结合纹理的SVM遥感影像分类研究

针对传统统计模式识别分类方法分类精度不高,分类时未加入像元灰度的空间分布和结构特征以及分类时样本不足等缺陷,采用一种结合纹理的支持向量机(SVM)遥感图像分类方法。该方法在对Landsat7 ETM遥感影像进行纹理特征提取的基础上,构建了结合纹理的SVM分类模型。以河南省汝阳县为试验区,利用此模型对该区域的土地利用类型进行分类研究,并将分类结果与最大似然法和单源数据(光谱)SVM分类结果进行定性和定量比较分析。研究结果表明:该方法能够有效地解决单数据源分类效果破碎、分类精度不高等问题;对高维输入向量具有较高的推广能力;总精度达到90%,比单源信息的SVM分类法提高了6%,而与最大似然法相比,总精度提高了近9%,取得了良好的效果。

RS Image classification based on SVM method with texture

In order to overcome the shortage of low accuracy,the absence of pixels,spatial distribution and structure,and insufficient samples in the traditional statistical pattern recognition classification,a new method of classification using SVM based on texture is presented.In this method,the SVM classification model combined with texture analysis is established on the basis of texture extraction from Landsat7 ETM RS Image.RuYang country in Henan province is the test area.According to the model,the type of landuse in the area is classified. The classification result is compared with single data source(spectrum) SVM classification and maximum likelihood classification qualitatively and quantitatively.The research result shows that this method can effectively solve the problem of low accuracy and fracture classification result in single data source classification;it has high spread ability toward higher array input;the overall accuracy is 90%,which increases by 6% comparing with single data source SVM and increases by 9% comparing with maximum likelihood classification and thus acquires good effectiveness.