中文摘要：

影像分类问题是遥感影像信息提取研究领域的基本问题,同时也是十分关键的技术问题之一。随着国产遥感卫星数据种类和质量的丰富与提升,对其进行高精度分类以实现国产卫星数据的有效利用,成为非常重要且值得深入探讨的问题。文章以湖南省益阳市某区域高分一号遥感影像为例,在监督分类、非监督分类与面向对象分类三大分类方法中,各选取两种最具代表性的分类算法进行研究区影像分类,对比分析这六种分类方法对研究区影像分类的结果与精度,寻找适用于国产卫星高分一号影像的分类方法。研究结果表明：支持向量机-监督分类效果最佳,面向对象分类方法也有一定的优势;进一步研究发现,经过融合处理后的研究区影像,在分类效果和精度上,都有明显的改善与提升。

英文摘要：

Image classification plays a fundamental and crucial role in extracting remote sensing image information. With the enrichment of types and improvement of quality of domestic remote sensing satellite data, it has become a very important problem worthy of in-depth discussion to carry out high-precision classification so as to realize the effective utilization of domestic satellite data.In this article, a GF-1 image of a region in Yiyang, Hunan Province is taken as an example. Among the three classification methods of supervised classification, unsupervised classification and object-oriented classification, two representative classification algorithms are selected to classify images in the study area and the accuracy and precision of these six classification methods are compared and analyzed in a bid to find a suitable classification method for domestic high-resolution satellite images. The results show that SVM- supervised classification yields Further research shows that the best effect and object-oriented classification method also has certain advantages. image of fusion area has improved significantly in terms of classification effect and precision.