基于HJ-1A高光谱遥感数据的湟水流域典型农作物分类研究

发布时间：2018-01-19 07:30

  本文关键词： 高光谱遥感 HJ-A 农作物分类 支持向量机 遗传算法　出处：《遥感技术与应用》2017年02期 　论文类型：期刊论文

【摘要】：利用高光谱遥感技术识别农作物类型已经成为高光谱遥感研究的热点领域。以青海省湟水流域内油菜、小麦和青稞等典型农作物为分类对象,以HJ-1A HSI高光谱数据和GF-1 WFV高分辨率数据为数据源,探讨利用高光谱遥感影像进行农作物类型信息提取的方法。数据经预处理后,首先,利用WFV数据采用面向对象方法提取研究区农作物种植边界,并利用其对HSI高光谱影像进行种植区域提取;其次,将提取后的高光谱影像经数据形式变换获得包括:R、1/R、Log(R)、d(R)、d(Log(R))和CR共6种数据形式;最后,利用上述6种数据形式的全波段数据和经遗传算法GA-SVM进行光谱波段选取后的6种特征数据,采用支持向量机SVM方法进行农作物分类。结果表明:采用基于样本的面向对象分类方法提取耕地信息精度高且实现周期短;利用GA-SVM波段选取后的6种特征数据集进行农作物分类,其精度显著高于全波段数据集分类精度;6种数据变换形式中,d(Log(R))和CR是两种最优的高光谱分类数据形式,其全波段和特征波段数据进行农作物分类均能获得较好的分类精度,总体精度最高分别达88%和86%,而采用1/R、Log(R)和R数据形式需经GA-SVM光谱波段选取后才能获得较优分类精度。
[Abstract]:Using hyperspectral remote sensing technology to identify crop types has become a hot research area of hyperspectral remote sensing. Typical crops such as rape wheat and highland barley in Huangshui basin of Qinghai Province are taken as classification objects. Using HJ-1A HSI hyperspectral data and GF-1 WFV high-resolution data as data sources, the method of extracting crop type information from hyperspectral remote sensing images is discussed. Firstly, WFV data are used to extract the crop planting boundary in the study area, and the HSI hyperspectral image is used to extract the planting area. Secondly, the extracted hyperspectral images were transformed into six data forms, including 1 / 1 / R ~ (1 / R) / R ~ (1 / R) and CR (R ~ (1 / R) ~ (1 / R) / R ~ (1 / R) / R ~ (1 / R) / R ~ (1 / R)). Finally, the full band data of the above six data forms and the six characteristic data after the spectral band selection by genetic algorithm (GA-SVM) are used. The SVM method of support vector machine is used to classify crops. The results show that the method of object-oriented classification based on samples has high precision and short realization period. The precision of crop classification is significantly higher than that of the whole band data set by using the 6 characteristic data sets selected from GA-SVM band. Among the six kinds of data transformation, the two optimal hyperspectral classification data forms are Dendrogram (RU) and CR. The classification accuracy of the whole band and characteristic band data for crop classification can be obtained. The total precision is up to 88% and 86, respectively, and the best classification accuracy can be obtained by using 1 / R R) and R data after the GA-SVM spectral band is selected.
【作者单位】： 青海师范大学生命与地理科学学院青海省自然地理与环境过程重点实验室;
【基金】：国家自然科学基金项目(40861022、41550003) 青海省重点实验室发展专项(2014-Z-Y24、2015-Z-Y01)
【分类号】：S127;TP751
【正文快照】： 1引言农作物识别与分类是农业遥感的基础,是农情遥感监测的重要内容,高效、准确地提取农作物种植面积、结构和类型分布信息对于国家粮食安全、社会经济稳定、农业政策制定等均具有重要意义[1]。受我国农作物种植分散性、类型多样化和地域复杂性影响,采用常规的地面调查方法获，

本文编号：1443237