摘要
基于高分遥感影像进行震后建筑物检测,对开展应急响应救援等具有重要意义。在缺乏震前数据条件下,其核心在于构建高效的特征空间以进行震害建筑物特征建模。为此,提出了一种基于稀疏词典的震后建筑物检测方法。首先,联合光谱、纹理和几何形态学特征对震后建筑物进行多角度刻画;其次,通过构建相同、相异词对进一步引入空间上下文信息,从而构建多特征初始视觉词典;在此基础上,结合正交匹配追踪(OMP)和K奇异值分解(K-SVD)算法对视觉词典进行稀疏表示,以尽可能的减少冗余信息;最后,通过支持向量机获得最终检测结果。通过多组震后影像实验表明,所提出方法的总体精度可达到85%以上,且在目视解译和定量分析中均显著优于对比方法。
The post-earthquake building detection based on high-resolution remote sensing image is of great significance for emergency response and rescue. In the absence of pre-earthquake data, the key is to construct an efficient feature space for feature modeling of buildings damaged by earthquake. Therefore, a post-earthquake building detection method based on sparse dictionary is proposed in this study. Firstly, the post-earthquake buildings are depicted from multiple angles by combining spectral, textural and geometric morphological features. Secondly, the spatial context information is further introduced by constructing the same and different pairs of words to construct the multi-feature initial visual dictionary. On this basis, the orthogonal matching pursuit algorithm and the K-singular value decomposition algorithm are fused to perform sparse representation of the visual dictionary to reduce redundant information as much as possible. Finally, the detection result is achieved by support vector machine. Experimental results of multiple post-earthquake image show that the overall accuracy of the proposed method can reach over 85%. It is significantly superior to the comparison method in visual interpretation and quantitative analysis.
作者
石梵
王超
申祎
张艳
仇星
Shi Fan;Wang Chao;Shen Yi;Zhang Yan;Qiu Xing(School of Electronic and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China;Nanjing Yuntian Zhixin Information Technology Co.,Ltd.,Nanjing 211106,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2020年第7期205-213,共9页
Chinese Journal of Scientific Instrument
基金
江苏省“六大人才高峰”项目(XYDXX-135)
江苏省高等学校优势学科资助
关键词
高分辨率遥感影像
建筑物检测
视觉词典
稀疏
支持向量机
high-resolution remote sensing image
building inspection
visual dictionary
sparse
support vector machine(SVM)
