摘要
围绕形态学属性剖面和SVM分类算法相结合进行遥感图像特征匹配和检索展开研究,提出了基于属性剖面(attribute profiles,APs)和支持向量机(support vector machine,SVM)的遥感图像检索方法。首先,对遥感图像进行形态学属性滤波;接着对滤波后图像提取灰度特征和Gabor纹理特征;最后,在提取的特征基础上,使用SVM对候选图像进行预分类,判断待检索图像类别;再从该类别中根据距离度量选择相似度高的图像以完成检索。实验表明,所提方法具有较好的检索性能。
Combining the morphological attribute profiles and classification based on the support vector machine,the study is carried out on remote sensing image feature matching and retrieval.An novel remote sensing image retrieval method using morphological attribute profiles（APs）and support vector machine（SVM）is proposed.Firstly,remote sensing images are filtered by APs to obtain structure information of images.Secondly,features of intensity and Gabor texture are extracted from the filtered images.Finally,based on the extracted features,SVM is applied to classify the candidate set of images and query image.The candidate images which have the same labels with the query image are ranked according to the distance metric,and images with higher similarity values are selected as retrieval results.Experiments show the effectiveness of the proposed method.
出处
《电子测量技术》
2016年第8期96-99,共4页
Electronic Measurement Technology
关键词
遥感图像检索
属性剖面
支持向量机
纹理特征
remote sensing image retrieval
attribute profiles
support vector machine
texture features
