|
|
|
|
  • 软件名称:基于PSO-RBF神经网络模型反演闽江下游水体悬浮物浓度
  • 软件大小: 0.00 B
  • 软件评级: ★★★
  • 开 发 商: 谢旭, 陈芸芝
  • 软件来源: 《遥感技术与应用》
  • 解压密码:www.gissky.net

资源简介

摘要: 总悬浮物浓度是水环境重要参数之一,二类水体光谱特征复杂,光谱特征与悬浮物浓度之间关系不能用简单的线性模型来表示。利用2017年7月12日~13日2 d间对闽江40点位进行水质采样和光谱测量,结合光谱响应函数模拟GF-1 WFV1各波段遥感反射率,分析遥感因子与总悬浮物浓度相关性。利用相关系数较高的波段及组合b3、b3/b2和b3/b1,构建PSO-RBF和传统RBF神经网络总悬浮物浓度反演模型,同时建立以b3/b2为自变量的经验比值模型。结果表明:与传统RBF神经网络和经验模型相比,PSO\|RBF神经网络模型效果更佳,R2=0.890,RMSE=3.01 mg·L-1。基于训练好的PSO-RBF模型,应用GF-1 WFV1遥感影像对闽江下游水体总悬浮物浓度进行反演,影像反演的总悬浮物浓度RMSE=3.65 mg·L-1,MRE=14.11%,遥感影像反演结果精度明显高于克里金空间插值结果。分析其空间分布特征,从上游方向往下游方向呈现增加趋势,马尾至闽江入海口河段总悬浮物浓度增加明显。 关键词: 悬浮物浓度;  遥感反射率;  闽江;  粒子群;  神经网络     Abstract: Total suspended matter (TSM) is one of the important parameters of water environment.As the spectral characteristics of the Case-II water are complicated,so it is not suitable to represent the relationship between spectral characteristics andTSM by simple linear models.In this paper,the test data,which is acquired by water quality sampling and spectral measurement of 40 points from July 12 -13,2017,together with GF-1 WFV1 bands reflectance data are used to analysis the correlation between remote sensing factors and TSM.Taking advantage of high correlation coefficients between bands,such as b3,b3/b2 and b3/b1,we construct PSO-RBF and RBF neural network model to inverse TSM.At the same time,a empirical b3/b2 ratio model is also proposed.The result shows that PSO-RBF neural network model’s performance is better than traditional RBF neural network and the empirical model,whose R2=0.890,RMSE=3.01 mg/L.On this basis,the GF-1 WFV1 remote sensing image is used to inverse TSM of Minjiang River,which is calculated by the well-trained PSO-RBF model.Furthermore,the spatial distribution characteristics of TSM is also studied.The result of TSM inversion comes to RMSE=3.65 mg·L-1,MRE=14.11% respectively,and remote sensing image retrieval results accuracy was significantly higher than that of Kriging interpolation results,and there is

下载说明

·如果您发现该资源不能下载,请通知管理员.gissky@gmail.com

·为确保下载的资源能正常使用,请使用[WinRAR v3.8]或以上版本解压本站资源,缺省解压密码www.gissky.net ,如果是压缩文件为分卷多文件,请依次下载每一个文件,并按照顺序命名为1.rar,2.rar,3.rar...,然后鼠标右击1.rar解压.

·为了保证您快速的下载速度,我们推荐您使用[网际快车]等专业工具下载.

·站内提供的资源纯属学习交流之用,如侵犯您的版权请与我们联系.