摘要:
为了提高遥感图像检索的效率和准确性,提出了一种融合相容粒计算模型的遥感图像检索方法.首先,根据相容粒理论定义了区域相容粒、图像相容粒和区域相容粒信息表等相关概念,将遥感图像粒化;然后,计算出图像区域相容粒的相似度;最后,结合综合区域匹配算法,提出融合相容粒理论的遥感图像相似性度量算法,并利用IKONOS影像进行对比实验.实验结果表明,融合相容粒理论的检索算法能够提高遥感图像检索的查准率,与综合区域匹配算法相比,本文算法查准率提高了12.08%,基本满足用户需求.
Abstract:
In order to improve efficiency and accuracy of remote sensing image retrieval, this paper proposes a remote sensing image retrieval approach based on granular computing model. Firstly, according to the tolerance granular computing theory, a series of concepts are defined, such as region tolerance granule, image tolerance granule and regional tolerance granular information table, and remote sensing images are granulated. Secondly, the region tolerance granular similarity is calculated. Finally, the remote sensing image similarity model is built combining tolerance granular computing and image integrated region matching algorithm. Using IKONOS data, the authors verified the two retrieval algorithms. The experimental results show that the precision of proposed approach is increased by 12. 08% in comparison with original integrated region matching algorithm. Therefore, it can be concluded that the proposed approach can meet the users' requirements.
下载: