中文摘要：

结合支持向量机技术与基于粗糙集的粒度计算,提出一种新的高分辨率遥感影像面向对象分类方法。首先,采用相位一致模型得到QuickBird全色影像的梯度图,并利用扩展最小变换技术获取地物目标的前景标识,进而采用强制最小技术重建梯度。在此基础上,采用分水岭变换得到较佳分割效果。然后,从多光谱波段数据中提取对象的光谱特征,并用Gabor小波产生纹理特征,利用多核支持向量机进行初步的面向对象分类,对分类结果进行求交后则生成信息颗粒。最后,比较颗粒的特征均值与各样本中心的欧氏距离区分颗粒的类别,通过定量分析颗粒间的空间相邻关系判断待定类别的颗粒,利用少量人工交互的识别处理得到最终分类结果。与基于高斯径向基核函数的支持向量机和神经网络两种方法进行对比分析,试验结果表明本文所提方法能够取得更好的分类效果。

英文摘要：

A new object-oriented method for classification of high resolution remotely sensed imagery is proposed,which integrates support vector machine（SVM） technique with rough-set-based granular computing（RSBGC）.First,gradient image is obtained by applying phase congruency model to the QuickBird panchromatic image.Extended minima transform and minima imposition are used to get foreground marking of geo-objects and implement gradient reconstruction respectively.Based on these improvement measures,better segmentation is achieved using watershed transform.Second,spectral characteristic is got from multi-spectral data and texture feature is extracted by Gabor wavelet.Multi-kernel SVM is used to present preparatory object-oriented classification,and information granularities are obtained through intersection of the classification results.Third,granularities are differentiated by means of comparing the Euclidean distance between average value of granularity and every sample central moment.Spatial adjacency relation among the granularities is quantitative analyzed in order to classify the uncertain granularities after the former clustering.The resulting classification is achieved by little artificial interaction identification.A comparative experiment is performed with both SVM and neural network methods based on RBF-kernel function.It is shown that the proposed method can obtain better classification results.