【摘要】 为高效准确地监测冬小麦种植面积，本文以“珠海一号”高光谱影像(OrbitaHyperSpectral,简称OHS)为数据源，以河北省安新县为研究区，使用支持向量机(Support vector machine,SVM)、决策树、面向对象3种分类方法结合光谱特征、植被指数特征和纹理特征提取冬小麦种植分布信息。结果表明：3种方法提取冬小麦效果都比较理想，支持向量机(SVM)方法更适用于“珠海一号”高光谱数据冬小麦面积监测，总体精度、Kappa系数和解译精度分别为0.95、0.89和93.08%;基于“珠海一号”高光谱数据监测冬小麦面积可行性程度较高，可有效促进遥感大尺度高精准提取农作物应用。更多还原

【Abstract】 In order to efficiently and accurately monitor the planting area of winter wheat, this study took the Zhuhai-1 Orbita hyperspectral image(OHS) as the data source and Anxin county in Hebei province as the research area, support vector machine(SVM), decision tree, and object-oriented classification methods were used to extract the planting distribution and area information of winter wheat by combining spectral characteristics, vegetation indices characteristics, and texture characteristics. The results showed that the three methods were all effective in extracting winter wheat, and the support vector machine(SVM) method was more suitable for monitoring the area of winter wheat in the hyperspectral data of Zhuhai-1 satellite. The overall accuracy, Kappa coefficient, and interpretation accuracy were 0.95, 0.89 and 93.08%, respectively. It is feasible to monitor winter wheat area based on the hyperspectral data of Zhuhai-1 satellite, which can effectively promote the application of large-scale and high-accuracy crop extraction by remote sensing.更多还原