CAJ | 학술논문

高光谱遥感数据的聚类是定量遥感研究的热点。为提高其聚类精度，将光谱相似性测度与模糊动态聚类法有机结合，构建四种模糊聚类新方法：光谱角度-模糊聚类法、光谱信息熵-模糊聚类法、光谱角度与光谱信息熵混合-模糊聚类法与海明距离-模糊聚类法。以实测植被高光谱信息为数据，利用该四种方法进行比较分析其聚类精度。实验结果表明光谱角度与光谱信息熵混合模糊聚类法以及光谱角度模糊聚类法的聚类精度均相对其他几种方法较好，是最有效高光谱遥感聚类方法。为今后该领域的方法的进一步研究提供一种借鉴与参考。
고광보요감수거적취류시정량요감연구적열점。위제고기취류정도，장광보상사성측도여모호동태취류법유궤결합，구건사충모호취류신방법：광보각도-모호취류법、광보신식적-모호취류법、광보각도여광보신식적혼합-모호취류법여해명거리-모호취류법。이실측식피고광보신식위수거，이용해사충방법진행비교분석기취류정도。실험결과표명광보각도여광보신식적혼합모호취류법이급광보각도모호취류법적취류정도균상대기타궤충방법교호，시최유효고광보요감취류방법。위금후해영역적방법적진일보연구제공일충차감여삼고。
Hyperspectral remote sensing data clustering is one of the hot spot of the quantitative remote sensing. In order to improve the clustering accuracy of the hyperspectral remote sensing data, the spectral similarity measure and the fuzzy dynamic clustering algorithm are combined to produce four new fuzzy clustering algorithms,which are spectral angle mapper-fuzzy clustering algorithm, spectrum information divergence-fuzzy clustering algorithm combination of spectral angle mapper and spectral information divergence-fuzzy clustering algorithm and hamming distance-fuzzy clustering algorithm. The clustering results of the four new algorithms are compared by used of the actual hyperspectral data of vegetation. It is concluded that clustering accuracy of the combination of the spectral angle mapper and spectral information divergence-fuzzy clustering algorithm, spectral angle mapper-fuzzy clustering algorithm are higher than the other methods, which is the most effective clustering algorithm of hyperspectral remote sensing data. It will provide a reference for further study in this field.