基于多光谱影像和专家决策法的作物分类研究

摘要

[Objective] The aim was to explore the feasibility of using single spectrum image to classify crops based on multi-spectral image and Decision Tree Method. [Method] Taking the typical agriculture plantation area in Hulunbeier area, according to field measured spectrum data, the optimum time of main crops, barley, wheat, rapeseed, based on crops spectrum characteristics, by dint of decision-making tree method, and considering spectral matching method, classification of crops was studied such as SAM. [Result] By dint of Landsat TM image gained in the first half of August, based on geographic and atmospheric proof-reading, decision-making tree was constructed. Plantation information about wheat, barley, and rapeseed and plantation grassland was extracted successfully. The general classification accuracy reached 86.90%. Kappa coefficient was 0.831 1. [Conclusion] Taking typical spectrum image as data source, and applying Decision Tree Method to get crops type＇s information had fine application future.%[目的]探讨基于多光谱影像和专家决策法的作物分类,验证利用单时相多光谱影像区分农作物的可行性。[方法]以呼伦贝尔地区典型农业种植区为研究区,根据野外实测光谱数据,寻找区分研究区主要作物大麦、小麦、油菜的最佳时间,根据作物波谱特征,采用决策树方法,结合光谱角度制图（SAM）等光谱匹配方法,开展了作物分类研究。[结果]利用8月上旬获取的LandsatTM影像,在对影像进行几何校正、大气校正的基础上,构建决策树,成功提取了小麦、大麦、油菜、种植草场的种植信息,分类总体精度达到86.90%,Kappa系数达到0.8311。[结论]以典型时相的多光谱影像为数据源,应用决策树方法提取作物类型信息,具有较好的应用前景。