摘要
针对地球观测系统/中分辨率成像光谱仪EOS/MODIS(Earth Observation System/Moderate Resolution Imaging Spectroradiometer)影像资料中的积雪检测,提出了基于支持向量机SVMs(Sup-port Vector Machines)的遥感影像分类方法。分析了积雪检测过程中的特征选择和提取,建立基于支持向量机的遥感影像分类模型,并进行了积雪检测试验。结果表明,特征选择对积雪检测起到了积极的作用,同时也证明了支持向量机方法在遥感影像分类中的优势,值得进一步推广使用。
Focusing on the detection of snow cover from the image data of Earth Observation System/ Moderate Resolution Imaging Spectroradiometer(EOS/MODIS) ,this paper introduces a new classification model of remote sensing images based on Support Vector Machines (SVMs). Firstly, the feature extraction and selection are analyzed of snow cover detection process, then a new classification method is established for remote sensing images based on SVMs, and at last a snow cover detection experiment is performed to validate the model. The results suggest that the feature selection is effective for the snow cover detection, which also certifies the predominance of SVMs method in remote sensing image classification. The new method is worth to further popularize and apply.
出处
《南京气象学院学报》
CSCD
北大核心
2009年第1期134-139,共6页
Journal of Nanjing Institute of Meteorology
基金
国家重点基础研究发展计划(973计划)项目(2006CB400505)
国家自然科学基金资助项目(40675040)