【摘要】 以太湖流域作为研究对象,基于环境减灾卫星HJ-1B影像数据,比较利用最大似然法和支持向量机法进行土地利用/覆盖分类的效果。结果显示,SVM法在总体分类精度和Kappa系数上较传统最大似然法有所提高。SVM分类方法对于有限样本的分类表现出优越的性能,改善了传统分类方法的局限性,具有很大的应用潜力。更多还原

【Abstract】 With TaihuLake Basin as a case study,we aim to compare maximum likelihood classification(MLC) and support vector machine(SVM) methods for land use/cover classification based on HJ-1B satellite images.The results show that SVM method has better performance in overall accuracy and Kappa coefficient compared to MLC method.SVM method has superior performance in the classification problem for limited samples,and it may overcome the limitations of traditional methods and hold great potentials in remote sensing classification.更多还原