论文部分内容阅读
提出一种基于贝叶斯理论和马尔科夫随机场MRF(Markov Random Fields)的主被动遥感数据协同分类方法。该方法依据光学与微波遥感数据在地物提取中的各自优势,首先对ASAR后向散射系数进行入射角归一化,然后构建一种基于贝叶斯理论和MRF的分类器,以归一化后的ASAR双极化数据与TM7个波段共同参与分类。分别对ASAR入射角归一化的有效性和主被动协同的必要性进行验证,结果表明,采用本文方法的分类精度达到89.4%,较未进行角度校正的主被动数据协同分类的精度提高了4.1%,较单独TM分类的精度提高了11.5%,体现出主被动遥感数据协同在分类上的潜力。
This paper proposes a collaborative classification method of active and passive remote sensing data based on Bayesian theory and Markov Random Fields (MRF). Based on the respective advantages of optical and microwave remote sensing data in the feature extraction, firstly, the ASAR backscattering coefficient is normalized to the incident angle, and then a classifier based on Bayesian theory and MRF is constructed to normalize After the ASAR dual polarization data and TM7 band jointly participate in classification. The effectiveness of normalization of ASAR incident angles and the necessity of active and passive coordination respectively are verified. The results show that the classification accuracy of this method is 89.4%, which improves the precision of coordinated classification of active and passive data without angle correction by 4.1 %, Which is 11.5% higher than that of the single TM classification, which shows the potential of the classification of the active and passive remote sensing data.