期刊名称：International Journal of Signal Processing, Image Processing and Pattern Recognition
印刷版ISSN：2005-4254
出版年度：2015
卷号：8
期号：8
页码：361-372
DOI：10.14257/ijsip.2015.8.8.37
出版社：SERSC
摘要：High resolution remote sensing image contains abundant information, and the amount of data expands rapidly, it is challenge and advantage to the corn information extraction. For high resolution images, the spectral characteristics are the most important characteristics, but remote sensing classification only according to the spectral ignores the texture information as positive factors. This paper introduced texture features, establishes joint spectrum - texture feature set. Support Vector Machine (SVM) classification method relies on the selected parameters, this paper used Particle swarm optimization (PSO) to optimize penalty and kernel function parameters. Under the condition of a large number of high resolution remote sensing data, the increase of samples will influence the algorithm efficiency greatly. This paper improved SVM, namely Adaboost_SVM, extracted corn planting information. The method can be used as support for corn area prediction and yield estimation. The experimental results show that the proposed method improves extraction accuracy, improves efficiency compared with classic algorithms significantly
关键词：remote sensing; region extraction; PSO; parameter optimization; ; Adaboost_SVM
