光谱学与光谱分析, 2017, 37 (11): 3386, 网络出版: 2018-01-04  

近红外光谱主成分分析与模糊聚类的典型地面目标物识别

Typical Ground Object Recognition Based on Principle Component Analysis and Fuzzy Clustering with Near-Infrared Diffuse Reflectance Spectroscopy
作者单位
1 天津大学精密测试技术及仪器国家重点实验室, 天津 300072
2 天津大学精密仪器与光电子工程学院, 天津 300072
摘要
近红外光谱技术在遥感监测领域中应用广泛, 针对典型地面目标物遥感监测识别需要, 提出了光谱主成分分析(PCA)与模糊聚类结合的分类识别方法, 提高了识别算法效率及准确性。 以四类典型地面目标物作为研究对象, 分别测量其在1 100~2 500 nm范围内漫反射光谱, 首先对漫反射光谱进行主成分分析, 得到代表光谱特征的主成分分量, 然后将其作为模糊聚类分析模型输入参数, 计算样品主成分集合之间贴合度, 最后利用择近原则对样品进行匹配分类。 结果表明, 主成分分析可以有效提取光谱特征并且降低数据维度, 结合基于择近原则的模糊分类方法, 可有效提高算法准确性与效率, 为遥感光谱在地面目标物识别应用提供了有益的参考。
Abstract
Near infrared spectroscopy is widely applied in remote sensing and recognition. In this manuscript, the classification method which is based on principal component analysis (PCA) and fuzzy clustering is applied to recognize typical objects with near-infrared diffuse reflectance spectra. The diffuse reflectance spectra of four types of ground targets are measured in the range of 1 100~2 500 nm. Firstly, the spectral features are extractedwith PCA analysis. Secondly, the principal components is set to be the input to the fuzzy clustering model to calculate the closeness degree between different samples. Finally, the typical objects were classified based on principle of fuzzy closeness optimization. The results indicated that the proposed method is beneficial for automatic recognition and classificationof typical ground. Thenear-infrared diffuse reflectance spectroscopy reflects the feature of typical ground object. Taking the advantage of spectral features extracting and data dimensions reduction, the PCA algorithm can effectively improve the recognition efficiency. The fuzzy classification method also improve the robustness of the model. This method provides a new concept for the analysis and processing of remote sensing spectroscopy.

李晨曦, 孙哲, 蒋景英, 刘蓉, 陈文亮, 徐可欣. 近红外光谱主成分分析与模糊聚类的典型地面目标物识别[J]. 光谱学与光谱分析, 2017, 37(11): 3386. LI Chen-xi, SUN Zhe, JIANG Jing-ying, LIU Rong, CHEN Wen-liang, XU Ke-xin. Typical Ground Object Recognition Based on Principle Component Analysis and Fuzzy Clustering with Near-Infrared Diffuse Reflectance Spectroscopy[J]. Spectroscopy and Spectral Analysis, 2017, 37(11): 3386.

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!