论文部分内容阅读
以渭干河——库车河三角洲绿洲为例,利用SAR数据,采用不同的分类方法来提取该研究区盐渍化土地覆盖信息。首先用Enhanced frost滤波算法对SAR图像进行去噪处理。然后基于灰度共生矩阵理论提取去噪后的SAR图像4种纹理特征,并在不同窗口大小下筛选出有效的纹理特征。最后结合纹理特征分别采用最大似然分类法和SVM分类法对SAR图像进行分类。研究结果表明:基于纹理特征的SVM分类方法,能够有效解决单源数据信息图像分类效果破碎问题;13×13窗口的总精度达到98.2456%,Kappa系数达到0.9763,更有利于遥感图像分类和盐渍化信息监测,是地物遥感信息提取的有效途径。
Taking the Weigan-Kuqa River delta oasis as an example, different classification methods were used to extract salinized land cover information from the study area using SAR data. Firstly, the enhanced frost filtering algorithm is used to denoise SAR images. Then, based on the gray level co-occurrence matrix theory, four texture features of the denoised SAR image are extracted, and valid texture features are screened at different window sizes. Finally, the SAR images are classified by maximum likelihood classification and SVM classification according to texture features respectively. The results show that the SVM classification method based on texture feature can effectively solve the problem of fragmentation of single-source data image classification effect. The total accuracy of 13 × 13 window reaches 98.2456%, and the Kappa coefficient reaches 0.9763, which is more beneficial to remote sensing image classification and salty The monitoring of information is an effective way to extract remote sensing information of land object.