一种改进低层特征算子的遥感影像检索方法

• 作 者: 周维勋 邵振峰
• 机 构: 武汉大学测绘遥感信息工程国家重点实验室,
• 出 处: 测绘科学
• 发表日期: 2015年03期
• 关键词: 低层特征 纹理特征 颜色特征 T_Meta C_Meta MF 遥感影像 颜色直方图 检索方法 二次检索

在遥感数据检测中,现阶段往往将C_Meta(颜色特征计算得到)和T_Meta(纹理特征计算得到)作为两种不同的Meta-feature特征,使得对于同类地物影像利用C_Meta和T_Meta检索的结果不理想且有较大差异。为了改善检索结果,该文提出通过初步检索选择检索结果较好的C_Meta或T_Meta作为不同类别影像最终的Meta-feature特征;并利用新的影像特征库进行第二次检索。实验结果表明,此方法的检索结果好于原始的C_Meta和T_Meta以及传统的颜色直方图和小波纹理。

In the detection of remote sensing data, C_Meta (color feature calculation) and T_Meta (texture feature calculation) are often taken as two different Meta-feature features at this stage, so that the results of C_Meta and T_Meta retrieval are not ideal for the same type of feature images And there is a big difference. In order to improve the retrieval results, this paper proposes to select the best C_Meta or T_Meta as the final Meta-feature of different categories of images through the preliminary retrieval and use the new image feature library for the second retrieval. The experimental results show that the retrieval result of this method is better than the original C_Meta and T_Meta, as well as the traditional color histogram and wavelet texture.