010年
3 刘伟东;高光谱遥感土壤信息提取与挖掘研究[D];中国科学院研究生院(遥感应用研究所);2002年
4 李裕元;坡地土壤磷素与水分迁移试验研究[D];西北农林科技大学;2002年
5 张富仓;土壤-根系统养分迁移机制及其数值模拟[D];西北农林科技大学;2001年
相关硕士学位论文 前5条
1 李曦;基于高光谱遥感的土壤有机质预测建模研究[D];浙江大学;2013年
2 严加亮;武夷山不同海拔土壤磷素的空间异质性研究[D];福建农林大学;2012年
3 刘青;人工神经网络遥感影像分类系统的设计与实现[D];昆明理工大学;2012年
4 鲍晨光;森林类型遥感分类研究[D];东北林业大学;2010年
5 李建勇;测土配方施肥中土壤有效氮钾测定方法研究[D];西南大学;2008年
,本文编号:2175392
本文链接:https://www.wllwen.com/kejilunwen/nykj/2175392.html