摘要
以环境卫星为数据源对泰兴市水稻面积进行提取,选择水稻与其他作物光谱差别最大的时期作为水稻识别的最佳时期,在利用影像原始光谱信息的基础上,分析和提取多种分类特征,运用支持向量机法、CART决策树法和最大似然法进行分类,提取水稻面积。结果表明支持向量机的分类精度最高,总体精度为80.38%,Kappa系数为0.74。说明当样本量较少时,支持向量机可以更好地利用多源信息,具有更高的分类精度。
Based on the satellite images corrected at the period during which paddy rice and other crops could be differentiated due to their spectra characteristics, the planting areas of paddy rice in Taixing county of Jiangsu province were extracted by SVM （support vector machine ） , CART （classifcation and regression tree ） and MLC （maximum likelihood clas- sifier）. The method of SVM displayed the highest classification accuracy, with overall accuracy being 80.38% and Kappa coefficient being 0.74, indicating that SVM could make better use of multi-information, and it had higher classifyication ac-curacy when the amount of sample is small.
出处
《江苏农业学报》
CSCD
北大核心
2012年第4期728-732,共5页
Jiangsu Journal of Agricultural Sciences
基金
江苏省农业三项工程项目[SX(2011)392]
关键词
遥感
水稻种植面积
环境卫星
分类
remote sensing
planting area of paddy rice
environmental satellite
classification
