论文部分内容阅读
以研究区2012年的HJ卫星CCD影像为数据源,通过物候历和主要农作物的光谱特征分析,确定棉花识别最佳时相。采用分类回归树分析(CART)的决策树方法提取棉花种植面积信息,并以农田实地调查样点和统计数据为参考对提取的棉花种植面积结果进行评价。结果表明,基于HJ-CCD数据,使用CART算法的决策树可以较好地提取棉花覆盖信息,最终提取的棉花种植面积总量精度为94.29%,位置精度为88.57%;本研究采用的决策树方法,操作方便、容易实现,分类结果较为实际,基本满足棉花种植面积遥感监测的需求,可对棉花种植面积估算和种植结构分析提供一定的参考。
Based on the 2012 HJ satellite CCD image data in the research area, the best phase of cotton identification was determined through the analysis of the phenological characteristics of the phenological calendar and the main crops. The information of cotton acreage was extracted by the method of decision tree based on Categorical Regression Tree Analysis (CART), and the results of acreage of cotton acreage were evaluated with field sampling points and statistical data of farmland as reference. The results showed that the decision tree based on CART algorithm could extract the cotton cover information well based on HJ-CCD data. The final total precision of cotton acreage was 94.29% and the position precision was 88.57%. The decision tree method , Easy to operate, easy to implement, the classification results are more realistic, and basically meet the needs of remote sensing monitoring of cotton acreage, which can provide some reference for cotton acreage estimation and planting structure analysis.