摘要： 介绍了一种基于Gabor纹理特征的遥感图像检索方法。利用Gabor滤波器对图像进行多分辨率分解，提出了一种改进的基于滤波图像及图像能量的纹理特征提取方法，利用尺度、方向、纹理结构构成特征向量，使用归一化欧氏距离进行了相似性判断。该方法计算过程简单、快速，提高了方向判别的可靠性。仿真实验和实际运用验证了该算法的有效性。

关键词: Gabor小波, 基于内容的图像检索, 纹理描述符, 遥感图像

Abstract: Using Gabor wavelet texture features, this paper proposes a content-based retrieval algorithm on remote sensing images. A descriptor of sensing image texture feature, in which the scale, direction and structure of texture are combined, is introduced, and guarantees the reliability of distinguishing on texture structure direction. The similarity between query and remote sensing images is computed by using normalized Euclidean distance. The algorithm can be executed faster than others based on Gabor wavelet because of the feature representation and similarity computing. The emulation and application show its effect.

Key words: Gabor wavelet, Content-based image retrieva1, Texture descriptor, Remote sensing image

中图分类号: 

• TP18