土地利用/覆被遥感分类评价与季节变化分析--以连云港为例  

Accuracy assessment of land use/cover classification from different seasons based on remote sensing imagery:A case study of Lianyungang city,China

作　　者：陈龙高[1] 李英奎 杨小艳[1] 陈龙乾[3] 

机构地区：[1]江苏师范大学土地资源研究所,江苏徐州221116 [2]Department of Geography,University of Tennessee,Knoxville,Tennessee 37996 [3]中国矿业大学环境与测绘学院,江苏徐州221116

出　　处：《中国矿业大学学报》2013年第5期873-879,886,共8页Journal of China University of Mining & Technology

基　　金：国家自然科学基金项目(41271121);江苏省高校自然科学基金项目(09KJB170006)

摘　　要：基于遥感影像地物的季节变化特征对土地利用/覆被(LUC)信息提取的重要性,使用中巴卫星冬末、春末及秋季的影像,采用最大似然法对连云港沿海区域进行了土地利用/覆被分类,评估了其精度与误差,并对3个时期的土地利用/覆被空间分布特征及季节变化进行了研究.结果表明:分类精度最高的季节为秋季,其总体精度(85.2%)和总体Kappa系数(0.81)可以满足土地利用环境影响分析的要求;不同季节土地利用/覆被类型转移主要涉及高密度建设用地与水体之间、低密度建设用地与耕地及高密度植被之间、高密度植被与高低密度建设用地之间等;沿海地区土地利用/覆被季节变化特征可用于辅助公路等线状地物的提取、耕地利用季节变化的检测、沿海滩涂空间布局的调查分析等.Seasonal change analysis of different land use/cover （LUC） types is of great impor- tance for the LUC classification using remote sensing imagery. Based on the LUC classification of Lianyungang City with CBERS-02 imagery taken from three seasons in 2005 （the end of win- ter, the end of spring and fall） using Maximum Likelihood Classification Method （MLC） and the accuracy assessment of each seasonal result, the LUC spatial distribution and seasonal changes in the study area were studied. The results show that： 1） The image from the fall sea- son has the highest classification accuracy with an overall percentage correctly classified of 85.2~ and a Kappa coefficient of 0.81; 2） The transition of LUC types in the three seasons was mostly occurred between the high density residential and water, the low density residential and cropland or high density vegetation, and the high density vegetation and high/low densityresidential; 3） The characteristics and spatial patterns of LUC transitions among different sea- sons could be used to extract linear land use features （such as road）, assess the seasonal change of cropland, and analyze the spatial distribution of coastal intertidal land.

关 键 词：土地利用 覆被 CBERS-02遥感影像 季节变化 分类精度与误差 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] P962[自动化与计算机技术—控制科学与工程]