论文部分内容阅读
为了进一步提高遥感图像分类的精度,提出了一种基于Log-Gabor小波和Krawtchouk矩的遥感图像分类算法。首先利用Log-Gabor小波对遥感图像进行多方向、多分辨率滤波,提取遥感图像的纹理特征;同时计算遥感图像的Krawtchouk矩不变量,作为遥感图像的边缘形状特征,并与基于Log-Gabor小波提取的纹理特征构成完整的特征向量;最后依据所提取的特征向量利用支持向量机(support vector machine,SVM)分类器对待分类图像进行分类,得到最终的遥感图像分类结果。实验结果表明,与近年来提出的基于Gabor小波、基于Log-Gabor小波、基于Krawtchouk矩等3种遥感图像分类算法相比,本文算法在主观视觉效果和分类精度等客观定量评价指标上都有了明显的改善,是一种行之有效的遥感图像分类算法。
In order to further improve the accuracy of remote sensing image classification, a remote sensing image classification algorithm based on Log-Gabor wavelet and Krawtchouk moment is proposed. Firstly, Log-Gabor wavelet is applied to multi-directional and multi-resolution filtering of remote sensing images to extract the texture features of remote sensing images. Simultaneously, Krawtchouk moment invariant of remote sensing images is calculated as the edge shape feature of remote sensing images and compared with Log-Gabor wavelet The extracted texture features constitute a complete eigenvector. Finally, the classification of the image to be classified is carried out by using a support vector machine (SVM) classifier based on the extracted eigenvectors to obtain the final remote sensing image classification result. The experimental results show that compared with the three remote sensing image classification algorithms based on Gabor wavelet, Log-Gabor wavelet and Krawtchouk moment proposed in recent years, the proposed algorithm has objective and quantitative evaluation indexes such as subjective visual effect and classification accuracy Obvious improvement, is an effective remote sensing image classification algorithm.