基于改进支持向量机的盐碱地信息精确提取方法研究

徐存东; 张锐; 王荣荣; 程慧; 刘璐瑶; 王燕; 朱兴林

引用本文:	徐存东,张锐,王荣荣,等.基于改进支持向量机的盐碱地信息精确提取方法研究[J].灌溉排水学报,2018,37(9):62-68.
	XU Cundong,ZHANG Rui,WANG Rongrong,et al.基于改进支持向量机的盐碱地信息精确提取方法研究[J].灌溉排水学报,2018,37(9):62-68.

【打印本页】【下载PDF全文】【查看/发表评论】【EndNote】【RefMan】【BibTex】

←前一篇|后一篇→

过刊浏览高级检索

本文已被：浏览 1534次下载 1183次	码上扫一扫！
分享到：微信更多字体:加大+\|默认\|缩小-
基于改进支持向量机的盐碱地信息精确提取方法研究
徐存东，张锐，王荣荣，程慧，刘璐瑶，王燕，朱兴林
1.华北水利水电大学，郑州 450045； 2.水资源高效利用与保障工程河南省协同创新中心，郑州 450045； 3.河南省水工结构安全工程技术研究中心，郑州 450046

摘要:

【目的】探索无人机遥感下可以精确提取盐碱地信息的方法。【方法】以甘肃省景泰川电力提灌灌区一期灌区为研究区，采用 Trimble UX5固定机翼无人机采集研究区遥感数据，提出了一种基于AdaBoost算法的改进支持向量机(SVM) 分类的新方法，以实现盐碱地信息的精确提取。【结果】与传统SVM分类相比，改进SVM分类精度提升显著，总精度最高达96.55%，Kappa系数为0.957 3。改进前后不同类型盐碱地提取面积与实测面积相比，最大误差为5.4%，平均误差为2.16%。【结论】该文提出的基于改进支持向量机分类方法可有效提高遥感影像的分类精度。

关键词: 高空无人机；盐碱地；AdaBoost；支持向量机；景电灌区

DOI：10.13522/j.cnki.ggps.20180091

分类号:

基金项目:

An Improved Support Vector Machine Method for Estimating Saline-alkali Soil from Remote Sensing Imagery

XU Cundong，ZHANG Rui， WANG Rongrong，CHENG Hui，LIU Luyao，WANG Yan，ZHU Xinglin

1. School of Water Conservancy, North China University of Water Resources and Electric Power, Zhengzhou 450045, China; . Collaborative Innovation Center of Water Resources Efficiency and Protection Engineering, Zhengzhou 450045, China；3. Henan Provincial Hydraulic Structure Safety Engineering Research Center, Zhengzhou 450046, China

Abstract:

【Objective】 Remote sensing technology has been increasingly used in agriculture and this paper presents an improved support vector machine method to estimate saline-alkali soil from imagery obtained from UAV.【Method】 We took the first irrigation district at the Jingtai Electric Power Irrigation District of Gansu Province as an example, and the Trimble UX5 fixed-wing UAV was used to collect the remote sensing imageries of the area. A new method based on the AdaBoost algorithm for classification of SVM was proposed to estimate the area of saline-alkali soil. 【Result】Compared with traditional classification methods, the proposed SVM classification method significantly improved the calculation with an overall accuracy of up to 96.55% and Kappa coefficient of 0.957 3. Compared with the measured data, maximum and average errors of the improved methods were 5.4% and 2.16% respectively. 【Conclusion】 The improved SVM classification method proposed in this paper could effectively estimate saline-alkali soil based on remote sensing imageries.

Key words: UAVs at high altitude; saline alkali; AdaBoost; support vector machines; irrigation district