高分辨率遥感影像中的土地利用分类研究——以西南大学为例

大学校园是城市用地中的一个特殊群体,作为公共设施用地的一个重要部分,在城市土地利用日益紧张的条件下,对其进行土地利用分类研究有着重要的意义.本研究以西南大学为例,将最大似然监督分类法和决策树分类法相结合对其进行分类研究.研究表明:1)最大似然监督分类法和决策树法相结合对西南大学土地利用进行分类,取得了较好的效果,总分类精度达到了91.51%,Kappa系数达到了0.8831;2)分析出西南大学建筑分布特点:①校园内部建筑物趋于密集布局,提高了土地利用效率;②校园建筑物高层化,提高了建筑容积率.

Study of the Land Use Classification From High- resolution Multispectral Satellite Imagery A Case Study of Southwest University

The university campus is a special part of the land use in city.As an important part of the public-installation land,it is very helpful to study the classification of the land use in campus,especially when the situation of the city-land use is more and more severe.This paper takes the Southwest University as an example,and adopts the method of combining Maximum Likehood and Decision-tree to classify and study it.The results are:(1) the combination of Maximum Likehood and Decision-tree technologies which is adopted to classify the land of Southwest University has achieved a pretty good effect such as the Overall Classification Accuracy has reached 89.29% and Kappa coefficient reached 0.8653;(2) the characteristics of the distribution of the buildings in Southwest University are firstly,the buildings inside the campus tend to be densely distributed and this could improve the land use effficiency;secondly,the buildings tend to be more tall and therefore the volume is raised.