基于大数据的高分辨率遥感场景识别

Remote Sensing Scene Recognition Based on Big Data

文章基于遥感大数据和Matlab平台,对输入的遥感图像采用LBP、SIFT和CH等不同描述算子进行特征提取,而后通过支持向量机进行分类识别。通过应用标准遥感数据集AID和WHU-RS19进行测试比较,结果表明使用低层特征进行场景识别的方法能够获得较好的识别结果,且能满足实时性要求。

The technology of remote sensing scene recognition has been widely used in both commodity econo⁃my and military.Due to the slow start and immature technology in China,the current methods of remote sens⁃ing scene recognition are not widely used.Based on the big remote sensing data and MATLAB platform,this pa⁃per used LBP,sift,CH and other different description operators to extract the features of the input remote sens⁃ing images,and then used support vector machine for classification and recognition.Through the test and com⁃parison of standard remote sensing data set aid and whu-rs19,the results show that the method of scene recogni⁃tion using low-level features can obtain better recognition results and meet the real-time requirements.